Loudness adjustment for downmixed audio content

Abstract

To provide a loudness adjustment method for downmixed audio content.SOLUTION: A method for downmixing audio content coded for a reference speaker configuration to downmix audio content coded for a specific speaker configuration includes: performing one or more gain adjustments on individual portions of the downmix audio content coded for the specific speaker configuration; performing loudness measurements on the individual portions of the downmix audio content; and generating an audio signal that comprises the audio content coded for the reference speaker configuration and downmix loudness metadata, the downmix loudness metadata being created based at least in part on the loudness measurements on the individual portions of the downmix audio content.SELECTED DRAWING: Figure 3

Description

[Technical field]

[0001] CROSS-REFERENCE TO RELATED APPLICATIONS This application claims priority to U.S. Provisional Patent Application No. 61 / 877,230, filed September 12, 2013, U.S. Provisional Patent Application No. 61 / 891,324, filed October 15, 2013, U.S. Provisional Patent Application No. 61 / 938,043, filed February 10, 2014, and U.S. Provisional Patent Application No. 61 / 892,313, filed October 17, 2013, the contents of each of which are incorporated herein by reference in their entirety.

[0002] technology The present invention relates generally to processing audio signals, and more particularly to techniques that may be used to apply dynamic range control and other types of audio processing operations to audio signals in any of a wide variety of playback environments. [Background technology]

[0003] The growing popularity of media consumption devices has created new opportunities and challenges for creators and distributors of media content for playback on such devices, and for designers and manufacturers of such devices. Many consumer devices are capable of playing a wide range of media content types and formats, including those often associated with high quality, high bandwidth, and wide dynamic range audio content for HDTV, Blu-ray, or DVD. Media processing devices may be used to play this type of audio content over their own internal acoustic transducers or over external transducers such as headphones. However, media processing devices are generally unable to play this content with consistent loudness and intelligibility across a variety of media formats and content types.

[0004] The approaches described in this section are approaches that could be pursued, but not necessarily approaches that have been previously conceived or pursued. Thus, unless otherwise noted, it should not be assumed that any of the approaches described in this section qualify as prior art by virtue of their inclusion in this section. Similarly, unless otherwise noted, it should not be assumed that problems identified with one or more approaches have been recognized in any prior art based on this section. [Brief description of the drawings]

[0005] The present invention is illustrated by way of example, and not by way of limitation, in the accompanying figures in which like reference symbols refer to similar elements and in which: [Figure 1A] FIG. 2 illustrates an exemplary audio decoder. [Figure 1B] FIG. 2 illustrates an exemplary audio encoder. [Figure 2A] FIG. 2 illustrates an exemplary dynamic range compression curve. [Figure 2B] FIG. 2 illustrates an exemplary dynamic range compression curve. [Diagram 3] FIG. 1 illustrates an example processing logic for combined DRC and limiting gain determination / calculation. [Figure 4] FIG. 1 illustrates an exemplary differential encoding of gain. [Diagram 5] FIG. 1 illustrates an exemplary codec system having an audio encoder and an audio decoder. [Figure 6A] FIG. 1 illustrates an exemplary process flow. [Figure 6B] FIG. 1 illustrates an exemplary process flow. [Figure 6C] FIG. 1 illustrates an exemplary process flow. [Figure 6D] FIG. 1 illustrates an exemplary process flow. [Figure 7]FIG. 1 illustrates an exemplary hardware platform upon which a computer or computing device described herein may be implemented. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0006] Exemplary embodiments relating to applying dynamic range control and other types of audio processing operations to audio signals in any of a wide variety of playback environments are described herein. In the following description, for purposes of explanation, numerous specific details are set forth in order to provide a thorough understanding of the present invention. It will be apparent, however, that the present invention may be practiced without such specific details. In other instances, well-known structures and devices are not described in exhaustive detail in order to avoid unnecessarily obscuring, obscuring, or obscuring the present invention.

[0007] Exemplary embodiments are described herein according to the following outline. 1. General Overview 2. Dynamic range control 3. Audio Decoder 4. Audio Encoder 5. Dynamic range compression curve 6. DRC Gain, Gain Limiting and Gain Smoothing 7. Input and Gain Smoothing 8. DRC across multiple frequency bands 9. Volume adjustment in the loudness area 10. Downmix & Loudness Adjustment 11. Additional Gain-Related Actions 12. Specific and broadband loudness levels 13. Individual gains for individual subsets of channels 14. Auditory Scene Analysis 15. Loudness Level Transition 16. Reset 17. Gain provided by the encoder 18. Exemplary System and Process Flows 19. Implementation Mechanism - Hardware Overview 20. Equivalents, Extensions, Substitutions, etc.

[0008] 1. General Overview This overview provides a basic description of some aspects of the embodiments of the present invention. It should be noted that this overview is not a comprehensive or exhaustive summary of the aspects of the embodiments. Moreover, it should be noted that this overview is not intended to be understood as identifying any particularly significant aspects or elements of the embodiments, or as defining any scope of the present invention, particularly the embodiments, in general. This overview merely presents some concepts related to the exemplary embodiments in a condensed and simplified form, and should be understood merely as a conceptual introduction to the more detailed description of the exemplary embodiments that follow. It should be noted that although separate embodiments are discussed herein, any combination of the embodiments and / or sub-embodiments discussed herein may be combined to form further embodiments.

[0009] In some approaches, the encoder assumes that the audio content is encoded for a particular environment for the purposes of dynamic range control and determines audio processing parameters, such as gains for dynamic range control, for that particular environment. The gains determined by the encoder under these approaches are typically smoothed over some time interval, with some time constant (e.g., in an exponential decay function, etc.). Additionally, the gains determined by the encoder under these approaches may incorporate gain limiting to ensure that the loudness level does not exceed a clipping level for the assumed environment. Thus, the gains encoded by the encoder under these approaches into the audio signal along with the audio information are the result of many different effects and are lossy. A decoder receiving the gains under these approaches would not be able to distinguish which parts of the gain are for dynamic range control, which parts of the gain are for gain smoothing, which parts of the gain are for gain limiting, etc.

[0010] Under the techniques described herein, the audio encoder does not assume that a particular playback environment at the audio decoder only needs to be supported. In one embodiment, the audio encoder transmits an encoded audio signal having audio content from which a correct loudness level (e.g., no clipping, etc.) can be determined. The audio encoder may also transmit one or more dynamic range compression curves to the audio decoder. Any of the one or more dynamic range compression curves may be standard-based, proprietary, customized, content provider specific, etc. Reference loudness levels, attack times, release times, etc. may be transmitted by the audio encoder as part of or in association with the one or more dynamic range compression curves.

[0011] In some embodiments, the audio encoder implements an auditory scene analysis (ASA) technique, uses the ASA technique to detect auditory events in the audio content, and sends one or more ASA parameters describing the detected auditory events to the audio decoder.

[0012] In some embodiments, the audio encoder may also be configured to detect a reset event in the audio content and send an indication of the reset event together with the audio content in a time-synchronous manner to a downstream device, such as an audio decoder.

[0013] In some embodiments, an audio encoder may be configured to calculate one or more sets of gains (e.g., DRC gains) for individual portions of audio content (e.g., audio data blocks, audio data frames, etc.) and encode the set of gains together with the individual portions of audio content into an encoded audio signal. In some embodiments, the set of gains generated by the audio encoder corresponds to one or more gain profiles. In some embodiments, Huffman coding, differential coding, etc. may be used to code the set of gains into or read the set of gains from components, subdivisions, etc. of audio data frames. These components, subdivisions, etc. may be referred to as subframes of an audio data frame. Different sets of gains may correspond to different sets of subframes. Each set of gains or each set of subframes may have two or more temporal components (e.g., subframes, etc.). In some embodiments, a bitstream formatter in an audio encoder described herein may use one or more for loops to write one or more sets of gains together as differential data codes into one or more sets of subframes in an audio data frame, and correspondingly, a bitstream parser in an audio decoder described herein may read any of the one or more sets of gains encoded as the differential data codes from the one or more sets of subframes in an audio data frame.

[0014] In some embodiments, the audio encoder determines a dialogue loudness level in the audio content to be encoded into the encoded audio signal and sends the dialogue loudness level together with the audio content to the audio decoder.

[0015] In some embodiments, audio content is encoded for a reference speaker configuration (e.g., surround sound configuration, 5.1 speaker configuration, etc.) that has more audio channels or speakers in the encoded audio signal than many audio decoders (e.g., mobile phones, tablet computers, etc.) will operate with (e.g., a two-channel headset configuration, etc.) in. Loudness levels measured in the reference speaker configuration for individual portions of the audio content may differ from loudness levels measured in a particular speaker configuration, such as a two-channel configuration, for the same individual portions of the audio content, even if the same gain adjustments are made in both speaker configurations.

[0016] In some embodiments, the audio encoder described herein is configured to provide downmix-related metadata (e.g., including one or more downmix loudness parameters) to a downstream audio decoder. The downmix-related metadata from the audio encoder (150) can be used by the downstream audio decoder to efficiently and consistently perform additional downmix-related gain adjustment operations (e.g., in real-time, near real-time, etc.) to generate a relatively accurate target loudness level in the downmix sound output. The additional downmix-related gain adjustment operations can be used by the downstream audio decoder to prevent mismatches in measured loudness levels between a reference speaker configuration and a particular speaker configuration of the decoder.

[0017] The techniques described herein do not require that the audio decoder be locked-in with audio processing (e.g., lossy) that may have been performed by an upstream device such as an audio encoder, while assuming an assumed playback environment, scenario, etc. in an assumed audio decoder. The decoders described herein may be configured to customize audio processing operations based on a particular playback scenario, e.g., to distinguish between various loudness levels present in the audio content, minimize loss of audio perceptual quality at or near boundary loudness levels, maintain spatial balance among channels or subsets of channels, etc.

[0018] An audio decoder receiving an encoded audio signal with dynamic range compression curves, reference loudness levels, attack times, release times, etc. can determine the particular playback environment being used at the decoder and select a particular compression curve with a corresponding reference loudness level that corresponds to that particular playback environment.

[0019] The decoder may calculate / determine loudness levels for individual portions of the audio content (e.g., audio data blocks, audio data frames, etc.) extracted from the encoded audio signal, or obtain loudness levels for individual portions of the audio content if the audio encoder calculated and provided loudness levels in the encoded audio signal. Based on one or more of the loudness levels for the individual portions of the audio content, the loudness levels for previous portions of the audio content, the loudness levels for subsequent portions of the audio content if available, the particular compression curve, a particular profile related to the particular playback environment or scenario, etc., the decoder determines audio processing parameters such as gain for dynamic range control (DRC gain), attack time, release time, etc. The audio processing parameters may also include adjustments to align the dialogue loudness level to a particular reference loudness level for the particular playback environment (which may be user adjustable).

[0020] The decoder applies audio processing operations with the audio processing parameters, including dynamic range control (e.g., multi-channel, multi-band, etc.), dialogue level adjustment, etc. Audio processing operations performed by the decoder may further include, but are not limited to, gain smoothing based on attack and release times provided as part of or in conjunction with a selected dynamic range compression curve, gain limiting to prevent clipping, etc. Different audio processing operations may be performed with different (e.g., adjustable, threshold-dependent, controllable, etc.) time constants. For example, gain limiting to prevent clipping may be applied to individual audio data blocks, individual audio data frames, etc. with a relatively short time constant (e.g., instantaneous, on the order of about 5.3 milliseconds, etc.).

[0021] In some embodiments, the decoder may be configured to extract ASA parameters (e.g., temporal positions of auditory event boundaries, time-dependent values ​​of event certainty measures, etc.) from metadata in the encoded audio signal and to control the rate of gain smoothing at auditory events based on the extracted ASA parameters (e.g., using a short time constant for the attack at auditory event boundaries, a long time constant to slow down gain smoothing within auditory events, etc.).

[0022] In some embodiments, the decoder also maintains a histogram of instantaneous loudness levels over a time interval or window and uses the histogram to control the rate of gain change at loudness level transitions between programs, between programs and commercials, etc., e.g. by modifying the time constant.

[0023] In some embodiments, the decoder supports two or more speaker configurations (e.g., portable mode with speakers, portable mode with headphones, stereo mode, multi-channel mode, etc.). The decoder may be configured to maintain the same loudness level between two different speaker configurations (e.g., between stereo mode and multi-channel mode), for example, when playing the same audio content. The audio decoder may use one or more downmix equations to downmix multi-channel audio content received from an encoded audio signal for a reference speaker configuration to a particular speaker configuration in the audio decoder, the multi-channel audio content being encoded for the reference speaker configuration.

[0024] In some embodiments, automatic gain control (AGC) may be disabled in the audio decoders described herein.

[0025] In some embodiments, the device may be part of a media processing system, including, but not limited to, audiovisual devices, flat panel TVs, handheld devices, game consoles, televisions, home theater systems, tablets, mobile devices, laptop computers, netbook computers, cellular wireless telephones, e-readers, point-of-sale terminals, desktop computers, computer workstations, computer kiosks, and various other types of terminals and media processing units.

[0026] Various modifications to the preferred embodiment and general principles and features described herein will be readily apparent to those skilled in the art, and thus the present disclosure is not intended to be limited to the embodiments shown but is to be accorded the widest scope consistent with the principles and features described herein.

[0027] 2. Dynamic range control Without customized dynamic range control, the input audio information (e.g., PCM samples, time-frequency samples in a QMF matrix, etc.) is often reproduced at a loudness level that is inappropriate for the particular playback environment of the playback device (i.e., including the physical and / or mechanical playback limitations of the device), which may differ from the playback environment that was targeted when the encoded audio content was coded in the encoding device.

[0028] The techniques described in this document can be used to support dynamic range control of a wide variety of audio content customized to any of a wide variety of playback environments, while maintaining the perceptual quality of the audio content.

[0029] Dynamic range control (DRC) refers to a time-dependent audio processing operation that changes (e.g., compresses, cuts, expands, boosts, etc.) an input dynamic range of loudness levels in audio content to an output dynamic range that is different from the input dynamic range. For example, in a dynamic range control scenario, soft sounds may be mapped (e.g., boosted, etc.) to higher loudness levels and loud sounds may be mapped (e.g., cut, etc.) to lower loudness values. As a result, in the loudness domain, in this example, the output range of loudness levels is smaller than the input range of loudness levels. However, in some embodiments, the dynamic range control may be reversible such that the original range is restored. For example, an expansion operation may be performed to restore the original range as long as the mapped loudness levels in the output dynamic range mapped from the original loudness levels are below the clipping level, each unique original loudness level is mapped to a unique output loudness level, etc.

[0030] The DRC techniques described herein can be used to provide a better listening experience in certain playback environments or situations. For example, soft sounds in a noisy environment may be masked by noise that makes the soft sounds inaudible. Conversely, loud sounds may be undesirable in some situations, such as noisy neighbors. Many devices, typically with small form factor loudspeakers, cannot reproduce sounds at high output levels. In some cases, lower signal levels may be reproduced below the human hearing threshold. The DRC techniques may perform a mapping of input loudness levels to output loudness levels based on DRC gains (e.g., scaling factors for scaling audio amplitudes, boost ratios, cut ratios, etc.) found using a dynamic range compression curve.

[0031] A dynamic range compression curve is a function (e.g., a look-up table, a curve, a multi-segment piecewise linear function, etc.) that maps individual input loudness levels (e.g., non-dialogue sounds) determined from individual audio data frames to individual gains or gains for dynamic range control. Each individual gain indicates the amount of gain to be applied to the corresponding individual input loudness level. The output loudness level after applying the individual gains represents the target loudness level for the audio content in that individual audio data frame in a particular playback environment.

[0032] In addition to specifying a mapping between gains and loudness levels, the dynamic range compression curve may include or be provided with specific release and attack times when applying a particular gain. Attack refers to an increase in signal energy (or loudness) between successive time samples, whereas release refers to a decrease in signal energy (or loudness) between successive time samples. The attack time (e.g., 10 ms, 20 ms, etc.) refers to a time constant used in smoothing the DRC gain when the corresponding signal is in attack mode. The release time (e.g., 80 ms, 100 ms, etc.) refers to a time constant used in smoothing the DRC gain when the corresponding signal is in release mode. In some embodiments, these time constants are additionally, optionally, or alternatively used for smoothing the signal energy (loudness) before determining the DRC gain.

[0033] Different playback environments may correspond to different dynamic range compression curves. For example, a dynamic range compression curve for a playback environment of a flat panel TV may be different from a dynamic range compression curve for a playback environment of a portable device. In some embodiments, a playback device may have more than one playback environment. For example, a first dynamic range compression curve for a first playback environment of a portable device using speakers may be different from a second dynamic range compression curve for a second playback environment of the same portable device using a headset.

[0034] 3. Audio Decoder FIG. 1A shows an exemplary audio decoder 100 having a data extractor 104, a dynamic range controller 106, an audio renderer 108, etc.

[0035] In some embodiments, the data extractor (104) is configured to receive the encoded input signal 102. The encoded input signal as described herein may be a bitstream including encoded (e.g., compressed) input audio data frames and metadata. The data extractor (104) is configured to extract / decode the input audio data frames and metadata from the encoded input signal (102). Each of the input audio data frames comprises a number of encoded audio data blocks, each of which represents a number of audio samples. Each frame represents a (e.g., fixed) time interval that includes a certain number of audio samples. The frame size may vary with the sample rate and the encoding data rate. An audio sample is a quantized audio data element (e.g., an input PCM sample, an input time-frequency sample in a QMF matrix, etc.) that represents a spectral content in one, two or more (audio) frequency bands or frequency ranges. The quantized audio data elements in the input audio data frames may represent pressure waves in the digital (quantized) domain. Quantized audio data elements may cover a finite range of loudness levels up to a maximum possible value (e.g. clipping level, maximum loudness level, etc.).

[0036] The metadata can be used by a wide variety of receiving decoders to process the input audio data frames. The metadata may include various operational parameters related to one or more operations to be performed by the decoder (100), normalization parameters related to the dialogue loudness level represented in the input audio data frames, etc. Dialogue loudness level may refer to the dialogue loudness, program loudness, average dialogue loudness, etc. (e.g., psychoacoustic, perceptual, etc.) level of an entire program (e.g., a movie, television program, radio broadcast, etc.), a portion of a program, dialogue in a program, etc.

[0037] The operation and functionality of the decoder (100) or some or all of its modules (e.g., data extractor 104, dynamic range controller 106, etc.) may be adapted in response to metadata extracted from the encoded input signal (102). For example, the metadata, including but not limited to dynamic range compression curves, dialogue loudness levels, etc., may be used by the decoder (100) to generate output audio data elements in the digital domain (e.g., output PCM samples, output time-frequency samples in a QMF matrix, etc.). The output data elements can then be used to drive audio channels or speakers to achieve a specified loudness or reference playback level during playback in a particular playback environment.

[0038] In some embodiments, the dynamic range controller (106) is configured to receive some or all of the audio data elements and metadata in the input audio data frames, and to perform audio processing operations (e.g., dynamic range control operations, gain smoothing operations, gain limiting operations, etc.) on the audio data elements in the input audio data frames based at least in part on the metadata extracted from the encoded audio signal (102).

[0039] In some embodiments, the dynamic range controller (106) may include a selector 110, a loudness calculator 112, a DRC gain unit 114, etc. The selector (110) may be configured to determine a speaker configuration associated with a particular playback environment in the decoder (100) (e.g., flat panel mode, portable device with speakers, portable device with headphones, 5.1 speaker configuration, 7.1 speaker configuration, etc.), select a particular dynamic range compression curve from the dynamic range compression curves extracted from the encoded input signal (102), etc.

[0040] The loudness calculator (112) may be configured to calculate one or more types of loudness levels represented by audio data elements in the input audio data frame. Exemplary types of loudness levels include, but are not limited to, any of the following: individual loudness levels across individual frequency bands in individual channels across individual time intervals, broadband loudness levels across a wide frequency range in individual channels, loudness levels determined from or smoothed across an audio data block or frame, loudness levels determined from or smoothed across two or more audio data blocks or frames, loudness levels smoothed across one or more time intervals, etc. Zero, one or more of these loudness levels may be modified by the decoder (100) for dynamic range control.

[0041] To determine the loudness level, the loudness calculator (112) can determine one or more time-dependent physical wave attributes, such as spatial pressure levels at specific audio frequencies, represented by the audio data elements in the input audio data frames. The loudness calculator (112) can use the one or more time-varying physical wave attributes to derive one or more types of loudness levels based on one or more psychoacoustic functions that model human loudness perception. The psychoacoustic functions can be nonlinear functions, such as those constructed based on a model of the human auditory system, that convert specific spatial pressure levels at specific audio frequencies into specific loudness for the specific audio frequencies.

[0042] A loudness level across multiple (audio) frequencies or frequency bands (e.g., broadband, wideband, etc.) may be derived through integration of a specific loudness level across multiple (audio) frequencies or frequency bands. A time-averaged, smoothed, etc. loudness level across one or more time intervals (e.g., longer than represented by an audio data element in an audio data block or frame) may be obtained using one or more smoothing filters implemented as part of the audio processing operations in the decoder (100).

[0043] In an example embodiment, specific loudness levels for different frequency bands may be calculated for each audio data block of a certain (e.g., 256) samples. A pre-filter may be used to apply frequency weighting (e.g., similar to IEC B weighting) to the specific loudness levels in integrating them into a broadband loudness level. A summation of broad loudness levels across two or more channels (e.g., left front, right front, center, left surround, right surround, etc.) may be performed to provide an overall loudness level for the two or more channels.

[0044] In some embodiments, an overall loudness level may refer to a broadband loudness level in a channel (e.g., center) of a speaker configuration. In some embodiments, an overall loudness level may refer to a broadband loudness level in multiple channels. The multiple channels may be all channels in a speaker configuration. Additionally, optionally, or alternatively, the multiple channels may include a subset of channels in a speaker configuration (e.g., a subset of channels including left front, right front, and low frequency effects (LFE), a subset of channels including left surround and right surround, etc.).

[0045] The loudness levels (e.g. broadband, wideband, global, specific, etc.) may be used as input to find the corresponding DRC gains (e.g. static, pre-smoothed, pre-limited, etc.) from the selected dynamic range compression curve. The loudness levels used as input to find the DRC gains may first be adjusted or normalized with respect to the dialogue loudness levels from the metadata extracted from the encoded audio signal (102).

[0046] In some embodiments, the DRC gain unit (114) may be configured with a DRC algorithm to generate a gain (e.g., for dynamic range control, for gain limiting, for gain smoothing, etc.) and apply the gain to one or more loudness levels at one or more types of loudness levels represented by audio data elements in the input audio data frame to achieve a target loudness level for the particular playback environment, etc. The application of a gain (e.g., DRC gain, etc.) as described herein may occur in the loudness domain, although this is not required. In some embodiments, the gain may be generated based on a loudness calculation (which may be expressed in terms of SPL compensated for sones or simply, for example, dialogue loudness levels without transformation), smoothed, and applied directly to the input signal. In some embodiments, the techniques as described herein may apply a gain to a signal in the loudness domain, then convert the signal back from the loudness domain to the (linear) SPL domain, and calculate the corresponding gain to be applied to the signal by evaluating the signal before and after the gain has been applied to the signal in the loudness domain. The ratio (or difference when expressed in log dB terms) then determines the corresponding gain for that signal.

[0047] In some embodiments, the DRC algorithm works with multiple DRC parameters, including a dialogue loudness level that has already been calculated and embedded in the encoded audio signal (102) by an upstream encoder (e.g., 150) and can be obtained by the decoder (100) from metadata in the encoded audio signal (102). The dialogue loudness level from the upstream encoder indicates an average dialogue loudness level (e.g., per program, relative to the energy of a full-scale 1 kHz sine wave, relative to the energy of a reference square wave, etc.). In some embodiments, the dialogue loudness level extracted from the encoded audio signal (102) may be used to reduce loudness level differences between programs. In one embodiment, the reference dialogue loudness level may be set to the same value between different programs in the same specific playback environment in the decoder (100). Based on the dialogue loudness level from the metadata, a DRC gain unit (114) can apply a dialogue loudness related gain to each audio data block in the program such that the output dialogue loudness level averaged across multiple audio data blocks of the program is raised / lowered to a reference dialogue loudness level (e.g. pre-configured, system default, user configurable, profile dependent, etc.) for that program.

[0048] In some embodiments, DRC gains may be used to address differences in loudness levels within a program by boosting or cutting input loudness levels in soft and / or loud sounds according to a selected dynamic range compression curve. One or more of these DRC gains may be calculated / determined by a DRC algorithm based on a selected dynamic range compression curve and loudness levels (broadband, wideband, global, specific, etc.) determined from one or more corresponding audio data blocks, audio data frames, etc.

[0049] The loudness level used to determine the DRC gain (e.g., static, unsmoothed, ungain-limited, etc.) by searching the selected dynamic range compression curve may be calculated over a short interval (e.g., about 5.3 ms). The integration time of the human auditory system can be much longer (e.g., about 200 ms). The DRC gain obtained from the selected dynamic range compression curve may be smoothed with a time constant to take into account the long integration time of the human auditory system. To implement a fast rate of change (increase or decrease) in loudness level, a short time constant may be used to cause the loudness level to change over a short time interval corresponding to the short time constant. Conversely, to implement a slow rate of change (increase or decrease) in loudness level, a long time constant may be used to cause the loudness level to change over a long time interval corresponding to the long time constant.

[0050] The human auditory system may respond to increasing and decreasing loudness levels with different integration times. In some embodiments, different time constants may be used to smooth the static DRC gain retrieved from the selected dynamic range compression curve depending on whether the loudness level is increasing or decreasing. For example, attacks (increases in loudness level) may be smoothed with a relatively short time constant (e.g., attack time) while releases (decrease in loudness level) may be smoothed with a relatively long time constant (e.g., release time) to match the characteristics of the human auditory system.

[0051] A DRC gain for a portion of audio content (e.g., one or more of an audio data block, an audio data frame, etc.) may be calculated using a loudness level determined from said portion of audio content. The loudness level to be used for the lookup in the selected dynamic range compression curve may first be adjusted with respect to (e.g., relative to) a dialogue loudness level (e.g., of a program of which the audio content is a part) in metadata extracted from the encoded audio signal (102).

[0052] Reference dialogue loudness level (e.g. -31dB in "Line" mode) FS , -20dB in "RF" mode FS ) may be specified or established for a particular playback environment in the decoder (100). Additionally, alternatively or optionally, in some embodiments, a user may be given control over setting or changing a reference dialogue loudness level in the decoder (100).

[0053] The DRC gain unit (114) can be configured to determine a dialogue loudness related gain for the audio content to cause a change from an input dialogue loudness level to a reference dialogue loudness level as an output dialogue loudness level.

[0054] In some embodiments, the DRC gain unit (114) may be configured to handle peak levels in a particular playback environment in the decoder (100) and adjust the DRC gain to prevent clipping. In some embodiments, under the first approach, if the audio content extracted from the encoded audio signal (102) includes audio data elements for a reference multi-channel configuration with more channels than the channels of the particular speaker configuration in the decoder, a particular speaker configuration downmix may be performed from the reference multi-channel configuration before determining and processing the peak levels to prevent clipping. Additionally, optionally or alternatively, in some embodiments, under the second approach, if the audio content extracted from the encoded audio signal (102) includes audio data elements for a reference multi-channel configuration with more channels than the channels of the particular speaker configuration in the decoder, a downmix formula (e.g., ITU stereo downmix, matrixed-surround compatible downmix, etc.) may be used to obtain peak levels for the particular speaker configuration in the decoder (100). The peak level may be adjusted to reflect the change from the input dialogue loudness level to the reference dialogue loudness level as the output dialogue loudness level. The maximum allowable gain without causing clipping (e.g., for an audio data block, for an audio data frame, etc.) may be determined based at least in part on the inverse of the peak level (e.g., multiplied by −1). In this manner, an audio decoder under the techniques described herein can be configured to accurately determine the peak level and apply clipping prevention specifically for the decoder-side playback configuration. Neither the audio decoder nor the audio encoder need make assumptions about the worst-case scenario at a hypothetical decoder.In particular, the decoder in the first approach above can accurately determine peak levels and apply post-downmix clipping prevention without using the downmix equations, downmix channel gains, etc. (which are used under the second approach as described above).

[0055] In some implementations, the combination of the dialogue loudness level and adjustments to the DRC gain prevents peak level clipping even in the worst case downmix (e.g., one that produces the largest peak level after the downmix, one that produces the largest downmix channel gain, etc.). However, in some other embodiments, the combination of the dialogue loudness level and adjustments to the DRC gain may not be sufficient to prevent peak level clipping. In these embodiments, the DRC gain may be replaced (e.g., capped, etc.) by the highest gain that prevents clipping at the peak levels.

[0056] In some embodiments, the DRC gain unit (114) is configured to derive time constants (e.g., attack times, release times, etc.) from metadata extracted from the encoded audio signal (102). The DRC gains, time constants, maximum allowed gain, etc. may be used by the DRC gain unit (114) to perform DRC, gain smoothing, gain limiting, etc.

[0057] For example, the application of the DRC gain may be smoothed with a filter controlled by a time constant. The gain limiting operation may be implemented by a min() function that takes the smaller of the gain to be applied and the maximum allowed gain for the gain. Through this function, the gain (e.g., before limiting, DRC, etc.) may be replaced by the maximum allowed gain immediately, over a relatively short time interval, etc., thereby preventing clipping.

[0058] In some embodiments, the audio renderer (108) is configured to generate channel-specific (e.g., multi-channel) audio data (116) for the particular speaker configuration after applying a gain determined based on DRC, gain limiting, gain smoothing, etc. to the input audio data extracted from the encoded audio signal (102). The channel-specific audio data (118) may be used to drive speakers, headphones, etc. represented in the speaker configuration.

[0059] Additionally and / or optionally, in some embodiments, the decoder (100) may be configured to perform one or more other operations related to pre-processing, post-processing, rendering, etc., relating to the input audio data.

[0060] The techniques described herein can be used with a variety of speaker configurations corresponding to a variety of different surround sound configurations (e.g., 2.0, 3.0, 4.0, 4.1, 4.1, 5.1, 6.1, 7.1, 7.2, 10.2, 10-60 speaker configurations, 60+ speaker configurations, etc.) and a variety of different rendering environment configurations (e.g., movie theaters, parks, opera houses, concert halls, bars, homes, auditoriums, etc.).

[0061] 4. Audio Encoder 1B illustrates an exemplary encoder 150. The encoder (150) may include an audio content interface 152, a dialogue loudness analyzer 154, a DRC reference store 156, an audio signal encoder 158, etc. The encoder 150 may be part of a broadcast system, an Internet-based content server, an over-the-air network operator system, a movie production system, etc.

[0062] In some embodiments, the audio content interface (152) is configured to receive audio content 160, audio content control inputs 162, etc., and to generate an encoded audio signal (e.g., 102) based at least in part on the audio content (160), the audio content control inputs (162), etc. For example, the audio content interface (152) may be used to receive the audio content (160), the audio content control inputs (162), etc. from a content creator, a content provider, etc.

[0063] The audio content may be part of or in whole of an overall media data set that may include audio only, audiovisual, etc. The audio content (160) may include one or more of program portions, a program, several programs, one or more commercials, etc.

[0064] In some embodiments, the dialogue loudness analyzer (154) is configured to determine / establish one or more dialogue loudness levels for one or more portions of the audio content (152) (e.g., one or more programs, one or more commercials, etc.). In some embodiments, the audio content is represented by one or more collections of audio tracks. In some embodiments, dialogue audio content of the audio content is in a separate audio track. In some embodiments, at least a portion of the audio content is in an audio track that includes non-dialogue audio content.

[0065] The audio content control inputs (162) may include some or all of the following: user control inputs, control inputs provided by systems / devices external to the encoder (150), control inputs from a content creator, control inputs from a content provider, etc. For example, a user, such as a mixing engineer, may provide / specify one or more dynamic range compression curve identifiers, which may be used to retrieve one or more dynamic range compression curves that best fit the audio content (160) from a data repository, such as the DRC reference repository (156).

[0066] In some embodiments, the DRC reference repository (156) is configured to store DRC reference parameter sets, etc. The DRC reference parameter sets may include definition data for one or more dynamic range compression curves, etc. In some embodiments, the encoder (150) may encode (e.g., in parallel, etc.) two or more dynamic range compression curves into the encoded audio signal (102). Zero, one, or more of the dynamic range compression curves may be standard-based, proprietary, customized, decoder-modifiable, etc. In an example embodiment, both the dynamic range compression curves of FIG. 2A and FIG. 2B may be encoded (e.g., in parallel, etc.) into the encoded audio signal (102).

[0067] In some embodiments, the audio signal encoder (158) may be configured to receive audio content from the audio content interface (152), dialogue loudness levels from the dialogue loudness analyzer (154), etc., retrieve one or more DRC reference parameter sets from the DRC reference store (156), format the audio content into audio data blocks / frames, format the dialogue loudness levels, the DRC reference parameter sets, etc. into metadata (e.g., metadata containers, metadata fields, metadata structures, etc.), encode the audio data blocks / frames and metadata into the encoded audio signal (102), etc.

[0068] Audio content to be encoded into an encoded audio signal as described herein may be received in one or more of a variety of ways, such as wirelessly, via a wired connection, through a file, via Internet download, etc., in one or more of a variety of source audio formats.

[0069] The encoded audio signals described herein can be part of an overall media data bitstream (e.g., for an audio broadcast, an audio program, an audiovisual program, an audiovisual broadcast, etc.). The media data bitstream can be accessed from a server, a computer, a media storage device, a media database, a media file, etc. The media data bitstream can be broadcast, transmitted or received over one or more wireless or wired network links. The media data bitstream can be communicated over an intermediary such as one or more of a network connection, a USB connection, a wide area network, a local area network, a radio connection, an optical connection, a bus, a crossbar connection, a serial connection, etc.

[0070] Any of the components depicted (e.g., in Figures 1A, 1B, etc.) may be implemented as one or more processes and / or one or more IC circuits (e.g., ASICs, FPGAs, etc.) in hardware, software, or a combination of hardware and software.

[0071] 5. Dynamic range compression curve 2A and 2B show exemplary dynamic range compression curves that can be used by the DRC gain unit (104) in the decoder (100) to derive DRC gains from input loudness levels. As shown, the dynamic range compression curves may be centered around a reference loudness level in the program to provide an appropriate overall gain for a particular playback environment. Exemplary definition data (e.g., in the metadata of the encoded audio signal 102) of the dynamic range compression curves (e.g., including but not limited to boost ratios, cut ratios, attack times, release times, etc.) are shown in the table below. Here, each profile in a plurality of profiles (e.g., film standard, film light, music standard, music light, speech, etc.) represents a particular playback environment (e.g., in the decoder 100).

[0072] [Table 1] Some embodiments are SPL or dB FS Loudness level expressed in dB SPL The DRC gain may be expressed in dB. SPLA different loudness representation (e.g., sones) that has a non-linear relationship with loudness level may be implemented, and the compression curve used in the DRC gain calculation may then be transformed to be described using the different loudness representation (e.g., sones).

[0073] 6. DRC Gain, Gain Limiting and Gain Smoothing 3 shows exemplary processing logic for combined DRC and limiting gain determination / calculation. The processing logic may be implemented by a decoder (100), an encoder (150), etc. For illustrative purposes only, a DRC gain unit (e.g., 114) in a decoder (e.g., 100) may be used to implement the processing logic.

[0074] A DRC gain for a portion of audio content (e.g., one or more audio data blocks, frames of audio data, etc.) may be calculated using a loudness level determined from the portion of audio content. The loudness level may first be adjusted with respect to (e.g., relative to) a dialogue loudness level (e.g., of a program of which the audio content is part) in metadata extracted from the encoded audio signal (102). In the example shown in Figure 3, the difference between the loudness level of the portion of audio content and the dialogue loudness level ("dialnorm") may be used as input for finding a DRC gain from a selected dynamic range compression curve.

[0075] To prevent clipping of the output audio data elements in that particular playback environment, the DRC gain unit (114) may be configured to handle peak levels in a particular playback scenario (e.g., specific to a particular combination of the encoded audio signal 102 and the playback environment in the decoder 100), which may be one of a variety of possible playback scenarios (e.g., a multi-channel scenario, a downmix scenario, etc.).

[0076] In some embodiments, individual peak levels for individual portions of the audio content at a particular time resolution (e.g., an audio data block, several audio data blocks, audio data frames, etc.) may be provided as part of the metadata extracted from the encoded audio signal (102).

[0077] In some embodiments, the DRC gain unit (114) can be configured to determine peak levels in these scenarios and adjust the DRC gain if necessary. A parallel process may be used by the DRC gain unit (114) to determine peak levels of the audio content during the calculation of the DRC gain. For example, the audio content may be encoded for a reference multi-channel configuration that has more channels than the channels of a particular speaker configuration used by the decoder (100). The audio content for the more channels of the reference multi-channel configuration may be converted to downmixed audio data (e.g., an ITU stereo downmix, a matrixed-surround compatible downmix, etc.) to derive fewer channels for the particular speaker configuration in the decoder (100). In some embodiments, under the first approach, a downmix from the reference multi-channel configuration to the particular speaker configuration may be performed before determining and processing the peak levels to prevent clipping. Additionally, optionally, or alternatively, in some embodiments, under the second approach, downmix channel gains associated with downmixing the audio content may be used as part of the input for adjusting, deriving, calculating, etc., the peak levels for that particular speaker configuration. In an example embodiment, the downmix channel gains may be derived based at least in part on one or more downmix equations used to perform a downmix operation from a reference multi-channel configuration to a particular speaker configuration in the playback environment at the decoder (100).

[0078] Some media applications use a reference dialogue loudness level (e.g. -31dB in "Line" mode). FS , -20dB in "RF" mode FS) may be specified or assumed for a particular playback environment in the decoder (100). In some embodiments, a user may be given control to set or change a reference dialogue loudness level in the decoder (100).

[0079] A dialogue loudness related gain may be applied to the audio content to adjust the (e.g. output) dialogue loudness level to a reference dialogue loudness level. To reflect this adjustment, the peak levels should be adjusted accordingly. In one example, the (input) dialogue loudness level is -23dB. FS The reference dialogue loudness level is -31dB. FS In "line" mode, the adjustment to the (input) dialogue loudness level is -8dB to result in an output dialogue loudness level of the reference dialogue loudness level. In this "line" mode, the adjustment to the peak level is also -8dB, the same as the adjustment to the dialogue loudness level. If the reference dialogue loudness level is -20dB FS In "RF" mode, the adjustment to the (input) dialogue loudness level is 3dB to result in an output dialogue loudness level of the reference dialogue loudness level. In this "RF" mode, the adjustment to the peak level is also 3dB, the same as the adjustment to the dialogue loudness level.

[0080] The peak level and the sum of the difference between a reference dialogue loudness level (denoted as "dialref") and the dialogue loudness level in the metadata from the encoded audio signal (102) ("dialnorm") may be used as input to calculate the maximum (e.g., allowed) gain for the DRC gain. The adjusted peak level is (0 dB FS (relative to clipping level) dB FS, so the maximum allowable gain without clipping (e.g. for the current audio data block, for the current audio data frame, etc.) is simply the inverse of the adjusted peak level (e.g. multiplied by -1).

[0081] In some embodiments, the peak level may exceed the clipping level (0 dB), even though the dynamic range compression curve from which the DRC gain is derived is designed to cut some loud sounds. FS In some embodiments, the combination of the dialogue loudness level and adjustments to the DRC gain may prevent peak level clipping, even in a worst-case downmix (e.g., one that produces the largest downmix channel gain). However, in some other embodiments, the combination of the dialogue loudness level and adjustments to the DRC gain may not be sufficient to prevent peak level clipping. In these embodiments, the DRC gain may be replaced (e.g., capped, etc.) by the highest gain that prevents clipping at the peak levels.

[0082] In some embodiments, the DRC gain unit (114) is configured to derive time constants (e.g., attack time, release time, etc.) from metadata extracted from the encoded audio signal (102). These time constants may or may not vary with one or more of the dialogue loudness level or the current loudness level of the audio content. The DRC gain, time constant, and maximum gain retrieved from the dynamic range compression curve may be used to perform gain smoothing and limiting operations.

[0083] In some embodiments, the DRC gain, which may possibly be gain limited, does not exceed the maximum peak loudness level in a particular playback environment. The static DRC gain derived from the loudness level may be smoothed with a filter controlled by a time constant. The limiting operation may be implemented by one or more min() functions, through which the DRC gain (e.g. before limiting) may be replaced by the maximum allowed gain immediately, over a relatively short time interval, etc., thereby preventing clipping. The DRC algorithm may be configured to smoothly release from said clipping gain to a lower gain as the peak level of the incoming audio content transitions from above the clipping level to below the clipping level.

[0084] One or more different (e.g., real-time, two-pass, etc.) implementations may be used to perform the DRC gain determination / calculation / application shown in FIG. 3. For illustrative purposes only, the adjustments to the dialogue loudness level, the DRC gain (e.g., static, etc.), the time-dependent gain variations due to smoothing, the gain clipping due to limiting, etc. have been described as a combined gain from the DRC algorithm above. However, in various embodiments, other approaches may be used to apply gain to audio content for controlling the dialogue loudness level (e.g., between different programs), for dynamic range control (e.g., for different portions of the same program), to prevent clipping, for gain smoothing, etc. For example, some or all of the adjustments to the dialogue loudness level, the DRC gain (e.g., static, etc.), the time-dependent gain variations due to smoothing, the gain clipping due to limiting, etc. may be applied partially / individually, applied in series, applied in parallel, applied partially in series and partially in parallel, etc.

[0085] 7. Input Smoothing and Gain Smoothing In addition to DRC gain smoothing, other smoothing processes may be implemented under the techniques described herein in various embodiments. In one example, input smoothing may be used, where the input audio data extracted from the encoded audio signal (102) may be smoothed, for example with a simple single-pole smoothing filter, to obtain a spectrum of specific loudness levels with better temporal characteristics (e.g., smoother in time, less spikes in time, etc.) than the spectrum of specific loudness levels without input smoothing.

[0086] In some embodiments, the different smoothing processes described herein may use different time constants (e.g., 1 second, 4 seconds, etc.). In some embodiments, two or more smoothing processes may use the same time constant. In some embodiments, the time constants used in the smoothing processes described herein may be frequency dependent. In some embodiments, the time constants used in the smoothing processes described herein may be frequency independent.

[0087] One or more smoothing processes may be connected to a reset process that supports automatic or manual resetting of the one or more smoothing processes. In some embodiments, when a reset occurs in the reset process, the smoothing process may speed up the smoothing operation by switching or transitioning to a smaller time constant. In some embodiments, when a reset occurs in the reset process, the memory of the smoothing process may be reset to a value. This value may be the last input sample to the smoothing process.

[0088] 8. DRC across multiple frequency bands In some embodiments, specific loudness levels in specific frequency bands can be used to derive corresponding DRC gains in those specific frequency bands, however this can lead to changes in timbre, as those specific loudness levels can vary significantly in different bands and therefore suffer from different DRC gains, even when the broadband loudness level across all frequency bands remains constant.

[0089] In some embodiments, rather than applying DRC gains that vary with individual frequency bands, DRC gains that do not vary with frequency bands but vary with time are instead applied. The same time-varying DRC gain is applied across all frequency bands. The time-averaged DRC gain of the time-varying DRC gain may be set equal to a static DRC gain derived from the selected dynamic range compression curve based on broadband (or wideband) range or broadband, wideband and / or overall loudness levels across multiple frequency bands. As a result, changes to timbre effects that may be caused by applying different DRC gains in different frequency bands in other approaches can be avoided.

[0090] In some embodiments, the DRC gains in the individual frequency bands are controlled using broadband DRC gains determined based on broadband loudness levels. The DRC gains in the individual frequency bands may operate around the broadband DRCs found in the dynamic range compression curve based on the broadband loudness levels. Thus, the DRC gains in the individual frequency bands time-averaged over a time interval (e.g., longer than 5.3 ms, 20 ms, 50 ms, 80 ms, 100 ms, etc.) are the same as the broadband levels shown in the dynamic range compression curve. In some embodiments, loudness level fluctuations over short time intervals relative to the time interval that deviate from the time-averaged DRC gains are tolerable among channels and / or frequency bands. This approach ensures application of the correct multi-channel and / or multi-band time-averaged DRC gains as shown in the dynamic range compression curve and prevents the DRC gains in short time intervals from deviating too greatly from such time-averaged DRC gains as shown in the dynamic range compression curve.

[0091] 9. Volume adjustment in the loudness range Applying linear processing for volume adjustment to an audio excitation signal under other approaches that do not implement the techniques described herein may cause low audible signal levels to be inaudible (e.g., below the frequency-dependent hearing threshold of the human auditory system).

[0092] Under the technique described in this paper, volume adjustment of audio content is performed in the physical domain (e.g., dB SPLIn some embodiments, the loudness level of all bands is scaled by the same factor in the loudness domain to maintain the perceptual quality and / or integrity of the loudness level relationships among all bands at all volume levels. The volume adjustments based on setting and adjusting gains in the loudness domain described herein may be translated back to and implemented through nonlinear processing in the physical domain (or in a digital domain representing the physical domain) that applies different scaling factors to the audio excitation signal in different frequency bands. The nonlinear processing in the physical domain, translated from the volume adjustments in the loudness domain under the techniques described herein, attenuates or enhances the loudness level of the audio content with a DRC gain that prevents most or all of the lower audible levels in the audio content from being inaudible. In some embodiments, the difference in loudness levels between loud and soft sounds in a program is reduced--but not perceptually eliminated--with these DRC gains maintaining low audible signal levels above the hearing threshold of the human auditory system. In some embodiments, to maintain similarities in spectral perception and perceived timbre over a large range of volume levels, frequencies or frequency bands with excitation signal levels near the hearing threshold are attenuated less and therefore perceptually audible at low volume levels.

[0093] The techniques described herein may implement transformations (e.g., back and forth transformations) between signal levels, gains, etc. in the physical domain (or a digital domain representing the physical domain) and loudness levels, gains, etc. in the loudness domain. These transformations may be based on forward and inverse versions of one or more nonlinear functions (e.g., mappings, curves, piecewise linear segments, lookup tables, etc.) constructed based on models of the human auditory system.

[0094] 10. Downmix and loudness adjustment In some embodiments, the audio content (152) is encoded in the encoded audio signal (102) for a reference speaker configuration that includes multiple audio channels or speakers (e.g., a surround sound configuration, a 5.1 speaker configuration, etc.).

[0095] A receiving decoder operating with a particular speaker configuration having a smaller number of audio channels or speakers (e.g., a two-channel headset configuration) is expected to downmix the audio content (152) received from the encoded audio signal (102) from the multiple audio channels in the reference speaker configuration to the fewer audio channels in the decoder's particular speaker configuration (e.g., using one or more downmix equations) to generate a downmix output sound output.

[0096] A loudness level measured in the reference speaker configuration for an individual portion of the audio content (152) may differ from a loudness level measured in the particular speaker configuration, such as a two-channel configuration, for the same individual portion of the audio content (152). For example, if a portion of the audio content (152) before downmixing has a particular channel-dependent sound distribution centered in the left-front and right-front channels of the reference speaker configuration, then the loudness level of the same portion of the audio content (152) after downmixing to a two-channel configuration may be higher or louder than the loudness level of the same portion of the audio content (152) in the reference speaker configuration before downmixing. On the other hand, if a portion of the audio content (152) before downmixing has a particular channel-dependent sound distribution concentrated in channels other than the left front and right front channels of the reference speaker configuration, the loudness level of the same portion of the audio content (152) after downmixing to a two-channel configuration may be lower or quieter than the loudness level of the same portion of the audio content (152) in the reference speaker configuration before downmixing.

[0097] In some embodiments, an audio encoder (e.g., 150) described herein is configured to provide downmix-related metadata (e.g., including one or more downmix loudness parameters, etc.) to a downstream audio decoder. The downmix-related metadata from the audio encoder (150) can be used by the downstream audio decoder to efficiently and consistently perform downmix-related gain adjustment operations (e.g., in real-time, near real-time, etc.) to enable the downstream audio decoder to generate a relatively accurate actual target loudness level in the downmix sound output, to prevent mismatches in measured loudness levels between a reference speaker configuration and a decoder-specific speaker configuration, etc.

[0098] In some embodiments, the audio encoder (150) determines one or more downmix parameters based at least in part on a reference speaker configuration and the audio content (152) encoded for a particular speaker configuration (e.g., a two-channel configuration) different from the reference speaker configuration. In some embodiments, the downmix loudness parameters include one or more different sets of downmix loudness parameters for different types of downmix operations. The downmix loudness parameters may comprise a single set of downmix loudness parameters to be used by a downstream audio decoder to perform a particular type of downmix, such as an LtRt downmix, a LoRo downmix, etc. The downmix loudness parameters may comprise two or more sets of downmix loudness parameters to be used by a downstream audio decoder to perform any of two or more particular types of downmixes, such as an LtRt downmix, a LoRo downmix, etc. The downmix loudness data generated by the audio encoder (150) may carry downmix loudness parameters for one or more different types of downmix operations. The downmix loudness data may also include a preference flag indicating which type of downmix operation is preferred for the audio content to be downmixed. The downmix loudness parameters may be delivered to a downstream decoder as part of metadata delivered in an encoded audio signal (102) that includes audio content (152) encoded for a reference speaker configuration.

[0099] Examples of downmix loudness parameters described herein may include, but are not limited to: one or more downmix loudness metadata indicators, one or more downmix loudness data fields, etc. In an example embodiment, the downmix loudness parameters may have an indicator indicating whether downmix loudness offset data is present (e.g., a one-bit data field labeled "dmixloudoffste"), a data field indicating the downmix loudness offset (e.g., a five-bit data field labeled "5-bit_dmixloudoffst"), etc. In some embodiments, one or more instances of these indicators and data fields may be generated by the audio encoder (150) for one or more different types of downmix operations.

[0100] In some embodiments, the "dmixloudoffste" field may be set to 1 only when the encoded audio signal (102) carries audio data (e.g., audio samples) for more than two channels; the "dmixloudoffst" field may be carried if the "dmixloudoffste" field is set to 1. In one example where the encoded audio signal (102) is an AC-3 or E-AC-3 bitstream, the "dmixloudoffste" field may be set to 1 when the audio coding mode (e.g., "acmode") for the AC-3 or E-AC-3 bitstream is set to a value greater than 2. Such a value of the audio coding mode indicates that the reference speaker configuration is a multi-channel speaker configuration that includes more than two audio channels or speakers, and is not a center speaker only configuration (e.g., having a value of 1 for "acmod") or a left front and right front only speaker configuration (e.g., having a value of 2 for "acmod").

[0101] The “dmixloudoffst” field may be used to indicate the difference between an expected (e.g., expected, anticipated, etc.) loudness of a downmix sound output from an audio decoder (e.g., an AC-3 decoder, an E-AC-3 decoder, etc.) and a measured loudness of such downmix sound output. Some or all of the gain adjustments due to dialogue normalization, dynamic range compression, fixed attenuation to protect against downmix overload, etc. have been applied prior to performing the measurement that gives the measured loudness. In some embodiments, the measured loudness includes one or more different sets of downmix loudness measurements for one or more different types of downmix sound output with one or more different sets of gain adjustments. In some embodiments, the audio encoder (150) generates one or more downmixes based on one or more types of downmix operations (e.g., an LtRt downmix operation, a LoRo downmix operation, etc.). For example, the audio encoder (150) may apply one or more different sets of downmix coefficients / downmix equations (e.g., LtRt downmix coefficients / downmix equations, LoRo downmix coefficients / downmix equations, etc.) to audio content encoded for a reference speaker configuration (e.g., multi-channel, etc.) to generate the one or more downmixes. In some embodiments, the audio encoder (150) may apply one or more different sets of gain adjustments to one or more of the downmixes to generate the one or more different types of downmix sound outputs for loudness measurement.Examples of sets of gain adjustments include, but are not limited to, any of: a set of gain adjustments with a null gain, a set of gain adjustments including gain adjustments related to dynamic range compression, a set of gain adjustments including gain adjustments related to dialogue normalization, a set of gain adjustments excluding gain adjustments related to dynamic range compression, a set of gain adjustments excluding gain adjustments related to dialogue normalization, a set of gain adjustments including gain adjustments related to both dynamic range compression and dialogue normalization, etc. After the one or more different types of downmix sound output for loudness measurement are generated based on one or more different combinations of the one or more different sets of gain adjustments and the one or more downmixes, the audio encoder (150) can generate the measured loudness by performing one or more different sets of downmix loudness measurements on any, some or all of the one or more different types of downmix sound output. The measured loudness may be produced by the audio encoder (150) in any of a variety of loudness measurement standards (e.g., LKFS, LUFS, etc.), methods, tools, etc. For illustrative purposes only, the measured loudness may be represented by an LKFS value.

[0102] In some embodiments, the audio encoder (150) assumes that an audio decoder (e.g., 100) described herein that decodes the encoded audio signal (102) having a certain dialogue loudness level (e.g., "dialnorm") is expected to apply, upon decoding, an amount of attenuation (e.g., the difference between the reference loudness level and "dialnorm") to align / adjust the output dialogue loudness level of the downmixed sound output to the reference loudness level. For example, if a dialogue loudness level "dialnorm" (e.g., as determined from the encoded audio content (152) for a reference speaker configuration, such as a 5.1 speaker configuration) is greater than the value -24 dB, the audio decoder (100) may be expected to apply an amount of attenuation (e.g., the difference between the reference loudness level and "dialnorm") to align / adjust the output dialogue loudness level of the downmixed sound output to the reference loudness level. FSand the reference loudness level for the decoder's particular speaker configuration (e.g., a two-channel configuration to which the audio content (152) is to be downmixed) is -31 LKFS, then the audio decoder (100) is expected to apply 7 dB of attenuation to align / adjust the output dialogue loudness level to the reference loudness level. In some embodiments, the reference loudness level for the decoder's particular speaker configuration (e.g., -31 LKFS) represents the expected loudness level (e.g., a two-channel downmix sound output).

[0103] In some embodiments, the "dmixloudoffst" field may be used by the audio encoder (150) to indicate any loudness deviations between (1) the expected loudness level of the two-channel downmix sound output and (2) the measured loudness level of the two-channel downmix sound output, measured after some or all of the gain adjustments due to dialogue normalization, dynamic range compression, fixed attenuation to protect against downmix overload, etc. have been applied. The "dmixloudoffst" field may contain one or more instances of one or more different types of downmixes after applying one or more different sets of gain adjustments, etc. The loudness deviations indicated by the "dmixloudoffst" field may include, but are not limited to, loudness level differences caused by downmixing audio content from a reference speaker configuration to a particular speaker configuration, such as a two-channel configuration. The loudness deviation corresponds to (e.g. represents the inverse of) the loudness offset that should be applied by a decoder with a particular speaker configuration to which the audio content (152) is to be downmixed, in order to produce a reference loudness level in the downmix sound output.

[0104] In one example implementation, the "dmixloudoffst" field (e.g., an instance thereof) may be set to a value in the value range (e.g., integer) from 0 to 30, corresponding to a loudness offset range of -7.5LKFS to +7.5LKFS in increments of 0.5LKFS. Additionally, optionally, or alternatively, a value of 31 for the "dmixloudoffst" field may be specified as a reserved value, and if present, may be interpreted as a downmix loudness offset of 0LKFS.

[0105] In some embodiments, positive LKFS values ​​of the "dmixloudoffst" field (e.g., values ​​of 16, 17, ..., 30 for the "dmixloudoffst" field) indicate that the measured loudness level of the downmix sound output is louder than the expected loudness level of the downmix sound output by the indicated LKFS value. Negative LKFS values ​​of the "dmixloudoffst" field (e.g., values ​​of 0, 1, ..., 15 for the "dmixloudoffst" field) indicate that the measured loudness level of the downmix sound output is quieter or less loud than the expected downmix loudness by the indicated LKFS value.

[0106] Some or all of the downmix loudness parameters may be used (e.g., additionally, optionally, alternatively, etc.) by an audio decoder (e.g., 100) having a speaker configuration such as a particular speaker configuration to control one or more audio processing operations, algorithms, etc. acting on the audio content (152) in the encoded audio signal (102) to compensate for loudness level differences - of individual portions of the audio content (152) in the encoded audio signal (102) - caused by downmixing the audio content (152) from a reference speaker configuration to the particular speaker configuration.

[0107] In some embodiments, an audio decoder (e.g., 100) described herein is configured to decode (e.g., multi-channel) audio content from an encoded audio signal (102), such as extracting a dialogue loudness level (e.g., "dialnorm") from loudness metadata delivered with the audio content. The audio decoder (100) may operate with a particular speaker configuration (e.g., a two-channel configuration) that has fewer audio channels than a reference speaker configuration to which the audio content corresponds.

[0108] In some embodiments, the audio decoder (100) uses one or more downmix equations to downmix the multi-channel audio content received from the encoded audio signal (102) for a reference speaker configuration for which the multi-channel audio content was encoded to a particular speaker configuration at the audio decoder, performing one or more audio processing operations, algorithms, etc. on the downmixed audio content to generate a downmixed sound output, etc. The audio decoder (100) may be capable of performing one or more different types of downmix operations. The audio decoder (100) may be configured to determine and perform a particular type of downmix operation (e.g., LtRt downmix, LoRo downmix, etc.) based on one or more factors. These factors may include one or more of, but are not limited to, an input specifying a preference for a particular user-selected type of downmix operation, a user input specifying a preference for a system-selected type of downmix operation, the particular speaker configuration and / or capabilities of the audio decoder (100), the availability of downmix loudness metadata for the particular type of downmix operation, any encoder-generated preference flags for a type of downmix operation, etc. In some embodiments, the audio decoder (100) may implement one or more priority rules and / or seek further user input to determine the particular type of downmix operation when these factors conflict with each other.

[0109] The one or more audio processing operations, algorithms, etc. may include, but are not limited to: a loudness attenuation operation that applies an amount of attenuation (e.g., the difference between a reference loudness level and the "dialnorm") to align / adjust an output dialogue loudness level of the downmixed sound output to a reference loudness level based, at least in part, on the dialogue loudness level (e.g., "dialnorm") and a reference loudness level (e.g., -31LKFS). In some embodiments, the audio decoder (100) further performs some or all of the gain adjustments due to dialogue normalization, dynamic range compression, fixed attenuation to protect against downmix overload, etc. In some embodiments, these gain adjustments may correspond to - e.g., be the same or substantially the same as - those performed by the audio encoder (150) in determining the measured loudness level discussed above. One or more of these gain adjustments may be specific to the type of downmix operation (e.g., LtRt downmix, LoRo downmix, etc.) performed by the audio encoder (100).

[0110] Additionally, optionally or alternatively, in some embodiments, the audio decoder (100) is configured to extract downmix loudness metadata (e.g., “dmixloudoffste” field, “dmixloudoffst” field, etc.) from the encoded audio signal (102) as part of the metadata delivered together with the audio content. In some embodiments, the downmix loudness parameters in the extracted downmix loudness metadata include one or more different sets of downmix loudness parameters for different types of downmix operations indicated to be present by one or more flags carried in the downmix loudness metadata. In response to determining that the one or more sets of downmix loudness parameters are present, the audio decoder (100) can determine / select a set of downmix loudness parameters from the one or more different sets of downmix loudness parameters that corresponds to a particular type of downmix operation performed by the audio decoder (100) (e.g., LtRt downmix, LoRo downmix, etc.). The audio decoder (100) determines whether downmix loudness offset data is present in that particular set of downmix loudness parameters (e.g., based on whether the “dmixloudoffste” field has a value of 1 or 0).In response to determining that downmix loudness offset data is present in that particular set of downmix loudness parameters (e.g., based on the “dmixloudoffste” field having a value of 1), the audio decoder (100) performs a loudness adjustment operation based on the downmix loudness offset in the downmix loudness metadata (e.g., the “dmixloudoffst” field in the same set of downmix loudness parameters) extracted together with the audio content from the encoded audio signal (102). The downmix loudness metadata may include a “dmixloudoffst” field with one or more instances for one or more different types of downmixes, such as after applying one or more different sets of gain adjustments. Based on the actual downmix operation and the actual set of gain adjustments performed by the audio decoder (100) (e.g., no gain adjustments, gain adjustments excluding those related to DRC, gain adjustments including those related to DRC, gain adjustments excluding those related to dialogue normalization, gain adjustments including those related to dialogue normalization, gain adjustments including those related to both dialogue normalization and DRC, etc.), the audio decoder (100) can determine / select a particular instance of the one or more instances of the “dmixloudoffst” field in the downmix loudness metadata.

[0111] In response to determining that the “dmixloudoffst” field indicates a positive LKFS value (e.g., a value of 16, 17, …, 30 for the “dmixloudoffst” field), which means that the loudness level (as measured by an upstream audio encoder, such as 150) of the downmix sound output after applying some or all of the gain adjustments due to dialogue normalization, dynamic range compression, fixed attenuation to protect against downmix overload, etc., is louder than the expected loudness level of the downmix sound output by the magnitude of the indicated LKFS value, the audio decoder (100) performs a further gain adjustment with a negative gain value having the magnitude of the indicated LKFS value, which reduces or adjusts the loudness level of the downmix sound output to the expected loudness (e.g., the reference loudness level).

[0112] In response to determining that the “dmixloudoffst” field indicates a negative LKFS value (e.g., a value of 1, 2, …, 15 for the “dmixloudoffst” field), which means that the loudness level (as measured by an upstream audio encoder, such as 150) of the downmix sound output after applying some or all of the gain adjustments due to dialogue normalization, dynamic range compression, fixed attenuation to protect against downmix overload, etc., is quieter or less loud than the expected loudness level of the downmix sound output by the magnitude of the indicated LKFS value, the audio decoder (100) performs a further gain adjustment with a negative gain value having the magnitude of the indicated LKFS value, which raises or adjusts the loudness level of the downmix sound output to the expected loudness (e.g., a reference loudness level).

[0113] The LKFS value of the "dmixloudoffst" field (e.g., values ​​of 1, 2, ..., 15 for the "dmixloudoffst" field) indicates that the measured loudness level of the downmix sound output is quieter or less loud than the expected downmix loudness by the amount of the indicated LKFS value. In some embodiments, if a negative LKFS value is indicated / signaled in the encoded audio signal (102) to a receiving decoder, the receiving decoder (e.g., 150) can take steps to ensure that any positive gain applied to the two-channel downmix sound output to compensate for said negative LKFS value does not introduce clipping of the loudness level in the two-channel downmix sound output.

[0114] The further gain adjustment based on the loudness offset indicated in the downmix loudness metadata may be, but is not limited to, specific to the type of downmix operation performed by the audio decoder (100).

[0115] 11. Additional actions related to gains Under the techniques described herein, other processing such as dynamic equalization, noise compensation, etc. can also be performed in the loudness (eg, perceptual) domain rather than the physical domain (or a digital domain representing the physical domain).

[0116] In some embodiments, the gains from some or all of the various processes, such as DRC, equalization noise compensation, clip prevention, gain smoothing, etc., may be combined into the same gain in the loudness domain and / or may be applied in parallel. In some other embodiments, the gains from some or all of the various processes, such as DRC, equalization noise compensation, clip prevention, gain smoothing, etc., may be separate gains in the loudness domain and / or may be applied at least partially in series. In some other embodiments, the gains from some or all of the various processes, such as DRC, equalization noise compensation, clip prevention, gain smoothing, etc., may be applied in sequence.

[0117] 12. Specific and broadband loudness levels One or more audio processing elements, units, components, etc., such as transmission filters, acoustic filter banks, synthesis filter banks, short-time Fourier transforms, etc., may be used by an encoder or decoder to perform the audio processing operations described herein.

[0118] In some embodiments, one or more transmission filters that model the outer and middle ear filtering of the human auditory system may be used to filter the incoming audio signal (e.g., the encoded audio signal 102, audio content from a content provider, etc.). In some embodiments, an auditory filter bank may be used to model the frequency selectivity and frequency spread of the human auditory system. The excitation signal levels from some or all of these filters may be determined / calculated and smoothed with a frequency-dependent time constant that is shorter at higher frequencies to model the integration of energy in the human auditory system. A non-linear function (e.g., relationship, curve, etc.) between the excitation signal and the specific loudness level may then be used to obtain a frequency-dependent specific loudness level profile. A broadband (or wideband) loudness level may be obtained by integrating the specific loudness over frequency bands.

[0119] A straightforward summation / integration of the specific loudness levels (e.g., using equal weights for all frequency bands) may work well for broadband signals. However, such an approach may underestimate the (e.g., perceptual) loudness levels for narrowband signals. In some embodiments, the specific loudness levels at different frequencies or in different frequency bands are given different weights.

[0120] In some embodiments, the auditory filterbank and / or transmission filters as described above may be replaced by one or more short-time Fourier transforms (STFTs). The responses of the transmission filters and auditory filterbank may be applied in the Fast Fourier Transform (FFT) domain. In some embodiments, one or more inverse transmission filters are used, for example when one or more (e.g. forward) transmission filters are used in or before the transformation from the physical domain (or a digital domain representing the physical domain) to the loudness domain. In some embodiments, the inverse transmission filters are not used, for example when an STFT is used instead of the auditory filterbank and / or transmission filters. In some embodiments, the auditory filterbank is omitted; instead, one or more quadrature mirror filters (QMFs) are used. In these embodiments, the diffusion effect of the basilar membrane in the model of the human auditory system may be omitted without significantly affecting the subject matter of the audio processing operations described herein.

[0121] Under the techniques described herein, different embodiments may use different numbers of frequency bands (e.g., 20 frequency bands, 40 frequency bands, etc.) Additionally, optionally, or alternatively, different bandwidths may be used in different embodiments.

[0122] 13. Individual Gains for Individual Subsets of Channels In some embodiments, when a particular speaker configuration is a multi-channel configuration, the overall loudness level may be obtained by first summing the excitation signals of all channels before conversion from the physical domain (or a digital domain representing the physical domain) to the loudness domain. However, applying the same gain to all channels in a particular speaker configuration may not preserve the spatial balance between the different channels of that particular speaker configuration (balance in terms of relative loudness levels between the different channels, etc.).

[0123] In some embodiments, to preserve spatial balance such that relative perceptual loudness levels between different channels may be optimally or correctly maintained, a respective loudness level and a corresponding gain obtained based on the respective loudness level may be determined or calculated for each channel. In some embodiments, the corresponding gains obtained based on the respective loudness levels are not equal to the same overall gain. For example, some or all of the corresponding gains may each be equal to the overall gain plus a small (e.g., channel-specific) correction.

[0124] In some embodiments, to preserve spatial balance, the respective loudness levels and corresponding gains obtained based on the respective loudness levels may be determined or calculated for each subset of channels. In some embodiments, the corresponding gains obtained based on the respective loudness levels are not equal to the same overall gain. For example, some or all of the corresponding gains may each be equal to the overall gain plus a small (e.g., channel-specific) correction. In some embodiments, the subset of channels may include two or more channels that are a true subset of all channels in that particular speaker configuration (e.g., a subset of channels including left front, right front, and low frequency effects (LFE), a subset of channels including left surround, and right surround, etc.). The audio content for the subset of channels may form a submix of the overall mix carried in the encoded audio signal (102). The channels in the submix may have the same gain applied to them.

[0125] In some embodiments, to generate an actual loudness (e.g., as actually perceived) from a particular speaker configuration, the signal levels in the digital domain are converted to corresponding physical (e.g., dB) levels in the physical domain represented by the digital domain. SPL One or more calibration parameters may be used to relate the spatial pressure (by a spatial frequency, spatial pressure, etc.) levels to the acoustical characteristics of the sound source, which may be given values ​​specific to the physical sound installation in a particular speaker configuration.

[0126] 14. Auditory Scene Analysis In some embodiments, the encoders described herein may implement computer-based auditory scene analysis (ASA) to detect auditory event boundaries in audio content (e.g., as encoded in encoded audio signal 102), generate one or more ASA parameters, and format the one or more ASA parameters as part of the encoded audio signal (e.g., 102) that is delivered to a downstream device (e.g., decoder 100). The ASA parameters may include, but are not limited to, locations of auditory event boundaries, values ​​of auditory event certainty indicators (described further below), etc.

[0127] In some implementations, the location (e.g., in time) of an auditory event boundary may be indicated in metadata encoded within the encoded audio signal (102). Additionally, optionally, or alternatively, the location (e.g., in time) of an auditory event boundary may be indicated (e.g., using a flag, a data field, etc.) in the audio data block and / or frame in which the location of the auditory event boundary is detected.

[0128] As used herein, an auditory event boundary refers to the point where a preceding auditory event ends and / or a following auditory event begins. Each auditory event occurs between two successive auditory event boundaries.

[0129] In some embodiments, the encoder (150) is configured to detect auditory event boundaries by a difference in a specific loudness spectrum between two consecutive (e.g., in time) audio data frames, each specific loudness spectrum comprising an unsmoothed loudness spectrum calculated from a corresponding one of the consecutive audio data frames.

[0130] In some embodiments, the specific loudness spectrum N[b,t] is the normalized specific loudness spectrum N NORM It may be normalized to obtain [b,t].

[0131] N NORM [b,t]=N[b,t] / max b {N[b,t]} (1) where b is the bandwidth, t is the time or audio frame index, and max b {N[b,t]} is the maximum specific loudness level across all frequency bands.

[0132] The normalized specific loudness spectra are subtracted from each other and used to derive the sum of absolute differences D[t], as shown in the following equation:

[0133] D[t]=Σ b |N NORM [b,t]-N NORM [b,t-1] | (2) The sum of absolute differences is mapped to an auditory event certainty index with a value range from 0 to 1 as follows:

[0134]

number

[0135] In some embodiments, the encoder (150) may be configured to determine whether D[t] (e.g., at a particular t) is D min t)。 t is configured to detect an auditory event boundary (e.g., at said particular t) when t exceeds t.

[0136] In some embodiments, a decoder described herein (e.g., 100) extracts ASA parameters from an encoded audio signal (e.g., 102) and uses the ASA parameters to prevent unintentional boosting of soft sounds and / or unintentional cutting of loud sounds, which causes perceptual distortion of auditory events.

[0137] The decoder (100) may be configured to mitigate or prevent unintended distortion of auditory events by ensuring that the gain is more constant within the auditory event and constraining many of the gain changes to the vicinity of the auditory event boundaries. For example, the decoder (100) may be configured to use a relatively small time constant (e.g., comparable to or shorter than the minimum duration of the auditory events) in response to gain changes in the attacks (e.g., loudness level increases) of auditory event boundaries. Thus, gain changes in the attacks can be implemented relatively quickly by the decoder (100). On the other hand, the decoder (100) may be configured to use a relatively long time constant compared to the duration of the auditory event in response to gain changes in the releases (e.g., loudness level decreases) of auditory events. Thus, gain changes in the releases can be implemented relatively slowly by the decoder (100), so that sounds that should be perceived as constant or that should decay gradually may not be auditorily or perceptually disturbed. A fast response in the attack of auditory event boundaries and a slower response in the release of auditory events allows the fast perception of the arrival of auditory events while preserving the perceptual quality and / or integrity between auditory events, such as piano chords, which may contain loud and soft sounds linked by specific loudness level relationships and / or specific time relationships.

[0138] In some embodiments, auditory event boundaries indicated by auditory events and ASA parameters are used by the decoder (100) to control gain changes in one, two, some or all of the channels in a particular speaker configuration in the decoder (100).

[0139] 15. Loudness Level Transition A loudness level transition may occur, for example, between two programs, between a program and a loud commercial, etc. In some embodiments, the decoder (100) is configured to maintain a histogram of instantaneous loudness levels based on past audio content (e.g., received from the encoded audio signal 102 over the past 4 seconds). Over the time interval from before the loudness level transition to after the loudness level transition, two regions of increased probability may be recorded in the histogram. One of the regions is centered on the previous loudness level, while the other of the regions is centered on the new loudness level.

[0140] The decoder (100) may dynamically determine a smoothed loudness level as the audio content is processed and determine a corresponding bin of a histogram based on the smoothed loudness level (e.g., a bin of instantaneous loudness level that contains the same value as the smoothed loudness level). The decoder (100) may be further configured to compare the probability in the corresponding bin to a threshold (e.g., 6%, 7%, 7.5%, etc.), where the total area of ​​the histogram curve (e.g., the sum of all bins) represents a probability of 100%. The decoder may be configured to detect the occurrence of a loudness level transition by determining the probability that the corresponding bin falls below the threshold. In response, the decoder (100) may be configured to select a relatively small time constant to adapt relatively quickly to the new loudness level. As a result, the duration of a loud (or soft) onset in a loudness level transition may be reduced.

[0141] In some embodiments, the decoder (100) uses a silence / noise gate to prevent low instantaneous loudness levels from entering the histogram and becoming high probability bins in the histogram. Additionally, optionally or alternatively, the decoder (100) may be configured to use the ASA parameter to detect auditory events to be included in the histogram. In some embodiments, the decoder (100) uses a time-dependent value of the time-averaged auditory event certainty index (ASA) to detect auditory events to be included in the histogram.

number

number

number

[0142] In some embodiments, for loudness levels (e.g., instantaneous, etc.) that are allowed to be included in the histogram (e.g., have corresponding values ​​of A[t] above the histogram inclusion threshold), the loudness levels are assigned weights that are the same as or proportional to the time-dependent values ​​of the time-averaged auditory event certainty index A[t] concurrently with those loudness levels. As a result, loudness levels that are close to auditory event boundaries have more influence on the histogram (e.g., have relatively larger values ​​of A[t]) than other loudness levels that are not close to auditory event boundaries.

[0143] 16. Reset In some embodiments, an encoder (e.g., 150) described herein is configured to detect a reset event and include an indication of the reset event in the encoded audio signal (e.g., 102). In a first example, the encoder (150) detects a reset event in response to determining that a continuous period of relative silence occurs (e.g., 250 milliseconds, configurable by the system and / or user). In a second example, the encoder (150) detects a reset event in response to determining that a large momentary drop in excitation level occurs across all frequency bands. In a third example, the encoder is given an input (e.g., user input, system controlled metadata, etc.) where a content transition (e.g., program start / end, scene change, etc.) occurs that requires a reset.

[0144] In some embodiments, the decoders described herein (such as 100) implement a reset mechanism that can be used to speed up instantaneous gain smoothing, which may be useful and may be invoked when switching between channels and audiovisual inputs occurs.

[0145] In some embodiments, the decoder (100) can be configured to determine whether a reset event occurs by determining whether a continuous period of relative silence (e.g., 250 milliseconds, configurable by the system and / or user) occurs, whether a large momentary drop in excitation level across all frequency bands occurs, etc.

[0146] In some embodiments, the decoder (100) is configured to determine that a reset event has occurred in response to receiving an indication (e.g., of a reset event) provided in the encoded audio signal (102) by an upstream encoder (e.g., 150).

[0147] The reset mechanism may be configured to issue a reset when the decoder (100) determines that a reset event occurs. In some embodiments, the reset mechanism is configured to use a more aggressive cut behavior of the DRC compression curve to prevent hard starts (e.g., for loud programs / channels / audiovisual sources, etc.). Additionally, optionally, or alternatively, the decoder (100) may be configured to implement safeguards to recover gracefully when the decoder (100) detects that a reset has been falsely triggered.

[0148] 17. Gains Provided by the Encoder In some embodiments, the audio decoder may be configured to calculate one or more sets of gains (e.g., DRC gains, etc.) for individual portions (e.g., audio data blocks, audio data frames, etc.) of audio content to be encoded into an encoded audio signal. The sets of gains generated by the audio encoder may include a first set of gains including a single broadband (or wideband) gain for all channels (e.g., left front, right front, low frequency effects or LFE, center, left surround, right surround, etc.); a second set of gains including individual broadband (or wideband) gains for individual subsets of channels; a third set of gains including individual broadband (or wideband) gains for individual subsets of channels and for each of a first number (e.g., two, etc.) of individual bands (e.g., two bands in each channel); a fourth set of gains including individual broadband (or wideband) gains for individual subsets of channels and for each of a second number (e.g., four, etc.) of individual bands (e.g., four bands in each channel); etc. The subset of channels described herein may be one of a subset including a left front, right front and LFE channel, a subset including a center channel, a subset including a left surround and right surround channel, and the like.

[0149] In some embodiments, the audio encoder is configured to transmit one or more portions of audio content (e.g., audio data blocks, audio data frames, etc.) and one or more sets of gains calculated for the one or more portions of audio content in a time-synchronous manner. An audio decoder receiving the one or more portions of audio content can select and apply a set of gains from the one or more sets of gains with little or no delay. In some embodiments, the audio encoder can implement a subframing technique in which the one or more sets of gains are carried (e.g., using differential encoding, etc.) in one or more subframes as shown in FIG. 4. In one example, the subframes may be encoded within the audio data block or audio data frame for which the gains are calculated. In another example, the subframes may be encoded within audio data blocks or audio data frames preceding the audio data block or audio data frame for which the gains are calculated. In another non-limiting example, subframes may be encoded into audio data blocks or frames within a time from the audio data block or frame for which their gains are calculated. In some embodiments, Huffman and differential encoding may be used to populate and / or compress the subframes that carry their set of gains.

[0150] 18. Exemplary System and Process Flows FIG. 5 illustrates an exemplary codec system in a non-limiting exemplary embodiment. A content creator, which may be a processing unit in an audio encoder such as 150, is configured to provide audio content ("audio") to an encoder unit ("NGC Encoder"). The encoder unit formats the audio content into audio data blocks and / or frames and encodes the audio data blocks and / or frames into an encoded audio signal. The content creator is also configured to establish / generate one or more dialogue loudness levels ("dialnorms") and one or more dynamic range compression curve identifiers ("compression curve IDs") for one or more programs, commercials, etc. in the audio content. The content creator may determine the dialogue loudness levels from one or more dialogue audio tracks in the audio content. The dynamic range compression curve identifiers may be selected based at least in part on user input, system configuration setting parameters, etc. A content creator may be a human (eg, artist, audio engineer, etc.) who uses tools to generate audio content and dialnorms.

[0151] Based on the dynamic range compression curve identifier, the encoder (150) generates one or more DRC parameter sets, including but not limited to corresponding reference dialogue loudness levels ("reference levels") for a plurality of playback environments supported by the one or more dynamic range compression curves. These DRC parameter sets may be encoded in metadata of the encoded audio signal, such as in-band with the audio content, out-of-band with the audio content, etc. Operations such as compression, format multiplexing ("MUX"), etc. may be performed as part of generating an encoded audio signal that may be delivered to an audio decoder, such as 100. The encoded audio signal may be encoded with a syntax that supports carrying audio data elements, DRC parameter sets, reference loudness levels, dynamic range compression curves, functions, lookup tables, Huffman codes used in compression, subframes, etc. In some embodiments, the syntax allows an upstream device (e.g., encoder, decoder, transcoder, etc.) to transmit gains to a downstream device (e.g., decoder, transcoder, etc.). In some embodiments, the syntax used to encode and / or decode data into and / or from the encoded audio signal is configured to support backward compatibility, such that devices that rely on gains calculated by an upstream device may optionally continue to do so.

[0152] In some embodiments, the encoder (150) calculates two or more sets of gains (e.g., gain smoothing with an appropriate reference dialogue loudness level, DRC gains, etc.) for the audio content. These sets of gains may be provided in metadata encoded into an audio signal encoded together with the audio content, along with the one or more dynamic range compression curves. A first set of gains may correspond to a broadband (or wideband) gain for all channels in a (e.g., default) speaker configuration or profile. A second set of gains may correspond to a broadband (or wideband) gain for each of all channels in a speaker configuration or profile. A third set of gains may correspond to a broadband (or wideband) gain for each of two bands in each of all channels in a speaker configuration or profile. A fourth set of gains may correspond to a broadband (or wideband) gain for each of four bands in each of all channels in a speaker configuration or profile. In some embodiments, the set of gains calculated for a speaker configuration may be transmitted together with the (e.g., parameterized, etc.) dynamic range compression curve for the speaker configuration in metadata. In some embodiments, a set of gains calculated for a speaker configuration may replace a (e.g., parameterized, etc.) dynamic range compression curve for that speaker configuration in the metadata. Additional speaker configurations or profiles may be supported under the techniques described herein.

[0153] The decoder (100) is configured to extract audio data blocks and / or frames and metadata from the encoded audio signal, for example through operations such as decompression, deformatting, and demultiplexing ("DEMUX"). The extracted audio data blocks and / or frames may be decoded into audio data elements or samples by a decoder unit ("NGC Decoder"). The decoder (100) is further configured to determine a profile for a particular playback environment in which the audio content will be rendered in the decoder (100) and to select a dynamic range compression curve from the metadata extracted from the encoded audio signal. The digital audio processing unit ("DAP") is configured to apply DRC or other operations to the audio data elements or samples to generate an audio signal that drives an audio channel in a particular playback environment. The decoder (100) can calculate and apply a DRC gain and a selected dynamic range compression curve based on the audio data block or frame. The decoder (100) can also adjust the output dialogue loudness level based on a reference dialogue loudness level associated with the selected dynamic range compression curve and on the dialogue loudness level in the metadata extracted from the encoded audio signal. The decoder (100) can then apply a gain limiter specific to the audio content and the playback scenario associated with the particular playback environment. In this way, the decoder (100) can render / play back the audio content in a manner tailored to the playback scenario.

[0154] 6A-6D illustrate an exemplary process flow that, in some embodiments, one or more computing devices or units in a media processing system may perform the process flow.

[0155] 6A illustrates an exemplary process flow that may be implemented by an audio decoder described herein. In block 602 of FIG. 6A, a first device (such as, for example, audio decoder 100 of FIG. 1A) receives an audio signal that includes audio content and definition data for one or more dynamic range compression curves.

[0156] In block 604, the first device determines a particular playback environment.

[0157] In block 606, the first device establishes a specific dynamic range compression curve for the specific playback environment based on the definition data for the one or more dynamic range compression curves extracted from the audio signal.

[0158] At block 608, the first device performs one or more dynamic range control (DRC) operations on one or more portions of audio content extracted from the audio signal, the one or more DRC operations being based, at least in part, on one or more DRC gains derived from a particular dynamic range compression curve.

[0159] In one embodiment, the definition data for the one or more dynamic range compression curves comprises an attack time, a release time or a reference loudness level related to at least one of the one or more dynamic range compression curves.

[0160] In one embodiment, the first device is further configured to perform the steps of: calculating one or more loudness levels for the one or more portions of audio content; determining the one or more DRC gains based on the particular dynamic range compression curve and the one or more loudness levels for the one or more portions of the audio content;

[0161] In some embodiments, at least one of the loudness levels calculated for the one or more portions of the audio content is one or more of: a specific loudness level relating to one or more frequency bands, a broadband loudness level across a broadband range, a wideband loudness level across a wideband range, a broadband loudness level across multiple frequency bands, a wideband loudness level across multiple frequency bands, etc.

[0162] In some embodiments, at least one of the loudness levels calculated for the one or more portions of the audio content is one or more of an instantaneous loudness level or a loudness level smoothed over one or more time intervals.

[0163] In one embodiment, the one or more operations relate to one or more of adjusting a dialogue loudness level, gain smoothing, gain limiting, dynamic equalization, noise compensation, and the like.

[0164] In one embodiment, the first apparatus is further configured to: extract one or more dialogue loudness levels from the encoded audio signal; adjust the one or more dialogue loudness levels to one or more reference dialogue loudness levels; etc.

[0165] In one embodiment, the first apparatus is further configured to: extract one or more auditory scene analysis (ASA) parameters from the encoded audio signal; vary one or more time constants used in smoothing a gain applied to the audio content, the gain related to one or more of the one or more DRC gains; perform gain smoothing or gain limiting, etc.

[0166] In one embodiment, the first apparatus is further configured to: determine that a reset event occurs in the one or more portions of the audio content based on an indication of a reset event, the indication of the reset being extracted from the encoded audio signal; and in response to determining that the reset event occurs in the one or more portions of the audio content, performing one or more actions on one or more gain smoothing operations that are running at the time of determining that the reset event occurs in the one or more portions of the audio content.

[0167] In one embodiment, the first apparatus is further configured to perform the following steps: maintaining a histogram of instantaneous loudness levels, the histogram being populated with instantaneous loudness levels calculated from a time interval in the audio content; determining whether a specific loudness level is above a threshold in a high probability region of the histogram, the specific loudness level being calculated from a portion of the audio content; and in response to determining that the specific loudness level is above the threshold in the high probability region of the histogram, determining that a loudness transition is occurring and, e.g., shortening a time constant used in gain smoothing to speed up the loudness transition.

[0168] 6B illustrates an example process flow that may be implemented by an audio encoder described herein. In block 652 of FIG. 6B, a second device (such as audio encoder 150 of FIG. 1B) receives audio content in a source audio format.

[0169] At block 654, the second device obtains definition data for one or more dynamic range compression curves.

[0170] In block 656, the second device generates an audio signal including the audio content and the definition data for the one or more dynamic range compression curves.

[0171] In one embodiment, the second device is further configured to perform the steps of: determining one or more identifiers for the one or more dynamic range compression curves; and retrieving the definition data for the one or more dynamic range compression curves from a reference data store based on the one or more identifiers.

[0172] In one embodiment, the second device is further configured to: calculate one or more dialogue loudness levels for the one or more portions of the audio content; and encode the one or more dialogue loudness levels together with the one or more portions of the audio content into the encoded audio signal.

[0173] In one embodiment, the second device is configured to: perform an auditory event scene (ASA) on the one or more portions of the audio content; generate one or more ASA parameters based on results of the ASA on the one or more portions of the audio content; encode the one or more ASA parameters together with the one or more portions of the audio content into the encoded audio signal; and the like.

[0174] In one embodiment, the second device is further configured to: determine that one or more reset events occur in the one or more portions of the audio content; and encode one or more indications of the one or more reset events together with the one or more portions of the audio content into the encoded audio signal.

[0175] In an embodiment, the second device is further configured to encode the one or more portions of the audio content into one or more audio data frames or audio data blocks.

[0176] In one embodiment, a first DRC gain of the one or more DRC gains applies to each channel in a first real subset of the set of all channels in a particular speaker configuration corresponding to the particular playback environment, while a second, different DRC gain of the one or more DRC gains applies to each channel in a second real subset of the set of all channels in the particular speaker configuration corresponding to the particular playback environment.

[0177] In one embodiment, a first DRC gain of the one or more DRC gains applies to a first frequency band and a second, different DRC gain of the one or more DRC gains applies to a second, different frequency band.

[0178] In some embodiments, the one or more portions of the audio content include one or more audio data frames or audio data blocks. In some embodiments, the encoded audio signal is part of an audiovisual signal.

[0179] In one embodiment, the one or more DRC gains are defined in the loudness domain.

[0180] 6C illustrates an example process flow that may be implemented by an audio encoder as described herein. In block 662 of FIG. 6C, a third device (such as, for example, audio encoder 150 of FIG. 1B) generates audio content encoded for a reference speaker configuration.

[0181] In block 664, the second device downmixes the audio content encoded for the reference speaker configuration into downmix audio content encoded for the specific speaker configuration.

[0182] In block 666, the second device performs one or more gain adjustments on respective portions of the downmix audio content encoded for the particular speaker configuration.

[0183] In block 668, the second device performs loudness measurements on the individual portions of the downmix audio content.

[0184] In block 670, the second device generates an audio signal including audio content encoded for the reference speaker configuration and downmix loudness metadata generated based at least in part on loudness measurements for individual portions of the downmix audio content.

[0185] In some embodiments, the loudness measurement for the individual parts of the downmix audio content is performed after the one or more gain adjustments have been applied to the individual parts of the downmix audio content. In some embodiments, the loudness measurement is based on the Loudness-K-weighted-Full-Scale (LKFS) standard. In some other embodiments, the loudness measurement is based on a loudness standard other than the Loudness-K-weighted-Full-Scale (LKFS) standard.

[0186] In an embodiment, the audio content encoded for a reference speaker configuration is downmixed to downmix audio content encoded for a specific speaker configuration based on one or more types of downmix operation: loudness measurements for individual portions of the downmix audio content include loudness measurements for individual portions of the downmix audio content related to each of said one or more types of downmix operation.

[0187] In one embodiment, the third device is configured to prevent downmixed audio content for the particular speaker configuration from being encoded in the audio signal.

[0188] 6D illustrates an exemplary process flow that may be implemented by an audio decoder described herein. In block 682 of FIG. 6D, a fourth device (such as audio decoder 100 of FIG. 1A) operating with a particular speaker configuration receives an audio signal including audio content and downmix loudness metadata encoded for a reference speaker configuration.

[0189] In block 684, the first device downmixes the audio content encoded for the reference speaker configuration to a downmix audio content encoded for the particular speaker configuration.

[0190] In block 686, the first device performs one or more gain adjustments on respective portions of the downmix audio content encoded for a particular speaker configuration, where the one or more gain adjustments are not based on the downmix loudness metadata but correspond to one or more gain adjustments performed by an upstream audio encoder prior to generating the downmix loudness metadata by the upstream audio encoder.

[0191] In block 688, the first device performs one or more additional gain adjustments on respective portions of the downmix audio content encoded for a particular speaker configuration, the one or more additional gain adjustments being based on the downmix loudness metadata.

[0192] In an embodiment, the first device is further configured to: determine a specific type of a downmix operation based on one or more selection factors; apply the specific type of downmix operation in downmixing the audio content encoded for the reference speaker configuration to a downmix audio content encoded for the specific speaker configuration; determine, from one or more sets of downmix loudness parameters in the downmix loudness metadata, a specific set of downmix loudness parameters to which the specific type of downmix operation corresponds; and perform the one or more additional gain adjustments on individual portions of the downmix audio content encoded for the specific speaker configuration based at least in part on the specific set of downmix loudness parameters.

[0193] In an embodiment, the one or more gain adjustments do not generate an expected loudness in a downmix sound output for at least one individual portion of the one or more individual portions of the downmix audio content, and the one or more additional gain adjustments are performed to generate an expected loudness in a downmix sound output for the at least one individual portion of the one or more individual portions of the downmix audio content.

[0194] In one embodiment, the reference speaker configuration is a surround speaker configuration and the specific speaker configuration is a two-channel configuration.

[0195] In some embodiments, the audio content encoded for the reference speaker configuration is downmixed to downmix audio content encoded for the specific speaker configuration based on one or more downmix equations.

[0196] In an embodiment, the downmix loudness metadata comprises one or more sets of downmix loudness parameters, each set of the one or more sets of downmix loudness parameters corresponding to a distinct type of downmix operation of the one or more types of downmix operation to which the one or more sets of downmix loudness parameters correspond.

[0197] In an embodiment, the one or more types of downmix operations include at least one of an LtRt downmix operation or a LoRo downmix operation.

[0198] In one embodiment, the one or more gain adjustments include at least one gain adjustment related to one or more of dialogue normalization, dynamic range compression, or a fixed attenuation to protect against downmix overload.

[0199] In one embodiment, the one or more gain adjustments use different gain adjustment parameter values ​​for at least two of the respective portions of the audio content.

[0200] In one embodiment, the downmix loudness metadata represents a portion of overall audio metadata that is encoded in the audio signal. In one embodiment, the downmix loudness metadata comprises a data field indicating a downmix loudness offset. In one embodiment, the encoded audio signal is part of an audiovisual signal.

[0201] In one embodiment, an apparatus having a processor and configured to perform any of the methods described herein.

[0202] In one embodiment, a non-transitory computer-readable storage medium containing software instructions that, when executed by one or more processors, cause the performance of any of the methods described herein. It is noted that although separate embodiments are discussed herein, any combination of the embodiments and / or sub-embodiments discussed herein may be combined to form further embodiments.

[0203] 19. Implementation Mechanism - Overview of Hardware According to an embodiment, the techniques described herein are implemented by one or more special purpose computing devices. The special purpose computing devices may be hard-wired to perform the techniques, or may include digital electronic devices, such as one or more application specific integrated circuits (ASICs) or field programmable gate arrays (FPGAs), that are persistently programmed to perform the techniques, or may include one or more general purpose hardware processors that are programmed to perform the techniques according to program instructions in firmware, memory, other storage, or combinations. Such special purpose computing devices may combine custom hard-wired logic, ASICs, or FPGAs with custom programming to accomplish the techniques. The special purpose computing devices may be desktop computer systems, portable computer systems, handheld devices, networking devices, or any other devices incorporating hard-wired and / or program logic to implement the techniques.

[0204] For example, Figure 7 is a block diagram illustrating a computer system 700 in which an embodiment of the invention may be implemented. Computer system 700 includes a bus 702 or other communication mechanism for communicating information, and a hardware processor 704 coupled to bus 702 for processing information. Hardware processor 704 may be, for example, a general-purpose microprocessor.

[0205] Computer system 700 also includes a main memory 706, coupled to bus 702, such as a random access memory (RAM) or other dynamic storage device, for storing information and instructions to be executed by processor 704. Main memory 706 may also be used for storing temporary variables or other intermediate information during execution of instructions to be executed by processor 704. Such instructions, when stored in a non-transitory storage medium accessible to processor 704, cause computer system 700 to become a device-specific, special-purpose machine for performing operations specified in the instructions.

[0206] Computer system 700 further includes a read only memory (ROM) 708 or other static storage device coupled to bus 702 for storing static information and instructions for processor 704. A storage device 710, such as a magnetic disk or optical disk, is provided and coupled to bus 702 for storing information and instructions.

[0207] Computer system 700 may be coupled via bus 702 to a display 712, such as a liquid crystal display (LCD), for displaying information to a computer user. An input device 714, including alphanumeric and other keys, is coupled to bus 702 for communicating information and command selections to processor 704. Another type of user input device is a cursor control 716, such as a mouse, trackball, or cursor direction keys, for communicating directional information and command selections to processor 704 and for controlling cursor movement on display 712. This input device typically has two degrees of freedom in two axes, a first axis (e.g., x) and a second axis (e.g., y), allowing the device to specify a position in a plane.

[0208] Computer system 700 may use device-specific hard-wired logic, one or more ASICs or FPGAs, firmware and / or program logic that in combination with the computer system configures or programs computer system 700 as a special purpose machine to implement the techniques described herein. According to one embodiment, the techniques are performed by computer system 700 in response to processor 704 executing one or more sequences of one or more instructions contained in main memory 706. Such instructions may be read into main memory 706 from another storage medium, such as storage device 710. Execution of the sequences of instructions contained in main memory 706 causes processor 704 to perform the process steps described herein. In alternative embodiments, hard-wired circuitry may be used in place of or in combination with software instructions.

[0209] The term "storage media" as used herein refers to any non-transitory medium that stores data and / or instructions that cause a machine to operate in a specific manner. Such storage media may include non-volatile media and / or volatile media. Non-volatile media include, for example, optical or magnetic disks, such as storage device 710. Volatile media include dynamic memory, such as main memory 706. Common forms of storage media include, for example, floppy disks, flexible disks, hard disks, solid state drives, magnetic tape or any other magnetic data storage medium, CD-ROMs, any other optical data storage medium, any physical medium with a pattern of holes, RAM, PROMs and EPROMs, flash EPROMs, NVRAMs, any other memory chips or cartridges.

[0210] Storage media is distinct from, but may be used in conjunction with, transmission media, which participate in transferring information between storage media. For example, transmission media include coaxial cables, copper wire and fiber optics, including the wires that comprise bus 702. Transmission media can also take the form of acoustic or light waves, such as those generated during radio wave and infrared data communications.

[0211] Various forms of media may be involved in carrying one or more sequences of one or more instructions to the processor 704 for execution. For example, the instructions may initially be carried on a magnetic disk or solid state drive of a remote computer. The remote computer can load the instructions into its dynamic memory and send the instructions over a telephone line using a modem. A modem local to the computer system 700 can receive the data on the telephone line and use an infrared transmitter to convert the data to an infrared signal. An infrared detector can receive the data carried in the infrared signal and appropriate circuitry can place the data on the bus 702. The bus 702 carries the data to main memory 706, from which the processor 704 retrieves and executes the instructions. The instructions received by main memory 706 may optionally be stored on storage device 710 either before or after execution by processor 704.

[0212] Computer system 700 also includes a communication interface 718 coupled to bus 702. The communication interface 718 provides a two-way data communication coupling to a network link 720 that is connected to a local network 722. For example, the communication interface 718 may be an Integrated Services Digital Network (ISDN) card, cable modem, satellite modem, or a modem to provide a data communication connection to a corresponding type of telephone line. As another example, the communication interface 718 may be a local area network (LAN) card to provide a data communication connection to a compatible LAN. A wireless link may also be implemented. In any such implementation, the communication interface 718 sends and receives electrical, electromagnetic or optical signals that carry digital data streams representing various types of information.

[0213] Network link 720 typically provides data communication through one or more networks to other data devices. For example, network link 720 may provide a connection through a local network 722 to a host computer 724 or to data equipment operated by an Internet Service Provider (ISP) 726. ISP 726 provides data communication services through the world-wide packet data communication network now commonly referred to as the "Internet" 728. Local network 722 and Internet 728 both use electrical, electromagnetic or optical signals that carry digital data streams. The signals through the various networks and the signals on network link 720 and through communication interface 718, which carry the digital data to and from computer system 700, are exemplary forms of transmission media.

[0214] Computer system 700 can send messages and receive data, including program code, through the network(s), network link 720 and communication interface 718. In the Internet example, a server 730 might transmit a requested code for an application program through Internet 728, ISP 726, local network 722 and communication interface 718.

[0215] The received code may be executed by processor 704 as it is received, and / or stored in storage device 710, or other non-volatile storage for later execution.

[0216] 20. Equivalents, extensions, substitutions, etc. In the foregoing specification, exemplary embodiments of the invention have been described with reference to numerous specific details that may vary from implementation to implementation. Thus, the sole and exclusive indication of what is, and what is intended by the applicant to be, the invention is the claims of the patent issued on this application in the specific form in which such claims are patented, including any subsequent amendments thereto. Any definitions expressly set forth herein for terms contained in such claims shall govern the meaning of such terms as used in the claims. Thus, no limitations, elements, attributes, features, advantages or characteristics not expressly set forth in the claims should limit the scope of such claims in any manner. Thus, the specification and drawings are to be regarded in an illustrative and not restrictive sense.

[0217] Several aspects will be described. [Aspect 1] generating encoded audio content for a reference loudspeaker configuration; downmixing the audio content encoded for the reference speaker configuration to a downmix audio content encoded for a specific speaker configuration; performing one or more gain adjustments on individual portions of the downmix audio content encoded for the particular speaker configuration; performing a loudness measurement on the individual portions of the downmix audio content; generating an audio signal comprising the audio content encoded for the reference speaker configuration and downmix loudness metadata generated at least in part based on the loudness measurements for the individual portions of the downmix audio content. A method performed by one or more computing devices. [Aspect 2] 2. The method of claim 1, wherein the reference speaker configuration is a surround speaker configuration and the specific speaker configuration is a two-channel configuration. [Aspect 3] 3. The method of claim 1 or 2, wherein the audio content encoded for the reference speaker configuration is downmixed to the downmix audio content encoded for the specific speaker configuration based on one or more downmix equations. Aspect 4 4. The method of any one of aspects 1 to 3, wherein the downmix loudness metadata comprises one or more sets of downmix loudness parameters, each set of the one or more sets of downmix loudness parameters corresponding to a distinct type of downmix operation of one or more types of downmix operation to which the one or more sets of downmix loudness parameters correspond. Aspect 5 5. The method of embodiment 4, wherein the one or more types of downmix operations include at least one of an LtRt downmix operation or a LoRo downmix operation. Aspect 6 downmixing the audio content encoded for the reference speaker configuration into a downmix audio content encoded for a specific speaker configuration based on one or more types of downmix operations, 6. The method of any one of aspects 1 to 5, wherein performing loudness measurements on the individual portions of the downmix audio content includes performing loudness measurements on the individual portions of the downmix audio content associated with each of the one or more types of downmix operations. Aspect 7 7. The method of any one of aspects 1 to 6, wherein the one or more gain adjustments include at least one gain adjustment related to one or more of dialogue normalization, dynamic range compression, or fixed attenuation to protect against downmix overload. Aspect 8 8. The method of any one of aspects 1 to 7, wherein the loudness measurement for the individual portions of the downmix audio content is performed after the one or more gain adjustments have been applied to the individual portions of the downmix audio content. Aspect 9 9. The method of any one of aspects 1 to 8, wherein the one or more gain adjustments use different gain adjustment parameter values ​​for at least two of the individual portions of the audio content. Aspect 10 10. The method of claim 9, wherein the at least two of the individual portions of the audio content represent audio content portions at at least two different times. Aspect 11 11. The method of any one of aspects 1 to 10, further comprising: preventing the downmixed audio content for the particular speaker configuration from being encoded in the audio signal. Aspect 12 12. The method of any one of the preceding aspects, wherein the loudness measurement is based on the Loudness K-weighted Full Scale (LKFS) standard. Aspect 13 12. The method of any one of aspects 1 to 11, wherein the loudness measurement is based on a loudness standard other than the Loudness K-weighted Full Scale (LKFS) standard. Aspect 14 14. The method of any one of aspects 1-13, wherein the downmix loudness metadata represents a portion of overall audio metadata encoded in the audio signal. Aspect 15 15. The method of any one of aspects 1 to 14, wherein the downmix loudness metadata comprises a data field indicating a downmix loudness offset. Aspect 16 16. The method of any one of aspects 1-15, wherein the encoded audio signal is part of an audiovisual signal. Aspect 17 receiving, by an audio decoder operating with a particular speaker configuration, an audio signal comprising audio content and downmix loudness metadata encoded for a reference speaker configuration; downmixing the audio content encoded for the reference speaker configuration to a downmix audio content encoded for the specific speaker configuration; performing one or more gain adjustments on individual portions of the downmix audio content encoded for the particular speaker configuration, the one or more gain adjustments not being based on downmix loudness metadata; performing one or more additional gain adjustments on the respective portions of the downmix audio content encoded for the particular speaker configuration, the one or more additional gain adjustments being based on the downmix loudness metadata. A method performed by one or more computing devices. Aspect 18 20. The method of claim 17, wherein the reference speaker configuration is a surround speaker configuration and the specific speaker configuration is a two-channel configuration. Aspect 19 determining a particular type of downmix operation based on one or more selection factors; applying the particular type of downmix operation in downmixing audio content encoded for the reference speaker configuration to a downmix audio content encoded for the particular speaker configuration; determining, from one or more sets of downmix loudness parameters in the downmix loudness metadata, a particular set of downmix loudness parameters to which the particular type of downmix operation corresponds; and performing the one or more additional gain adjustments on the respective portions of the downmix audio content encoded for the particular speaker configuration based at least in part on the particular set of downmix loudness parameters. 19. The method of embodiment 17 or 18. Aspect 20 20. The method of any one of aspects 17 to 19, wherein the audio content encoded for the reference speaker configuration is downmixed to the downmix audio content encoded for the specific speaker configuration based on one or more downmix equations, the one or more downmix equations being the same downmix equations used by an audio encoder generating the audio signal. Aspect 21 21. The method of any one of aspects 17 to 20, wherein the one or more gain adjustments include at least one gain adjustment related to one or more of dialogue normalization, dynamic range compression, or fixed attenuation to protect against downmix overload. Aspect 22 22. The method of any one of aspects 17 to 21, wherein the one or more gain adjustments represent a particular set of gain adjustments determined from one or more of a set of null gains, a set of gain adjustments including gain adjustments related to dynamic range compression (DRC), a set of gain adjustments excluding gain adjustments related to DRC, a set of gain adjustments including gain adjustments related to dialogue normalization, a set of gain adjustments excluding gain adjustments related to dialogue normalization, or a set of gain adjustments including gain adjustments related to both DRC and dialogue normalization. Aspect 23 23. The method of any one of aspects 17 to 22, wherein the one or more gain adjustments use different gain adjustment parameter values ​​for at least two of the individual portions of the audio content. Aspect 24 24. The method of any one of aspects 17 to 23, wherein the downmix loudness metadata represents a portion of overall audio metadata encoded in the audio signal. Aspect 25 20. The method of claim 17, wherein the downmix loudness metadata has a data field indicating a downmix loudness offset, and the one or more additional gain adjustments are made based at least in part on the downmix loudness offset. Aspect 26 26. The method of any one of aspects 17 to 25, wherein the one or more gain adjustments do not generate an expected loudness in a downmix sound output for at least one individual portion of the one or more individual portions of the downmix audio content, and the one or more additional gain adjustments are performed to generate an expected loudness in a downmix sound output for the at least one individual portion of the one or more individual portions of the downmix audio content. Aspect 27 27. The method of any one of aspects 17-26, wherein the encoded audio signal is part of an audiovisual signal. Aspect 28 28. A media processing system configured to perform any of the methods described in any one of aspects 1 to 27. Aspect 29 28. An apparatus having a processor configured to perform any of the methods described in any one of aspects 1 to 27. Aspect 30 A non-transitory computer-readable storage medium comprising software instructions that, when executed by one or more processors, cause performance of the method of any one of embodiments 1 to 27.

Claims

1. 1. A method for gain adjusting an audio signal based on loudness metadata generated by an encoder, comprising: receiving, by an audio decoder operating in a playback channel configuration different from a reference channel configuration, an audio signal for the reference channel configuration, the audio signal including audio sample data for each channel of the reference channel configuration and loudness metadata generated by the encoder, the loudness metadata including loudness metadata for a plurality of channel configurations including the playback channel configuration and the reference channel configuration; selecting loudness metadata for the playback channel configuration from loudness metadata for the plurality of channel configurations; determining a loudness adjustment gain from loudness metadata for the playback channel configuration; applying the loudness adjustment gain as a portion of an overall gain applied to the audio sample data to generate output audio sample data for each channel of the playback channel configuration; the loudness adjustment gain depends on a loudness level indicated by loudness metadata for the playback channel configuration and on a reference loudness level; The playback channel configuration is a two-channel configuration. method.

2. A non-transitory computer readable storage medium storing software instructions that, when executed by one or more processors, perform the following: receiving, by an audio decoder operating in a playback channel configuration different from a reference channel configuration, an audio signal for the reference channel configuration, the audio signal including audio sample data for each channel of the reference channel configuration and encoder-generated loudness metadata, the encoder-generated loudness metadata including loudness metadata for a plurality of channel configurations including the playback channel configuration and the reference channel configuration; selecting loudness metadata for the playback channel configuration from loudness metadata for the plurality of channel configurations; determining a loudness adjustment gain from loudness metadata for the playback channel configuration; and applying the loudness adjustment gain as part of an overall gain applied to the audio sample data to generate output audio sample data for each channel of the playback channel configuration; the loudness adjustment gain depends on a loudness level indicated by loudness metadata for the playback channel configuration and on a reference loudness level; The playback channel configuration is a two-channel configuration. Medium.

3. 1. An audio signal processing apparatus for gain adjusting an audio signal based on loudness metadata generated by an encoder, the audio signal processing apparatus comprising: receiving, by an audio decoder operating in a playback channel configuration different from a reference channel configuration, an audio signal for the reference channel configuration, the audio signal including audio sample data for each channel of the reference channel configuration and loudness metadata generated by the encoder, the loudness metadata including loudness metadata for a plurality of channel configurations including the playback channel configuration and the reference channel configuration; selecting loudness metadata for the playback channel configuration from loudness metadata for the plurality of channel configurations; determining a loudness adjustment gain from loudness metadata for the playback channel configuration; applying the loudness adjustment gain as part of an overall gain applied to the audio sample data to generate output audio sample data for each channel of the playback channel configuration; the loudness adjustment gain depends on a loudness level indicated by loudness metadata for the playback channel configuration and on a reference loudness level; The playback channel configuration is a two-channel configuration. Audio signal processing device.

4. A computer program product for causing a computer to carry out the method of claim 1.

Images