Multi-channel signal generator, audio encoder and related methods relying on mixed noise signals

Abstract

A multi-channel signal generator (200) is provided for generating a multi-channel signal (204) having a first channel (201) and a second channel (203). The multi-channel signal generator (200) includes: a first audio source (211) for generating a first audio signal (221); a second audio source (213) for generating a second audio signal (223); a mixed noise source (212) for generating a mixed noise signal (222); and a mixer (206) for mixing the mixed noise signal (222) with the first audio signal (221) to obtain the first channel (201), and mixing the mixed noise signal (222) with the second audio signal (222) to obtain the second channel (203). The present invention also provides an audio encoder, comprising: an activity detector (380) for analyzing a multichannel signal (304) to determine whether a frame in a frame sequence (381) is an inactive frame (308); and a noise parameter calculator (3040) for calculating first parameter noise data (p_noise, v_v) of the first channel (301, 201) of the multichannel signal (304). m，ind ), and calculate the second parameter noise data (p_noise, v) of the second channel (303) of the multi-channel signal (320). s，ind The coherence calculator (320) is used to calculate coherence data (404, c) indicating the coherence between the first channel (301, 201) and the second channel (303, 203) in the inactive frame (308); and the output interface (310) is used to generate an encoded multi-channel audio signal (232) having encoded audio data of the active frame (306) and first parametric noise data (p_noise, v_v) of the inactive frame (308). m，ind ), second parameter noise data (p_noise, v s，ind ), and / or a first linear combination of the first parameter noise data and the second parameter noise data and a second linear combination of the first parameter noise data and the second parameter noise data, and coherence data (c, 404).

Description

[0001] manual

[0002] This invention relates in particular to soft noise generation (CNG) for implementing discontinuous transmission (DTX) in a stereo codec. It also relates to multi-channel signal generators, audio encoders, and related methods, such as those relying on mixed noise signals. This invention can be implemented in devices, apparatuses, systems, methods, non-transitory storage units recording instructions, and encoded multi-channel audio signals, wherein when a computer (processor, controller) executes the aforementioned instructions, it can cause the computer (processor, controller) to perform a specific method.

[0003] introduce

[0004] Soft noise generators are commonly used for discontinuous transmission (DTX) of audio signals, especially those containing speech. In this mode, the audio signal is first divided into active and inactive frames by a speech activity detector (VAD). Based on the VAD results, only active speech frames are encoded and transmitted at the nominal bit rate. During long pauses where only background noise exists, the bit rate is reduced or reduced to zero, and the background noise is parameterized using silence insertion descriptor frames (SID frames), thereby significantly reducing the average bit rate.

[0005] Noise is generated by a soft noise generator (CNG) during inactive frames at the decoder end, and the size of the SID frame is practically very limited. Therefore, the number of parameters describing the background noise must be as small as possible. To achieve this, noise estimation is not applied directly to the output of the spectral transform; instead, it is applied at a lower spectral resolution by averaging the input power spectrum across band groups (e.g., following the Barkscale). Averaging can be achieved through arithmetic or geometric methods. Unfortunately, the limited number of parameters transmitted in the SID frame does not allow for capturing the fine spectral structure of the background noise; therefore, the CNG can only reproduce a smoothed spectral packet of noise. When the VAD triggers a CNG frame, the difference between the reconstructed smoothed spectrum of the soft noise and the spectrum of the actual background noise becomes very noticeable at the transition between the active frame (involving the regular encoding and decoding of the noisy speech portion of the signal) and the CNG frame.

[0006] Some typical CNG techniques can be found in 3GPP specifications such as ITU-T Recommendations G.729B[1], G.729.1C[2], G.718[3], or AMR[4] and AMR-WB[5]. All of these techniques generate soft noise (CN) by using linear prediction (LP) analysis / synthesis methods.

[0007] To further reduce the transmission rate, the 3GPP telecom codec for Enhanced Voice Service (EVS) in LTE[6] is equipped with a Discontinuous Transmission (DTX) mode for applying Soft Noise Generation (CNG) to inactive frames, which are defined as frames consisting only of background noise. For these frames, the low rate parameter of the signal indicates that at most 8 frames (160 milliseconds) are transmitted by a Silence Insertion Descriptor (SID) frame, which allows the CNG in the decoder to generate an artificial noise signal similar to the actual background noise. In EVS, CNG can be implemented using either a linear prediction scheme (LP-CNG) or a frequency domain scheme (FD-CNG), depending on the spectral characteristics of the background noise.

[0008] The LP-CNG method in EVS[7] operates on a segmented basis, with encoding including low-band and high-band analysis / synthesis coding stages. Unlike low-band coding, no parametric modeling of the high-band noise spectrum is performed on the high-band signal. Only the energy of the high-band signal is encoded and transmitted to the decoder, and the high-band noise spectrum is generated purely on the decoder side. Both low-band and high-band CN are synthesized by filtering the excitation through a synthesis filter. The low-band excitation is derived from the received low-band excitation energy and low-band excitation frequency packet. The low-band synthesis filter is derived from the LP parameters in the form of the received line spectrum frequency (LSF) coefficients. The high-band excitation is obtained using the energy extrapolated from the low-band energy, and the high-band synthesis filter is derived by interpolating the LSF from the decoder side. The high-band synthesis is flipped in the spectrum and added to the low-band synthesis to form the final CN signal.

[0009] The FD-CNG method [8], [9] uses a frequency domain noise estimation algorithm and then performs vector quantization on the smooth spectral packets of background noise. The decoded packets are refined in the decoder by running a second frequency domain noise estimator. Since pure parametric representation is used during inactive frames, the decoder cannot obtain the noise signal in this case. In FD-CNG, noise estimation is performed in each frame (active and inactive) at both the encoder and decoder ends based on a minimum statistics algorithm.

[0010] A method for generating soft noise in the case of two (or more) channels is described in

[10] . In

[10] , a system for stereo DTX and CNG is described, which combines a mono SID with a per-band coherence metric calculated on the two input stereo channels in the encoder. At the decoder, mono CNG information and coherence values ​​are decoded from the bitstream, and the target coherence in multiple bands is synthesized. To reduce the bit rate of the resulting stereo SID frame, the coherence values ​​are encoded using a prediction scheme, followed by entropy coding with a variable bit rate. Soft noise is generated for each channel using the method described in the preceding paragraphs, and then the two CNGs are band-mixed using a formula weighted based on the transmission band coherence values ​​contained in the SID frame.

[0011] Motivation / Disadvantages of existing technology

[0012] In stereo systems, generating background noise in isolation results in completely unrelated noise that sounds unpleasant and is very different from actual background noise. This causes abrupt, audible transitions when switching to Active Mode background or from Active Mode background to DTX Mode background. Furthermore, it is impossible to preserve a stereo image of the background using only two completely unrelated noise sources. Finally, if there is a background noise source and the speaker moves around the source with a handheld device, the spatial image of the background noise will change over time, which cannot be replicated when reconstructing the background noise independently for each channel. Therefore, a new method is needed to address the problems of stereo signals.

[0013] This is also addressed in

[10] , however, in the embodiments, inserting a common noise source for both channels to mimic related noise to generate the final soft noise plays an important role in mimicking stereo background noise recording.

[0014] Current communication voice codecs typically encode only mono signals; therefore, most existing DTX systems are designed for mono CNG. Simply applying DTX operations independently to the two channels of a stereo signal seems straightforward, but it presents several problems. First, the method requires transmitting two sets of parameters describing two background noise signals in both channels. This increases the data rate required for SID frame transmission, thus reducing the benefits of reduced network load. Another problematic aspect lies in the VAD decision, which must be synchronized between channels to avoid quirks and distortions in the spatial image of the stereo signal and optimize for a reduction in the system's bit rate. Furthermore, when CNG is applied independently to the two channels at the receiver, the two independent CNG algorithms often produce two random noise signals with zero or very low coherence, resulting in a very wide stereo image within the generated soft noise. On the other hand, applying only the noise generator and using the same soft noise signal in both channels results in very high coherence and a very narrow stereo image. However, for most stereo signals, the stereo image and its spatial impression will fall somewhere between these two extremes. Therefore, switching to the active frame or from the active frame to DTX mode introduces abrupt audible transitions. Furthermore, if a background noise source is present and the speaker moves around the source with a handheld device, the spatial image of the background noise will change over time, which cannot be reproduced when reconstructing the background noise independently for each channel. Therefore, a new approach is needed to address the problems with stereo signals.

[0015] The system described in

[10] addresses these issues by transmitting information about the mono CNG and parameter values ​​for resynthesizing the stereo image of the background noise in the decoder. This type of DTX system is well-suited for parametric stereo encoders that apply downmixing to both input channels before encoding and transmission, from which the mono CNG parameters can be derived. However, in discrete stereo coding schemes, both channels are typically still encoded jointly, and upmixing parameters such as fine-grained coherence metrics are usually not derived. Therefore, a different approach is needed for these types of stereo encoders.

[0016] Aspects of the present invention

[0017] This example demonstrates efficient transmission of stereo speech signals. Transmitting stereo signals improves user experience and speech clarity compared to transmitting only one audio channel (mono), especially when background noise or other sounds are added. Stereo signals can be encoded parametrically, where a mono downmix of both stereo channels is applied, and this single downmixed channel is encoded and transmitted to the receiver along with auxiliary information used to approximate the original stereo signal in the decoder. Another approach is to use discrete stereo coding, which aims to remove redundancy between channels through some signal preprocessing to achieve a more compact two-channel representation of the original signal. The two processed channels are then encoded and transmitted. At the decoder, inverse processing is applied. Nevertheless, auxiliary information related to stereo processing can be transmitted along both channels. Therefore, the main difference between parametric and discrete stereo coding methods lies in the number of channels transmitted.

[0018] Typically, in a conversation, there are periods when not all speakers are actively engaged. Therefore, the signal input to the speech encoder during these periods consists primarily of background noise or (near) silence. To conserve data rate and reduce network load, the speech encoder attempts to distinguish between frames containing speech (active frames) and frames primarily containing background noise or silence (inactive frames). For inactive frames, the data rate can be significantly reduced by deriving a parameterized low-bit-rate description of the current background noise as a Silence Insertion Descriptor (SID) frame, instead of encoding the audio signal as in active frames. This SID frame is periodically transmitted to the decoder to update the parameters describing the background noise, while for inactive frames in between, the bit rate is reduced, or even no information is transmitted. In the decoder, the background noise is reconstructed using the parameters transmitted in the SID frame via a Soft Noise Generation (CNG) algorithm. In this way, the transmission rate of inactive frames can be reduced to zero without the user interpreting it as a connection interruption or termination.

[0019] We describe a DTX system for discretely encoding stereo signals, comprising stereo SID and a CNG method. The CNG method generates stereo soft noise by modeling the spectral characteristics of background noise in the two channels and the degree of correlation between them, while maintaining an average bit rate comparable to that of mono applications.

[0020] Overview

[0021] According to one aspect, a multichannel signal generator is provided for generating multichannel signals having a first channel and a second channel, comprising:

[0022] The first audio source is used to generate the first audio signal;

[0023] A second audio source is used to generate a second audio signal;

[0024] A mixed noise source, used to generate mixed noise signals; and

[0025] A mixer for mixing a mixed noise signal with a first audio signal to obtain a first channel, and for mixing a mixed noise signal with a second audio signal to obtain a second channel.

[0026] According to one aspect, the first audio source is a first noise source and the first audio signal is a first noise signal, or the second audio source is a second noise source and the second audio signal is a second noise signal.

[0027] The first noise source or the second noise source is configured to generate a first noise signal or a second noise signal, such that the first noise signal or the second noise signal is decorrelated to the mixed noise signal.

[0028] According to one aspect, the mixer is configured to generate a first channel and a second channel such that the amount of the mixed noise signal in the first channel is equal to the amount of the mixed noise signal in the second channel, or is in the range of 80% to 120% of the amount of the mixed noise signal in the second channel.

[0029] According to one aspect, the mixer includes a control input for receiving control parameters, wherein the mixer is configured to control the amount of the mixed noise signal in the first channel and the second channel based on the control parameters.

[0030] According to one aspect, the first audio source, the second audio source, and the mixed audio source are Gaussian noise sources.

[0031] According to one aspect, the first audio source includes a first noise generator for generating a first audio signal as a first noise signal, wherein the second audio source includes a decorrelation unit for decorrelating the first noise signal to generate a second audio signal as a second noise signal, and wherein the mixed noise source includes the second noise generator, or

[0032] The first audio source includes a first noise generator for generating a first audio signal as a first noise signal; the second audio source includes a second noise generator for generating a second audio signal as a second noise signal; and the mixed noise source includes a decorrelation unit for decorrelating either the first noise signal or the second noise signal to generate a mixed noise signal.

[0033] The first audio source, the second audio source, and the mixed noise source each include a noise generator for generating a noise signal; another of the first audio source, the second audio source, and the mixed noise source includes a first decorrelation unit for decorrelating the noise signal; and yet another of the first audio source, the second audio source, and the mixed noise source includes a second decorrelation unit for decorrelating the noise signal. The first decorrelation unit and the second decorrelation unit are different from each other, such that their output signals are decorrelated.

[0034] The first audio source includes a first noise generator, the second audio source includes a second noise generator, and the mixed noise source includes a third noise generator. The first noise generator, the second noise generator, and the third noise generator are configured to generate mutually decorrelated noise signals.

[0035] According to one aspect, one of the first audio source, the second audio source, and the mixed noise source includes a pseudo-random number sequence generator for generating a pseudo-random number sequence based on a seed, and wherein at least two of the first audio source, the second audio source, and the mixed noise source are configured to initialize the pseudo-random number sequence generator using different seeds.

[0036] According to one aspect, one of the first audio source, the second audio source, and the mixed noise source is configured to operate using a pre-stored noise table, or

[0037] One of the first audio source, the second audio source, and the mixed noise source is configured to generate a complex spectrum for a frame, using the first noise value as the real part and the second noise value as the imaginary part.

[0038] Optionally, at least one noise generator is configured to generate complex noise spectrum values ​​for a frequency bin k, using a first random value at index k as one of the real and imaginary parts, and a second random value at index (k+M) as the other of the real and imaginary parts, wherein the first and second noise values ​​are included in a noise array, for example derived from a random number sequence generator, a noise table, or a noise process, ranging from a start index to an end index, wherein the start index is less than M, and the end index is equal to or less than 2M, where M and k are integers.

[0039] According to one aspect, the mixer includes:

[0040] A first amplitude element is used to influence the amplitude of a first audio signal;

[0041] A first adder is used to add at least a portion of the output signal of the first amplitude element and the mixed noise signal;

[0042] The second amplitude element is used to influence the amplitude of the second audio signal;

[0043] The second adder is used to add the output of the second amplitude element to at least a portion of the mixed noise signal.

[0044] Wherein, the influence amount performed by the first amplitude element is equal to the influence amount performed by the second amplitude element, or the difference between the influence amount performed by the second amplitude element and the influence amount performed by the first amplitude element is less than 20% of the influence amount performed by the first amplitude element.

[0045] According to one aspect, the mixer includes a third amplitude element for influencing the amplitude of the mixed noise signal.

[0046] The influence of the third amplitude element depends on the influence of the first or second amplitude element, such that when the influence of the first or second amplitude element decreases, the influence of the third amplitude element increases.

[0047] According to one aspect, the influence quantity performed by the third amplitude element is the value c. q The square root of the effect of the first amplitude element and the effect of the second amplitude element are 1 and c. q The square root of the difference between them.

[0048] According to one aspect, the input interface is used to receive encoded audio data in a frame sequence, the frame sequence including active frames and inactive frames following the active frames; and

[0049] The audio decoder is used to decode the encoded audio data of the active frame to generate the decoded multichannel signal of the active frame.

[0050] The first audio source, the second audio source, the mixed noise source, and the mixer are active in the inactive frame to generate the multi-channel signal for the inactive frame.

[0051] According to one aspect, the coded audio signal of the active frame has a first plurality of coefficients describing a first number of frequency windows; and

[0052] The encoded audio signal of inactive frames has a second set of coefficients describing a second number of frequency windows.

[0053] The first number of frequency windows is larger than the second number of frequency windows.

[0054] According to one aspect, the encoded audio data of the inactive frame includes silence insertion descriptor data, which includes soft noise data, indicating the signal energy of each of the two channels, or each of the first linear combination of the first and second channels and the second linear combination of the first and second channels, for the inactive frame, and indicating the coherence between the first and second channels in the inactive frame, and

[0055] The mixer is configured to mix a mixed noise signal and a first audio signal or a second audio signal based on soft noise data indicating coherence.

[0056] The multi-channel signal generator also includes a signal modifier for modifying the first and second channels, or the first audio signal, or the second audio signal, or the mixed noise signal. The signal modifier is configured to control the signal energy of the first and second audio channels, or the signal energy of a first linear combination of the first and second audio channels and a second linear combination of the first and second audio channels, using soft noise data control.

[0057] According to one aspect, the audio data of inactive frames includes:

[0058] A first mute insert descriptor frame for the first channel and a second mute insert descriptor frame for the second channel, wherein the first mute insert descriptor frame includes:

[0059] Soft noise parameter data for the first channel and / or for the first linear combination of the first and second channels, and

[0060] Used for soft noise generation auxiliary information for the first and second channels, and

[0061] The second mute insertion descriptor frame includes:

[0062] Soft noise parameter data for the second channel and / or for the second linear combination of the first and second channels, and

[0063] Coherence information indicating the coherence between the first and second channels of inactive frames, and

[0064] The multi-channel signal generator includes a controller for controlling the generation of multi-channel signals in inactive frames, using soft noise generation auxiliary information from a first silence insert descriptor frame to determine soft noise generation modes for the first and second channels, and / or for a first linear combination of the first and second channels and a second linear combination of the first and second channels, using coherence information from a second silence insert descriptor frame to set the coherence between the first and second channels in inactive frames, and using soft noise parameter data from the first silence insert descriptor frame and soft noise parameter data from the second silence insert descriptor frame to set the energy status of the first and second channels.

[0065] According to one aspect, the audio data of inactive frames includes:

[0066] At least one mute insert description frame for a first linear combination of the first and second channels and for a second linear combination of the first and second channels.

[0067] At least one mute insert descriptor frame includes:

[0068] Soft noise parameter data (p_noise) for the first linear combination of the first and second channels, and

[0069] The soft noise used for the second linear combination of the first and second channels generates auxiliary information.

[0070] The multi-channel signal generator includes a controller for controlling the generation of multi-channel signals in inactive frames using soft noise generation auxiliary information for a first linear combination of the first and second channels and a second linear combination of the first and second channels, setting the coherence between the first and second channels in inactive frames using coherence information in a second silence insert descriptor frame, and setting the energy status of the first and second channels using soft noise parameter data from at least one silence insert descriptor frame and soft noise parameter data from at least one silence insert descriptor frame.

[0071] According to one aspect, the spectrum-to-time converter is used to convert the adjusted first and second channels, which have undergone spectrum adjustment and coherence adjustment, into corresponding time-domain representations, so as to combine or connect them with the time-domain representations of the corresponding channels of the decoded multi-channel signal of the active frame.

[0072] According to one aspect, the audio data of inactive frames includes:

[0073] A mute insert descriptor frame, wherein the mute insert descriptor frame includes soft noise parameter data for the first and second channels, soft noise generation auxiliary information for the first and second channels, and / or for a first linear combination and a second linear combination of the first and second channels, and coherence information indicating the coherence between the first and second channels in an inactive frame, and

[0074] The multi-channel signal generator includes a controller for controlling the generation of multi-channel signals in inactive frames, using soft noise generation auxiliary information from the mute insert descriptor frame to determine the soft noise generation mode for the first and second channels, using coherence information from the mute insert descriptor frame to set the coherence between the first and second channels in inactive frames, and using soft noise parameter data from the mute insert descriptor frame to set the energy status of the first and second channels.

[0075] According to one aspect, the encoded audio data of the inactive frame includes silence insertion descriptor data, which includes soft noise data indicating the signal energy of each channel in the center / side representation, and coherence data indicating the coherence between the first and second channels in the left / right representation, wherein the multichannel signal generator is configured to convert the signal energy of the center / side representation in the first and second channels into the signal energy of the left / right representation.

[0076] The mixer is configured to mix a noise signal with a first audio signal and a second audio signal based on coherent data to obtain a first channel and a second channel.

[0077] The multi-channel signal generator also includes a signal modifier configured to modify the first and second channels by shaping the first and second channels based on the signal energy in the left / right domain.

[0078] According to one aspect, the multi-channel signal generator is configured to zero out the coefficients of the side channels when the audio data contains a signal indicating that the energy in the side channels is less than a predetermined threshold.

[0079] According to one aspect, the audio data of inactive frames includes:

[0080] At least one mute insert descriptor frame, wherein the at least one mute insert descriptor frame includes soft noise parameter data for the center and side channels, soft noise generation auxiliary information for the center and side channels, and coherence information indicating the coherence between the first and second channels in an inactive frame, and

[0081] The multi-channel signal generator includes a controller for controlling the generation of multi-channel signals in inactive frames, using soft noise generation auxiliary information from the mute insert descriptor frame to determine the soft noise generation mode for the first and second channels, using coherence information from the mute insert descriptor frame to set the coherence between the first and second channels in inactive frames, and using soft noise parameter data or its processed version from the mute insert descriptor frame to set the energy status of the first and second channels.

[0082] According to one aspect, the multi-channel signal generator is configured to scale the signal energy coefficients of the first and second channels using gain information encoded using soft noise parameter data of the first and second channels.

[0083] On one hand, the multi-channel signal generator is configured to convert the generated multi-channel signal from a frequency domain version to a time domain version.

[0084] According to one aspect, the first audio source is a first noise source and the first audio signal is a first noise signal, or the second audio source is a second noise source and the second audio signal is a second noise signal.

[0085] Wherein, the first noise source or the second noise source is configured to generate a first noise signal or a second noise signal, such that the first noise signal or the second noise signal is at least partially correlated, and

[0086] A mixed noise source is configured to generate a mixed noise signal having a first mixed noise component and a second mixed noise component, wherein the second mixed noise component is at least partially decorrelated with the first mixed noise component; and

[0087] The mixer is configured to mix a first mixed noise portion of a mixed noise signal with a first audio signal to obtain a first channel, and to mix a second mixed noise portion of a mixed noise signal with a second audio signal to obtain a second channel.

[0088] According to one aspect, a method for generating a multi-channel signal having a first channel and a second channel is provided, comprising:

[0089] A first audio signal is generated using a first audio source;

[0090] A second audio signal is generated using a second audio source;

[0091] Using a mixed noise source to generate a mixed noise signal; and

[0092] The mixed noise signal is mixed with a first audio signal to obtain a first channel, and the mixed noise signal is mixed with a second audio signal to obtain a second channel.

[0093] According to one aspect, an audio encoder is provided for generating encoded multichannel audio signals for a frame sequence including active frames and inactive frames, the audio encoder comprising:

[0094] An activity detector is used to analyze multi-channel signals to determine whether a frame in a frame sequence is an inactive frame;

[0095] A noise parameter calculator is used to calculate the first parameter noise data of the first channel of a multi-channel signal, and to calculate the second parameter noise data of the second channel of a multi-channel signal.

[0096] A coherence calculator is used to calculate coherence data indicating the coherence between the first and second channels in inactive frames; and

[0097] The output interface is used to generate coded multi-channel audio signals, which have coded audio data of active frames, and first parameter noise data, second parameter noise data, or a first linear combination of the first parameter noise data and the second parameter noise data, and a second linear combination of the first parameter noise data and the second parameter noise data, as well as coherence data of inactive frames.

[0098] According to one aspect, the coherence calculator is configured to calculate coherence values ​​and quantize the coherence values ​​to obtain quantized coherence values, wherein the output interface is configured to use the quantized coherence values ​​as coherence data in the encoded multichannel signal.

[0099] Based on one aspect, the coherence calculator is configured as follows:

[0100] Calculate the real and virtual intermediate values ​​from the complex spectrum values ​​of the first and second channels in the inactive frame;

[0101] Calculate the first energy value of the first channel and the second energy value of the second channel in the inactive frame; and

[0102] Coherence data are calculated using real intermediate values, imaginary intermediate values, first energy values, and second energy values, or

[0103] Smooth at least one of the real intermediate value, the imaginary intermediate value, the first energy value, and the second energy value, and use at least one smoothed value to calculate coherence data.

[0104] According to one aspect, the coherence calculator is configured to calculate the real intermediate value, which is the sum of the real parts of the product of the complex spectral values ​​of the corresponding frequency windows of the first and second channels of the inactive frame, or

[0105] Calculate the virtual intermediate value as the sum of the imaginary parts of the product of the complex spectrum values ​​of the corresponding frequency windows of the first and second channels in the inactive frame.

[0106] According to one aspect, the coherence calculator is configured to square the smoothed real intermediate value and the smoothed imaginary intermediate value, and then sum the squared values ​​to obtain the first component number.

[0107] The coherence calculator is configured to multiply a smoothed first energy value by a second energy value to obtain a second component number, and to combine the first component number with the second component number to obtain a result number of coherence values, with the coherence data based on the result number.

[0108] On one hand, the coherence calculator is configured to calculate the square root of the result number to obtain the coherence value, and the coherence data is based on the coherence value.

[0109] According to one aspect, the coherence calculator is configured to quantize the coherence values ​​using a uniform quantizer to obtain quantized coherence values, which are then used as n-bit coherence data.

[0110] According to one aspect, the output interface is configured to generate a first silence insert descriptor frame for the first channel and a second silence insert descriptor frame for the second channel, wherein the first silence insert descriptor frame includes soft noise parameter data for the first channel and soft noise generation auxiliary information for the first and second channels, and wherein the second silence insert descriptor frame includes soft noise parameter data for the second channel and coherence information indicating the coherence between the first and second channels in an inactive frame, or

[0111] The output interface is configured to generate a silence insertion descriptor frame, which includes soft noise parameter data for the first and second channels, soft noise generation auxiliary information for the first and second channels, and coherence information indicating the coherence between the first and second channels in an inactive frame, or

[0112] The output interface is configured to generate a first mute insert descriptor frame for the first channel and a second mute insert descriptor frame for the second channel. The first mute insert descriptor frame includes soft noise parameter data for the first channel and the second channel, as well as soft noise generation auxiliary information for the first channel and the second channel. The second mute insert descriptor frame includes soft noise parameter data for the first channel and the second channel, as well as coherence information indicating the coherence between the first channel and the second channel in an inactive frame.

[0113] According to one aspect, the uniform quantizer is configured to calculate n bits such that the value of n is equal to the value of the bits occupied by the soft noise generation auxiliary information of the first silent insertion descriptor frame.

[0114] Based on one aspect, the activity detector is configured to,

[0115] Analyze the first channel of a multi-channel signal to classify it as active or inactive, and

[0116] Analyze the second channel of a multi-channel signal to classify it as active or inactive, and

[0117] If both the first and second channels are classified as inactive, then the frames in the frame sequence are determined to be inactive; otherwise, the frames are determined to be active.

[0118] According to one aspect, the noise parameter calculator is configured to calculate the first gain information of the first channel and the second gain information of the second channel, and provide parameter noise data as the first gain information and the second gain information of the first channel.

[0119] According to one aspect, the noise parameter calculator is configured to convert at least some of the first parameter noise data and the second parameter noise data from a left / right representation to a center / side representation having a center channel and side channels.

[0120] According to one aspect, the noise parameter calculator is configured to reconvert the center / side representation of at least some of the first parameter noise data and the second parameter noise data into a left / right representation.

[0121] The noise parameter calculator is configured to calculate the first gain information of the first channel and the second gain information of the second channel from the reconverted left / right representation, and to provide the first gain information of the first channel included in the first parameter noise data, and the second gain information included in the second parameter noise data.

[0122] Based on one aspect, the noise parameter calculator is configured to calculate:

[0123] First gain information, obtained through comparison:

[0124] The first-parameter noise data for the first channel was reconverted from the center / side representation to a left / right representation version; with

[0125] The first parameter noise data of the first channel was converted from the previous version, from center / side representation to left / right representation; and / or

[0126] The second gain information is obtained through comparison:

[0127] The second-parameter noise data for the second channel was reconverted from the center / side representation to a left / right representation version; with

[0128] The second-parameter noise data for the second channel was converted from the previous version, from the center / side representation to the left / right representation.

[0129] According to one aspect, the noise parameter calculator is configured to compare the energy of a second linear combination between first parameter noise data and second parameter noise data with a predetermined energy threshold, and:

[0130] When the energy of the second linear combination between the first parameter noise data and the second parameter noise data exceeds a predetermined energy threshold, the coefficients of the side channel noise shape vector are set to zero; and

[0131] When the energy of the second linear combination between the first parameter noise data and the second parameter noise data is less than a predetermined energy threshold, the coefficients of the side channel noise shape vector are maintained.

[0132] According to one aspect, the audio encoder is configured to encode a second linear combination between the first parameter noise data and the second parameter noise data using fewer bits than the number of bits used to encode a first linear combination between the first parameter noise data and the second parameter noise data.

[0133] Based on one aspect, the output interface is configured as follows:

[0134] A coded multichannel audio signal with active frames is generated using a first plurality of coefficients for a first number of frequency windows; and

[0135] A second set of coefficients, used to describe a second quantity of frequency windows, are used to generate first parameter noise data, second parameter noise data, or a first linear combination of first parameter noise data and second parameter noise data, and a second linear combination of first parameter noise data and second parameter noise data.

[0136] The first number of frequency windows is larger than the second number of frequency windows.

[0137] According to one aspect, an audio encoding method is provided for generating encoded multichannel audio signals for a frame sequence including active and inactive frames, the method comprising:

[0138] Analyze the multi-channel signal to determine which frames in the frame sequence are inactive;

[0139] Calculate first parameter noise data for the first channel of the multi-channel signal, and / or for the first linear combination of the first and second channels of the multi-channel signal, and calculate second parameter noise data for the second channel of the multi-channel signal, and / or for the second linear combination of the first and second channels of the multi-channel signal.

[0140] Calculate coherence data indicating the coherence between the first and second channels in inactive frames; and

[0141] Generate an encoded multichannel audio signal. The encoded multichannel audio signal has encoded audio data of active frames, and first parameter noise data, second parameter noise data, and coherence data of inactive frames.

[0142] According to one aspect, a computer program is provided for performing the methods described above or below when running on a computer or processor.

[0143] According to one aspect, an encoded multichannel audio signal is provided, which is organized in a frame sequence, said frame sequence including active frames and inactive frames, the encoded multichannel audio signal including:

[0144] Encoded audio data of the active frame;

[0145] The first parameter noise data of the first channel in the inactive frame;

[0146] Second-parameter noise data of the second channel in inactive frames; and

[0147] Coherence data indicating the coherence between the first and second channels in an inactive frame.

[0148] According to one aspect, the first audio source is a first noise source and the first audio signal is a first noise signal, or the second audio source is a second noise source and the second audio signal is a second noise signal.

[0149] The first noise source or the second noise source is configured to generate a first noise signal or a second noise signal, such that the first noise signal or the second noise signal is decorrelated to the mixed noise signal.

[0150] According to one aspect, the mixer is configured to generate a first channel and a second channel such that the amount of the mixed noise signal in the first channel is equal to the amount of the mixed noise signal in the second channel, or is in the range of 80% to 120% of the amount of the mixed noise signal in the second channel.

[0151] According to one aspect, the mixer includes a control input for receiving control parameters, wherein the mixer is configured to control the amount of the mixed noise signal in the first channel and the second channel based on the control parameters.

[0152] According to one aspect, the first audio source, the second audio source, and the mixed audio source are Gaussian noise sources.

[0153] According to one aspect, the first audio source includes a first noise generator for generating a first audio signal as a first noise signal, and the second audio source includes a decorrelation unit for decorrelating the first noise signal to generate a second audio signal as a second noise signal, and wherein the mixed noise source includes the second noise generator, or

[0154] The first audio source includes a first noise generator for generating a first audio signal as a first noise signal; the second audio source includes a second noise generator for generating a second audio signal as a second noise signal; and the mixed noise source includes a decorrelation unit for decorrelating either the first noise signal or the second noise signal to generate a mixed noise signal.

[0155] The first audio source, the second audio source, and the mixed noise source each include a noise generator for generating a noise signal; another of the first audio source, the second audio source, and the mixed noise source includes a first decorrelation unit for decorrelating the noise signal; and yet another of the first audio source, the second audio source, and the mixed noise source includes a second decorrelation unit for decorrelating the noise signal. The first and second decorrelation units are different from each other, such that their output signals are decorrelated.

[0156] The first audio source includes a first noise generator, the second audio source includes a second noise generator, and the mixed noise source includes a third noise generator, wherein the first noise generator, the second noise generator, and the third noise generator are configured to generate mutually decorrelated noise signals.

[0157] According to one aspect, one of the first audio source, the second audio source, and the mixed noise source includes a pseudo-random number sequence generator for generating a pseudo-random number sequence based on a seed, and

[0158] At least two of the first audio source, the second audio source, and the mixed noise source are configured to initialize the pseudo-random number sequence generator with different seeds.

[0159] According to one aspect, one of the first audio source, the second audio source, and the mixed noise source is configured to operate using a pre-stored noise table, or

[0160] At least one of the first audio source, the second audio source, and the mixed noise source is configured to generate a complex spectrum for a frame, using a first noise value as the real part and a second noise value as the imaginary part.

[0161] Optionally, at least one noise generator is configured to generate complex noise spectrum values ​​for frequency window k, using a first random value at index k as one of the real and imaginary parts, and using a second random value at index (k+M) as the other of the real and imaginary parts.

[0162] The first noise value and the second noise value are included in the noise array, for example, derived from a random number sequence generator, a noise table, or a noise process, and range from the start index to the end index, where the start index is less than M and the end index is equal to or less than 2M, and M and k are integers.

[0163] According to one aspect, the mixer includes:

[0164] A first amplitude element is used to influence the amplitude of a first audio signal;

[0165] A first adder is used to add at least a portion of the output signal of the first amplitude element and the mixed noise signal;

[0166] The second amplitude element is used to influence the amplitude of the second audio signal;

[0167] The second adder is used to add the output of the second amplitude element to at least a portion of the mixed noise signal.

[0168] The influence amount exerted by the first amplitude element is equal to the influence amount exerted by the second amplitude element, or the difference between them is less than 20% of the influence amount exerted by the first amplitude element.

[0169] According to one aspect, the mixer includes a third amplitude element for influencing the amplitude of the mixed noise signal, wherein the amount of influence performed by the third amplitude element depends on the amount of influence performed by the first amplitude element or the second amplitude element, such that when the amount of influence performed by the first amplitude element or the second amplitude element decreases, the amount of influence performed by the third amplitude element increases.

[0170] According to one aspect, the multi-channel signal generator further includes:

[0171] An input interface is used to receive encoded audio data from a frame sequence, the frame sequence including the active frame and the inactive frames following the active frame; and

[0172] An audio decoder is used to decode the coded audio data of the active frame to generate the decoded multichannel signal of the active frame.

[0173] The first audio source, the second audio source, the mixed noise source, and the mixer are active in the inactive frame to generate the multi-channel signal for the inactive frame.

[0174] According to one aspect, the encoded audio data of the inactive frame includes silence insertion descriptor data, which includes soft noise data, indicating the signal energy for each of the two channels in the inactive frame, and indicating the coherence between the first and second channels in the inactive frame, and

[0175] The mixer is configured to mix a mixed noise signal and a first audio signal or a second audio signal based on soft noise data indicating coherence.

[0176] The multi-channel signal generator also includes a signal modifier for modifying the first and second channels, or the first audio signal, or the second audio signal, or a mixed noise signal.

[0177] The signal modifier is configured to be controlled by soft noise data indicating the signal energy of the first and second audio channels.

[0178] According to one aspect, the audio data of inactive frames includes:

[0179] A first mute insert descriptor frame for the first channel and a second mute insert descriptor frame for the second channel, wherein the first mute insert descriptor frame includes soft noise parameter data for the first channel and soft noise generation auxiliary information for the first and second channels, wherein the second mute insert descriptor frame includes soft noise parameter data for the second channel and coherence information indicating the coherence between the first and second channels in an inactive frame, and

[0180] The multi-channel signal generator includes a controller for controlling the generation of multi-channel signals in inactive frames, using soft noise generation auxiliary information from a first silence insert descriptor frame to determine soft noise generation modes for the first and second channels, using coherence information from a second silence insert descriptor frame to set the coherence between the first and second channels in inactive frames, and using soft noise parameter data from the first silence insert descriptor frame and soft noise parameter data from the second silence insert descriptor frame to set the energy conditions of the first and second channels.

[0181] According to one aspect, it also includes a spectrum-to-time converter for converting the adjusted first and second channels, which have undergone spectrum and coherence adjustments, into corresponding time-domain representations, to be combined or connected with the time-domain representations of the corresponding channels of the decoded multichannel signal of the active frame.

[0182] According to one aspect, the audio data of inactive frames includes:

[0183] A mute insert descriptor frame, wherein the mute insert descriptor frame includes soft noise parameter data for the first and second channels, soft noise generation auxiliary information for the first and second channels, and coherence information indicating the coherence between the first and second channels in an inactive frame, and

[0184] The multi-channel signal generator includes a controller for controlling the generation of multi-channel signals in inactive frames, using soft noise generation auxiliary information from the mute insert descriptor frame to determine the soft noise generation mode for the first and second channels, using coherence information from the second mute insert descriptor frame to set the coherence between the first and second channels in inactive frames, and using soft noise parameter data from the mute insert descriptor frame to set the energy status of the first and second channels.

[0185] According to one aspect, the first audio source is a first noise source and the first audio signal is a first noise signal, or the second audio source is a second noise source and the second audio signal is a second noise signal.

[0186] Wherein, the first noise source or the second noise source is configured to generate a first noise signal or a second noise signal, such that the first noise signal or the second noise signal is at least partially correlated, and

[0187] The mixed noise source is configured to generate a mixed noise signal having a first mixed noise component and a second mixed noise component, wherein the second mixed noise component is at least partially decorrelated with the first mixed noise component; and

[0188] The mixer is configured to mix a first mixed noise portion of the mixed noise signal with a first audio signal to obtain a first channel, and to mix a second mixed noise portion of the mixed noise signal with a second audio signal to obtain a second channel.

[0189] According to one aspect, a method for generating a multi-channel signal having a first channel and a second channel includes:

[0190] A first audio signal is generated using a first audio source;

[0191] A second audio signal is generated using a second audio source;

[0192] Using a mixed noise source to generate a mixed noise signal; and

[0193] The mixed noise signal is mixed with a first audio signal to obtain a first channel, and the mixed noise signal is mixed with a second audio signal to obtain a second channel.

[0194] According to one aspect, an audio encoder is provided for generating encoded multichannel audio signals for a frame sequence including active frames and inactive frames, the audio encoder comprising:

[0195] An activity detector is used to analyze multi-channel signals to determine whether a frame in a frame sequence is an inactive frame;

[0196] The noise parameter calculator is used to calculate the first parameter noise data of the first channel of a multi-channel signal and to calculate the second parameter noise data of the second channel of a multi-channel signal.

[0197] A coherence calculator is used to calculate coherence data indicating the coherence between the first and second channels in inactive frames; and

[0198] The output interface is used to generate coded multi-channel audio signals, which have coded audio data of active frames, and first parameter noise data, second parameter noise data, and coherence data of inactive frames.

[0199] According to one aspect, the coherence calculator is configured to calculate coherence values ​​and quantize the coherence values ​​to obtain quantized coherence values, wherein the output interface is configured to use the quantized coherence values ​​as coherence data in the encoded multichannel signal.

[0200] Based on one aspect, the coherence calculator is configured as follows:

[0201] Calculate the real and virtual intermediate values ​​from the complex spectrum values ​​of the first and second channels in the inactive frame;

[0202] Calculate the first energy value of the first channel and the second energy value of the second channel in the inactive frame; and

[0203] Coherence data are calculated using real intermediate values, imaginary intermediate values, first energy values, and second energy values, or

[0204] Smooth at least one of the real intermediate value, the imaginary intermediate value, the first energy value, and the second energy value, and use at least one smoothed value to calculate coherence data.

[0205] According to one aspect, the coherence calculator is configured to calculate the real intermediate value, which is the sum of the real parts of the product of the complex spectral values ​​of the corresponding frequency windows of the first and second channels of the inactive frame, or

[0206] Calculate the virtual intermediate value as the sum of the imaginary parts of the product of the complex spectrum values ​​of the corresponding frequency windows of the first and second channels in the inactive frame.

[0207] According to one aspect, the coherence calculator is configured to square the smoothed real intermediate value and the smoothed imaginary intermediate value, and then sum the squared values ​​to obtain the first component number.

[0208] The coherence calculator is configured to multiply the smoothed first and second energy values ​​to obtain a second component number, and combine the first component number and the second component number to obtain a result number of coherence values, with the coherence data based on the result number.

[0209] According to one aspect, an audio encoder is provided, wherein a coherence calculator is configured to calculate the square root of the result number to obtain a coherence value, and the coherence data is based on the coherence value.

[0210] According to one aspect, the coherence calculator is configured to quantize the coherence values ​​using a uniform quantizer to obtain quantized coherence values, which are then used as N-bit coherence data.

[0211] Based on one aspect, an audio encoder is provided.

[0212] The output interface is configured to generate a first silence insert descriptor frame for the first channel and a second silence insert descriptor frame for the second channel. The first silence insert descriptor frame includes soft noise parameter data for the first channel and soft noise generation auxiliary information for both the first and second channels. The second silence insert descriptor frame includes soft noise parameter data for the second channel and coherence information indicating the coherence between the first and second channels in inactive frames.

[0213] The output interface is configured to generate a mute insert descriptor frame, which includes soft noise parameter data for the first and second channels, soft noise generation auxiliary information for the first and second channels, and coherence information indicating the coherence between the first and second channels in an inactive frame.

[0214] According to one aspect, the uniform quantizer is configured to calculate N bits such that the value of N is equal to the value of the bits occupied by the soft noise generation auxiliary information of the first silent insertion descriptor frame.

[0215] According to one aspect, an audio coding method for generating encoded multichannel audio signals for a frame sequence including active and inactive frames, the method comprising:

[0216] Analyze the multi-channel signal to determine which frames in the frame sequence are inactive;

[0217] Calculate the first parameter noise data for the first channel of the multi-channel signal, and calculate the second parameter noise data for the second channel of the multi-channel signal.

[0218] Calculate coherence data indicating the coherence between the first and second channels in an inactive frame; and generate an encoded multichannel audio signal having encoded audio data of the active frame, and first parameter noise data, second parameter noise data, and coherence data of the inactive frame.

[0219] According to one aspect, the encoded multi-channel audio signal is organized in a frame sequence, which includes active frames and inactive frames. The encoded multi-channel audio signal includes:

[0220] Encoded audio data of the active frame;

[0221] The first parameter noise data of the first channel in the inactive frame;

[0222] Second-parameter noise data of the second channel in inactive frames; and

[0223] Coherence data indicating the coherence between the first and second channels in an inactive frame.

[0224] Attached Figure

[0225] Figure 1 Show examples of encoders, specifically classifying frames as active or inactive.

[0226] Figure 2 Show examples of encoders and decoders.

[0227] Figures 3a to 3f This shows an example of a multichannel signal generator that can be used in a decoder.

[0228] Figure 4 Show examples of encoders and decoders.

[0229] Figure 5 This shows an example of the noise parameter quantization stage.

[0230] Figure 6 This shows an example of the noise parameter dequantization stage.

[0231] Some aspects that can be implemented in the example

[0232] In this specification, we specifically describe a new technique, such as DTX and CNG for discretely encoded stereo signals. Instead of operating a mono downmixing of the stereo signal, we derive, jointly encode, and transmit noise parameters for both channels. In the decoder (or more generally in a multichannel generator), three independent soft noise signals can be mixed based on a single wideband inter-channel coherence value, which, for example, is transmitted along with two sets of noise parameters. Some aspects of the examples may encompass at least one of the following aspects in some examples:

[0233] The CNG in the decoder, for example, is generated by mixing three independent noise signals. After decoding the stereo SID and reconstructing the noise parameters of the left and right channels, two noise signals may be generated, for example, as a mixture of correlated and uncorrelated noise. For this purpose, a common noise source for both channels (used as a correlated noise source) and two separate noise sources (providing uncorrelated noise) can be mixed together. The mixing process can be controlled by the inter-channel coherence value transmitted in the stereo SID. After mixing, the two mixed noise signals are spectrally shaped separately using the reconstructed noise parameters of the left and right channels.

[0234] The joint encoding of noise parameters can be derived from both channels of the stereo signal. To maintain a low bit rate for the stereo SID, the noise parameters can be further compressed before being encoded into the stereo SID. This can be achieved, for example, by converting the left / right channel representation of the noise parameters to a center / side representation and encoding the side noise parameters with fewer bits than the center noise parameters.

[0235] • SID (Stereo SID) for two-channel DTX. This SID may contain the noise parameters of the two channels of the stereo signal, as well as the coherence value between a single broadband channel and a flag indicating equal noise parameters between the two channels.

[0236] The examples shown below in this specification can be implemented in apparatuses, devices, systems, methods, controllers, and non-transitory storage units that store instructions, which, when executed by a processor, cause the processor to perform the techniques (e.g., methods, such as sequences of operations) described in this specification.

[0237] In particular, at least one of the following blocks can be controlled by the controller.

[0238] Example

[0239] Before discussing the various aspects of this example in detail, let's quickly outline some of the most important ones:

[0240] 1) Figures 3a-3f This illustrates an example of a multichannel signal generator (e.g., formed by at least one first signal or channel and one second audio signal or channel) used to generate multichannel audio signals (e.g., at a decoder). The multichannel audio signal (initially in the form of multiple decorrelated channels) may be affected (e.g., scaled) by an amplitude element. The amount of influence can be based on coherence data between the first and second audio signals estimated at the encoder. The first and second audio signals can be mixed with a co-mixed signal (which can also be decorrelated and affected (e.g., scaled) by coherence data). The amount of influence on the mixed signal can be such that when the mixed signal is scaled with low weights (e.g., 0 or greater than but, for example, close to 0), the first and second audio signals are scaled with high weights (e.g., 1 or less than but, for example, close to 1), and vice versa. The amount of influence on the mixed signal can be such that high coherence measured at the encoder causes the first and second audio signals to be scaled with low weights (e.g., 0 or greater than but, for example, close to 0), and high coherence measured at the encoder causes the first and second audio signals to be scaled with high weights (e.g., 1 or less than but, for example, close to 1). Figures 3a-3f The technology shown can be used to implement soft noise generators (CNG).

[0241] 2) Figure 1 , 2 And 4 shows an example of an encoder. The encoder can classify audio frames as active or inactive. If the audio frame is inactive, only some parametric noise data is encoded in the bitstream (e.g., providing the parametric noise shape, which gives a parametric representation of the noise shape without providing the noise signal itself), and coherence data between the two channels can also be provided.

[0242] 3) Figure 2and 4 This shows an example of a decoder. The decoder can generate audio signals (with soft noise), for example, via:

[0243] a. Use as Figures 3a-3f One of the techniques shown (point 1 above) (especially considering the coherence value provided by the encoder and applied as a weight to the amplitude element); and

[0244] b. Shape the generated audio signal (soft noise) using parametric noise data encoded in the bitstream.

[0245] It is worth noting that the encoder does not need to provide the complete audio signal for inactive frames, but only provides a parameter representation of the coherence value and the noise shape, thereby reducing the number of bits to be encoded in the bitstream.

[0246] Signal generator (e.g., decoder side), CNG

[0247] Figures 3a-3f An example of CNG is shown, or more generally, a multi-channel signal generator 200 for generating a multi-channel signal 204 having a first channel 201 and a second channel 203 (in this specification, the generated audio signals 221 and 223 are considered noise, but may also be different types of signals that are not noise). First refer to... Figure 3f This shows a general example, while Figures 3a-3e Then a specific example will be displayed.

[0248] The first audio source 211 can be a first noise source, which may be indicated to generate a first audio signal 221, which may be a first noise signal. The mixed noise source 212 can generate a mixed noise signal 222. The second audio source 213 can generate a second audio signal 223, which may be a second noise signal. The multichannel signal generator 200 can mix the first audio signal (first noise signal) 221 with the mixed noise signal 222, and mix the second audio signal (second noise signal) 223 with the mixed noise signal 222 (alternatively or alternatively, the first audio signal 221 can be mixed with version 221a of the mixed noise signal 222, and the second audio signal 223 can be mixed with version 221b of the mixed noise signal 222, wherein versions 221a and 221b can be different, for example, differing from each other by 20%; each of versions 221a and 221b can be, for example, an amplified and / or downscaled version of the common signal 222). Therefore, the first channel 201 of the multi-channel signal 204 can be obtained from the first audio signal (first noise signal) 221 and the mixed noise signal 222. Similarly, the second channel 203 of the multi-channel signal 204 can be obtained by mixing the mixed noise signal 222 with the second audio signal 223. Note that the signals here can be in the frequency domain, and k represents a specific index or coefficient (associated with a specific frequency window).

[0249] from Figures 3a-3f As can be seen, the first audio signal 221, the mixed noise signal 222, and the second audio signal 223 can be decorrelated with each other. This can be achieved, for example, by decorrelateding the same signal (e.g., at a decorrelator) and / or by generating noise independently (as in the examples provided below).

[0250] Mixer 208 can be implemented to mix the first audio signal 221 and the second audio signal 223 with the mixed noise signal 222. This mixing can be of the type where the first audio signal 221, the mixed noise signal 222, and the second audio signal 223 are weighted by scaling (e.g., at amplitude elements 208-1, 208-2, 208-3) and then the signals are summed (e.g., at adder stages 206-1 and 206-3). The mixing is of the "weighted and then added together" type. Figures 3a-3f This demonstrates the actual signal processing used to generate the noise signal N. l [k] and N r [k], where the adder (+) element represents the sampled addition of two signals (k is the index of the frequency window).

[0251] Amplitude elements (or weighting elements, or scaling elements) 208-1, 208-2, and 208-3 can be obtained, for example, by scaling the first audio signal 221, the mixed noise signal 222, and the second audio signal 223 using appropriate coefficients, and can output a weighted version 221' of the first audio signal 221, a weighted version 222' of the mixed noise signal 222, and a weighted version 223' of the second audio signal 223. Appropriate coefficients can be sqrt(coh) and sqrt(1-coh), and can be obtained, for example, from coherence information encoded in a signaling-specific descriptor frame (also seen below) (sqrt here refers to the square root operation). The coherence "coh" will be discussed in detail below, and can be, for example, represented by "c" or "c" below. ind "or "c q The meaning, for example, is encoded in the coherence information 404 of bitstream 232 (see below, in conjunction with...). Figure 2 and 4 It is worth noting that the mixed noise signal 222 can be scaled, for example, by weighting the square root of the coherence value, while the first audio signal 221 and the second audio signal 222 can be scaled by weighting the square root of the complementary values ​​of coherence coh and 1. However, the mixed noise signal 222 can be considered a common-mode signal, a portion of which is mixed into a weighted version 221' of the first audio signal 221 and a weighted version 223' of the second audio signal 223 to obtain the first channel 201 and the second channel 203 of the multi-channel signal 204, respectively. In some cases, the first noise source 211 or the second noise source 213 can be configured to generate the first noise signal 221 or the second noise signal 223 such that the first noise signal 221 and / or the second noise signal 223 are decorrelated with the mixed noise signal 222 (see the following references). Figures 3b-3e (The narrative).

[0252] At least one (or each) of the first audio source 211, the second audio source 213, and the mixed noise source 212 can be a Gaussian noise source.

[0253] In example 3a, the first audio source 211 (referred to herein as 211a) may include or be connected to the first noise generator, and the second audio source 213 (213a) may include or be connected to the second noise generator. The mixed noise source 212 (212a) may include or be connected to the third noise generator. The first noise generator 211 (211a), the second noise generator 213 (213a), and the third noise generator 212 (212a) may generate mutually decorrelated noise signals.

[0254] In the example, at least one of the first audio source 211 (211a), the second audio source 213 (213a), and the mixed noise source 212 (212a) can be operated using a pre-stored noise table, thus providing a random sequence.

[0255] In some examples, at least one of the first audio source 211, the second audio source 213, and the mixed noise source 212 can generate a complex spectrum for the frame, using a first noise value as the real part and a second noise value as the imaginary part. Optionally, at least one noise generator can generate complex noise spectrum values ​​(e.g., coefficients) for a frequency window k, using a first random value at index k as one of the real and imaginary parts, and a second random value at index (k+M) as the other of the real and imaginary parts. The first and second noise values ​​can be included in a noise array, for example derived from a random number sequence generator, a noise table, or a noise process, ranging from a start index to an end index, where the start index is less than M and the end index is equal to or less than 2×M (i.e., twice M). M and k can be integers (k is the index of the specific frequency window in the frequency domain representation of the signal).

[0256] Each audio source 211, 212, 213 may include at least one audio source generator (noise generator) that generates noise, for example, according to N1[k], N2[k], N3[k].

[0257] Figures 3a-3f The multi-channel signal generator 200 can be used, for example, in decoders 200a, 200b (200'). Specifically, the multi-channel signal generator 200 can be considered as... Figure 4 This is part of the soft noise generator (CNG) 220. The decoder 200 is typically used to decode signals already encoded by the encoder, or to generate an audio signal corresponding to the original input audio signal input to the encoder by generating a signal that will be shaped by energy information obtained from the bitstream. In some examples, a classification is made between frames with speech (or typically non-empty audio signals) and silence insert descriptor frames. As explained in this specification, silence insert descriptor frames (SIDs) (also known as “inactive frames 308,” for example, they may be encoded as SID frames 241 and / or 243) are generally provided with low bit rate information and are therefore provided at a lower frequency than normal speech frames (so-called “active frames 306,” see also below). Furthermore, the information present in the silence insert descriptor frames (SIDs, inactive frames 308) is typically limited (and may substantially correspond to the energy information of the signal).

[0258] Nevertheless, it should be understood that the content of the SID frame can be supplemented with multichannel noise 204 generated by a multichannel signal generator. Essentially, audio sources 211, 212, and 213 can process signals that are independent and unrelated to each other (e.g., noise). Although the first audio signal 221, the mixed noise signal 222, and the second audio signal 223 can be scaled and inserted into the bitstream using coherence information provided by the encoder. Figures 3a-3f As can be seen, the coherence values ​​of the mixed noise signal 222 can be the same, providing a common-mode signal for the first audio signal 221 and the second audio signal 223, thus allowing the acquisition of a multi-channel signal 204 from the first channel 201 and the second channel 203. The coherence signal is typically a value between 0 and 1.

[0259] - Coherence equal to 0 means that the original first audio channel (e.g., L, 301) and the second audio channel (e.g., R, 303) are completely uncorrelated with each other, and the amplitude element 208-2 of the mixed noise signal 222 scales the mixed noise signal 222 to 0. This will result in the first audio signal 221 and the second audio signal 223 not being mixed with any common-mode signal (by being mixed with a signal that is constant to 0), and the output channels 201, 203 will be substantially the same as the first noise signal 221 and the second noise signal 223 of the multi-channel signal 204.

[0260] - Coherence equal to 1 means that the original first audio channel (e.g., L, 301) and second audio channel (e.g., R, 303) should be the same, and the amplitude elements 208-1 and 208-3 have a scaling of 0 on the input signal. Then the first and second channels are equal to the mixed noise signal 222 (which has a scaling of 1 at the amplitude element 208-2).

[0261] - Coherence between 0 and 1 will result in an intermediate mixture between the two cases mentioned above.

[0262] Now we will discuss some aspects and variations of the mixer 206 and / or CNG 220.

[0263] The first audio source (211) can be a first noise source, and the first audio signal (221) can be a first noise signal; or the second audio source (213) can be a second noise source, and the second audio signal (223) can be a second noise signal. The first noise source (211) or the second noise source (213) can be configured to generate the first noise signal (221) or the second noise signal (223) such that the first noise signal (221) or the second noise signal (223) is decorrelated with the mixed noise signal (222).

[0264] The mixer (206) can be configured to generate a first channel (201) and a second channel (203) such that the amount of the mixed noise signal (222) in the first channel (201) is equal to the amount of the mixed noise signal (222) in the second channel (203), or is in the range of 80% to 120% of the amount of the mixed noise signal (222) in the second channel (203) (e.g., its portions 221a and 221b are different from each other and different from the original mixed noise signal 222 in the range of 80% to 120%).

[0265] In some cases,

[0266] The influence exerted by the first amplitude element (208-1) and the influence exerted by the second amplitude element (208-3) are equal to each other (e.g., when there is no difference between parts 221a and 221b), or

[0267] The difference between the influence amount performed by the second amplitude element (208-3) and the influence amount performed by the first amplitude element (208-1) is less than 20% of the influence amount performed by the first amplitude element (208-1) (e.g., when the difference between parts 221a and 221b is less than 20%).

[0268] The mixer (206) and / or CNG 220 may include control inputs for receiving control parameters (404, c). Thus, the mixer (206) may be configured to control the amount of mixed noise signal (222) in the first channel (201) and the second channel (203) in response to the control parameters (404, c).

[0269] exist Figures 3a-3f The mixed noise signal 222 is shown to undergo the coefficient sqrt(coh), and the first signal 221 and the second audio signal 223 are shown to undergo the coefficient sqrt(1-coh).

[0270] As mentioned above, Figure 3a The CNG 220a is shown, in which the first source 211a (211), the second source 213a (213), and the mixed noise source 212a (212) include different generators. This is not absolutely necessary and there can be many variations.

[0271] More generally:

[0272] 1. The first variant of CNG 220b (e.g.) Figure 3b ):

[0273] a. The first audio source 211b (211) may include a first noise generator for generating a first audio signal (221) as a first noise signal.

[0274] b. The second audio source 213b (213) may include a decorrelation unit for decorrelating the first noise signal (221) to generate a second audio signal (213) as the second noise signal (e.g., the second audio signal obtained from the first audio signal after decorrelation), and

[0275] c. The hybrid noise source 212b(212) may include a second noise generator (which is independent of the first noise generator itself);

[0276] 2. The second variant of CNG 220c (e.g.) Figure 3c ):

[0277] a. The first audio source 211c (211) may include a first noise generator for generating a first audio signal (221) as a first noise signal.

[0278] b. The second audio source 213c (213) may include a second noise generator for generating a second audio signal (223) as a second noise signal (e.g., the second noise generator is independent of the first noise generator itself), and

[0279] c. The mixed noise source 212c (212) may include a decorrelation unit for decorrelating the first noise signal (221) or the second noise signal (223) to generate a mixed noise signal (222);

[0280] 3. The third variant of CNG 220d (e.g.) Figure 3d and 3e ):

[0281] a. One of the first audio source 211d or 211e (211), the second audio source 213d or 213e (213), and the mixed noise source 212d or 212e (212) may include a noise generator for generating a noise signal.

[0282] b. Another of the first audio source 211d or 211e (211), the second audio source 213d or 213e (213), and the mixed noise source 212d or 212e (212) may include a first decorrelation unit for decorrelation of the noise signal, and

[0283] c. Another one of the first audio source 211d or 211e (211), the second audio source 213d or 213e (213), and the mixed noise source 212d or 212e (212) may include a second decorrelation unit for decorrelation of the noise signal.

[0284] d. The first decorreductor and the second decorreductor can be different from each other, so that the output signals of the first decorreductor and the second decorreductor are uncorrelated.

[0285] 4. The fourth variant of CNG 220 (e.g.) Figure 3a ):

[0286] a. The first audio source 211a(211) includes a first noise generator,

[0287] b. The second audio source 213a (213) includes a second noise generator.

[0288] c. The mixed noise source 212a(212) includes a third noise generator.

[0289] d. The first noise generator, the second noise generator, and the third noise generator can generate mutually decorrelated noise signals (e.g., the three generators are not correlated with each other).

[0290] 5. Fifth variant:

[0291] a. One of the first audio source (211), the second audio source (213), and the mixed noise source (212) may include a pseudo-random number sequence generator for generating a pseudo-random number sequence based on a seed.

[0292] b. At least two of the first audio source (211), the second audio source (213), and the mixed noise source (212) can be initialized with different seeds to generate a pseudo-random number sequence.

[0293] 6. The sixth variant:

[0294] a. At least one of the first audio source (211), the second audio source (213), and the mixed noise source (212) can be operated using a pre-stored noise table.

[0295] b. Optionally, at least one of the first audio source (211), the second audio source (213), and the mixed noise source (212) can generate the complex spectrum of the frame, using the first noise value as the real part and the second noise value as the imaginary part.

[0296] c. Optionally, at least one noise generator can generate complex noise spectrum values ​​for a frequency window k, using a first random value at index k as one of the real and imaginary parts, and a second random value at index (k+M) as the other of the real and imaginary parts (the first and second noise values ​​are included in a noise array, for example derived from a random number sequence generator, a noise table, or a noise process, ranging from the start index to the end index, where the start index is less than M).

[0297] The ending index is equal to or less than 2×M, where M and k are integers.

[0298] As from Figure 4As can be seen, in addition to the CNG 220 shown in Figure 3, the decoder 200' (200a, 200b) may also include: an input interface 210 for receiving encoded audio data from a frame sequence, the frame sequence including active frames and inactive frames following the active frames; and an audio decoder for decoding the encoded audio data of the active frames to generate decoded multi-channel signals of the active frames, wherein the first audio source 211, the second audio source 213, the mixed noise source 212, and the mixer 206 are active in the inactive frames to generate multi-channel signals of the inactive frames.

[0299] It should be noted that active frames are those that the encoder classifies as having speech (or any other type of non-noise sound), while inactive frames are those that are classified as having silence or only noise.

[0300] Any example of CNG 220 (220a-220e) can be controlled by a suitable controller.

[0301] encoder

[0302] The encoder is now discussed. It can encode both active and inactive frames. For inactive frames, the encoder can encode parametric noise data (e.g., noise shape and / or coherence values) but not the entire audio signal. It's important to note that the encoding of inactive audio frames can be reduced relative to active audio frames to decrease the amount of information to be encoded in the bitstream. Furthermore, the parametric noise data (e.g., noise shape) for each frequency band in inactive frames can have less information and / or a smaller window compared to the information encoded in active frames. The parametric noise data can be given in the left / right domain or another domain (e.g., the center / side domain), for example, by providing a first linear combination between the parametric noise data of the first and second channels and a second linear combination between the parametric noise data of the first and second channels (in some cases, gain information not associated with the first and second linear combinations but given in the left / right domain can also be provided). The first and second linear combinations are typically linearly independent of each other.

[0303] The encoder may include an activity detector, which classifies frames as active or inactive.

[0304] Figure 1 , 2Figure 4 shows examples of encoders 300a and 300b (also referred to as 300 when there is no need to distinguish between encoders 300a and 300b). Each audio encoder 300 can generate an encoded multi-channel audio signal 232 for a frame sequence of input signal 304. Input signal 304 is here considered to be distinguishable as a first channel 301 (also represented as the left channel or "l", where the uppercase letter "l" is "L", the first letter of the English word "left") and a second channel 303 (or "r", where the uppercase letter "r" is "R", the first letter of the English word "right").

[0305] The encoded multi-channel audio signal 232 can be defined in a frame sequence, which can be, for example, in the time domain (e.g., each sample "n" can refer to a specific moment and the samples of the frame can form a sequence, such as a sampled sequence of the input audio signal or a filtered sequence of the input audio signal).

[0306] Encoder 300 (300a, 300b) may include activity detector 380, which is not in Figure 2 and 4 As shown (although implemented in some of the examples), but Figure 1 As shown in the figure, Figure 1 Each frame of the input signal 304 can be classified as an "active frame 306" or an "inactive frame 308". An inactive frame 308 makes the signal appear as silent (and, for example, only silent or noise), while an active frame 306 may have some detection of noise-free audio signals (such as speech, music, etc.).

[0307] In the encoded multichannel audio signal 232 (e.g., bitstream) encoded by encoder 300, information about whether a frame is an active frame 306 or a silent frame 308 can be signaled, for example, in the so-called “soft noise generation auxiliary information” 402 (p_frame), also known as “auxiliary information”.

[0308] Figure 1 The preprocessing stage 360 ​​displays information that determines (e.g., classifies) whether a frame is an active frame 306 or a silent frame 308. Note that the channels 301 and 303 of the input signal 304 are represented by uppercase letters, such as L (301, left channel) and R (303, right channel), to indicate that they are in the frequency domain. From Figure 1As can be seen, spectrum analysis steps 370 can be applied (first spectrum analysis 370-1 for the first channel 301, L; and second stage 370-3 for the second channel 303, R). Spectrum analysis stage 370 can be performed for each frame of the input signal 304 and can be based, for example, on harmonic measurements. It is worth noting that in some examples, the spectrum analysis performed by stage 370 on the first channel 301 can be performed separately from the spectrum analysis performed on the second channel 303 in the same frame. In some cases, spectrum analysis stage 370 may include the calculation of energy-related parameters, such as the average energy over a predetermined frequency band and the total average energy.

[0309] An activity detection phase 380 can be performed (which can be considered as speech activity detection in the case of searching speech). A first activity detection phase 380-1 can be applied to the first channel 301 (and specifically to measurements performed on the first channel), and a second activity detection phase 380-3 can be applied to the second channel 303 (and specifically to measurements performed on the second channel). In the example, activity detection phase 380 can estimate the energy of background noise in the input signal 304 and use the estimate to calculate the signal-to-noise ratio (SNR), which is compared to an SNR threshold to determine whether a frame is classified as an active frame or an inactive frame (i.e., a calculated SNR exceeding the SNR threshold indicates that the frame is classified as active; and a calculated SNR below the SNR threshold indicates that the frame is classified as inactive). In the example, phase 380 can compare the harmonics obtained by the spectrum analysis phases 370-1 and 370-3, respectively, with one or two harmonic thresholds (e.g., a first threshold for the first channel 301 and a second threshold for the second channel 303). In both cases, not only can each frame be classified, but each channel of each frame can also be classified as an active channel or an inactive channel.

[0310] A decision 381 can be executed, and based on this decision, it can be determined (e.g., marked as switch 381') whether to execute discrete stereo processing 306a or stereo discontinuous transmission processing (stereo DTX) 306b. It is worth noting that in the case of an active frame (and discrete stereo processing 306a), encoding can be performed according to any strategy, processing standard, or procedure; therefore, it will not be analyzed in further detail here. Most of the following discussion will relate to stereo DTX 306b.

[0311] It is worth noting that, in the example, a frame (in stage 381) is classified as an inactive frame only if both channels 301 and 303 are classified as inactive by stages 380-1 and 380-3, respectively. Therefore, the problems in activity detection decisions described above can be avoided. In particular, it is not necessary to signal each channel of each frame to indicate its activity / inactivity classification (thus reducing signal notification), and synchronization between channels is inherently achieved. Furthermore, in the decoder discussed in this specification, the coherence between the first channel 301 and the second channel 303 can be used to generate some noise signals, which are correlated or decorrelated based on the coherence obtained for signal 304. Here, the elements of the encoder 300 (300a, 300b) for encoding inactive frames will be discussed in detail. As explained, any other techniques can be used to encode active frames 308, and therefore will not be discussed here.

[0312] Generally, encoders 300a, 300b (300) may include a noise parameter calculator 3040 for calculating parametric noise data 401, 403 for the first channel 301 and the second channel 303. The noise parameter calculator 3040 can calculate parametric noise data 401, 403 (e.g., index and / or gain) for the first channel 301 and the second channel 303. Therefore, the noise parameter calculator 3040 can provide encoded audio data 232 in a frame sequence, which may include an active frame 306 and an inactive frame 308 (which may follow the active frame 306). Specifically, in the case of an inactive frame 308, the encoded audio data 232 may be encoded as one or two silence insertion descriptor frames (SIDs) 241, 243. In some examples (e.g., in...), Figure 2 In some cases, there is only a single SID frame; in others, there may be two SID frames (e.g., in [the context of the previous sentence]). Figure 4 middle).

[0313] Inactive frame 308 may specifically include at least one of the following:

[0314] - Gentle noise generates auxiliary information (e.g., 402, p_frame);

[0315] - The soft noise parameter data 401 of the first channel 301 or the first linear combination (v) of the soft noise parameter data of the first channel 301 and the soft noise parameter data of the second channel. l,ind v m,ind p_noise, gain g l,q );

[0316] - A second linear combination (v) of the soft noise parameter data 403 of the second channel 303 or the soft noise parameter data of the first channel 301 with the soft noise parameter data of the second channel.r,ind v s,ind p_noise, gain g r,q );

[0317] - Coherence information (coherence data) (c, 404).

[0318] In some examples, the first silence insert descriptor frame 241 may include the first two items of the list above, and the second silence insert descriptor frame 243 may include the last two features of a specific data field. Nevertheless, different protocols may provide different data fields or different bitstream organization. However, in some cases (e.g., in...) Figure 2 In the middle, the noise parameters of the two channels may only be a single inactive frame.

[0319] It will be shown that coherence information (e.g., a portion of a “mute insert descriptor”) may include a single value (e.g., encoded in several bits, such as four bits) indicating coherence information (e.g., coherence data such as coherence data), such as the coherence between the first channel 301 and the second channel 303 of the same inactive frame 308. On the other hand, soft noise parameter data 401, 403 may indicate the signal energy of the inactive frame 308 for each channel 301, 303 (e.g., it may substantially provide a packet), or in any case, may provide noise shape information. The packet or noise shape information may take the form of multiple coefficients of a frequency window and the gain of each channel. Noise shape information can be obtained using the original input channels (301, 303) in stage 312 (see below), and then the noise shape parameter vector is mid / side encoded. It will be shown that some noise channels (201, 203 as shown in Figure 3) may be generated in the decoder that may be affected by coherence information 404. Therefore, the noise channels 201 and 203 generated by CNG 220 (220a-220) can be modified by a signal modifier 250 controlled by control noise data (soft noise parameter data 401, 403, 2312), which indicates the noise for the first audio channel L. out Second audio channel R out The signal energy.

[0320] Audio encoder 300 (300a, 300b) may include coherence calculator 320, which can obtain coherence information (404) for encoding in a bitstream (e.g., signal 232, frame 241 or 243). The coherence information (c, 404) can indicate the coherence between a first channel 301 (e.g., left channel) and a second channel 303 (e.g., right channel) in an inactive frame 308. Examples of this will be discussed later.

[0321] Encoder 300 (300a, 300b) may include output interface 310 configured to generate multichannel audio signal 232 (bitstream) having coded audio data of active frame 306 and first parameter data (soft noise parameter data) 401 (p_noise, left) and second parameter noise data (p_noise, right, 403) of inactive frame 308, as well as coherence data c (404). The first parameter data 401 may be parameter data of a first channel (e.g., left channel) or a first linear combination of the first and second channels (e.g., center channel). The second parameter data 403 may be parameter data of a second channel (e.g., right channel) or a second linear combination of the first and second channels (e.g., side channel), wherein the second linear combination differs from the first linear combination.

[0322] In bitstream 232, there may also be auxiliary information 402, which includes an indication of whether the current frame is an active frame 306 or an inactive frame 308, for example to inform the decoder of the decoding technique to be used.

[0323] In particular, Figure 4 Display noise parameter calculator (noise parameter calculation stage) 3040, which includes a first noise parameter calculator stage 304-1 for calculating soft noise parameter data 401 for the first channel 301, and a second noise parameter calculator stage 304-3 for calculating the second soft noise parameter 403 for the second channel 303. Figure 2 An example is shown where noise parameters are jointly processed and quantized. Internal parts (such as converting the noise shape vector to an M / S representation) are as follows... Figure 5 As shown. Basically, we may have the noise shape of the first channel M and the noise shape of the second channel S, which can be encoded as center index and side index, while the gain of the noise shape of the left channel 301 and the gain of the noise shape of the right channel 303 can also be encoded.

[0324] The coherence calculator 320 can calculate coherence data (coherence information) c(404) indicating the coherence between the first channel L and the second channel R. In this case, the coherence calculator 320 can operate in the frequency domain.

[0325] As can be seen, the coherence calculator 320 may include a channel coherence calculation stage 320', which obtains a coherence value c(404). Then, downstream of this, a uniform quantizer stage 320' can be used. Therefore, a quantized version c of the coherence value c can be obtained. ind .

[0326] The following will explain how to obtain coherence and how to quantify it.

[0327] In some examples, the coherence calculator 320 can:

[0328] Calculate the real and virtual intermediate values ​​from the complex spectrum values ​​of the first and second channels (303) in the inactive frame;

[0329] Calculate the first energy value of the first channel and the second energy value of the second channel (303) in the inactive frame; and

[0330] Coherence data (404, c) are calculated using real intermediate values, imaginary intermediate values, first energy values, and second energy values, and / or

[0331] Smooth at least one of the real intermediate value, the imaginary intermediate value, the first energy value, and the second energy value, and use at least one smoothed value to calculate coherence data.

[0332] The coherence calculator 320 can square the smoothed real intermediate value and the smoothed imaginary intermediate value, and add the squared values ​​to obtain a first component number. The coherence calculator 320 can multiply the smoothed first and second energy values ​​to obtain a second component number, and combine the first and second component numbers to obtain a result number of coherence values, based on which the coherence data is derived. The coherence calculator 320 can calculate the square root of the result number to obtain the coherence value as the basis of the coherence data. Several examples of formulas are provided below.

[0333] Now, let’s explain how to obtain the shape of the noise (or other signal energy) to be presented at the decoder. What will be encoded is essentially the shape of the noise (or other energy-related information) of the original input signal 302, which will be applied at the decoder to the generated noise 203 and shaped to present noise 252 (output audio signal), which is similar to the original noise of signal 304.

[0334] First, it should be noted that the aforementioned signal 304 is not encoded in bitstream 232 by the encoder. However, noise information (such as energy information and packet information) can be encoded in bitstream 232 to subsequently generate a noise signal with the noise shape encoded by the encoder.

[0335] The noise shape acquisition block 312 can be applied to the encoder's input signal 304. The noise shape acquisition block 312 calculates a low-resolution parametric representation 1312 of the spectral packets of noise in the input signal 304. This can be done, for example, by calculating the energy values ​​in the frequency bands of the frequency domain representation of the input signal 304. The energy values ​​can be converted to a logarithmic representation (if needed) and can be compressed into a lower number (N) of parameters that are later used in the decoder to generate soft noise. These low-resolution representations of the noise are referred to herein as the "noise shape" 1312. Therefore, the downstream of the noise shape acquisition block 312 should not be understood as representing the input signal 304, but rather its noise shape (the parametric representation of the noise spectral packets in each channel). This is important because the encoder may only transmit this low-resolution representation of the noise spectral packets in the SID frame. Therefore, in Figure 2 In the text, all sections of the "Noise Parameter Calculator" (3040) can be understood as referring only to these noise-related parameter vectors (e.g., denoted as v). l v r v m,ind and v s,ind It performs operations on the signal representation of signal 304, but does not operate on the signal representation of signal 304.

[0336] Figure 5 This shows an example of section 3040 (Joint Noise Shape Quantization) in the "Noise Parameter Calculator". The L / R to M / S converter stage 314 can be applied to obtain the middle channel representation of the noise shape 1312. m (The first linear combination of the noise shapes of channels L and R) and the side channel representation of noise shape 1312 v r (A second linear combination of the noise shapes of channels L and R). The following will show how to obtain it. Therefore, noise shape 304 can be divided into two channels v. m and v r .

[0337] Next, in the normalization stage 316, the mid-channel representation of the noise shape 1312 is v m The side channel representation of noise shape 1312 v r At least one of them can be normalized to obtain a mid-channel representation of noise shape 1312. m Normalized version v m,n And / or the side channel representation of noise shape 1312 v r Normalized version v r,n .

[0338] Next, the quantization stage (e.g., vector quantization, VQ) 318 can be applied to a normalized version of the signal 1304, for example, in the form of a normalized mid-channel representation of noise shape 1312.m,n Quantized version v m,ind Normalized side channel representation of noise shape 1312 v s,n Quantized version v s,ind In the form of v. Vector quantization can be used (e.g., via a multi-stage vector quantizer), therefore, the index v m,ind [k] (where k is the index of a specific frequency window) can describe the mid-representation of the noise shape, and the index v s,ind [k] can describe the lateral representation of the noise shape. Therefore, the index v m,ind [k] and v s,ind [k] can be encoded in bitstream 232 as a first linear combination of soft noise parameter data of the first channel and soft noise parameter data of the second channel, and a second linear combination of soft noise parameter data of the first channel and soft noise parameter data of the second channel.

[0339] In the dequantization stage 322, the normalized midchannel representation v of the noise shape 1312 can be obtained. m,n Quantized version v m,ind Normalized side channel representation of noise shape 1312 v s,n Quantized version v s,ind Perform dequantization.

[0340] The M / S to L / R converter 324 can be applied to the dequantization representation of noise shape 1312. m,q And side represents v s,q The dequantized version to obtain the original (left and right) channel v' l and v' r The noise shape version 1312 in the text.

[0341] Subsequently, in stage 326, the gain g can be calculated. l and g r It is worth noting that the gain for the same channel (v') in the same inactive frame 306 l and v' r All samples of the noise shape are valid. Gain g l and g r v' can be represented by considering the shape of the noise. l and v' r It is obtained by taking the total (or almost total) frequency window in the range.

[0342] Gain g l This can be obtained by comparing the following two:

[0343] The value of the frequency window representing the noise shape of the first channel 301 in the L / R domain (upstream of the L / R to M / S converter 314); with

[0344] - Once reconverted in the L / R domain, the value of the frequency window of the noise shape 1312 of the first channel 301 (downstream of the M / S to L / R converter 324).

[0345] Similarly, the gain g r This can be obtained by comparing the following two:

[0346] The value of the noise shape coefficient of the second channel 303 in the L / R domain (upstream of the L / R to M / S converter 314); with

[0347] - The value of the coefficient of the noise shape 1312 of the second channel 303 (downstream of the M / S to L / R converter 324) that is reconverted in the L / R domain.

[0348] The following presents an example of how gain can be obtained. However, in the linear domain, gain can, for example, be proportional to the geometric mean of multiple fractions, each fraction being the ratio between the coefficients of the noise shape of a particular channel in the L / R domain (upstream of the L / R to M / S converter 314) and the coefficients of the same channel when it is re-converted downstream in the L / R domain to the M / S to L / R converter 324. In the logarithmic domain, for each channel, gain can be obtained as a ratio to an algebraic mean, the difference between the coefficients of the FD version of the noise shape in the L / R domain (upstream of the L / R to M / S converter 314) and the coefficients of the noise shape when it is re-converted downstream in the L / R domain to the M / S to L / R converter 324. Typically, in the logarithmic or scalar domain, gain can provide the relationship between the version of the noise shape of the left or right channel before L / R to M / S conversion and quantization and the version of the noise shape of the left or right channel after dequantization and M / S to L / R re-conversion.

[0349] The quantization stage 328 can be applied to the gain g. l To obtain its identifier as g l,q The quantized version, and applied to the gain g r To obtain its identifier as g r,q The quantized version, which can be derived from the non-quantized gain g r Gain. Gain g l,q and g r,q It can be encoded in bitstream 232 (e.g., as soft noise parameter data 401 and / or 403) for read by the decoder.

[0350] In some examples, the energy of the side channel noise shape vector (e.g., before normalization, such as between stages 314 and 316) can also be compared with a predetermined energy threshold α (which can be a positive real value) (0.1 in this example, but can also be different values, such as values ​​between 0.05 and 0.15). In comparison block 435, the side representation v of the noise shape of the inactive frame 308 can be determined. s Does it have enough energy? If the side of the noise shape represents v s If the energy is less than the energy threshold α, then the binary result (“no side flag”) is signaled in bitstream 232 as auxiliary information 402. Here, it is assumed that if the side representation of the noise shape is v... s If the energy is less than the energy threshold α, then the no-side flag = 1; if the side of the noise shape represents v... s If the energy is greater than the energy threshold α, then the unside flag = 0. In some cases, where the energy is exactly equal to the energy threshold, the flag can be either 1 or 0, depending on the specific application. Block 436 negates the binary value of the unside flag 436' (if the input to block 436 is 1, then the output 436' is 0; if the input to block 436 is 0, then the output 436' is 1). Block 436 is displayed as the output 436' used to provide the opposite value of the flag. Therefore, if the side of the noise shape represents v s If the energy is greater than the energy threshold, then the value 436' can be 1, if the side of the noise shape represents v. s If the energy is less than a predetermined threshold, then the value 436' is 0. Note that the dequantized value v... s,q This can be multiplied by a binary value of 436'. This is just one possible way to obtain the following information: if the side of the noise shape represents the energy v... s If the energy is less than a predetermined energy threshold α, then the dequantized side of the noise shape represents v. s,q The window can be manually zeroed out (the output 437' of block 437 will be 0). On the other hand, if the side of the noise shape represents v... s If the energy is large enough (>α), then the output 437' of block 437 (multiplier) may be related to v. s,q Exactly the same. Therefore, if the side of the noise shape represents the energy v s If the energy is less than the predetermined energy threshold α, then the side representation v, regardless of the noise shape, is used. s (especially its dequantized version v) s,q This is to obtain a left / right representation of the noise shape (it will be shown that, alternatively, the decoder may also have a similar mechanism to zero out the coefficients of the side representation of the noise shape). Note that a sideless flag can also be encoded in bitstream 232 as part of auxiliary information 402.

[0351] It should be noted that the energy represented by the side of the noise shape is shown as measured (by block 435) before the noise shape is normalized (in block 316), and the energy is not normalized before it is compared with a threshold. In principle, it could also be measured by block 435 after the noise shape has been normalized (e.g., block 435 could be measured by v). s,n Input instead of v s enter).

[0352] A reference energy threshold α is used to compare the side representation of the noise shape; this value is 0.1, and it can be arbitrarily chosen in some examples. In the examples, the threshold α can be selected after experimentation and adjustment (e.g., by calibration). In some examples, in principle, any number suitable for the digital format (floating-point or fixed-point) or the precision of the individual implementation can be used. Therefore, the threshold α can be an implementation-specific parameter that can be entered after calibration.

[0353] Note that the output interface (310) can be configured as follows:

[0354] A coded multichannel audio signal (232) with coded audio data of an active frame (306) is generated using a first plurality of coefficients for a first number of frequency windows; and

[0355] A second set of coefficients, used to describe a second quantity of frequency windows, are used to generate first parameter noise data, second parameter noise data, or a first linear combination of first parameter noise data and second parameter noise data, and a second linear combination of first parameter noise data and second parameter noise data.

[0356] The first number of frequency windows is larger than the second number of frequency windows.

[0357] In fact, a lower resolution can be used for inactive frames, further reducing the number of bits used to encode them into a bitstream. The same applies to the decoder.

[0358] Any example of an encoder can be controlled by a suitable controller.

[0359] decoder

[0360] Now, let's discuss the decoder based on the example. The decoder may include, for example, the soft noise generator 220 (220a-220e) discussed above, such as... Figures 3a-3f As shown. Soft noise 204 (multi-channel audio signal) can be shaped at signal modifier 250 to obtain output signal 252. We are interested here in showing the operation used to generate noise in inactive frame 308, not for active frame 306.

[0361] Figure 4The first example shown is of decoder 200', denoted here as 200' (200b). Note that decoder 200' includes a soft noise generator 220, which may include, according to... Figures 3a-3f Any of the generators 220 (220a-220e) shown. Downstream of generators 220 (220a-220e), a signal modifier 250 (not shown, but...) may exist. Figure 4 (As shown in the diagram), it is used to shape the generated multichannel noise 204 according to the energy parameters encoded in the soft noise parameter data (401, 403). Through the decoder input interface 210, the decoder 200' can obtain the soft noise parameter data (401, 403) from the bitstream 232, which may include soft noise parameter data describing the signal energy (e.g., for the first channel and the second channel, or for a first linear combination and a second linear combination of the first and second channels, where the first and second linear combinations are linearly independent of each other). Through the decoder input interface 210, the decoder 200' can obtain coherence data 404, which indicates the coherence between different channels. Figure 4 As shown in bitstream 232, two different silence descriptor frames 241 and 243 are provided for encoding inactive frames, but it is possible to use more than two descriptor frames, or only a single descriptor frame. The output of decoder 200b is a multi-channel output.

[0362] refer to Figure 2 As shown, we will now discuss decoder 200' (referred to as 200a here) as an example of decoder 200, which can be used to generate output signal 252, which may be in the form of noise, for example.

[0363] First, decoder 200a (200') may include input interface 210 for receiving encoded audio data 232 (bitstream) from frame sequences 306, 308, which is encoded, for example, by encoder 300a or 300b. Decoder 200a (200') may be a multichannel signal generator 200, or more generally, part of a multichannel signal generator 200, which may be or include, for example... Figures 3a-3f One of the soft noise generators 220 (220a-220e).

[0364] first, Figure 2 The stereo soft noise generator (CNG) 220 (220a-220e) is shown. Specifically, the soft noise generator 220 (220a-220e) can be similar to... Figures 3a-3f The illustrated soft noise generator or one of its variants. Here, the coherence information 404 (e.g., c, or more precisely c) obtained from encoder 300a or 300b qYou can also use "coh" or c ind The multichannel signal 204 (in channels 201, 203) generated by the CNG 220 (220a-220e) can be further modified, for example, by taking into account the soft noise parameter data 401 and 403, such as the noise shape information of the first (left) channel and the second (right) channel of the multichannel signal to be shaped. In particular, the index v generated by the encoder 300a (and especially by the noise parameter calculator 3040) at stages 316 and / or 318 will be shown here. m,ind (401) and side index v s,ind (403), and the gain g obtained at stage 326 and / or 328. l,q and g r,q .

[0365] like Figure 2 As shown, auxiliary information 402 allows determination of whether the current frame is an active frame 306 or an inactive frame 308. For example... Figure 2 The components shown refer to the processing of inactive frame 308, and are intended to be used to generate the output signal in active frame 306 using any technique; therefore, they are not the subject of this specification.

[0366] like Figure 2 The diagram illustrates several examples of obtaining soft noise data from bitstream 232. As described above, the soft noise data may include coherence information (data) 404, parameters 401 and 403 (v) representing the noise shape. m,ind and v s,ind ) and / or gain (g l,q and g r,q ).

[0367] Stage 212-C can quantize version c of coherence information 404. ind Dequantization is performed to obtain the dequantized relevant information c. q .

[0368] Stage 2120 (Joint Noise Shape Dequantization) allows for the dequantization of additional soft noise data obtained from bitstream 232. See reference... Figure 6 The dequantization stage 212 is formed by other dequantization stages, denoted here as 212-M, 212-S, 212-R, and 212-L. Stage 212-M can dequantize the mid-channel noise shape parameters 401 and 403 to obtain the dequantized noise shape parameter v. m,q and v s,q Stage 212-S can provide side channel noise shape parameters of 403 (v). s,ind The dequantized version vs,q In some examples, a sideless flag can be used so that the noise shape vector v s The zeroing phase 212-S outputs when the energy is identified as less than a predetermined threshold α by block 435 at encoder 300a. The noise shape vector v is determined when the energy is less than the predetermined threshold α and signaled with a sideless flag. s Dequantized version v s,q It can be zeroed out (conceptually shown as multiplying by the flag 536' obtained from block 536, which has the same function as block 436 of the encoder, even though block 536 actually reads the sideless flag encoded in the auxiliary information of bitstream 232 without performing any comparison with the threshold α). Therefore, if the energy of the side channel at the encoder is determined to be less than the predetermined threshold α, the noise shape vector v s Dequantized version v s,q The value is artificially zeroed out, and the output 537' of scaler block 537 is zero. Otherwise, if the energy is greater than a predetermined threshold, the output 537' is zeroed out along with the side index 403 (v) of the noise shape of the side channel. s,ind Quantized version v s,q Same. In other words, when the energy of the side channel is below a predetermined energy threshold α, the noise shape vector v s,ind The value is ignored.

[0369] In the M / S to L / R stage 516, an M / S to L / R conversion is performed to obtain the L / R version v' of the parameter data (noise shape). l v' r Subsequently, gain stage 518 (formed by stages 518-L and 518-R) can be used to make channel v' at stage 518-L. l From the gain g l,d Scaling, and at stage 518-R, channel v' r From the gain g r,q Scaling. Therefore, the energy channel v can be obtained. l,q With v r,q This is the output of gain stage 518. Stage blocks 518-L and 518-R are indicated by "+" because the value transformation is imagined to be in the logarithmic domain, thus additionally indicating the scaling of the values. However, gain stage 518 indicates the reconstructed noise shape vector v. l,q and v r,q Scaled. Reconstructed noise shape vector v l,q and v r,qHere, 2312 is used to indicate the noise shape in complexity and is a reconstructed version of the noise shape 1312, as originally obtained by the "Get Noise Shape" block 312 at the encoder. In general, for all indices (coefficients) of the same channel in the same inactive frame, each gain is constant.

[0370] Note that index v m,ind v s,ind and gain g l,q g r,q These are coefficients representing the noise shape and providing information about the frame energy. They essentially refer to parameter data associated with the input signal 304 used to generate signal 252, but do not represent signal 304 or the signal 252 to be generated. In other words, the noise channel v r,q and v l,q The packet to be applied to the multi-channel signal 204 generated by CNG 220 is described.

[0371] Back Figure 2 The reconstructed noise shape vector v is used at signal modifier 250. l,q and v r,q (2312) to obtain a modified signal 252 by shaping the noise 204. Specifically, the first channel 201 of the generated noise 204 can be controlled by channel v at stage 250-L. l,q The noise 204 generated can be shaped at stage 250-R to obtain the output multi-channel audio signal 252 (L). out and R out ).

[0372] In the example, the soft noise signal 204 itself is not generated in the logarithmic domain: only the noise shape can be represented logarithmically. A transformation from the logarithmic domain to the linear domain can be performed (although not shown in the figure).

[0373] It can also perform conversions from the frequency domain to the time domain (although not shown in the figure).

[0374] Decoder 200' (200a, 200b) may also include a spectrum-to-time converter (e.g., signal modifier 250) for converting the spectrally adjusted and coherently adjusted first channel 201 and second channel 203 into corresponding time-domain representations for combination or concatenation with the time-domain representations of the corresponding channels of the decoded multichannel signal of the active frame. The conversion of generated soft noise into a time-domain signal occurs during... Figure 2 The signal modifier block 250 shown below. The "Combined or Connected" part basically means before or after an inactive frame using these CNG techniques, or before or after an active frame. Figure 1Other processing paths are shown), and in order to generate a continuous output without any gaps or audible clicking sounds, multiple frames need to be properly connected.

[0375] In some examples:

[0376] The coded audio signal (232) for the active frame (306) has a first plurality of coefficients describing a first number of frequency windows; and

[0377] The encoded audio signal (232) for the inactive frame (308) has a second plurality of coefficients describing a second number of frequency windows.

[0378] The first number of frequency windows can be larger than the second number of frequency windows.

[0379] Any example of a decoder can be controlled by a suitable controller.

[0380] Processing steps: Version 1

[0381] The noise parameters encoded in the two SID frames of the two channels are calculated according to the method in EVS[6], such as LP-CNG or FD-CNG, or both. The noise energy shaping in the decoder is also the same as in EVS, such as LP-CNG or FD-CNG, or both.

[0382] In the encoder, the coherence of the two channels is calculated separately, using four-bit uniform quantization and transmitted in a bitstream of 232. In the decoder, CNG operation can then be controlled via the transmitted coherence value 404. This can be achieved using methods such as... Figures 3a-3f The three Gaussian noise sources N1, N2, and N3 (211a, 212a, 213a; 211b, 212b, 213b; 211c, 212c, 213c; 211d, 212d, 213d; 211e, 212e, 213e) are shown. When the channel coherence is high, the main correlated noise can be added to channels 221' and 223', while when the coherence is low, more uncorrelated noise is added.

[0383] For all inactive frames 306, parameters (noise parameters) for soft noise generation can be continuously estimated in the encoder (e.g., 300, 300a, 300b). This can be done, for example, by applying a frequency domain noise estimation algorithm (e.g., [8]), for example, by calculating two sets of noise parameters (e.g., 401, 403) on the two input channels (e.g., 301, 303) as described in [6], which are also interpreted as parametric noise data. Furthermore, the coherence (c, 404) of the two channels can be calculated as follows (e.g., at coherence calculator 320): given two input channels The M-point DFT-spectrum (L and R can be 301 and 303) can be used to calculate four intermediate values, for example...

[0384]

[0385]

[0386] And the energy of the two channels

[0387]

[0388]

[0389] Here, M = 256, Represents the real part of a complex number. Let {·} represent the imaginary part of a complex number, and let {·} * This indicates complex conjugation. These intermediate values ​​can then be smoothed, for example, using the corresponding value from the previous frame:

[0390]

[0391]

[0392]

[0393]

[0394] This section could be part of the "Calculate Channel Coherence" block 320' at the encoder. This is a temporal smoothing of the internal parameters to avoid sudden large jumps in parameters between frames. In other words, a low-pass filter is applied to the parameters here.

[0395] The range 0.95±0.03 and... Other constants within the range are used to replace the constants 0.95 and 0.05.

[0396] Alternatively, it can be defined as:

[0397]

[0398]

[0399]

[0400]

[0401] Where β, γ ∈ [0, 1], and β + γ = 1, for example β = 0.95 and γ = 0.05.

[0402] The coherence (c, 404) can then be calculated (potentially between 0 and 1), for example, at the coherence calculator (320) as follows:

[0403]

[0404] And use, for example, four-bit uniform quantization (e.g., at quantizer 320"), as follows:

[0405] c ind =0,min(15,floor(15×c+0.5))

[0406] The estimated noise parameters 1312 and 2312 for the two channels can be encoded separately, for example, as described in [6]. The two SID frames 241 and 243 can then be encoded and sent to the decoder. The first SID frame 241 may contain the estimated noise parameters 401 for channel L and several bits (e.g., four bits) of auxiliary information 402, for example, as described in [6]. In the second SID frame 243, the noise parameters 403 for channel R can be sent along with four-bit quantized coherence values ​​c and 404 (different bit amounts can be selected in different examples).

[0407] In the decoder (e.g., 200', 200a, 200b), the noise parameters (401, 403) of the two SID frames and the auxiliary information 402 of the first frame can be decoded, as described in [6]. The coherent value 404 in the second frame can be dequantized in stage 212-C as follows:

[0408]

[0409] (exist Figure 2 middle, be c q replace).

[0410] For the generation of soft noise (e.g., in any of generator 220 or generators 220a-220e, it may include...) Figures 3a-3e (Any one of them), according to the example, three Gaussian noise sources 211, 212, and 213 as shown in Figure 3 can be used. Noise sources 211, 212, and 213 can be adaptively added together, for example, based on the coherence value (c, 404) (e.g., at adder stages 206-1 and 206-3). Left and right channel noise signals N l [k]、N r The DFT spectrum of [k] can be calculated as follows:

[0411]

[0412]

[0413] Where k∈{0,1,…,M-1} (this is the index of a specific frequency window, and each channel has M frequency windows), j 2 =-1 (i.e., j is the imaginary unit), "×" represents normal multiplication. Here, "frequency window" refers to the spectrum N. l and N r The number of complex values. M is the transform length of the FFT or DFT used, so the length of the spectrum is M.

[0414] It should be noted that the noise from real part insertion and the noise from imaginary part insertion may be different. Therefore, for a spectral length M...

[0415] In other words, we need to generate 2×M values ​​(one real number and one imaginary number) from each noise source. Or, in other words:

[0416] N l and N r N1 is a complex numerical vector of length M, while N2 and N3 are real numerical vectors of length 2×M.

[0417] Subsequently, the noise signals 204 in both channels are spectrally shaped using the corresponding noise parameters (2312) decoded from the corresponding SID frames (as shown in the image). Figure 2 The process is performed in stages 250-L and 250-R, and then transformed back to the time domain (as described in [6]) for frequency domain soft noise generation.

[0418] Any example of processing can be performed by the appropriate controller.

[0419] Processing steps: Version 2

[0420] The processing steps described above can be integrated with at least one of the following aspects. This is primarily referenced here. Figure 2 and 5 However, it can also be used as a reference. Figure 4 .

[0421] A block diagram of the general framework of an encoder is shown below. Figure 1 As shown. For each frame in the encoder, as described in [6], the current signal can be classified as active or inactive by running VAD separately on each channel. The VAD decision can then be synchronized between the two channels. In the example, the frame is classified as inactive frame 308 only if both channels are classified as inactive. Otherwise, the frame is classified as active, and both channels are jointly encoded in the band M / S manner described in

[10] in an MDCT-based system. When switching from an active frame to an inactive frame, the signal may enter the SID encoding path shown in Figure 3.

[0422] The parameters (e.g., 1312, 401, 403, q) used for soft noise generation can be continuously estimated in the encoder (e.g., 300, 300a, 300b) for active and inactive frames (306, 308). l,q g r,q (e.g., noise parameters). This can be accomplished, for example, by applying a frequency domain noise estimation process as discussed in [8] and / or described in [6], for example, calculating two sets of noise parameters on the two input channels 301, 303 respectively, which include, for example, the spectral noise shape (M) in the logarithmic domain of each channel. i 401 and / or I s Or 403).

[0423] Furthermore, the coherence (404, c) of the two channels can be calculated as follows (e.g., in the coherence calculator 320): given the M-point DFT-spectrum L of the two input channels, The four intermediate values ​​can be calculated as follows:

[0424]

[0425]

[0426] And the energy of the two channels

[0427]

[0428]

[0429] Here, M can be 256 (M can use other values). Represents the real part of a complex number. The imaginary part of a complex number is represented by {·}. * This represents the complex conjugate. These intermediate values ​​are then smoothed over a 10-millisecond subframe, where {·} previous This represents the corresponding value from the previous subframe. The smoothed value can be calculated as follows:

[0430]

[0431]

[0432]

[0433]

[0434] The range 0.95±0.03 and... Other constants within the range are used to replace the constants 0.95 and 0.05.

[0435] Alternatively, it can be defined as:

[0436]

[0437]

[0438]

[0439]

[0440] Where β, γ ∈ [0, 1], and β + γ = 1, for example β = 0.95 and γ = 0.05 (β > γ, for example β > 3 × γ or β > 6 × γ).

[0441] The coherence c∈[0,1] (e.g., at 320') can then be calculated as follows:

[0442]

[0443] And use four bits (but possibly a different number of bits) to uniformly quantize (e.g., in 320") as follows:

[0444]

[0445] in, This indicates rounding down to the nearest integer (round down function).

[0446] The coding of the estimated noise shape for both channels can be done jointly. From left (v l ) and right (v r The shape of the channel noise can be used to obtain different channels (e.g., through linear combination), for example, the middle channel (v) can be calculated. m Noise shape and side channel (v) s The noise shape (e.g., in block 314) is as follows:

[0447]

[0448]

[0449] Wherein, for example in the frequency domain, N represents the length of the noise shape vector (e.g., 308 for each inactive frame). As estimated in EVS[6], N represents the length of the noise shape vector, which can be between 17 and 24. The noise shape vector can be viewed as a more compact representation of the spectral envelope of noise in the input frame. Or, more abstractly, a parameterized spectral description of the noise signal using N parameters. N is independent of the transform length of the FFT or DFT.

[0450] These noise shapes can then be normalized (e.g., in stage 316) and / or quantized. For example, they can be vector quantized (e.g., in stage 318), for example using a multi-stage vector quantizer (MSVQ) (an example is described in [6, p 442]).

[0451] Used to quantize v at stage 318 m Shape (to obtain v) m,ind MSVQ of 401 can have 6 stages (but may be other numbers of stages) and / or use 37 bits (but may be other numbers of bits), as implemented for mono in [6], while stage 318 is used for quantization v. s Shape (to obtain v) s,ind MSVQ (403) may have been reduced to 4 stages (or in any case, fewer stages than used in stage 318), and / or a total of 25 bits are used (or in any case, fewer bits than used in stage 318 for encoding shape v). m (Number of bits).

[0452] The MSVQ codebook index can be transmitted in the bitstream (e.g., in data 232, more specifically in the soft noise parameter data 401, 403). The index is then dequantized to produce a dequantized noise shape v. m,q and v m,q .

[0453] When the background noise is a single noise source at the center of the stereo image, the estimated noise shape v for both channels. m v s The expected shapes are very similar, or even equal. The resulting S-channel noise shape will then contain only zeros. However, the vector quantizer (stage 322) used to quantize the current implementation may not be able to model an all-zero vector, and after dequantization, the dequantized v... s Noise shape (v) s,q The vector quantizer 322 may no longer be all zeros. This could lead to perceptual problems in representing this central background noise. To circumvent this drawback of the unquantized v, we can use the unquantized v as a reference. s The energy of the shape vector (e.g., v after stage 314 and / or before stage 316) s The energy of the noise shape vector is calculated (and can also be signaled in the bitstream) without side values ​​(without side flags), where the without side flags can be:

[0454]

[0455] For example, the energy threshold α could be 0.1 or another value in the range [0.05, 0.15]. However, the threshold α can be arbitrary and, in implementation, can depend on the digital format used (e.g., fixed-point or floating-point) and / or the signal normalization that might be used. In the example, a positive real value could be used, depending on the severity of the definition employed for the "mute" S-channel. Therefore, this range could be (0, 1). No lateral value can be used to indicate whether v should be used. s noise shape to reconstruct v l and v r Channel noise shape (e.g., at the decoder). If the side value is 1, then the dequantized v... s The shape is set to 0 (e.g., by...). Figure 2 The value of 436' in the vocal tract v s,q Scaling is a logical inverse (NOT) (no side value). The no side value is transmitted in bitstream 232 (signaling), for example, in auxiliary information 402. Subsequently, an inverse M / S transform (e.g., stage 324) can be applied to the dequantized noise shape vector v. m,q and v s,q (When the energy is low, the latter is, for example, replaced with 0, therefore in) Figure 2 (represented by 437' in Chinese), to obtain the intermediate vector v' l and v' r as follows:

[0456]

[0457]

[0458] Using these intermediate vectors v' l and v' r and the dequantized noise shape vector v l and v r The two gain values ​​were calculated as follows:

[0459]

[0460]

[0461] The two gain values ​​can then be linearly quantized (e.g., in stage 328) as follows:

[0462]

[0463] (Other quantifications are also possible.)

[0464] Quantization gain can be encoded in the SID bitstream (e.g., as part of the soft noise parameter data 401 or 403, more specifically, g). l,qIt can be part of the first parameter noise data, and g r,q It can be part of the second parameter noise data, for example, for the gain value g. l,q Use seven bits, and / or for the gain value g r,q Use seven bits (different numbers of bits can also be used for each gain value).

[0465] In the decoder (e.g., 200', 200a, 200b), the quantized noise shape vector (e.g., a portion of the soft noise parameter data 401 or 403, and more specifically a portion of the first parameter noise data and the second parameter noise data) can be dequantized, for example, in stage 212 (in particular, in any of the sub-stages 212-M, 212-S).

[0466] The gain value can be dequantized, for example, in stage 212 (specifically, in any of sub-stages 212-L and 212-R) as follows:

[0467]

[0468]

[0469] (The value 45 depends on the quantization and may vary depending on the quantization.) (In Figure 2 In the middle, use g l,d and g r,d Replace g l,deq and g r,deq ).

[0470] The coherence value 404 can be dequantized (e.g., in stage 212-C) as follows:

[0471] c q =15×c ind .

[0472] If the unbiased flag (in auxiliary information 402) is 1, then the intermediate vector v' is computed. l and v' r Previously (e.g., in stage 516), the dequantized v was... s Shape v s,q Set to 0 (value 537'). Then add the corresponding gain value to all elements of the corresponding intermediate vector to generate the dequantized noise shape v. l,q and v r,q Its complex form is 522, as follows:

[0473] v l,q ={v′ l,1 +g L,deq ,…,v′ l,N +g L,deq}

[0474] v r,q ={v′ r,1 +g R,deq ,…,v′ r,N +g R,deq}

[0475] (Addition is because we are in the logarithmic field and correspond to the product of factors in the linear field.)

[0476] For the generation of soft noise, such as Figures 3a-3f Any one of them (or any other technique may be used) can be used with three Gaussian noise sources N1, N2, N3 (e.g., Figure 3a The figures 211a, 212a, and 213a shown are... Figure 3b As shown in 211b, 212b, 212c, etc., when the channel coherence is high, mainly related noise is added to the two channels, while if the coherence is low, more unrelated noise is added.

[0477] Using three noise sources, the noise signals N from the left and right channels. l (201) and N r The DFT spectrum of (203) can be calculated as follows:

[0478]

[0479]

[0480] Where k∈{0,1,…,M-1} and j 2 = -1. Here, M represents the block length of the DFT. In order to generate independent noise in the real and imaginary parts of the complex spectrum, each noise source must generate 2×M values ​​per frame (two values ​​per frequency window). Therefore, N1, N2, and N3 (located in...) Figure 3f 211, 212, 213) in the vector can be viewed as a real-valued noise vector of length 2×M, while N r and N k (located at 201 and 203 respectively) is a complex numerical vector of length M.

[0481] Then, the noise signals in both channels can be decoded using their corresponding noise shapes (v) from the bitstream 232. l,q or v r,q The frequency domain is spectral sculpted (e.g. at signal modifier 252) and then transformed back from the logarithmic domain to the scalar domain and back from the frequency domain to the time domain, as described in [6], in order to generate a stereo soft noise signal.

[0482] Any example of this process can be executed by a suitable controller.

[0483] Some advantages

[0484] This invention provides a stereo soft noise generation technique particularly suitable for discrete stereo coding schemes. By jointly encoding and transmitting the noise shape parameters of the two channels, stereo CNG can be applied without mono downmixing.

[0485] Along with two independent sets of noise parameters, the mixing of a common noise source controlled by a single coherence value and two independent noise sources allows for faithful reconstruction of stereo images with background noise without transmitting fine-grained stereo parameters that typically only exist in parametric audio encoders. Because only this one parameter is used, SID encoding is straightforward, requiring no complex compression methods, while still keeping SID frames at a low size.

[0486] Some important aspects:

[0487] In some examples, at least one of the following aspects can be obtained:

[0488] 1. Generating soft noise for the stereo signal by mixing three Gaussian noise sources (one for each channel) and a third common noise source to create relevant background noise.

[0489] 2. Control the mixing of noise sources with coherent values ​​transmitted with the SID frame.

[0490] 3. By jointly encoding the noise shape in an M / S manner, independent noise shape parameters are transmitted for both stereo channels. The SID frame bit rate is reduced by encoding the S shape with fewer bits than M.

[0491] Other technologies

[0492] Another method for generating a multichannel signal having a first channel and a second channel can be implemented, including:

[0493] A first audio signal is generated using a first audio source;

[0494] A second audio signal is generated using a second audio source;

[0495] A mixed noise signal is generated using a mixed noise source; and

[0496] The mixed noise signal is mixed with a first audio signal to obtain a first channel, and the mixed noise signal is mixed with a second audio signal to obtain a second channel.

[0497] An audio encoding method can also be implemented to generate encoded multichannel audio signals for a frame sequence including active and inactive frames, the method comprising:

[0498] Analyze the multi-channel signal to determine which frames in the frame sequence are inactive;

[0499] Calculate the first parameter noise data for the first channel of the multi-channel signal, and calculate the second parameter noise data for the second channel of the multi-channel signal.

[0500] Calculate coherence data indicating the coherence between the first and second channels in inactive frames; and

[0501] Generate coded multichannel audio signals, which have coded audio data of active frames, and first parameter noise data, second parameter noise data, and coherence data of inactive frames.

[0502] The present invention can also be implemented in a non-transitory storage unit that stores instructions, which, when executed by a computer (or processor, or controller), cause the computer (or processor, or controller) to perform the above-described method.

[0503] This invention can also be implemented in multichannel audio signals organized in a frame sequence, the frame sequence including active and inactive frames, the encoded multichannel audio signal including:

[0504] Encoded audio data of the active frame;

[0505] The first parameter noise data of the first channel in the inactive frame;

[0506] Second-parameter noise data of the second channel in inactive frames; and

[0507] Coherence data indicating the coherence between the first and second channels in an inactive frame. Multichannel audio signals can be obtained using one of the techniques disclosed above and / or below.

[0508] Advantages of the embodiments

[0509] Inserting a common noise source into both channels to simulate related noise and produce the final smooth noise is of great importance for simulating stereo background noise recording.

[0510] Embodiments of the present invention can also be considered as a process of generating soft noise for a stereo signal by mixing three Gaussian noise sources (one for each channel) and a third common noise source to create relevant background noise, or additionally or separately controlling the mixing of noise sources based on the coherence value transmitted with the SID frame, or additionally or separately, as follows: In a stereo system, generating background noise alone would result in completely irrelevant noise, which sounds unpleasant and is very different from the actual background, causing abrupt audio transitions when switching to the active mode background or from the active mode background to the DTX mode background. In the embodiment, on the encoder side, in addition to noise parameters, the coherence of the two channels is calculated, uniformly quantized, and added to the SID frame. In the decoder, the transmitted coherence value is then used to control CNG operation. Three Gaussian noise sources N_1, N_2, and N_3 are used; when the channel coherence is high, primarily relevant noise is added to the two channels, while when the coherence is low, more irrelevant noise is added.

[0511] It should be noted here that all the alternatives or aspects discussed above, as well as all aspects defined by the independent claims in the following claims, can be used individually; that is, there are no other alternatives or objectives other than the contemplated alternatives, objectives, or independent claims. However, in other embodiments, two or more alternatives or aspects or independent claims can be combined with each other, and in other aspects, all aspects or alternatives and all independent claims can be combined with each other.

[0512] The encoded signal of this invention can be stored on a digital storage medium or a non-transitory storage medium, or it can be transmitted on a transmission medium such as a wireless or wired transmission medium (e.g., the Internet).

[0513] Although some aspects have been described in the description of the equipment, it is clear that these aspects also represent the description of the corresponding method, where blocks or devices correspond to method steps or features of method steps. Similarly, the aspects described in the description of method steps also represent the description of corresponding blocks, items, or features of the corresponding equipment.

[0514] According to certain implementation requirements, embodiments of the present invention may be implemented using hardware or software, and the implementation may be performed using a digital storage medium, such as a floppy disk, DVD, CD, ROM, PROM, EPROM, EEPROM, or FLASH memory, having electronically readable control signals stored thereon, which cooperate with or are capable of cooperating with a programmable computer system to perform the corresponding methods.

[0515] Some embodiments of the present invention include a data carrier having electronically readable control signals that are capable of cooperating with a programmable calculator system to perform one of the methods described herein.

[0516] Typically, embodiments of the present invention can be implemented as a computer program product having program code that, when run on a computer, is operable to perform one of the methods. The program code may, for example, be stored on a machine-readable medium.

[0517] Other embodiments include a computer program for performing one of the methods described herein, stored on a machine-readable carrier or non-transitory storage medium.

[0518] In other words, embodiments of the method of the present invention are therefore computer programs with program code that, when run on a computer, perform one of the methods described.

[0519] Therefore, another embodiment of the method of the present invention is a data carrier (or digital storage medium, or computer-readable medium) that records a computer program for performing one of the methods.

[0520] Therefore, another embodiment of the method of the present invention is a data stream or signal sequence, which represents a computer program for performing one of the methods described. The data stream or signal sequence may, for example, be configured to be transmitted via a data communication connection (such as via the Internet).

[0521] Another embodiment includes a processing means, such as a computer or a programmable logic device, configured or adapted to perform one of the methods.

[0522] Another embodiment includes a computer having a computer program installed for performing one of the methods described.

[0523] In some embodiments, a programmable logic device (e.g., a field-programmable gate array) may be used to perform some or all of the functions described herein. In some embodiments, the field-programmable gate array may cooperate with a microprocessor to perform one of the methods described herein. Generally, these methods are preferably performed by any hardware device.

[0524] The above embodiments are merely illustrative of the principles of the present invention. It should be understood that the modifications and variations described herein will be readily apparent to those skilled in the art. Therefore, the scope of the present invention is limited by the scope of the claims that follow, and not solely by the specific details presented in the description and illustration of the embodiments.

[0525] Bibliography or references

[0526] [1]ITU-T G.729Annex B A silence compression scheme for G.729optimizedfor terminals conforming to ITU-T Recommendation V.70.InternationalTelecommunication Union(ITU)Series G,2007.

[0527] [2]ITU-T G.729.1Annex C DTX / CNG scheme.InternationalTelecommunication Union (ITU)Series G,2008.

[0528] [3]ITU-T G.718 Frame error robust narrow-band and wideband embeddedvariable bit-rate coding of speech and audio from 8-32 kbit / s.InternationalTelecommunication Union(ITU)Series G,2008.

[0529] [4]Mandatory Speech Codec speech processing functions;Adaptive Multi-Rate(AMR)speech codec;Transcoding functions,3GPP Technical Specification TS26.090,2014.

[0530] [5]Adaptive Multi-Rate-Wideband(AMR-WB)speech codec;Transcodingfunctions,3GPP,2014.

[0531] [6]3GPP TS 26.445,Codec for Enhanced Voice Services(EVS);Detailedalgorithmic description.

[0532] [7]Z.Wang and e.al,"Linear prediction based comfort noise generationin the EVS codec,"in IEEE International Conference on Acoustics,Speech andSignal Processing(ICASSP),Brisbane,QLD,2015.

[0533] [8]A.Lombard,S.Wilde,E.Ravelli,S. G.Fuchs and M.Dietz,"Frequency-domain Comfort Noise Generation for Discontinuous Transmission inEVS,"in IEEE International Conference on Acoustics,Speech and SignalProcessing(ICASSP),Brisbane,QLD,2015.

[0534] [9]A.Lombard,M.Dietz,S.Wilde,E.Ravelli,P.Setiawan and M.Multrus,"Generation of a comfort noise with high spectro-temporal resolution indiscontinuous transmission of audio signals".United States of America Patent9583114B2,19 June 2015.

[0535]

[10] E.NORVELL and F.JANSSON,"SUPPORT FOR GENERATION OF COMFORTNOISE.AND GENERATION OF COMFORT NOISE".WO Patent WO 2019 / 193149 A1,5 April2019.

Claims

1. A multi-channel signal generator (200) for generating a multi-channel signal (204) having a first channel (201) and a second channel (203), comprising: The first audio source (211) is used to generate the first audio signal (221). The second audio source (213) is used to generate the second audio signal (223); A mixed noise source (212) is used to generate a mixed noise signal (222); and A mixer (206) is used to mix the mixed noise signal (222) with the first audio signal (221) to obtain the first channel (201), and to mix the mixed noise signal (222) with the second audio signal (223) to obtain the second channel (203). The mixer (206) mentioned above includes: The first amplitude element (208-1) is used to influence the amplitude of the first audio signal (221); A first adder (206-1) is used to add at least a portion of the output signal (221) of the first amplitude element and the mixed noise signal (222); The second amplitude element (208-3) is used to influence the amplitude of the second audio signal (223); The second adder (206-3) is used to add at least a portion of the output (223) of the second amplitude element (208-3) and the mixed noise signal (222). Specifically, the influence exerted by the first amplitude element (208-1) is equal to the influence exerted by the second amplitude element (208-3), or the difference between the influence exerted by the second amplitude element (208-3) and the influence exerted by the first amplitude element (208-1) is less than 20% of the influence exerted by the first amplitude element (208-1). The mixer (206) includes a third amplitude element (208-2) for influencing the amplitude of the mixed noise signal (222). The amount of influence exerted by the third amplitude element (208-2) depends on the amount of influence exerted by the first amplitude element (208-1) or the second amplitude element (208-3), such that when the amount of influence exerted by the first amplitude element or the second amplitude element (208-3) decreases, the amount of influence exerted by the third amplitude element (208-2) increases.

2. The multi-channel signal generator as claimed in claim 1, wherein the first audio source (211) is a first noise source and the first audio signal (221) is a first noise signal, and / or the second audio source (213) is a second noise source and the second audio signal (223) is a second noise signal. in, The first noise source and / or the second noise source (213) are configured to generate the first noise signal (221) and / or the second noise signal (223) such that the first noise signal (221) and / or the second noise signal (223) are decorrelated with the mixed noise signal (222).

3. The multi-channel signal generator of claim 1, wherein the mixer (206) is configured to generate the first channel (201) and the second channel (203) such that the amount of the mixed noise signal (222) in the first channel (201) is equal to the amount of the mixed noise signal (222) in the second channel (203), or is in the range of 80% to 120% of the amount of the mixed noise signal (222) in the second channel (203).

4. The multichannel signal generator of claim 1, wherein the mixer (206) includes a control input for receiving control parameters (404, c), wherein the mixer (206) is configured to control the amount of the mixed noise signal (222) in the first channel (201) and in the second channel (203) according to the control parameters (404, c).

5. The multi-channel signal generator as claimed in claim 1, wherein each of the first audio source (211), the second audio source (213), and the mixed noise source (212) is a Gaussian noise source.

6. The multi-channel signal generator as described in claim 1, The first audio source (211) includes a first noise generator for generating the first audio signal (221) as a first noise signal, wherein the second audio source includes a decorrelation unit for decorrelation of the first noise signal (221) to generate the second audio signal as a second noise signal, and wherein the mixed noise source (212) includes a second noise generator.

7. The multi-channel signal generator as described in claim 1, The first audio source (211) includes a first noise generator for generating the first audio signal (221) as a first noise signal, the second audio source (213) includes a second noise generator for generating the second audio signal (223) as a second noise signal, and the mixed noise source (212) includes a decorrelation unit for decorrelating the first noise signal (221) or the second noise signal (223) to generate the mixed noise signal (222).

8. The multi-channel signal generator as described in claim 1, One of the first audio source (211), the second audio source (213), and the mixed noise source (212) includes a noise generator for generating a noise signal, and another of the first audio source (211), the second audio source (213), and the mixed noise source (212) includes a first decorrelation for decorrelation of the noise signal, and yet another of the first audio source (211), the second audio source (213), and the mixed noise source (212) includes a second decorrelation for decorrelation of the noise signal, wherein the first decorrelation and the second decorrelation are different from each other, such that the output signals of the first decorrelation and the second decorrelation are decorrelated to each other.

9. The multi-channel signal generator as described in claim 1, The first audio source (211) includes a first noise generator, the second audio source (213) includes a second noise generator, and the mixed noise source (212) includes a third noise generator, wherein the first noise generator, the second noise generator, and the third noise generator are configured to generate mutually decorrelated noise signals.

10. The multi-channel signal generator as described in claim 1, in, One of the first audio source (211), the second audio source (213), and the mixed noise source (212) includes a pseudo-random number sequence generator for generating a pseudo-random number sequence based on a seed, and at least two of the first audio source (211), the second audio source (213), and the mixed noise source (212) are configured to initialize the pseudo-random number sequence generator with different seeds.

11. The multi-channel signal generator as described in claim 1, At least one of the first audio source (211), the second audio source (213), and the mixed noise source (212) is configured to operate using a pre-stored noise table.

12. The multi-channel signal generator as described in claim 1, At least one of the first audio source (211), the second audio source (213), and the mixed noise source (212) is configured to generate a complex spectrum for a frame using a first noise value as the real part and a second noise value as the imaginary part.

13. The multi-channel signal generator as described in claim 1, The influence amount performed by the third amplitude element (208-2) is a predetermined value (c) q The square root of ), the influence amount performed by the first amplitude element (208-1) and the influence amount performed by the second amplitude element (208-3) are 1 and the predetermined value (c q The square root of the difference between .

14. The multi-channel signal generator as claimed in claim 1, further comprising: An input interface (210) is configured to receive encoded audio data (232) in a frame sequence (306, 308), the frame sequence (306, 308) including an active frame (306) and inactive frames (308) following the active frame (306); and An audio decoder (200', 200a, 200b) is used to decode the encoded audio data of the active frame (306) to generate a decoded multichannel signal of the active frame. The first audio source (211), the second audio source (213), the mixed noise source (212), and the mixer (206) are active in the inactive frame (308) to generate the multi-channel signal (204) of the inactive frame.

15. The multi-channel signal generator as claimed in claim 14, wherein: The coded audio data (232) of the active frame (306) has a first plurality of coefficients describing a first number of frequency windows; and The encoded audio data (232) of the inactive frame (308) has a second plurality of coefficients describing a second number of frequency windows. The first number of frequency windows is greater than the second number of frequency windows.

16. The multichannel signal generator of claim 14, wherein the audio data (232) of the inactive frame comprises: A first mute insert descriptor frame (241) for the first channel (201) and a second mute insert descriptor frame (243) for the second channel (203). The first mute insertion descriptor frame (241) includes: Soft noise parameter data (p_noise) for the first channel (201) and / or the first linear combination of the first channel and the second channel, and Soft noise generation auxiliary information (p_frame) for the first and second channels (203), and The second mute insertion descriptor frame (243) includes Soft noise parameter data (p_noise) for the second channel (203) and / or the second linear combination of the first channel and the second channel, and Coherence information (404, c) indicating the coherence between the first channel (201) and the second channel (203) of the inactive frame, and The multi-channel signal generator includes a controller for determining a soft noise generation mode for the first channel (201) and the second channel (203), and / or for a first linear combination of the first channel and the second channel and a second linear combination of the first channel and the second channel, using the soft noise generation auxiliary information (p_frame) of the first silence insertion descriptor frame (241); setting the coherence (404, c) between the first channel (201) and the second channel (203) in the inactive frame using the coherence information (404, c) in the second silence insertion descriptor frame (243); and setting the energy condition (v) of the first channel using the soft noise parameter data (p_noise) from the first silence insertion descriptor frame (241) and the soft noise parameter data (p-noise) from the second silence insertion descriptor frame (243). l, q ) and the energy status of the second channel (v r, q ), to control the generation of the multi-channel signal (204) in the inactive frame.

17. The multichannel signal generator of claim 16, wherein the audio data (232) of the inactive frame comprises: At least one mute insert descriptor frame (241) for a first linear combination of the first channel and the second channel and a second linear combination of the first channel and the second channel. Wherein, the at least one mute insertion descriptor frame (241) includes The soft noise parameter data (p_noise) used for the first linear combination of the first channel and the second channel, and The soft noise used for the second linear combination of the first channel and the second channel generates auxiliary information (p_frame). The multi-channel signal generator includes a controller for generating auxiliary information (p_frame) using the soft noise of the first linear combination of the first channel and the second channel and the second linear combination of the first channel and the second channel, using the coherence information (404, c) in the second silence insertion descriptor frame (243) to set the coherence between the first channel (201) and the second channel (203) in the inactive frame, and using the soft noise parameter data (p_noise) from the at least one silence insertion descriptor frame (241) and the soft noise parameter data (p_noise) from the at least one silence insertion descriptor frame (243) to set the energy condition (v) of the first channel. l, q and the energy status of the second channel (v) r, q ), to control the generation of the multi-channel signal (204) in the inactive frame.

18. The multichannel signal generator of claim 14, wherein the audio data of the inactive frame comprises: A mute insert descriptor frame (241, 243), wherein the mute insert descriptor frame (241, 243) includes soft noise parameter data (p_noise) for the first and second channels (201, 203) and soft noise generation auxiliary information (p_frame) for the first channel (201) and the second channel (203), and / or for a first linear combination of the first channel and the second channel and a second linear combination of the first channel and the second channel, and coherence information (404, c) indicating the coherence between the first channel (201) and the second channel (203) in the inactive frame, and The multi-channel signal generator (200) includes a controller for determining soft noise generation modes for the first channel (201) and the second channel (203) using the soft noise generation auxiliary information (p_frame) of the mute insertion descriptor frames (241, 243), setting the coherence (404, c) between the first channel (201) and the second channel (203) in the inactive frame using the coherence information (404, c) in the mute insertion descriptor frame (241), and setting the energy status (v) of the first channel using the soft noise parameter data (p_noise) from the mute insertion descriptor frames (241, 243). l, q ) and the energy status of the second channel (v r, q This is used to control the generation of the multi-channel signal in the inactive frame.

19. The multi-channel signal generator as described in claim 14, in, The encoded audio data (232) of the inactive frame includes silence insertion descriptor data (p_noise, c), which includes soft noise data (c, p_noise) indicating the signal energy of each channel in the center / side representation, and coherence data (404, c) indicating the coherence between the first channel and the second channel in the left / right representation, wherein the multichannel signal generator is configured to convert the signal energy in the center / side representation of the first channel and the second channel into the signal energy in the left / right representation. The mixer (206) is configured to mix the mixed noise signal (222) (206-1, 206-3) into the first audio signal (221) and the second audio signal (223) based on the coherence data (404, c) to obtain the first channel (201) and the second channel (203), and The multi-channel signal generator further includes a signal modifier (250) configured to modify the first and second channels (201, 203) by shaping the first and second channels (201, 203) based on the signal energy in the left / right domain.

20. The multichannel signal generator of claim 19, configured to, when the audio data contains a signal indicating that the energy in the side channel in the center / side representation is less than a predetermined threshold, generate the side channel (v s, q The coefficient of ) is zero.

21. The multichannel signal generator of claim 19, wherein the audio data of the inactive frame comprises: At least one mute insert descriptor frame (241, 243), wherein the at least one mute insert descriptor frame (241, 243) includes the center channel and side channel (v) for the center / side representation. m, q , v s, q The soft noise parameter data (p_noise, v) m, ind , q l,q , q r,q , v s, ind ) and for the middle channel and the side channels (v m, q , v s, q The soft noise generates auxiliary information (p_frame), and coherence information (404, c) indicating the coherence between the first channel (201) and the second channel (203) in the inactive frame, and The multi-channel signal generator (200) includes a controller for determining soft noise generation modes for the first channel (201) and the second channel (203) using the soft noise generation assistance information (p_frame) of the mute insert descriptor frames (241, 243), setting the coherence (404, c) between the first channel (201) and the second channel (203) in the inactive frame using the coherence information (404, c) in the mute insert descriptor frames (241), and setting the energy condition (v) of the first channel using the soft noise parameter data (p_noise) or its processed version from the mute insert descriptor frames (241, 243). l, q ) and the energy status of the second channel (v r, q This is used to control the generation of the multi-channel signal in the inactive frame.

22. The multichannel signal generator of claim 14 is further configured to transmit gain information (g) l,q , q r,q ) Scaling the signal energy coefficients of the first and second channels (v') l , v' r The gain information is encoded using soft noise parameter data (401, 403) of the first and second channels.

23. The multi-channel signal generator of claim 1, configured to convert the generated multi-channel signal (252) from a frequency domain version to a time domain version.

24. The multi-channel signal generator as claimed in claim 1, wherein the first audio source (211) is a first noise source and the first audio signal (221) is a first noise signal, or the second audio source (213) is a second noise source and the second audio signal (223) is a second noise signal. in, The first noise source or the second noise source is configured to generate the first noise signal or the second noise signal such that the first noise signal or the second noise signal is at least partially correlated. The mixed noise source (212) is configured to generate a mixed noise signal (222) having a first mixed noise component (221a) and a second mixed noise component (221b), wherein the second mixed noise component (221b) is at least partially decorrelated with the first mixed noise component (221a); ​​and The mixer (206) is configured to mix the first mixed noise portion (221a) of the mixed noise signal (222) with the first audio signal (221) to obtain the first channel (201), and to mix the second mixed noise portion (221b) of the mixed noise signal (222) with the second audio signal (223) to obtain the second channel (203).

25. A method for generating a multi-channel signal having a first channel and a second channel (203), comprising: A first audio signal (221) is generated using a first audio source (211); The second audio signal (223) is generated using the second audio source (213); A mixed noise signal (222) is generated using a mixed noise source (212); and The method includes mixing (206) the mixed noise signal (222) with the first audio signal (221) to obtain the first channel (201), and mixing the mixed noise signal (222) with the second audio signal (223) to obtain the second channel, the method comprising: The amplitude of the first audio signal (221) is affected by the first amplitude element (208-1); The first adder (206-1) is used to add at least a portion of the output signal (221) of the first amplitude element and the mixed noise signal (222); The amplitude of the second audio signal (223) is affected by the second amplitude element (208-3); The output (223) of the second amplitude element (208-3) and at least a portion of the mixed noise signal (222) are added using a second adder (206-3). Specifically, the influence exerted by the first amplitude element (208-1) is equal to the influence exerted by the second amplitude element (208-3), or the difference between the influence exerted by the second amplitude element (208-3) and the influence exerted by the first amplitude element (208-1) is less than 20% of the influence exerted by the first amplitude element (208-1). The mixing (206) uses a third amplitude element (208-2) to affect the amplitude of the mixed noise signal (222). The amount of influence exerted by the third amplitude element (208-2) depends on the amount of influence exerted by the first amplitude element (208-1) or the second amplitude element (208-3), such that when the amount of influence exerted by the first amplitude element or the second amplitude element (208-3) decreases, the amount of influence exerted by the third amplitude element (208-2) increases.

Images