Processing spatially diffuse or large audio objects

Abstract

To provide a processing of a spatially spread or large audio object.SOLUTION: A spreaded or spatially large audio object may be specified for special processing. A decorrelation process may be performed on an audio signal corresponding to the large audio object to generate an audio signal of the decorrelated large audio object. The audio signal of these decorrelated large audio objects may be associated with object positions. The object position can be a static or time-varying position. For example, the audio signal of the decorrelated large audio object may be rendered to a virtual or actual speaker position. The decorrelation, association, and / or scene simplification process may be performed prior to a process of encoding audio data.SELECTED DRAWING: Figure 6A

Description

[Technical field]

[0001] CROSS-REFERENCE TO RELATED APPLICATIONS This application claims priority from Spanish Patent Application No. P201331193, filed July 31, 2013, and U.S. Provisional Application No. 61 / 885,805, filed October 2, 2013, the contents of each of which are incorporated herein by reference in their entirety.

[0002] Technical Field The present disclosure relates to processing audio data, and in particular to processing audio data corresponding to diffuse or spatially large audio objects. [Background technology]

[0003] Since the introduction of sound to cinema in 1927, there has been a steady advancement in the technology used to capture and reproduce the artistic intent of a film soundtrack. In the 1970s, Dolby introduced a cost-effective means of encoding and distributing three screen channels mixed with a mono surround channel. In the 1990s, Dolby brought digital sound to cinemas with the 5.1 channel format, which gave discrete left, center and right screen channels, left and right surround arrays and a subwoofer channel for low-frequency effects. Dolby Surround 7.1, introduced in 2010, increased the number of surround channels by splitting the existing left and right surround channels into four "zones".

[0004] Both cinema and home theater audio playback systems are becoming more and more versatile and complex. Home theater audio playback systems include more and more speakers. As the number of channels increases and loudspeaker layouts move from planar, two-dimensional (2D) arrays to three-dimensional (3D) arrays that include height, reproducing sound in the playback environment becomes an increasingly complex process. Improved audio processing methods would be desirable. [Prior art documents] [Non-patent literature]

[0005] [Non-Patent Document 1] V. Pulkki, Compensating Displacement of Amplitude-Panned Virtual Sources, Audio Engineering Society (AES) International Conference on Virtual, Synthetic and Entertainment Audio [Non-Patent Document 2] Robinson and Vinton, "Automated Speech / Other Discrimination for Loudness Monitoring", Audio Engineering Society, Preprint number 6437 of Convention 118, May 2005 Summary of the Invention [Problem to be solved by the invention]

[0006] Improved methods are provided for processing diffuse or spatially large audio objects. As used herein, the term "audio object" may refer to an audio signal (also referred to herein as "audio object signal") and associated metadata. The associated metadata may be generated or "authored" without reference to any particular playback environment. The associated metadata may include audio object position data, audio object gain data, audio object size data, audio object trajectory data, and the like. As used herein, the term "rendering" may refer to a process of converting an audio object into speaker feed signals for a particular playback environment. The rendering process may be performed, at least in part, in accordance with said associated metadata and in accordance with playback environment data. The playback environment data may include an indication of a number of speakers in the playback environment and an indication of a position of each speaker within the playback environment.

[0007] Spatially large audio objects are not intended to be perceived as point sources, but instead should be perceived as covering a large spatial region. In some cases, large audio objects should be perceived as surrounding the listener. Such audio effects may not be achievable by simple panning, but rather may require additional processing. To produce convincing spatial object size or spatial diffuseness, a significant proportion of the speaker signals in the playback environment should be independent of each other, or at least uncorrelated (e.g., independent with respect to first order cross-correlation or covariance). Sufficiently complex rendering systems, such as those intended for theaters, may be able to provide such decorrelation. However, less complex rendering systems, such as those intended for home theater systems, may not be able to provide sufficient decorrelation. [Means for solving the problem]

[0008] Some implementations described herein may involve identifying diffuse or spatially large audio objects for special processing. A decorrelation process may be performed on the audio signals corresponding to the large audio objects to generate decorrelated large audio object audio signals. These decorrelated large audio object audio signals may be associated with object positions. The object positions may be static or time-varying positions. The association process may be independent of the actual playback speaker configuration. For example, decorrelated large audio object audio signals may be rendered to virtual speaker positions. In some implementations, the output of such a rendering process may be input to a scene simplification process.

[0009] Thus, at least some aspects of the present disclosure may be implemented in a method that may involve receiving audio data including an audio object, which may include an audio object signal and associated metadata, which may include at least audio object size data.

[0010] The method may involve determining a large audio object having an audio object size greater than a threshold size based on the audio object size data, and performing a decorrelation process on the audio signal of the large audio object to generate a decorrelated audio signal of the large audio object. The method may involve associating the decorrelated audio signal of the large audio object with an object position. The association process may be independent of an actual playback speaker configuration. The actual playback speaker configuration may finally be used to render the decorrelated audio signal of the large audio object to speakers of the playback environment.

[0011] The method may involve receiving decorrelated metadata for a large audio object. The decorrelation process may be performed, at least in part, in accordance with the decorrelation metadata. The method may involve encoding audio data output from the correlation process. In some implementations, the encoding process may not involve encoding the decorrelated metadata for the large audio object.

[0012] The object positions may include positions corresponding to at least a portion of the audio object position data of the received audio object. At least a portion of the object positions may be static. However, in some implementations, at least a portion of the object positions may change over time.

[0013] The associating process may involve rendering the decorrelated audio signals of the large audio object according to the virtual speaker positions. In some examples, the receiving process may involve receiving one or more audio bed signals corresponding to the speaker positions. The method may involve mixing the decorrelated audio signals of the large audio object with at least a portion of the received audio bed signal or the received audio object signal. The method may involve outputting the decorrelated audio signals of the large audio object as an additional audio bed signal or audio object signal.

[0014] The method may involve applying a level adjustment process to the decorrelated audio signal of the large audio object. In some implementations, the metadata of the large audio object may include audio object position metadata, and the level adjustment process may depend at least in part on the audio object size metadata and the audio object position metadata of the large audio object.

[0015] The method may involve attenuating or removing the audio signal of the large audio object after the decorrelation process has been performed, however, in some implementations the method may involve retaining the audio signal corresponding to the point source contribution of the large audio object after the decorrelation process has been performed.

[0016] The large audio object metadata may include audio object position metadata. In some such implementations, the method may involve calculating contributions from virtual sources within an audio object area or volume defined by the large audio object position data and the large audio object size data. The method may involve determining a set of audio object gain values ​​for each of a plurality of output channels based at least in part on the calculated contributions. The method may involve mixing the decorrelated audio signal of the large audio object with an audio signal for an audio object that is spatially separated from the large audio object by a threshold amount of distance.

[0017] In some implementations, the method may involve performing an audio object clustering process after the decorrelation process. In some such implementations, the audio object clustering process may be performed after the association process.

[0018] The method may further involve evaluating the audio data to determine a content type. In some such implementations, the decorrelation process may be selectively performed as a function of the content type. For example, the amount of decorrelation to be performed may depend on the content type. The decorrelation process may involve delays, all-pass filters, pseudo-random filters, and / or reverberation algorithms.

[0019] The methods disclosed herein may be implemented via hardware, firmware, software stored on one or more non-transitory media, and / or combinations thereof. For example, at least some aspects of the present disclosure may be implemented in an apparatus including an interface system and a logic system. The interface system may include a user interface and / or a network interface. In some implementations, the apparatus may include a memory system. The interface system may include at least one interface between the logic system and the memory system.

[0020] A logic system may include at least one processor, such as a general purpose single-chip or multi-chip processor, a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic device, discrete gate or transistor logic, discrete hardware components, and / or combinations thereof.

[0021] In some implementations, the logic system may be capable of receiving audio data including an audio object via an interface system. The audio object may include an audio object signal and associated metadata. In some implementations, the metadata may include at least audio object size data. The logic system may be capable of determining a large audio object having an audio object size greater than a threshold size based on the audio object size data, and performing a decorrelation process on the audio signal of the large audio object to generate a decorrelated audio signal of the large audio object. The logic system may be capable of associating the decorrelated audio signal of the large audio object with an object position.

[0022] The correlation process may be independent of the actual playback speaker configuration. For example, the correlation process may involve rendering the decorrelated audio signals of the large audio object to virtual speaker positions. The actual playback speaker configuration may finally be used to render the decorrelated audio signals of the large audio object to speakers of the playback environment.

[0023] The logic system may be capable of receiving decorrelation metadata for the large audio object via the interface system, and the decorrelation process may be performed at least in part according to the decorrelation metadata.

[0024] The logic system may be capable of encoding the audio data output from the correlation process. In some implementations, the encoding process may not involve encoding decorrelation metadata about larger audio objects.

[0025] At least a portion of the object position may be static. The metadata for the large audio object may include audio object position metadata. The object position may include a position corresponding to at least a portion of the audio object position metadata for the received audio object.

[0026] The receiving process may involve receiving one or more audio bed signals corresponding to the speaker positions. The logic system may be capable of mixing the decorrelated audio signal of the large audio object with the received audio bed signal or at least a portion of the received audio object signal. The logic system may be capable of outputting the decorrelated audio signal of the large audio object as an additional audio bed signal or audio object signal.

[0027] The logic system may be capable of applying a level adjustment process to the decorrelated audio signal of the large audio object, the level adjustment process may depend at least in part on audio object size metadata and audio object position metadata of the large audio object.

[0028] The logic system may be capable of attenuating or removing audio signals of large audio objects after the decorrelation process has been performed, however, the apparatus may be capable of retaining audio signals corresponding to point source contributions of large audio objects after the decorrelation process has been performed.

[0029] The logic system may be capable of calculating contributions from virtual sources within an audio object area or volume defined by the large audio object position data and the large audio object size data. The logic system may be capable of determining a set of audio object gain values ​​for each of a plurality of output channels based at least in part on the calculated contributions. The logic system may be involved in mixing the decorrelated audio signal of the large audio object with an audio signal for an audio object that is spatially separated from the large audio object by a threshold amount of distance.

[0030] The logic system may be capable of performing an audio object clustering process after the decorrelation process. In some implementations, the audio object clustering process may be performed after the association process.

[0031] The logic system may be capable of evaluating the audio data to determine a content type. The decorrelation process may be selectively performed depending on the content type. For example, the amount of decorrelation to be performed may depend on the content type. The decorrelation process may involve delays, all-pass filters, pseudo-random filters, and / or reverberation algorithms.

[0032] Details of one or more implementations of the subject matter described herein are set forth in the accompanying drawings and the description below. Other features, aspects, and advantages will become apparent from the description, drawings, and claims. Please note that the relative dimensions of the following figures may not be drawn to scale. [Brief description of the drawings]

[0033] [Figure 1] This figure shows an example of a playback environment with Dolby Surround 5.1 configuration. [Diagram 2] This figure shows an example of a playback environment with Dolby Surround 7.1 configuration. [Diagram 3] 1A and 1B show two example home theater playback environments including height speaker configurations. [Figure 4A] FIG. 1 shows an example of a graphical user interface (GUI) for depicting speaker zones at various heights in a virtual playback environment. [Figure 4B] FIG. 13 is a diagram illustrating another example of a playback environment. [Diagram 5] 1 is a flow diagram giving an example of audio processing for spatially large audio objects. [Figure 6A] FIG. 2 illustrates an example of components of an audio processing device capable of processing large audio objects. [Figure 6B] FIG. 2 illustrates an example of components of an audio processing device capable of processing large audio objects. [Figure 6C]FIG. 2 illustrates an example of components of an audio processing device capable of processing large audio objects. [Figure 6D] FIG. 2 illustrates an example of components of an audio processing device capable of processing large audio objects. [Figure 6E] FIG. 2 illustrates an example of components of an audio processing device capable of processing large audio objects. [Figure 6F] FIG. 2 illustrates an example of components of an audio processing device capable of processing large audio objects. [Figure 7] FIG. 1 is a block diagram illustrating an example of a system that can perform a clustering process. [Figure 8] FIG. 1 is a block diagram illustrating an example of a system capable of clustering objects and / or beds in an adaptive audio processing system. [Figure 9] A block diagram giving an example of the clustering process after decorrelation processing for large objects. [Figure 10A] FIG. 2 shows an example of a virtual source position relative to the playback environment. [Figure 10B] FIG. 13 shows alternative examples of virtual source positions for the playback environment. [Figure 11] 1 is a block diagram providing examples of components of an audio processing apparatus.Like reference numbers and designations in the various drawings indicate like elements. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0034] The following description is directed to certain implementations for purposes of describing some novel aspects of the disclosure and examples of contexts in which these novel aspects may be implemented. However, the teachings of the present disclosure may be applied in a variety of different ways. For example, while various implementations are described using a specific playback environment, the teachings of the present disclosure are broadly applicable to other known and future playback environments. Furthermore, the described implementations may be implemented, at least in part, in a variety of devices and systems, such as hardware, software, firmware, cloud-based systems, and the like. Thus, the teachings of the present disclosure are not intended to be limited to the implementations shown in the drawings and / or described herein, but rather are of broad applicability.

[0035] FIG. 1 shows an example of a playback environment with a Dolby Surround 5.1 configuration. In this example, the playback environment is a movie theater playback environment. Although Dolby Surround 5.1 was developed in the 1990s, this configuration is still widely deployed in home and movie theater playback environments. In the movie theater playback environment, a projector 105 may be configured to project a video image, for example for a movie, onto a screen 150. Audio data may be synchronized with the video image and processed by a sound processor 110. A power amplifier 115 may provide speaker feed signals to speakers of the playback environment 100.

[0036] The Dolby Surround 5.1 configuration includes a left surround channel 120 for the left surround array 122 and a right surround channel 125 for the right surround array 127. The Dolby Surround 5.1 configuration also includes a left channel 130 for the left speaker array 132, a center channel 135 for the center speaker array 137, and a right channel 140 for the right speaker array 142. In a cinema environment, these channels might be referred to as the left screen channel, the center screen channel, and the right screen channel, respectively. A separate low-frequency effects (LFE) channel 144 is provided for the subwoofer 145.

[0037] In 2010, Dolby provided an improvement to digital cinema sound by introducing Dolby Surround 7.1. FIG. 2 shows an example of a playback environment with a Dolby Surround 7.1 configuration. Digital projector 205 may be configured to receive digital video data and project a video image onto screen 150. Audio data may be processed by sound processor 210. Power amplifiers 215 may provide speaker feed signals to speakers in playback environment 200.

[0038] Similar to Dolby Surround 5.1, the Dolby Surround 7.1 configuration includes a left channel for the left speaker array 132, a center channel 135 for the center speaker array 137, a right channel 140 for the right speaker array 142, and an LFE channel 144 for the subwoofer 145. The Dolby Surround 7.1 configuration also includes a left side surround (Lss) array 220 and a right side surround (Rss) array 225, each of which may be driven by a single channel.

[0039] However, Dolby Surround 7.1 increases the number of surround channels by splitting the left and right surround channels of Dolby Surround 5.1 into four zones: left lateral surround array 220 and right lateral surround array 225, plus separate channels for left rear surround (Lrs) speaker 224 and right rear surround (Rrs) speaker 226. Increasing the number of surround zones in the playback environment 200 can significantly improve sound localization.

[0040] In an effort to create a more immersive environment, some playback environments may be configured with an increased number of speakers driven by an increased number of channels. Additionally, some playback environments may include speakers that are deployed at various heights, some of which may be "height speakers" configured to generate sound from an area above the seating area of ​​the playback environment.

[0041] 3A and 3B show two examples of home theater playback environments including height speaker configurations. In these examples, playback environments 300a and 300b include the main features of a Dolby Surround 5.1 configuration, including a left surround speaker 322, a right surround speaker 327, a left speaker 332, a right speaker 342, a center speaker 337, and a subwoofer 145. However, playback environment 300 includes an extension of the Dolby Surround 5.1 configuration for height speakers, which is sometimes referred to as a Dolby Surround 5.1.2 configuration.

[0042] FIG. 3A shows an example of a playback environment with height speakers mounted on the ceiling 360 of the home theater playback environment. In this example, the playback environment 300a includes a height speaker 352 in a left top middle (Ltm) position and a height speaker 357 in a right top middle (Rtm) position. In the example shown in FIG. 3B, the left speaker 332 and the right speaker 342 are Dolby Elevation speakers configured to reflect sound from the ceiling 360. If properly configured, the reflected sound can be perceived by a listener 365 as if the sound source was emanating from the ceiling 360. However, these speaker numbers and configurations are given by way of example only. Some current home theater implementations provide up to 34 speaker positions, and envisioned home theater implementations may allow even more speaker positions.

[0043] So the current trend is not only to include more speakers and more channels, but also speakers of different heights. As the number of channels increases and speaker layouts move from 2D to 3D, the task of positioning and rendering the sound becomes more difficult.

[0044] Accordingly, Dolby has developed a variety of tools, including but not limited to user interfaces, that enhance the capabilities and / or reduce the authoring complexity for 3D audio sound systems. Some such tools may be used to generate audio objects and / or metadata for audio objects.

[0045] 4A shows an example of a graphical user interface (GUI) depicting speaker zones at various heights in a virtual playback environment. GUI 400 may be displayed on a display device, for example, according to instructions from a logic system, according to signals received from a user input device, etc. Some such devices are described below with reference to FIG.

[0046] As used herein to refer to a virtual playback environment such as virtual playback environment 404, the term "speaker zone" refers generally to a logical construct that may or may not have a one-to-one correspondence with a playback speaker in a real playback environment. For example, a "speaker zone location" may or may not correspond to a specific playback speaker location in a movie theater playback environment. Instead, the term "speaker zone location" may generally refer to a zone in a virtual playback environment. In some implementations, a speaker zone in a virtual playback environment may correspond to a virtual speaker through the use of a virtualization technology such as, for example, Dolby Headphone™ (sometimes referred to as Mobile Surround™), which generates a virtual surround sound environment in real time using a set of two-channel stereo headphones. GUI 400 has seven speaker zones 402a at a first elevation and two speaker zones 402b at a second elevation, for a total of nine speaker zones in virtual playback environment 404. In this example, speaker zones 1-3 are in a front region 405 of virtual playback environment 404. The front area 405 may for example correspond to the area in a cinema playback environment where the screen 150 is located, or in a home where a television screen is located, etc.

[0047] Here, speaker zone 4 generally corresponds to speakers in left region 410, and speaker zone 5 corresponds to speakers in right region 415 of virtual playback environment 404. Speaker zone 6 corresponds to left rear region 412, and speaker zone 7 corresponds to right rear region 414 of virtual playback environment 404. Speaker zone 8 corresponds to speakers in top region 420a, and speaker zone 9 corresponds to speakers in top region 420b, which may be a virtual ceiling region. Thus, the locations of speaker zones 1-9 shown in FIG. 4A may or may not correspond to the locations of playback speakers in the actual playback environment. Additionally, other implementations may include more or fewer speaker zones and / or heights.

[0048] In various implementations described herein, a user interface such as GUI 400 may be used as part of an authoring tool and / or a rendering tool. In some implementations, the authoring tool and / or the rendering tool may be implemented via software stored on one or more non-transitory media. The authoring tool and / or the rendering tool may be implemented (at least in part) by hardware, firmware, etc., such as the logic system and other devices described below with reference to FIG. 11. In some authoring implementations, an associated authoring tool may be used to generate metadata about associated audio data. The metadata may include, for example, data indicating the position and / or trajectory of audio objects in three-dimensional space, speaker zone constraint data, etc. The metadata may be generated with respect to speaker zones 402 of the virtual playback environment 404, rather than with respect to a particular speaker layout of the actual playback environment. The rendering tool may receive the audio data and associated metadata and may calculate audio gain and speaker feed signals for the playback environment. Such audio gain and speaker feed signals may be calculated according to an amplitude panning process. The amplitude panning process can create the perception that a sound comes from a position P in the playback environment. For example, a speaker feed signal can be expressed as x i (t)=g i x(t) i=1,…,N (Equation 1) may be provided to playback speakers 1 to N of the playback environment according to

[0049] In formula (1), x i (t) represents the speaker feed signal applied to speaker i, and g irepresents the gain factor of the corresponding channel, x(t) represents the audio signal, and t represents time. The gain factor may be determined according to the amplitude panning methods described in, for example, Section 2, pp. 3-4 of "Phases 1 and 2", incorporated herein by reference. In some implementations, the gain may be frequency dependent. In some implementations, a time delay may be introduced by replacing x(t) with x(t-Δt).

[0050] In some rendering implementations, the audio playback data generated with reference to speaker zones 402 may be mapped to speaker locations of a wide range of playback environments, which may be in a Dolby Surround 5.1 configuration, a Dolby Surround 7.1 configuration, a Hamasaki 22.2 configuration, or other configurations. For example, with reference to FIG. 2, the rendering tool may map the audio playback data for speaker zones 4 and 5 to the left lateral surround array 220 and the right lateral surround array 225 of a playback environment having a Dolby Surround 7.1 configuration. The audio playback data for speaker zones 1, 2, and 3 may be mapped to the left screen channel 230, the right screen channel 240, and the center screen channel 235, respectively. The audio playback data for speaker zones 6 and 7 may be mapped to the left rear surround speaker 224 and the right rear surround speaker 226.

[0051] 4B shows another example of a playback environment. In some implementations, the rendering tool may map audio playback data for speaker zones 1, 2, and 3 to corresponding screen speakers 455 of the playback environment 450. The rendering tool may map audio playback data for speaker zones 4 and 5 to left lateral surround array 460 and right lateral surround array 465, and may map audio playback data for speaker zones 8 and 9 to left overhead speaker 470a and right overhead speaker 470b. Audio playback data for speaker zones 6 and 7 may be mapped to left rear surround speaker 480a and right rear surround speaker 480b.

[0052] In some authoring implementations, the authoring tool may be used to generate metadata about audio objects. The metadata may indicate the object's 3D position, rendering constraints, content type (e.g., dialogue, effects, etc.), and / or other information. Depending on the implementation, the metadata may include other types of data, such as width data, gain data, trajectory data, etc. Some audio objects may be static, while other audio objects may move.

[0053] Audio objects are rendered according to associated metadata, which typically includes positional metadata that describes the location of the audio object in three-dimensional space at a given time. When the audio objects are monitored or played in a playback environment, they are rendered using the speakers present in the playback environment according to said positional metadata, rather than being output to pre-defined physical channels as is the case in traditional channel-based systems such as Dolby 5.1 and Dolby 7.1.

[0054] In addition to position metadata, other types of metadata may be required to generate the intended audio effect. For example, in some implementations, metadata associated with an audio object may indicate the audio object size, sometimes referred to as "width". The size metadata may be used to indicate the spatial area or volume that the audio object occupies. Spatially large audio objects should be perceived as covering a large spatial area, rather than simply as a point source with a position defined only by the audio object position metadata. For example, in some cases, a large audio object should be perceived as occupying a significant portion of the playback environment, potentially even surrounding the listener.

[0055] The human auditory system is highly sensitive to changes in the correlation or coherence of the signals arriving at both ears, and maps this correlation to a perceived object size attribute if the normalized correlation is less than a value of +1. Therefore, to produce convincing spatial object size or spatial diffuseness, a significant proportion of the speaker signals in the playback environment should be mutually independent or at least uncorrelated (e.g., independent in terms of first-order cross-correlation or covariance). A satisfactory decorrelation process is typically fairly complex and usually involves time-varying filters.

[0056] A cinema soundtrack may contain hundreds of objects, each with associated position metadata, size metadata and possibly other spatial metadata. Furthermore, a cinema sound system may contain hundreds of speakers, which may be individually controlled to give a satisfying perception of audio object position and size. Thus, in a cinema, hundreds of objects may be reproduced by hundreds of speakers, and the object-to-speaker signal mapping consists of a very large matrix of pan coefficients. This matrix has up to N×N elements, when the number of objects is given by M and the number of speakers is given by N. This concerns the reproduction of diffuse or large sized objects. To create convincing spatial object size or spatial diffuseness, a significant proportion of the N speaker signals should be mutually independent or at least uncorrelated. This generally involves the use of a large number (up to N) of independent decorrelation processes, which causes a significant processing load for the rendering process. Furthermore, the amount of decorrelation may be different for each object, which further complicates the rendering process. Sufficiently complex rendering systems, such as those for commercial theaters, may be able to provide such decorrelation.

[0057] However, less complex rendering systems, such as those intended for home theater systems, may not be able to provide sufficient decorrelation. Some such rendering systems may not provide any decorrelation at all. Decorrelation programs that are simple enough to run on home theater systems may introduce artifacts. For example, comb filter artifacts may be introduced when a low-complexity decorrelation process is used following the downmix process.

[0058] Another potential problem is that in some applications, object-based audio is transmitted in the form of a backward-compatible mix (such as Dolby Digital or Dolby Digital Plus) augmented with additional information to extract one or more objects from the backward-compatible mix. Backward-compatible mixes usually do not include the effect of decorrelation. In some such systems, object reconstruction works reliably only if the backward-compatible mix was generated using a simple panning procedure. The use of a decorrelator in such a process can impair, sometimes severely, the audio object reconstruction process. In the past, this meant that one could either choose not to apply decorrelation in the backward-compatible mix, thereby compromising the artistic intent of the mix, or accept degradation in the object reconstruction process.

[0059] To address such potential problems, some implementations described herein involve identifying diffuse or spatially large audio objects for special processing. Such methods and apparatus may be particularly suitable for audio data to be rendered in a home theater. However, these methods and apparatus are not limited to home theater applications and have broad applicability.

[0060] Due to their spatially diffuse nature, objects with large size are not perceived as point sources with compact and concise locations. Therefore, multiple speakers are used to reproduce such spatially diffuse objects. However, the exact location of the speakers in the reproduction environment used to reproduce large audio objects is not as critical as the location of the speakers used to reproduce compact and small sized audio objects. Thus, high quality reproduction of large audio objects is possible without prior knowledge of the actual reproduction speaker configuration used to finally render the decorrelated signals of the large audio objects to the actual speakers of the reproduction environment. As a result, the decorrelation process for large audio objects can be performed "upstream", before the process of rendering the audio data for reproduction to a listener in a reproduction environment such as a home theater system. In some examples, the decorrelation process for large audio objects is performed before encoding the audio data for transmission to such a reproduction environment.

[0061] Such an implementation does not require the renderer of the playback environment to have high-complexity decorrelation capabilities, thereby allowing a rendering process that may be relatively simpler, more efficient, and less expensive. A backward-compatible downmix may include decorrelation effects to preserve the best possible artistic intent without the need to reconstruct objects for render-side decorrelation. A high-quality decorrelator can be applied to large audio objects upstream of the final rendering process, for example during the authoring or post-production process in a sound studio. Such a decorrelator may be robust with respect to downmixing and / or other downstream audio processing.

[0062] FIG. 5 is a flow diagram providing an example of audio processing for spatially large audio objects. The operations of method 500, as well as other methods described herein, are not necessarily performed in the order shown. Additionally, the methods may include more or fewer blocks than shown and / or described. The methods may be implemented, at least in part, by a logic system such as logic system 1110 shown in FIG. 11 and described below. Such a logic system may be a component of an audio processing system. Alternatively or additionally, such methods may be implemented via a non-transitory medium having software stored thereon. The software may include, at least in part, instructions for controlling one or more devices to perform the methods described herein.

[0063] In this example, method 500 begins at block 505 with receiving audio data including an audio object. The audio data may be received by an audio processing system. In this example, the audio object includes an audio object signal and associated metadata, where the associated metadata includes audio object size data. The associated metadata may also include audio object position data indicating the position of the audio object in three-dimensional space, decorrelation metadata, audio object gain information, etc. The audio data may also include one or more audio bed signals corresponding to speaker positions.

[0064] In this implementation, block 510 involves determining, based on the audio object size data, a large audio object having an audio object size greater than a threshold size. For example, block 510 may involve determining whether a numerical audio object size value exceeds a predetermined level. The numerical audio object size value may, for example, correspond to a portion of the playback environment that the audio object occupies. Alternatively or additionally, block 510 may involve determining whether another type of indication, such as a flag, decorrelation metadata, etc., indicates that the audio object has an audio object size greater than a threshold size. Although much of the discussion of method 500 involves processing a single large audio object, it will be appreciated that the same (or similar) processing may be applied to multiple large audio objects.

[0065] In this example, block 515 involves performing a decorrelation process on the audio signal of the large audio object to generate a decorrelated audio signal of the large audio object. In some implementations, the decorrelation process may be performed at least in part according to the received decorrelation metadata. The decorrelation process may involve delays, all-pass filters, pseudo-random filters, and / or reverberation algorithms.

[0066] Here, in block 520, the decorrelated audio signals of the large audio object are associated with object positions. In this example, the association process is independent of the actual playback speaker configuration that may be used to ultimately render the decorrelated audio signals of the large audio object on the actual playback speakers of the playback environment. However, in some alternative implementations, the object positions may correspond to the actual playback speaker positions. For example, according to some such alternative implementations, the object positions may correspond to the playback speaker positions of a commonly used playback speaker configuration. If an audio bed signal is received in block 505, the object positions may correspond to the playback speaker positions corresponding to at least some of the audio bed signal. Alternatively or additionally, the object positions may be positions corresponding to at least some of the audio object position data of the received audio objects. Thus, at least some of the object positions may be static and at least some of the object positions may change over time. In some implementations, block 520 may involve mixing the decorrelated audio signals of the large audio object with audio signals for audio objects that are spatially separated from the large audio object by a threshold distance.

[0067] In some implementations, block 520 may involve rendering the decorrelated large audio object audio signal according to the virtual speaker positions. Some such implementations may involve calculating contributions from virtual sources within an audio object area or volume defined by the large audio object position data and the large audio object size data. Such implementations may involve determining a set of audio object gain values ​​for each of a plurality of output channels based at least in part on those calculated contributions. Some examples are described below.

[0068] Some implementations may involve encoding audio data output from the correlation process. According to some such implementations, the encoding process involves encoding audio signals and associated metadata for audio objects. In some implementations, the encoding process includes a data compression process. The data compression process may be lossless or lossy. In some implementations, the data compression process involves a quantization process. According to some examples, the encoding process may not involve encoding decorrelation metadata for larger audio objects.

[0069] Some implementations involve performing an audio object clustering process, also referred to herein as a "scene simplification" process. For example, the audio object clustering process may be part of block 520. For implementations involving encoding, the encoding process may involve encoding audio data output from the audio object clustering process. In some such implementations, the audio object clustering process may be performed after a decorrelation process. Further examples of processes corresponding to the blocks of method 500, including the scene simplification process, are described below.

[0070] 6A-6F are block diagrams illustrating example components of an audio processing system capable of processing large audio objects as described herein. These components may correspond to modules of a logic system of the audio processing system, which may be implemented, for example, via hardware, firmware, software stored on one or more non-transitory media, and / or combinations thereof. The logic system may include one or more processors, such as general-purpose single-chip or multi-chip processors. The logic system may include a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field programmable gate array (FPGA), or other programmable logic device.

[0071] In FIG. 6A, an audio processing system 600 may detect a large audio object, such as a large audio object 605. The detection process may be substantially similar to one of the processes described with reference to block 510 of FIG. 5. In this example, an audio signal of the large audio object 605 is decorrelated by a decorrelation system 610 to generate a decorrelated large audio object signal 611. The decorrelation system 610 may perform the decorrelation process at least in part according to received decorrelation metadata for the large audio object 605. The decorrelation process may involve one or more of a delay, an all-pass filter, a pseudo-random filter, or a reverberation algorithm.

[0072] The audio processing system 600 may also receive other audio signals, which in this example are other audio objects and / or a bed 615, where the other audio objects are audio objects having a size below a threshold size for characterizing the audio object as a large audio object.

[0073] In this example, the audio processing system 600 can correlate the decorrelated audio signal 611 of the large audio object with other object positions. The object positions may be static or may change over time. The correlation process may be similar to one or more of the processes described above with reference to block 520 of FIG. 5.

[0074] The association process may involve a mixing process. The mixing process may be based, at least in part, on the distance between the large audio object position and another object position. In the implementation shown in FIG. 6A, the audio processing system 600 may mix the decorrelated large audio object signal 611 with at least some audio signals corresponding to the audio object and / or the bed 615. For example, the audio processing system 600 may be able to mix the decorrelated large audio object audio signal 611 with audio signals for other audio objects that are spatially separated from the large audio object by a certain threshold amount of distance.

[0075] In some implementations, the correlation process may involve a rendering process. For example, the correlation process may involve rendering the audio signals of the large audio object decorrelated according to the virtual speaker positions. After the rendering process, it may not be necessary to retain the audio signals corresponding to the large audio object received by the decorrelation system 610. Thus, the audio processing system 600 may be configured to attenuate or remove the audio signals of the large audio object 605 after the decorrelation process has been performed by the decorrelation system 610. Alternatively, the audio processing system 600 may be configured to retain at least a portion of the audio signals of the large audio object 605 (e.g., audio signals corresponding to point source contributions of the large audio object 605) after the decorrelation process has been performed.

[0076] In this example, the audio processing system 600 includes an encoder 620 capable of encoding the audio data, where the encoder 620 is configured to encode the audio data after the association process. In this implementation, the encoder 620 can apply an audio compression process to the audio data. The encoded audio data 622 can be stored and / or transmitted to other audio processing systems for downstream processing, playback, etc.

[0077] In the implementation shown in FIG. 6B, the audio processing system 600 has a level adjustment function. In this example, the level adjustment system 612 is configured to adjust the level of the output of the decorrelation system 610. The level adjustment process may depend on the metadata of the audio content in the original. In this example, the level adjustment process depends, at least in part, on the audio object size metadata and the audio object position metadata of the large audio object 605. Such level adjustment can be used to optimize the delivery of the decorrelator output to the audio object and / or other audio objects such as the bed 615. To improve the spatial spread of the resulting rendering, one may choose to mix multiple decorrelator outputs to other spatially distant object signals.

[0078] Alternatively or additionally, the level adjustment process may be used to ensure that sounds corresponding to the decorrelated large audio object 605 are only reproduced by the speakers from a certain direction. This may be achieved by only adding the decorrelator output to objects in the vicinity of the desired direction or position. In such an implementation, the position metadata of the large audio object 605 is taken into account in the level adjustment process to preserve information about the perceived direction from which the sound comes. Such an implementation may be appropriate for medium-sized objects, e.g. audio objects that are considered large, but whose size is not large enough to encompass the entire reproduction / playback environment.

[0079] In the implementation shown in FIG. 6C, the audio processing system 600 can generate additional object or bed channels during the decorrelation process. Such functionality can be desirable, for example, if the other audio objects and / or bed 615 are not preferred or optimal. For example, in some implementations, the decorrelated large audio object signal 611 may correspond to a virtual speaker position. If the other audio objects and / or bed 615 do not correspond to a position close enough to a desired virtual speaker position, the decorrelated large audio object signal 611 may correspond to a new virtual speaker position.

[0080] In this example, a large audio object 605 is first processed by a decorrelation system 610. Afterwards, an additional object or bed channel corresponding to a decorrelated audio object signal 611 is provided to an encoder 620. In this example, the decorrelated large audio object signal 611 is subjected to a level adjustment before being sent to the encoder 620. The decorrelated large audio object signal 611 may be a bed channel signal and / or an audio object signal, the latter of which may correspond to a static or moving object.

[0081] In some implementations, the audio signal output to the encoder 620 may include at least a portion of the original large audio object signal. As mentioned above, the audio processing system 600 may be able to retain the audio signal corresponding to the point source contribution of the large audio object 605 after the decorrelation process has been performed. This may be beneficial, for example, because various signals may be correlated to each other to different degrees. It may therefore be beneficial to pass through at least a portion of the original audio signal (e.g., the point source contribution) corresponding to the large audio object 605 intact and render it separately. In such implementations, it may be advantageous to smooth the decorrelated signals and the original signals corresponding to the large audio object 605.

[0082] One such example is shown in Fig. 6D. In this example, at least a part of the original large audio object signal 613 is subjected to a first equalization process by a level adjustment system 612a and the decorrelated large audio object signal 611 is subjected to an equalization process by a level adjustment system 612b. Here, the level adjustment system 612a and the level adjustment system 612b provide output audio signals to the encoder 620. The output of the level adjustment system 612b is in this example also mixed with said other audio objects and / or bed 615.

[0083] In some implementations, the audio processing system 600 may be capable of evaluating the input audio data to determine (or at least estimate) a content type. The decorrelation process may be based at least in part on the content type. In some implementations, the decorrelation process may be selectively performed depending on the content type. For example, the amount of decorrelation to be performed on the input audio data may depend at least in part on the content type. For example, it would generally be desirable to lower the amount of decorrelation for speech.

[0084] One example is shown in Figure 6E. In this example, a media intelligence system 625 can evaluate an audio signal to infer a content type. For example, the media intelligence system 625 may be able to evaluate an audio signal corresponding to a large audio object 605 to infer whether the content type is speech, music, sound effects, etc. In the example shown in Figure 6E, the media intelligence system 625 can send a control signal 627 to control the amount of decorrelation or sizing of the object depending on the content type estimation.

[0085] For example, if the media intelligence system 625 estimates that the audio signals of the large audio object 605 correspond to speech, the media intelligence system 625 may send a control signal 627 indicating that the amount of decorrelation for these signals should be reduced or that these signals should not be decorrelated. Various methods of automatically determining the likelihood that a signal is a speech signal may be used. According to an embodiment, the media intelligence system 625 may include a speech likelihood estimator that can generate a speech likelihood value based at least in part on audio information in the center channel. Some examples are described by [2].

[0086] In some implementations, the control signal 627 may indicate the amount of level adjustment and / or may indicate parameters for mixing the decorrelated large audio object signal 611 with the audio signals for the audio objects and / or bed 615.

[0087] Alternatively or additionally, the amount of decorrelation for large audio objects may be based on "stems," "tags," or other explicit indications of content type. Such explicit indications of content type may, for example, be generated by a content creator (e.g., during a post-production process) and transmitted as metadata together with the corresponding audio signal. In some implementations, such metadata may be human readable. For example, a human readable stem or tag may, in effect, explicitly indicate "this is dialogue," "this is a special effect," "this is music," etc.

[0088] Some implementations may involve a clustering process that combines objects that are similar in some respect, for example with respect to spatial location, spatial size, or content type. Some examples of clustering are described below with reference to Figures 7 and 8. In the example shown in Figure 6F, objects and / or beds 615a are input to a clustering process 630. The output of the clustering process 630 is a smaller number of objects and / or beds 615b. The audio data corresponding to the objects and / or beds 615b are mixed with the smoothed decorrelated large audio object signal 611. In some alternative implementations, the clustering process may be followed by a decorrelation process. One example is described below with reference to Figure 9. Such an implementation may, for example, prevent dialogue from being mixed into clusters with undesirable metadata, such as location not close to the center speaker or large cluster size.

[0089] Scene Simplification through Object Clustering For the purposes of the following description, the terms "clustering" and "grouping" or "combining" are used interchangeably to describe combining objects and / or beds (channels) to reduce the amount of data in a unit of adaptive audio content for transmission and rendering in an adaptive audio playback system; the term "reduction" may be used to refer to the process of performing adaptive audio scene simplification through such clustering of objects and beds. The terms "clustering", "grouping" or "combining" throughout this description are not limited to a strictly unique assignment of object or bed channels to a single cluster, but rather object or bed channels may be distributed across two or more output beds or clusters using weights or gain vectors that determine the relative contribution of the object or bed signals to the output cluster or output bed signal.

[0090] In one embodiment, an adaptive audio system includes at least one component configured to reduce bandwidth of object-based audio content through object clustering and a perceptually transparent simplification of the spatial scene created by the combination of channel beds and objects. The object clustering process performed by the component(s) reduces the complexity of the spatial scene by grouping similar objects into object clusters that replace the original objects using certain information about the objects, which may include spatial location, object content type, temporal attributes, object size, and / or other.

[0091] Additional audio processing for standard audio encoding to deliver and render a convincing user experience based on the original complex bed and audio track is commonly referred to as scene simplification and / or object clustering. The primary goal of this processing is to reduce the spatial scene through clustering or grouping techniques that reduce the number of individual audio elements (beds and objects) delivered to the playback device, but still retain enough spatial information so that the perceived differences between the originally authored content and the rendered output are minimized.

[0092] The scene simplification process can use information about the objects, such as spatial location, temporal attributes, content type, size, and / or other suitable characteristics, to dynamically cluster objects into a reduced number to facilitate rendering of the object+bed content in a reduced bandwidth channel or coding system. The process can reduce the number of objects by performing one or more of the following clustering operations: (1) clustering objects into objects; (2) clustering objects with beds; (3) clustering objects and / or beds into objects. Furthermore, objects can be distributed across two or more clusters. The process may use temporal information about the objects to control the clustering and de-clustering of objects.

[0093] In some implementations, an object cluster replaces the individual waveforms and metadata elements of the constituent objects with a single equivalent set of waveforms and metadata, such that data for N objects is replaced with data for a single object. This essentially compresses the object data from N to 1. Alternatively or additionally, an object or a bed channel may be distributed across two or more clusters (e.g., using an amplitude panning technique). This reduces the object data from N to M, where M < N. The clustering process may use an error metric based on distortion resulting from changes in the position, loudness, or other characteristics of the objects being clustered, to determine a trade-off between compression by clustering and degradation of the sound of the clustered objects. In some embodiments, the clustering process can be performed synchronously. Alternatively or additionally, the clustering process may be event-driven, such as by using auditory scene analysis (ASA) and / or event boundary detection to control object simplification through clustering.

[0094] In some embodiments, the process may utilize knowledge of the endpoint rendering algorithm and / or device to control clustering. In this way, certain characteristics or attributes of the playback device may be used to inform the clustering process. For example, different clustering schemes may be utilized for speakers versus headphones or other audio drivers, or different clustering schemes may be used for reversible versus irreversible encoding, and so on.

[0095] FIG. 7 is a block diagram illustrating an example of a system that may perform the clustering process. As shown in FIG. 7, the system 700 includes an encoder 704 and a decoder 706 stage that process an input audio signal to generate an output audio signal at a reduced bandwidth. In some implementations, the portion 720 and the portion 730 may be in different locations. For example, the portion 720 may correspond to a post-production authoring system, and the portion 730 may correspond to a playback environment such as a home theater system. In the example shown in FIG. 7, a portion 709 of the input signal is processed through known compression techniques to generate a compressed audio bitstream 705. This compressed audio bitstream 705 may be decoded by the decoder stage 706 to generate at least a portion of the output 707. Such known compression techniques may involve analyzing the input audio content 709, quantizing the audio data, and then performing compression techniques such as masking on the audio data itself. Compression techniques may be lossy or lossless and may be implemented in a system that may allow a user to select a compressed bandwidth such as 192 kbps, 256 kbps, 512 kbps, and so on.

[0096] In an adaptive audio system, at least a portion of the input audio includes an input signal 701 that includes audio objects, the audio objects including audio object signals and associated metadata. The metadata defines certain characteristics of the associated audio content, such as object spatial location, object size, content type, loudness, etc. Any practical number of audio objects (e.g., hundreds of objects) may be processed through the system for playback. To facilitate accurate playback of a large number of objects on a wide variety of playback systems and transmission media, the system 700 includes a clustering process or component 702 that reduces the number of objects to a smaller, more manageable number of objects by combining the original objects into a smaller number of object groups.

[0097] In this way, the clustering process builds groups of objects to generate a smaller number of output groups 703 from the original set of individual input objects 701. The clustering process 702 essentially processes the metadata of the objects in addition to the audio data itself to generate a reduced number of object groups. The metadata is analyzed to determine which object at any one time is best combined with other objects, and the corresponding audio waveforms for the objects to be combined may be summed to generate an alternative or combined object. In this example, the combined object groups are then input to an encoder 704, which is configured to generate a bitstream 705 including the audio and metadata for transmission to a decoder 706.

[0098] In general, an adaptive audio system incorporating the object clustering process 702 includes components that generate metadata from an original spatial audio format. The system 700 includes a portion of an audio processing system configured to process one or more bitstreams that include both conventional channel-based audio elements and audio object coding elements. An enhancement layer that includes the audio object coding elements may be added to the channel-based audio codec bitstream or the audio object bitstream. Thus, in this example, the bitstream 705 includes an enhancement layer to be processed by a renderer for use with existing speaker and driver designs or next generation speakers that utilize individually addressable drivers and driver definitions.

[0099] The spatial audio content from the spatial audio processor may include audio object, channel, and position metadata. When an object is rendered, it may be assigned to one or more speakers according to the position metadata and the position of the playback speakers. Additional metadata, such as size metadata, may be associated with the object to change the playback position or otherwise restrict the speakers used for playback. The metadata may be generated at the audio workstation in response to an engineer's mixing input that provides rendering cues that control spatial parameters (e.g., position, size, speed, intensity, timbre, etc.) and specifies which driver(s) or speaker(s) in the listening environment will play each sound during a presentation. The metadata may be associated with each audio data at the workstation for packaging and transport by the spatial audio processor.

[0100] Figure 8 is a block diagram illustrating an example of a system capable of clustering objects and / or beds in an adaptive audio processing system. In the example shown in Figure 8, an object processing component 806 capable of performing a scene simplification task reads any number of input audio files and metadata. The input audio files include an input object 802 and associated object metadata, and may also include a bed 804 and associated bed metadata. Thus, this input file / metadata corresponds to a "bed" or "object" track.

[0101] In this example, the object processing component 806 can combine media intelligence / content classification, spatial distortion analysis and object selection / clustering information to generate fewer output objects and bed tracks. Specifically, objects can be clustered together to generate new equivalent objects or object clusters 808 along with associated object / cluster metadata. These objects can also be selected for downmixing to the bed. This is shown in FIG. 8 as the output of downmixed objects 810 that are input to the renderer 816 for combination 818 with the bed 812 to form an output bed object and associated metadata 820. The output bed configuration 820 (e.g., Dolby 5.1 configuration) does not necessarily have to match the input bed configuration, which can be, for example, 9.1 for an Atmos cinema. In this example, by combining the metadata from the input tracks, new metadata is generated for the output track, and by combining the audio from the input tracks, new audio data is also generated for the output track.

[0102] In this implementation, the object processing component 806 can use certain processing configuration information 822. Such processing configuration information 822 can include the number of output objects, frame size, and certain media intelligence settings. Media intelligence can include parameters or characteristics of objects (or associated with objects), such as content type (i.e., dialogue / music / effects / etc.), regions (segments / classifications), preprocessing results, auditory scene analysis results, and other similar information. For example, the object processing component 806 can determine which audio signals correspond to speech, music, and / or special effects. In this implementation, the object processing component 806 can determine at least some such characteristics by analyzing the audio signals. Alternatively or additionally, the object processing component 806 can determine at least some such characteristics according to associated metadata, such as tags, labels, and the like.

[0103] In an alternative embodiment, audio generation can be deferred by keeping references to all the original tracks in addition to the simplification metadata (e.g., which objects belong to which clusters, which objects are rendered into beds, etc.). Such information can be useful, for example, to distribute the functionality of the scene simplification process between a studio and an encoding house, or in other similar scenarios.

[0104] 9 is a block diagram providing an example of a clustering process followed by a decorrelation process for large objects. The blocks of audio processing system 600 may be implemented via any suitable combination of hardware, firmware, software, etc. stored on a non-transitory medium. For example, the blocks of audio processing system 600 may be implemented via a logic system and / or other elements as described below with reference to FIG.

[0105] In this implementation, the audio processing system 600 includes audio objects O1 through O M 5, where the audio object includes an audio object signal and associated metadata including at least audio object position metadata. In this example, the large object detection module 905 may determine large audio objects 605 having a size greater than a certain threshold size based at least in part on the audio object size metadata. The large audio object detection module 905 may function, for example, as described above with reference to block 510 of FIG. 5.

[0106] In this implementation, the module 910 may perform a decorrelation process on the audio signal of the large audio object 605 to generate a decorrelated large audio object audio signal 611. In this example, the module 910 may also render the audio signal of the large audio object 605 to the virtual speaker positions. Thus, in this example, the decorrelated large audio object audio signal 611 output by the module 910 corresponds to the virtual speaker positions. Some examples of rendering audio object signals to virtual speaker positions will now be described with reference to Figures 10A and 10B.

[0107] 10A shows an example of a virtual source position relative to a playback environment. The playback environment can be a real playback environment or a virtual playback environment. The virtual source position 1005 and the speaker positions 1025 are merely examples. However, in this example, the playback environment is a virtual playback environment, and the speaker positions 1025 correspond to virtual speaker positions.

[0108] In some implementations, the virtual source positions 1005 may be uniformly spaced in all directions. In the example shown in FIG. 10A, the virtual source positions 1005 are uniformly spaced along the x, y, and z axes. The virtual source positions 1005 are N x Multiply N y Multiply N z The number of N positions may form a rectangular grid of virtual source positions 1005. In some implementations, the value of N may range from 5 to 100. The value of N may depend, at least in part, on the number of speakers in (or expected to be in) the playback environment. That is, it may be desirable to include more than one virtual source position 1005 between each speaker position.

[0109] However, in alternative implementations, the virtual source positions 1005 may be spaced differently. For example, in some implementations, the virtual source positions 1005 may have a first uniform spacing along the x and y axes and a second uniform spacing along the z axis. In other implementations, the virtual source positions 1005 may be non-uniformly spaced.

[0110] In this example, the audio object volume 1020a corresponds to the size of the audio object. The audio object 1010 may be rendered according to the virtual source positions 1005 that are enclosed by the audio object volume 1020a. In the example shown in FIG. 10A, the audio object volume 1020a occupies a portion, but not all, of the playback environment 1000a. A larger audio object may occupy a larger portion (or all) of the playback environment 1000a. In some examples, if the audio object 1010 corresponds to a point source, the audio object 1010 may have a size of 0 and the audio object volume 1020a may be set to 0.

[0111] According to some such implementations, the authoring tool may link audio object size with decorrelation by indicating (e.g., via a decorrelation flag included in associated metadata) that decorrelation should be turned on when the audio object size is equal to or greater than a certain size threshold, and that decorrelation should be turned off when the audio object size is below the size threshold. In some implementations, decorrelation may be controlled (e.g., increased, decreased or disabled) according to user input regarding the size threshold and / or other input values.

[0112] In this example, the virtual source positions 1005 are defined within the virtual source volume 1002. In some implementations, the virtual source volume may correspond to a volume within which an audio object can move. In the example shown in FIG. 10A, the playback environment 1000a and the virtual source volume 1002a are coextensive, and thus each of the virtual source positions 1005 corresponds to a position within the playback environment 1000a. However, in alternative implementations, the playback environment 1000a and the virtual source volume 1002 may not be coextensive.

[0113] For example, some of the virtual source positions 10005 may correspond to positions outside the playback environment. Figure 10B shows an alternative example of virtual source positions for a playback environment. In this example, a virtual source volume 1002b extends outside the playback environment 1000b. Some of the virtual source positions 1005 in the audio object volume 1020b are located inside the playback environment 1000b, and other virtual source positions 1005 in the audio object volume 1020b are located outside the playback environment 1000b.

[0114] In other implementations, the virtual source positions 1005 may have a first uniform spacing along the x and y axes and a second uniform spacing along the z axis. x Multiply N y Multiply N zThe N may form a rectangular grid of virtual source positions 1005. For example, in some implementations there may be fewer virtual source positions 1005 along the z-axis than along the x- or y-axis. In some such implementations, the value of N may range from 10 to 100, while the value of M may range from 5 to 10.

[0115] Some implementations involve calculating gain values ​​for each of the virtual source positions 1005 within the audio object volume 1020. In some implementations, a gain value for each of a plurality of output channels of a playback environment (which may be a real playback environment or a virtual playback environment) is calculated for each of the virtual source positions 1005 within the audio object volume 1020. In some implementations, the gain values ​​may be calculated by applying a vector-based amplitude panning (VBAP) algorithm, a pairwise panning algorithm, or similar algorithms to calculate gain values ​​for a point source located at each virtual source position 1005 within the audio object volume 1020. In other implementations, a separable algorithm is used to calculate gain values ​​for a point source located at each virtual source position 1005 within the audio object volume 1020. As used herein, a "separable" algorithm is one in which the gain of a given speaker can be expressed as a product of multiple factors (e.g., three factors), where each factor depends on only one of the coordinates of the virtual source position 1005. Examples include algorithms implemented in various existing mixing console panners, including but not limited to ProTools™ software, and panners implemented in digital film consoles offered by AMS Neve.

[0116] Returning again to FIG. 9, in this example, audio processing system 600 includes bed channels B1 through B NThe encoder 620 also receives a low frequency effects (LFE) channel, as well as audio objects and a bed channel. The audio objects and bed channel are processed according to a scene simplification or "clustering" process, e.g., as described above with reference to Figures 7 and 8. However, in this example, the LFE channel is not input to the clustering process, but instead is passed straight through to the encoder 620.

[0117] In this implementation, bed channels B1 through B N are converted by module 915 into static audio objects 917. Module 920 receives the static audio objects 917 in addition to the audio objects that the large object detection module 905 has determined not to be large audio objects, where module 920 also receives the decorrelated large audio object signals 611, which in this example correspond to the virtual speaker positions.

[0118] In this implementation, the module 920 divides the static objects 917, the received audio objects and the decorrelated large audio object signals 611 into clusters C1 to C2. P In general, the module 920 outputs a smaller number of clusters than the number of received audio objects. In this implementation, the module 920 can associate the decorrelated large audio object signal 611 with the appropriate cluster location, e.g., as described above with reference to block 520 of FIG. 5 .

[0119] In this example, clusters C1 to C PThe audio data for the LFE channel is encoded by the encoder 620 and transmitted to the playback environment 925. In some implementations, the playback environment 925 may include a home theater system. The audio processing system 930 can receive and decode the encoded audio data, and render the decoded audio data according to an actual playback speaker configuration of the playback environment 925, such as speaker positions, speaker capabilities (e.g., bass playback capabilities) of the actual playback speakers of the playback environment 925, etc.

[0120] 11 is a block diagram providing an example of components of an audio processing system. In this example, the audio processing system 1100 includes an interface system 1105. The interface system 1105 may include a network interface, such as a wireless network interface. Alternatively or additionally, the interface system 1105 may include a Universal Serial Bus (USB) interface or other such interface.

[0121] The audio processing system 1100 includes a logic system 1110. The logic system 1110 may include a processor, such as a general purpose single-chip or multi-chip processor. The logic system 1110 may include a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic device, discrete gate or transistor logic, or discrete hardware components, or a combination thereof. The logic system 1110 may be configured to control other components of the audio processing system 1100. Although interfaces between components of the audio processing system 1100 are not shown in FIG. 11, the logic system 1110 may be configured with interfaces for communication with other components, which may or may not be configured for communication with each other, as appropriate.

[0122] Logic system 1110 may be configured to perform audio processing functions, including but not limited to functions of the type described herein. In some such implementations, logic system 1110 may be configured to operate (at least in part) according to software stored on one or more non-transitory media. The non-transitory media may include memory associated with logic system 1110, such as random access memory (RAM) and / or read-only memory (ROM). The non-transitory media may include memory of memory system 1115. Memory system 1115 may include one or more suitable types of non-transitory storage media, such as flash memory, hard drives, etc.

[0123] Display system 1130 may include one or more suitable types of displays, depending on the implementation of audio processing system 1100. For example, display system 1130 may include a liquid crystal display, a plasma display, a bi-stable display, etc.

[0124] The user input system 1135 may include one or more devices configured to accept input from a user. In some implementations, the user input system 1135 may include a touch screen overlaying a display of the display system 1130. The user input system 1135 may include a mouse, a trackball, a gesture detection system, a joystick, one or more GUIs and / or menus presented on the display system 1130, buttons, keyboards, switches, etc. In some implementations, the user input system 1135 may include a microphone 1125: a user may provide voice commands for the audio processing system 1100 via the microphone 1125. A logic system may be configured for voice recognition and for controlling at least some operations of the audio processing system 1100 according to such voice commands. In some implementations, the user input system 1135 may be considered a user interface and thus part of the interface system 1105.

[0125] Power system 1140 may include one or more suitable energy storage devices, such as nickel-cadmium batteries or lithium-ion batteries. Power system 1140 may be configured to receive power from an electrical outlet.

[0126] Various modifications to the implementations described in this disclosure may be readily apparent to those skilled in the art. The general principles defined herein may be applied to other implementations without departing from the spirit or scope of the disclosure. Thus, the claims are not intended to be limited to the implementations shown herein, but are to be accorded the widest scope consistent with the disclosure, principles and novel features disclosed herein.

[0127] Several aspects will be described. [Aspect 1] receiving audio data including audio objects and including one or more audio bed signals corresponding to speaker positions, the audio objects including audio object signals and associated metadata, the metadata including at least audio object size data; determining a large audio object having an audio object size greater than a threshold size based on the audio object size data; performing a decorrelation process on the audio signal of the large audio object to generate a decorrelated audio signal of the large audio object; correlating the decorrelated audio signals of the large audio object with an object position, the correlating process being independent of the actual playback speaker configuration, and comprising mixing the decorrelated audio signals of the large audio object with the audio bed signal or at least a portion of the audio object signal; encoding audio data output from the correlating process, the encoding process including a data compression process and not including encoding decorrelated metadata for the large audio object; method. [Aspect 2] 2. The method of claim 1, further comprising receiving decorrelation metadata for the large audio object, wherein the decorrelation process is performed at least in part in accordance with the decorrelation metadata. [Aspect 3] 3. The method of embodiment 1 or 2, wherein at least some of the object positions are static. Aspect 4 4. The method of any one of aspects 1 to 3, wherein at least some of the object positions vary over time. Aspect 5 5. The method of any one of aspects 1 to 4, wherein the correlating process includes rendering the decorrelated audio signals of the large audio object according to virtual speaker positions. Aspect 6 6. The method of any one of aspects 1 to 5, wherein the actual playback speaker configuration is used to render the decorrelated audio signals of the large audio object to speakers of a playback environment. Aspect 7 7. The method of any one of aspects 1 to 6, further comprising the step of outputting the decorrelated audio signal of the large audio object as an additional audio bed signal or an audio object signal. Aspect 8 8. The method of any one of aspects 1 to 7, further comprising applying a level adjustment process to the decorrelated audio signals of the large audio objects. Aspect 9 The method of claim 8, wherein the metadata of the large audio object includes audio object position metadata, and the level adjustment process depends at least in part on the audio object size metadata and the audio object position metadata of the large audio object. Aspect 10 10. The method of any one of aspects 1 to 9, further comprising attenuating or removing the audio signal of the large audio object after the decorrelation process has been performed. Aspect 11 11. The method of any one of aspects 1 to 10, further comprising a step of retaining an audio signal corresponding to a point source contribution of the large audio object after the decorrelation process is performed. Aspect 12 The large audio object metadata includes audio object position metadata, the method further comprising: calculating contributions from virtual sources within an audio object area or volume defined by said large audio object position data and said large audio object size data; determining a set of audio object gain values ​​for each of the plurality of output channels based at least in part on the calculated contributions. 12. The method of any one of embodiments 1 to 11. Aspect 13 13. The method of any one of aspects 1 to 12, further comprising the step of performing an audio object clustering process after the decorrelation process. Aspect 14 14. The method of claim 13, wherein the audio object clustering process is performed after the associating process. Aspect 15 15. The method of any one of aspects 1 to 14, further comprising evaluating the audio data to determine a content type, and wherein the decorrelation process is selectively performed depending on the content type. Aspect 16 20. The method of embodiment 15, wherein the amount of decorrelation performed depends on the content type. Aspect 17 17. The method of any one of aspects 1 to 16, wherein the decorrelation process involves one or more of a delay, an all-pass filter, a pseudo-random filter, or a reverberation algorithm. Aspect 18 18. The method of any one of aspects 1 to 17, wherein the metadata of the large audio object includes audio object position metadata, and the method further includes mixing the decorrelated audio signal of the large audio object with an audio signal for an audio object that is spatially separated from the large audio object by a threshold amount of distance. Aspect 19 With the interface system; and a logic system, the logic system comprising: receiving, via the interface system, audio data including audio objects and including one or more audio bed signals corresponding to speaker positions, the audio objects including audio object signals and associated metadata, the metadata including at least audio object size data; determining a large audio object based on the audio object size data, the large audio object having an audio object size greater than a threshold size; performing a decorrelation process on the audio signal of the large audio object to generate a decorrelated audio signal of the large audio object; correlating the decorrelated audio signals of the large audio object with an object position, the correlating process being independent of the actual playback speaker configuration, and comprising mixing the decorrelated audio signals of the large audio object with the audio bed signal or at least a portion of the audio object signal; encoding audio data output from the correlating process, the encoding process including a data compression process and not including encoding decorrelated metadata for the large audio object. Device. Aspect 20 A non-transitory medium having software stored thereon, the software controlling at least one of: receiving audio data including audio objects and including one or more audio bed signals corresponding to speaker positions, the audio objects including audio object signals and associated metadata, the metadata including at least audio object size data; determining a large audio object having an audio object size greater than a threshold size based on the audio object size data; performing a decorrelation process on the audio signal of the large audio object to generate a decorrelated audio signal of the large audio object; correlating the decorrelated audio signals of the large audio object with an object position, the correlating process being independent of the actual playback speaker configuration, and comprising mixing the decorrelated audio signals of the large audio object with the audio bed signal or at least a portion of the audio object signal; encoding audio data output from the correlating process, the encoding process including a data compression process and not including encoding decorrelated metadata for the large audio object. Non-temporary medium.

Claims

1. receiving audio data including at least one audio object, the audio data including at least one audio signal and audio object metadata, the at least one audio signal associated with the at least one audio object, the audio object metadata associated with the at least one audio object, the audio object metadata including a size of the at least one audio object and a flag indicating whether the at least one audio object is spatially diffuse, the at least one audio object being spatially diffuse indicating that the at least one audio object has a perceived size greater than a threshold in a playback environment; based on determining that the at least one audio object is spatially diffuse, performing decorrelation filtering on the at least one audio object to determine decorrelated audio object audio signals, each of the decorrelated audio object audio signals corresponding to at least one playback speaker of a plurality of playback speakers; and outputting the decorrelated audio object audio signal. method.

2. The method of claim 1 , further comprising: rendering the decorrelated audio object audio signals to the plurality of playback speakers based on speaker zone constraints.

3. 2. The method of claim 1, wherein the at least one audio object is associated with at least one object position, at least one of the at least one object positions being static.

4. 2. The method of claim 1, wherein the at least one audio object is associated with at least one object position, at least one of the at least one object positions varying over time.

5. The method of claim 1 , further comprising: rendering the decorrelated audio object audio signals based on an actual playback speaker configuration of the playback environment.

6. The method of claim 1 , further comprising the step of applying a level adjustment process to the decorrelated audio object audio signals.

7. The method of claim 1 , wherein performing decorrelation includes at least one of a delay and a filter.

8. The method of claim 1 , wherein performing decorrelation includes at least one of an all-pass filter and a pseudo-random filter.

9. The method of claim 1 , wherein performing decorrelation comprises a reverberation process.

10. The method of claim 1 , further comprising: rendering the decorrelated audio object audio signals according to virtual speaker positions.

11. 2. The method of claim 1, further comprising: clustering the decorrelated audio object audio signals to generate one or more groups of the decorrelated audio object audio signals, the number of groups being less than the number of the decorrelated audio object audio signals.

12. A computer program product causing a processor to carry out the method of claim 1.

13. a receiver configured to receive audio data including at least one audio object, the audio data including at least one audio signal and audio object metadata, the at least one audio signal associated with the at least one audio object, the audio object metadata associated with the at least one audio object, the audio object metadata including a size of the at least one audio object and a flag indicating whether the at least one audio object is spatially diffuse, the at least one audio object being spatially diffuse indicating that the at least one audio object has a perceived size greater than a threshold in a playback environment; a decorrelator configured to perform, based on a determination that the at least one audio object is spatially diffuse, a decorrelation filtering on the at least one audio object to determine decorrelated audio object audio signals, each of the decorrelated audio object audio signals corresponding to at least one playback speaker of a plurality of playback speakers, and outputting the decorrelated audio object audio signals. Device.

14. The apparatus of claim 13 , further comprising a renderer for rendering the decorrelated audio object audio signals to the multiple playback speakers based on speaker zone constraints.

15. 14. The apparatus of claim 13, wherein the at least one audio object is associated with at least one object position, at least one of the at least one object positions being static.

16. 14. The apparatus of claim 13, wherein the at least one audio object is associated with at least one object position, at least one of the at least one object positions varying over time.

17. The apparatus of claim 13 , further comprising a renderer for rendering the decorrelated audio object audio signals based on an actual playback speaker configuration of the playback environment.

18. The apparatus of claim 13 , further comprising a level adjuster for applying a level adjustment process to the decorrelated audio object audio signals.

19. The apparatus of claim 13 , wherein the decorrelator comprises at least one of a delay and a filter.

20. The apparatus of claim 13 , wherein the decorrelator comprises at least one of an all-pass filter and a pseudo-random filter.

21. The apparatus of claim 13 , wherein the decorrelator comprises a reverberation process.

22. The apparatus of claim 13 , further comprising a renderer for rendering the decorrelated audio object audio signals according to virtual speaker positions.

Images