Patent Grant
9826328
One or more implementations relate generally to audio signal processing, and more specifically, to a system for rendering adaptive audio content through individually addressable drivers.
The subject matter discussed in the background section should not be assumed to be prior art merely as a result of its mention in the background section. Similarly, a problem mentioned in the background section or associated with the subject matter of the background section should not be assumed to have been previously recognized in the prior art. The subject matter in the background section merely represents different approaches, which in and of themselves may also be inventions.
Cinema sound tracks usually comprise many different sound elements corresponding to images on the screen, dialog, noises, and sound effects that emanate from different places on the screen and combine with background music and ambient effects to create the overall audience experience. Accurate playback requires that sounds be reproduced in a way that corresponds as closely as possible to what is shown on screen with respect to sound source position, intensity, movement, and depth. Traditional channel-based audio systems send audio content in the form of speaker feeds to individual speakers in a playback environment.
The introduction of digital cinema has created new standards for cinema sound, such as the incorporation of multiple channels of audio to allow for greater creativity for content creators, and a more enveloping and realistic auditory experience for audiences. Expanding beyond traditional speaker feeds and channel-based audio as a means for distributing spatial audio is critical, and there has been considerable interest in a model-based audio description that allows the listener to select a desired playback configuration with the audio rendered specifically for their chosen configuration. To further improve the listener experience, playback of sound in true three-dimensional (“3D”) or virtual 3D environments has become an area of increased research and development. The spatial presentation of sound utilizes audio objects, which are audio signals with associated parametric source descriptions of apparent source position (e.g., 3D coordinates), apparent source width, and other parameters. Object-based audio may be used for many multimedia applications, such as digital movies, video games, simulators, and is of particular importance in a home environment where the number of speakers and their placement is generally limited or constrained by the confines of a relatively small listening environment.
Various technologies have been developed to improve sound systems in cinema environments and to more accurately capture and reproduce the creator's artistic intent for a motion picture sound track. For example, a next generation spatial audio (also referred to as “adaptive audio”) format has been developed that comprises a mix of audio objects and traditional channel-based speaker feeds along with positional metadata for the audio objects. In a spatial audio decoder, the channels are sent directly to their associated speakers (if the appropriate speakers exist) or down-mixed to an existing speaker set, and audio objects are rendered by the decoder in a flexible manner. The parametric source description associated with each object, such as a positional trajectory in 3D space, is taken as an input along with the number and position of speakers connected to the decoder. The renderer then utilizes certain algorithms, such as a panning law, to distribute the audio associated with each object across the attached set of speakers. This way, the authored spatial intent of each object is optimally presented over the specific speaker configuration that is present in the listening room.
Current spatial audio systems have generally been developed for cinema use, and thus involve deployment in large rooms and the use of relatively expensive equipment, including arrays of multiple speakers distributed around the room. An increasing amount of cinema content that is presently being produced is being made available for playback in the home environment through streaming technology and advanced media technology, such as blu-ray, and so on. In addition, emerging technologies such as 3D television and advanced computer games and simulators are encouraging the use of relatively sophisticated equipment, such as large screen monitors, surround-sound receivers, and speaker arrays in home and other consumer (noncinema/theater) environments. However, equipment cost, installation complexity, and room size are realistic constraints that prevent the full exploitation of spatial audio in most home environments. For example, advanced object-based audio systems typically employ overhead or height speakers to play back sound that is intended to originate above a listener's head. In many cases, and especially in the home environment, such height speakers may not be available. In this case, the height information is lost if such sound objects are played only through floor or wall-mounted speakers.
What is needed therefore is a system that allows full spatial information of an adaptive audio system to be reproduced in various different listening environments, such as collocated speaker systems, headphones, and other listening environments that may include only a portion of the full speaker array intended for playback, such as limited or no overhead speakers.
Systems and methods are described for a spatial audio format and system that includes updated content creation tools, distribution methods and an enhanced user experience based on an adaptive audio system that includes new speaker and channel configurations, as well as a new spatial description format made possible by a suite of advanced content creation tools created for cinema sound mixers. Embodiments include a system that expands the cinema-based adaptive audio concept to other audio playback ecosystems including home theater (e.g., A/V receiver, soundbar, and blu-ray player), E-media (e.g., PC, tablet, mobile device, and headphone playback), broadcast (e.g., TV and set-top box), music, gaming, live sound, user generated content (“UGC”), and so on. The home environment system includes components that provide compatibility with the theatrical content, and features metadata definitions that include content creation information to convey creative intent, media intelligence information regarding audio objects, speaker feeds, spatial rendering information and content dependent metadata that indicate content type such as dialog, music, ambience, and so on. The adaptive audio definitions may include standard speaker feeds via audio channels plus audio objects with associated spatial rendering information (such as size, velocity and location in three-dimensional space). A novel speaker layout (or channel configuration) and an accompanying new spatial description format that will support multiple rendering technologies are also described. Audio streams (generally including channels and objects) are transmitted along with metadata that describes the content creator's or sound mixer's intent, including desired position of the audio stream. The position can be expressed as a named channel (from within the predefined channel configuration) or as 3D spatial position information. This channels plus objects format provides the best of both channel-based and model-based audio scene description methods.
Embodiments are specifically directed to a system for rendering adaptive audio content that includes overhead sounds that are meant to be played through overhead or ceiling mounted speakers. In a home or other small-scale listening environment that does not have overhead speakers available; the overhead sounds are reproduced by speaker drivers that are configured to reflect sound off of the ceiling or one or more other surfaces of the listening environment.
Each publication, patent, and/or patent application mentioned in this specification is herein incorporated by reference in its entirety to the same extent as if each individual publication and/or patent application was specifically and individually indicated to be incorporated by reference.
In the following drawings like reference numbers are used to refer to like elements. Although the following figures depict various examples, the one or more implementations are not limited to the examples depicted in the figures.
Systems and methods are described for an adaptive audio system that renders reflected sound for adaptive audio systems that lack overhead speakers. Aspects of the one or more embodiments described herein may be implemented in an audio or audio-visual system that processes source audio information in a mixing, rendering and playback system that includes one or more computers or processing devices executing software instructions. Any of the described embodiments may be used alone or together with one another in any combination. Although various embodiments may have been motivated by various deficiencies with the prior art, which may be discussed or alluded to in one or more places in the specification, the embodiments do not necessarily address any of these deficiencies. In other words, different embodiments may address different deficiencies that may be discussed in the specification. Some embodiments may only partially address some deficiencies or just one deficiency that may be discussed in the specification, and some embodiments may not address any of these deficiencies.
For purposes of the present description, the following terms have the associated meanings: the term “channel” means an audio signal plus metadata in which the position is coded as a channel identifier, e.g., left-front or right-top surround; “channel-based audio” is audio formatted for playback through a pre-defined set of speaker zones with associated nominal locations, e.g., 5.1, 7.1, and so on; the term “object” or “object-based audio” means one or more audio channels with a parametric source description, such as apparent source position (e.g., 3D coordinates), apparent source width, etc.; and “adaptive audio” means channel-based and/or object-based audio signals plus metadata that renders the audio signals based on the playback environment using an audio stream plus metadata in which the position is coded as a 3D position in space; and “listening environment” means any open, partially enclosed, or fully enclosed area, such as a room that can be used for playback of audio content alone or with video or other content, and can be embodied in a home, cinema, theater, auditorium, studio, game console, and the like. Such an area may have one or more surfaces disposed therein, such as walls or baffles that can directly or diffusely reflect sound waves.
Adaptive Audio Format and System
Embodiments are directed to a reflected sound rendering system that is configured to work with a sound format and processing system that may be referred to as a “spatial audio system” or “adaptive audio system” that is based on an audio format and rendering technology to allow enhanced audience immersion, greater artistic control, and system flexibility and scalability. An overall adaptive audio system generally comprises an audio encoding, distribution, and decoding system configured to generate one or more bitstreams containing both conventional channel-based audio elements and audio object coding elements. Such a combined approach provides greater coding efficiency and rendering flexibility compared to either channel-based or object-based approaches taken separately. An example of an adaptive audio system that may be used in conjunction with present embodiments is described in pending International Publication No. WO2013/006338 published on 10 Jan. 2013, which is hereby incorporated by reference.
An example implementation of an adaptive audio system and associated audio format is the Dolby® Atmos™ platform. Such a system incorporates a height (up/down) dimension that may be implemented as a 9.1 surround system, or similar surround sound configuration.
Audio objects can be considered groups of sound elements that may be perceived to emanate from a particular physical location or locations in the listening environment. Such objects can be static (that is, stationary) or dynamic (that is, moving). Audio objects are controlled by metadata that defines the position of the sound at a given point in time, along with other functions. When objects are played back, they are rendered according to the positional metadata using the speakers that are present, rather than necessarily being output to a predefined physical channel. A track in a session can be an audio object, and standard panning data is analogous to positional metadata. In this way, content placed on the screen might pan in effectively the same way as with channel-based content, but content placed in the surrounds can be rendered to an individual speaker if desired. While the use of audio objects provides the desired control for discrete effects, other aspects of a soundtrack may work effectively in a channel-based environment. For example, many ambient effects or reverberation actually benefit from being fed to arrays of speakers. Although these could be treated as objects with sufficient width to fill an array, it is beneficial to retain some channel-based functionality.
The adaptive audio system is configured to support “beds” in addition to audio objects, where beds are effectively channel-based sub-mixes or stems. These can be delivered for final playback (rendering) either individually, or combined into a single bed, depending on the intent of the content creator. These beds can be created in different channel-based configurations such as 5.1, 7.1, and 9.1, and arrays that include overhead speakers, such as shown in
An adaptive audio system effectively moves beyond simple “speaker feeds” as a means for distributing spatial audio, and advanced model-based audio descriptions have been developed that allow the listener the freedom to select a playback configuration that suits their individual needs or budget and have the audio rendered specifically for their individually chosen configuration. At a high level, there are four main spatial audio description formats: (1) speaker feed, where the audio is described as signals intended for loudspeakers located at nominal speaker positions; (2) microphone feed, where the audio is described as signals captured by 9 actual or virtual microphones in a predefined configuration (the number of microphones and their relative position); (3) model-based description, where the audio is described in terms of a sequence of audio events at described times and positions; and (4) binaural, where the audio is described by the signals that arrive at the two ears of a listener.
The four description formats are often associated with the following common rendering technologies, where the term “rendering” means conversion to electrical signals used as speaker feeds: (1) panning, where the audio stream is converted to speaker feeds using a set of panning laws and known or assumed speaker positions (typically rendered prior to distribution); (2) Ambisonics, where the microphone signals are converted to feeds for a scalable array of loudspeakers (typically rendered after distribution); (3) Wave Field Synthesis (WFS), where sound events are converted to the appropriate speaker signals to synthesize a sound field (typically rendered after distribution); and (4) binaural, where the L/R binaural signals are delivered to the LIR ear, typically through headphones, but also through speakers in conjunction with crosstalk cancellation.
In general, any format can be converted to another format (though this may require blind source separation or similar technology) and rendered using any of the aforementioned technologies; however, not all transformations yield good results in practice. The speaker-feed format is the most common because it is simple and effective. The best sonic results (that is, the most accurate and reliable) are achieved by mixing/monitoring in and then distributing the speaker feeds directly because there is no processing required between the content creator and listener. If the playback system is known in advance, a speaker feed description provides the highest fidelity; however, the playback system and its configuration are often not known beforehand. In contrast, the model-based description is the most adaptable because it makes no assumptions about the playback system and is therefore most easily applied to multiple rendering technologies. The model-based description can efficiently capture spatial information, but becomes very inefficient as the number of audio sources increases.
The adaptive audio system combines the benefits of both channel and model-based systems, with specific benefits including high timbre quality, optimal reproduction of artistic intent when mixing and rendering using the same channel configuration, single inventory with downward adaption to the rendering configuration, relatively low impact on system pipeline, and increased immersion via finer horizontal speaker spatial resolution and new height channels. The adaptive audio system provides several new features including: a single inventory with downward and upward adaption to a specific cinema rendering configuration, i.e., delay rendering and optimal use of available speakers in a playback environment; increased envelopment, including optimized downmixing to avoid inter-channel correlation (ICC) artifacts; increased spatial resolution via steer-thru arrays (e.g., allowing an audio object to be dynamically assigned to one or more loudspeakers within a surround array); and increased front channel resolution via high resolution center or similar speaker configuration.
The spatial effects of audio signals are critical in providing an immersive experience for the listener. Sounds that are meant to emanate from a specific region of a viewing screen or room should be played through speaker(s) located at that same relative location. Thus, the primary audio metadatum of a sound event in a model-based description is position, though other parameters such as size, orientation, velocity and acoustic dispersion can also be described. To convey position, a model-based, 3D audio spatial description requires a 3D coordinate system. The coordinate system used for transmission (e.g., Euclidean, spherical, cylindrical) is generally chosen for convenience or compactness; however, other coordinate systems may be used for the rendering processing. In addition to a coordinate system, a frame of reference is required for representing the locations of objects in space. For systems to accurately reproduce position-based sound in a variety of different environments, selecting the proper frame of reference can be critical. With an allocentric reference frame, an audio source position is defined relative to features within the rendering environment such as room walls and corners, standard speaker locations, and screen location. In an egocentric reference frame, locations are represented with respect to the perspective of the listener, such as “in front of me,” “slightly to the left,” and so on. Scientific studies of spatial perception (audio and otherwise) have shown that the egocentric perspective is used almost universally. For cinema, however, the allocentric frame of reference is generally more appropriate. For example, the precise location of an audio object is most important when there is an associated object on screen. When using an allocentric reference, for every listening position and for any screen size, the sound will localize at the same relative position on the screen, for example, “one-third left of the middle of the screen.” Another reason is that mixers tend to think and mix in allocentric terms, and panning tools are laid out with an allocentric frame (that is, the room walls), and mixers expect them to be rendered that way, for example, “this sound should be on screen,” “this sound should be off screen,” or “from the left wall,” and so on.
Despite the use of the allocentric frame of reference in the cinema environment, there are some cases where an egocentric frame of reference may be useful and more appropriate. These include non-diegetic sounds, i.e., those that are not present in the “story space,” e.g., mood music, for which an egocentrically uniform presentation may be desirable. Another case is near-field effects (e.g., a buzzing mosquito in the listener's left ear) that require an egocentric representation. In addition, infinitely far sound sources (and the resulting plane waves) may appear to come from a constant egocentric position (e.g., 30 degrees to the left), and such sounds are easier to describe in egocentric terms than in allocentric terms. In the some cases, it is possible to use an allocentric frame of reference as long as a nominal listening position is defined, while some examples require an egocentric representation that is not yet possible to render. Although an allocentric reference may be more useful and appropriate, the audio representation should be extensible, since many new features, including egocentric representation may be more desirable in certain applications and listening environments.
Embodiments of the adaptive audio system include a hybrid spatial description approach that includes a recommended channel configuration for optimal fidelity and for rendering of diffuse or complex, multi-point sources (e.g., stadium crowd, ambiance) using an egocentric reference, plus an allocentric, model-based sound description to efficiently enable increased spatial resolution and scalability.
The playback system 300 is configured to render and playback audio content that is generated through one or more capture, pre-processing, authoring and coding components. An adaptive audio pre-processor may include source separation and content type detection functionality that automatically generates appropriate metadata through analysis of input audio. For example, positional metadata may be derived from a multi-channel recording through an analysis of the relative levels of correlated input between channel pairs. Detection of content type, such as speech or music, may be achieved, for example, by feature extraction and classification. Certain authoring tools allow the authoring of audio programs by optimizing the input and codification of the sound engineer's creative intent allowing him to create the final audio mix once that is optimized for playback in practically any playback environment. This can be accomplished through the use of audio objects and positional data that is associated and encoded with the original audio content. In order to accurately place sounds around an auditorium, the sound engineer needs control over how the sound will ultimately be rendered based on the actual constraints and features of the playback environment. The adaptive audio system provides this control by allowing the sound engineer to change how the audio content is designed and mixed through the use of audio objects and positional data. Once the adaptive audio content has been authored and coded in the appropriate codec devices, it is decoded and rendered in the various components of playback system 300.
As shown in
Playback Application
As mentioned above, an initial implementation of the adaptive audio format and system is in the digital cinema (D-cinema) context that includes content capture (objects and channels) that are authored using novel authoring tools, packaged using an adaptive audio cinema encoder, and distributed using PCM or a proprietary lossless codec using the existing Digital Cinema Initiative (DCI) distribution mechanism. In this case, the audio content is intended to be decoded and rendered in a digital cinema to create an immersive spatial audio cinema experience. However, as with previous cinema improvements, such as analog surround sound, digital multi-channel audio, etc., there is an imperative to deliver the enhanced user experience provided by the adaptive audio format directly to listeners in their homes. This requires that certain characteristics of the format and system be adapted for use in more limited listening environments. For example, homes, rooms, small auditorium or similar places may have reduced space, acoustic properties, and equipment capabilities as compared to a cinema or theater environment. For purposes of description, the term “consumer-based environment” is intended to include any non-cinema environment that comprises a listening environment for use by regular consumers or professionals, such as a house, studio, room, console area, auditorium, and the like. The audio content may be sourced and rendered alone or it may be associated with graphics content, e.g., still pictures, light displays, video, and so on.
As shown in the example of system 420, the cinema-to-consumer translator 430 feeds sound for picture (e.g., broadcast, disc, OTT, etc.) and game audio bitstream creation modules 428. These two modules, which are appropriate for delivering cinema content, can be fed into multiple distribution pipelines 432, all of which may deliver to the consumer end points. For example, adaptive audio cinema content may be encoded using a codec suitable for broadcast purposes such as Dolby Digital Plus, which may be modified to convey channels, objects and associated metadata, and is transmitted through the broadcast chain via cable or satellite and then decoded and rendered in the home for home theater or television playback. Similarly, the same content could be encoded using a codec suitable for online distribution where bandwidth is limited, where it is then transmitted through a 3G or 4G mobile network and then decoded and rendered for playback via a mobile device using headphones. Other content sources such as TV, live broadcast, games and music may also use the adaptive audio format to create and provide content for a next generation spatial audio format.
The system of
The adaptive audio ecosystem is configured to be a fully comprehensive, end-to-end, next generation audio system using the adaptive audio format that includes content creation, packaging, distribution and playback/rendering across a wide number of end-point devices and use cases. As shown in
Current authoring and distribution systems for non-cinema audio create and deliver audio that is intended for reproduction to pre-defined and fixed speaker locations with limited knowledge of the type of content conveyed in the audio essence (i.e., the actual audio that is played back by the reproduction system). The adaptive audio system, however, provides a new hybrid approach to audio creation that includes the option for both fixed speaker location specific audio (left channel, right channel, etc.) and object-based audio elements that have generalized 3D spatial information including position, size and velocity. This hybrid approach provides a balanced approach for fidelity (provided by fixed speaker locations) and flexibility in rendering (generalized audio objects). This system also provides additional useful information about the audio content via new metadata that is paired with the audio essence by the content creator at the time of content creation/authoring. This information provides detailed information about the attributes of the audio that can be used during rendering. Such attributes may include content type (e.g., dialog, music, effect, Foley, background/ambience, etc.) as well as audio object information such as spatial attributes (e.g., 3D position, object size, velocity, etc.) and useful rendering information (e.g., snap to speaker location, channel weights, gain, bass management information, etc.). The audio content and reproduction intent metadata can either be manually created by the content creator or created through the use of automatic, media intelligence algorithms that can be run in the background during the authoring process and be reviewed by the content creator during a final quality control phase if desired.
Distributed/Centralized Rendering
In an embodiment the renderer 454 comprises a functional process embodied in a central processor associated with the network. Alternatively, the renderer may comprise a functional process executed at least in part by circuitry within or coupled to each driver of the array of individually addressable audio drivers. In the case of a centralized process, the rendering data is sent to the individual drivers in the form of audio signal sent over individual audio channels. In the distributed processing embodiment, the central processor may perform no rendering, or at least some partial rendering of the audio data with the final rendering performed in the drivers. In this case, powered speakers/drivers are required to enable the on-board processing functions. One example implementation is the use of speakers with integrated microphones, where the rendering is adapted based on the microphone data and the adjustments are done in the speakers themselves. This eliminates the need to transmit the microphone signals back to the central renderer for calibration and/or configuration purposes.
Listening Environments
Implementations of the adaptive audio system are intended to be deployed in a variety of different listening environments. These include three primary areas of consumer applications: home theater systems, televisions and soundbars, and headphones, but can also include cinema, theater, studios, and other large-scale or professional environments.
System 500 also includes a near field effect (NFE) speaker 512 that may be located right in front, or close in front of the listener, such as on table in front of a seating location. With adaptive audio it is possible to bring audio objects into the room and not have them simply be locked to the perimeter of the room. Therefore, having objects traverse through the three-dimensional space is an option. An example is where an object may originate in the L speaker, travel through the room through the NFE speaker, and terminate in the RS speaker. Various different speakers may be suitable for use as an NFE speaker, such as a wireless, battery powered speaker.
The adaptive audio renderer understands the spatial relationship between the mix and the playback system. In some instances of a playback environment, discrete speakers may be available in all relevant areas of the room, including overhead positions, as shown in
In many cases however, and especially in a home environment, certain speakers, such as ceiling mounted overhead speakers are not available. In this case, certain virtualization techniques are implemented by the renderer to reproduce overhead audio content through existing floor or wall mounted speakers. In an embodiment, the adaptive audio system includes a modification to the standard configuration through the inclusion of both a front-firing capability and a top (or “upward”) firing capability for each speaker. In traditional home applications, speaker manufacturers have attempted to introduce new driver configurations other than front-firing transducers and have been confronted with the problem of trying to identify which of the original audio signals (or modifications to them) should be sent to these new drivers. With the adaptive audio system there is very specific information regarding which audio objects should be rendered above the standard horizontal plane. In an embodiment, height information present in the adaptive audio system is rendered using the upward-firing drivers. Likewise, side-firing speakers can be used to render certain other content, such as ambience effects.
One advantage of the upward-firing drivers is that they can be used to reflect sound off of a hard ceiling surface to simulate the presence of overhead/height speakers positioned in the ceiling. A compelling attribute of the adaptive audio content is that the spatially diverse audio is reproduced using an array of overhead speakers. As stated above, however, in many cases, installing overhead speakers is too expensive or impractical in a home environment. By simulating height speakers using normally positioned speakers in the horizontal plane, a compelling 3D experience can be created with easy to position speakers. In this case, the adaptive audio system is using the upward-firing/height simulating drivers in a new way in that audio objects and their spatial reproduction information are being used to create the audio being reproduced by the upward-firing drivers.
In an embodiment, the adaptive audio system utilizes upward-firing drivers to provide the height element. In general, it has been shown that incorporating signal processing to introduce perceptual height cues into the audio signal being fed to the upward-firing drivers improves the positioning and perceived quality of the virtual height signal. For example, a parametric perceptual binaural hearing model has been developed to create a height cue filter, which when used to process audio being reproduced by an upward-firing driver, improves that perceived quality of the reproduction. In an embodiment, the height cue filter is derived from the both the physical speaker location (approximately level with the listener) and the reflected speaker location (above the listener). For the physical speaker location, a directional filter is determined based on a model of the outer ear (or pinna). An inverse of this filter is next determined and used to remove the height cues from the physical speaker. Next, for the reflected speaker location, a second directional filter is determined, using the same model of the outer ear. This filter is applied directly, essentially reproducing the cues the ear would receive if the sound were above the listener. In practice, these filters may be combined in a way that allows for a single filter that both (1) removes the height cue from the physical speaker location, and (2) inserts the height cue from the reflected speaker location.
Speaker Configuration
A main consideration of the adaptive audio system for home use and similar applications is speaker configuration. In an embodiment, the system utilizes individually addressable drivers, and an array of such drivers is configured to provide a combination of both direct and reflected sound sources. A bi-directional link to the system controller (e.g., A/V receiver, set-top box) allows audio and configuration data to be sent to the speaker, and speaker and sensor information to be sent back to the controller, creating an active, closed-loop system.
For purposes of description, the term “driver” means a single electroacoustic transducer that produces sound in response to an electrical audio input signal. A driver may be implemented in any appropriate type, geometry and size, and may include horns, cones, ribbon transducers, and the like. The term “speaker” means one or more drivers in a unitary enclosure.
For the embodiment of
In a typical adaptive audio environment, a number of speaker enclosures will be contained within the listening room.
The speakers used in an adaptive audio system for a home theater or similar environment may use a configuration that is based on existing surround-sound configurations (e.g., 5.1, 7.1, 9.1, etc.). In this case, a number of drivers are provided and defined as per the known surround sound convention, with additional drivers and definitions provided for the upward-firing sound components.
For the direct sub-channels, the speaker enclosure would contain drivers in which the median axis of the driver bisects the “sweet-spot”, or acoustic center of the room. The upward-firing drivers would be positioned such that the angle between the median plane of the driver and the acoustic center would be some angle in the range of 45 to 180 degrees. In the case of positioning the driver at 180 degrees, the back-facing driver could provide sound diffusion by reflecting off of a back wall. This configuration utilizes the acoustic principal that after time-alignment of the upward-firing drivers with the direct drivers, the early arrival signal component would be coherent, while the late arriving components would benefit from the natural diffusion provided by the room.
In order to achieve the height cues provided by the adaptive audio system, the upward-firing drivers could be angled upward from the horizontal plane, and in the extreme could be positioned to radiate straight up and reflect off of a reflective surface such as a flat ceiling, or an acoustic diffuser placed immediately above the enclosure. To provide additional directionality, the center speaker could utilize a soundbar configuration (such as shown in
The 5.1 configuration of
As an alternative to the n.1 configurations described above a more flexible pod-based system may be utilized whereby each driver is contained within its own enclosure, which could then be mounted in any convenient location. This would use a driver configuration such as shown in
In order to enhance the configurability and accuracy of the adaptive audio system using upward-firing addressable drivers, a number of sensors and feedback devices could be added to the enclosures to inform the renderer of characteristics that could be used in the rendering algorithm. For example, a microphone installed in each enclosure would allow the system to measure the phase, frequency and reverberation characteristics of the room, together with the position of the speakers relative to each other using triangulation and the HRTF-like functions of the enclosures themselves. Inertial sensors (e.g., gyroscopes, compasses, etc.) could be used to detect direction and angle of the enclosures; and optical and visual sensors (e.g., using a laser-based infra-red range finder) could be used to provide positional information relative to the room itself. These represent just a few possibilities of additional sensors that could be used in the system, and others are possible as well.
Such sensor systems can be further enhanced by allowing the position of the drivers and/or the acoustic modifiers of the enclosures to be automatically adjustable via electromechanical servos. This would allow the directionality of the drivers to be changed at runtime to suit their positioning in the room relative to the walls and other drivers (“active steering”). Similarly, any acoustic modifiers (such as baffles, horns or wave guides) could be tuned to provide the correct frequency and phase responses for optimal playback in any room configuration (“active tuning”). Both active steering and active tuning could be performed during initial room configuration (e.g., in conjunction with the auto-EQ/auto-room configuration system) or during playback in response to the content being rendered.
Bi-Directional Interconnect
Once configured, the speakers must be connected to the rendering system. Traditional interconnects are typically of two types: speaker-level input for passive speakers and line-level input for active speakers. As shown in
In an embodiment, each driver in each of the cabinets of the system is assigned an identifier (e.g., a numerical assignment) during system setup. Each speaker cabinet can also be uniquely identified. This numerical assignment is used by the speaker cabinet to determine which audio signal is sent to which driver within the cabinet. The assignment is stored in the speaker cabinet in an appropriate memory device. Alternatively, each driver may be configured to store its own identifier in local memory. In a further alternative, such as one in which the drivers/speakers have no local storage capacity, the identifiers can be stored in the rendering stage or other component within the sound source 1002. During a speaker discovery process, each speaker (or a central database) is queried by the sound source for its profile. The profile defines certain driver definitions including the number of drivers in a speaker cabinet or other defined array, the acoustic characteristics of each driver (e.g. driver type, frequency response, and so on), the x,y,z position of center of each driver relative to center of the front face of the speaker cabinet, the angle of each driver with respect to a defined plane (e.g., ceiling, floor, cabinet vertical axis, etc.), and the number of microphones and microphone characteristics. Other relevant driver and microphone/sensor parameters may also be defined. In an embodiment, the driver definitions and speaker cabinet profile may be expressed as one or more XML documents used by the renderer.
In one possible implementation, an Internet Protocol (IP) control network is created between the sound source 1002 and the speaker cabinet 1004. Each speaker cabinet and sound source acts as a single network endpoint and is given a link-local address upon initialization or power-on. An auto-discovery mechanism such as zero configuration networking (zeroconf) may be used to allow the sound source to locate each speaker on the network. Zero configuration networking is an example of a process that automatically creates a usable IP network without manual operator intervention or special configuration servers, and other similar techniques may be used. Given an intelligent network system, multiple sources may reside on the IP network as the speakers. This allows multiple sources to directly drive the speakers without routing sound through a “master” audio source (e.g. traditional A/V receiver). If another source attempts to address the speakers, communications is performed between all sources to determine which source is currently “active”, whether being active is necessary, and whether control can be transitioned to a new sound source. Sources may be pre-assigned a priority during manufacturing based on their classification, for example, a telecommunications source may have a higher priority than an entertainment source. In multi-room environment, such as a typical home environment, all speakers within the overall environment may reside on a single network, but may not need to be addressed simultaneously. During setup and auto-configuration, the sound level provided back over interconnect 1008 can be used to determine which speakers are located in the same physical space. Once this information is determined, the speakers may be grouped into clusters. In this case, cluster IDs can be assigned and made part of the driver definitions. The cluster ID is sent to each speaker, and each cluster can be addressed simultaneously by the sound source 1002.
As shown in
System Configuration and Calibration
As shown in
The microphone(s) are used to enable the automatic configuration and calibration of the renderer and post-processing algorithms. In the adaptive audio system, the renderer is responsible for converting a hybrid object and channel-based audio stream into individual audio signals designated for specific addressable drivers, within one or more physical speakers. The post-processing component may include: delay, equalization, gain, speaker virtualization, and upmixing. The speaker configuration represents often critical information that the renderer component can use to convert a hybrid object and channel-based audio stream into individual per-driver audio signals to provide optimum playback of audio content. System configuration information includes: (1) the number of physical speakers in the system, (2) the number individually addressable drivers in each speaker, and (3) the position and direction of each individually addressable driver, relative to the room geometry. Other characteristics are also possible.
The number of physical speakers in the system and the number of individually addressable drivers in each speaker are the physical speaker properties. These properties are transmitted directly from the speakers via the bi-directional interconnect 456 to the renderer 454. The renderer and speakers use a common discovery protocol, so that when speakers are connected or disconnected from the system, the render is notified of the change, and can reconfigure the system accordingly.
The geometry (size and shape) of the listening room is a necessary item of information in the configuration and calibration process. The geometry can be determined in a number of different ways. In a manual configuration mode, the width, length and height of the minimum bounding cube for the room are entered into the system by the listener or technician through a user interface that provides input to the renderer or other processing unit within the adaptive audio system. Various different user interface techniques and tools may be used for this purpose. For example, the room geometry can be sent to the renderer by a program that automatically maps or traces the geometry of the room. Such a system may use a combination of computer vision, sonar, and 3D laser-based physical mapping.
The renderer uses the position of the speakers within the room geometry to derive the audio signals for each individually addressable driver, including both direct and reflected (upward-firing) drivers. The direct drivers are those that are aimed such that the majority of their dispersion pattern intersects the listening position before being diffused by one or more reflective surfaces (such as floor, wall or ceiling). The reflected drivers are those that are aimed such that the majority of their dispersion patterns are reflected prior to intersecting the listening position such as illustrated in
Driver position and aiming is typically performed using manual or automatic techniques. In some cases, inertial sensors may be incorporated into each speaker. In this mode, the center speaker is designated as the “master” and its compass measurement is considered as the reference. The other speakers then transmit the dispersion patterns and compass positions for each off their individually addressable drivers. Coupled with the room geometry, the difference between the reference angle of the center speaker and each addition driver provides enough information for the system to automatically determine if a driver is direct or reflected.
The speaker position configuration may be fully automated if a 3D positional (i.e., Ambisonic) microphone is used. In this mode, the system sends a test signal to each driver and records the response. Depending on the microphone type, the signals may need to be transformed into an x, y, z representation. These signals are analyzed to find the x, y, and z components of the dominant first arrival. Coupled with the room geometry, this usually provides enough information for the system to automatically set the 3D coordinates for all speaker positions, direct or reflected. Depending on the room geometry, a hybrid combination of the three described methods for configuring the speaker coordinates may be more effective than using just one technique alone.
Speaker configuration information is one component required to configure the renderer. Speaker calibration information is also necessary to configure the post-processing chain: delay, equalization, and gain.
In the case of automatic calibration using multiple microphones, the delay, equalization, and gain are automatically calculated by the system using multiple omni-directional measurement microphones. The process is substantially identical to the single microphone technique, accept that it is repeated for each of the microphones, and the results are averaged.
Alternative Playback Systems
Instead of implementing an adaptive audio system in an entire room or theater, it is possible to implements aspects of the adaptive audio system in more localized applications, such as televisions, computers, game consoles, or similar devices. This case effectively relies on speakers that are arrayed in a flat plane corresponding to the viewing screen or monitor surface.
The television environment may also include an HRC speaker as shown within soundbar 1304. Such an HRC speaker may be a steerable unit that allows panning through the HRC array. There may be benefits (particularly for larger screens) by having a front firing center channel array with individually addressable speakers that allow discrete pans of audio objects through the array that match the movement of video objects on the screen. This speaker is also shown to have side-firing speakers. These could be activated and used if the speaker is used as a soundbar so that the side-firing drivers provide more immersion due to the lack of surround or back speakers. The dynamic virtualization concept is also shown for the HRC/Soundbar speaker. The dynamic virtualization is shown for the L and R speakers on the farthest sides of the front firing speaker array. Again, this could be used for creating the perception of objects moving along the sides on the room. This modified center speaker could also include more speakers and implement a steerable sound beam with separately controlled sound zones. Also shown in the example implementation of
With respect to headphone rendering, the adaptive audio system maintains the creator's original intent by matching HRTFs to the spatial position. When audio is reproduced over headphones, binaural spatial virtualization can be achieved by the application of a Head Related Transfer Function (HRTF), which processes the audio, and add perceptual cues that create the perception of the audio being played in three-dimensional space and not over standard stereo headphones. The accuracy of the spatial reproduction is dependent on the selection of the appropriate HR TF which can vary based on several factors, including the spatial position of the audio channels or objects being rendered. Using the spatial information provided by the adaptive audio system can result in the selection of one- or a continuing varying number- of HRTFs representing 3D space to greatly improve the reproduction experience.
The system also facilitates adding guided, three-dimensional binaural rendering and virtualization. Similar to the case for spatial rendering, using new and modified speaker types and locations, it is possible through the use of three-dimensional HRTFs to create cues to simulate sound coming from both the horizontal plane and the vertical axis. Previous audio formats that provide only channel and fixed speaker location information rendering have been more limited.
Headphone Rendering System
With the adaptive audio format information, a binaural, three-dimensional rendering headphone system has detailed and useful information that can be used to direct which elements of the audio are suitable to be rendering in both the horizontal and vertical planes. Some content may rely on the use of overhead speakers to provide a greater sense of envelopment. These audio objects and information could be used for binaural rendering that is perceived to be above the listener's head when using headphones.
The BRIR filter applied to each of the N bed signals is fixed to a specific location associated with a particular channel of the audio system. For instance, the BRIR filter applied to the center channel signal may correspond to a source located at 0 degrees azimuth and 0 degrees elevation, so that the listener gets the impression that the sound corresponding to the center channel comes from a source directly in front of the listener. Likewise, the BRIR filters applied to the left and right channels may correspond to sources located at+/−30 degree azimuth. The BRIR filter applied to each of the M object signals is time-varying and is adapted based on positional and/or trajectory data associated with each object. For example, the positional data for object 1 may indicate that at time t0 the object is directly behind the listener. In such case, a BRIR filter corresponding to a location directly behind the listener is applied to object 1. Furthermore, the positional data for object 1 may indicate that at time t1 the object is directly above the listener. In such case, an BRIR filter corresponding to a location directly above the listener is applied to object 1. Similarly, for each of the remaining objects 2-m, BRIR filters corresponding to the time-varying positional data for each object are applied.
With reference to
HRTF Filter Set
With respect to the actual listener in the listening environment, the human torso, head and pinna (outer ear) make up a set of boundaries that can be modeled using ray-tracing and other techniques to simulate the head-related transfer function (HRTF, in the frequency domain) or head-related impulse response (HRIR, in the time domain). These elements (torso, head and pinna) can be individually modeled in a way that allows them to be later structurally combined into a single HRIR. Such a model allows for a high degree of customization based on anthropomorphic measurements (head radius, neck height, etc.), and provides binaural cues necessary for localization in the horizontal (azimuthal) plane as well as weak low-frequency cues in the vertical (elevation) plane.
It is known that the pinna provides strong elevation cues, as well as front-to-back cues. These are typically described as spectral features in the frequency domain—often a set of notches that are related in frequency and move as the sound source elevation moves. These features are also present in the time domain by way of the HRIR. They can be seen as a set of peaks and dips in the impulse response that move in a strong, systematic way as elevation changes (there are also some weaker movements that correspond to azimuth changes).
In an embodiment, an HRTF filter set for use with the headphone rendering system is built using publically available HRTF databases to gather data on pinna features. The databases were translated to a common coordinate system and outlier subjects were removed. The coordinate system chosen was along the “inter-aural axis”, which allows for elevation features to be tracked independently for any given azimuth. The impulse responses were extracted, time aligned, and over-sampled for each spatial location. Effects of head shadow and torso reflections were removed to the extent possible. Across all subjects, for any given spatial location, a weighted averaging of the features was performed, with the weighting done in a way that the features that changed with elevation were given greater weights. The results were then averaged, filtered, and down-sampled back to a common sample rate. An average measurement for human anthropometry were used for the head and torso model and combined with the averaged pinna data.
In an embodiment, the headphone rendering system includes a method of compensating for the HETF for improved binaural rendering. This method involves modeling and deriving the compensation filter of HETFs in the Z domain. The HETF is affected by the reflections between the inner-surface of the headphone and the surface of the external ear involved. If the binaural recordings are made at the entrances to blocked ear canals as, for example, from a B&K 4100 dummy head, the HETF is defined as the transfer function from the input of the headphone to the sound pressure signal at the entrance to the blocked ear canal. If the binaural recordings are made at the eardrum as, for example, from a “HATS acoustic” dummy head, the HETF is defined as the transfer function from the input of the headphone to the sound pressure signal at the eardrum.
Considering that the reflection coefficient (R1) of the headphone inner-surface is frequency dependent, and that the reflection coefficient (R2) of external ear surface or eardrum is also frequency dependent, in the Z domain the product of the reflection coefficient from the headphone and the reflection coefficient from the external ear surface (i.e., R1*R2) can be modeled as a first order IIR (Infinite Impulse Response) filter. Furthermore, considering that there are time delays between the reflections from the inner surface of the headphone and the reflections from the surface of the external ear and that there are second-order and higher order reflections between them, the HETF in the Z domain is modeled as a higher order IIR filter H(z), which is formed by the summation of products of reflection coefficients with different time delays and orders. In addition, the inverse filter of the HETF is modeled using an IIR filter E(z), which is the reciprocal of the H(z).
From the measured impulse response of HETF, the process obtains e(n), the time domain impulse response of the inverse filter of the HETF, such that both the phase and the magnitude spectral responses of HETF are equalized. It further derives the parameters of the inverse filter E(z) from the e(n) sequence using Pony's method, as an example. In order to obtain a stable E(z), the order of E(z) is set to a proper number, and only the first M samples of e(n) are chosen in deriving the parameters of E(z).
This headphone compensation method equalizes both phase and magnitude spectra of the HETF. Moreover, by using the described IIR filter E(z) as the compensation filter, instead of a FIR filter to achieve equivalent compensation, it imposes less computational cost as well a shorter time delay, as compared to other methods.
Metadata Definitions
In an embodiment, the adaptive audio system includes components that generate metadata from the original spatial audio format. The methods and components of system 300 comprise an audio rendering system configured to process one or more bitstreams containing both conventional channel-based audio elements and audio object coding elements. A new extension layer containing the audio object coding elements is defined and added to either one of the channel-based audio codec bitstream or the audio object bitstream. This approach enables bitstreams, which include the extension layer to be processed by renderers for use with existing speaker and driver designs or next generation speakers utilizing individually addressable drivers and driver definitions. The spatial audio content from the spatial audio processor comprises audio objects, channels, and position metadata. When an object is rendered, it is assigned to one or more speakers according to the position metadata, and the location of the playback speakers.
Additional metadata may be associated with the object to alter the playback location or otherwise limit the speakers that are to be used for playback. Metadata is generated in the audio workstation in response to the engineer's mixing inputs to provide rendering queues that control spatial parameters (e.g., position, velocity, intensity, timbre, etc.) and specify which driver(s) or speaker(s) in the listening environment play respective sounds during exhibition. The metadata is associated with the respective audio data in the workstation for packaging and transport by spatial audio processor.
Upmixing
Embodiments of the adaptive audio rendering system include an upmixer based on factoring audio channels into reflected and direct sub-channels. A direct sub-channel is that portion of the input channel that is routed to drivers that deliver early-reflection acoustic waveforms to the listener. A reflected or diffuse sub-channel is that portion of the original audio channel that is intended to have a dominant portion of the driver's energy reflected off of nearby surfaces and walls. The reflected sub-channel thus refers to those parts of the original channel that are preferred to arrive at the listener after diffusion into the local acoustic environment, or that are specifically reflected off of a point on a surface (e.g., the ceiling) to another location in the room. Each sub-channel would be routed to independent speaker drivers, since the physical orientation of the drivers for one sub-channel relative to those of the other sub-channel, would add acoustic spatial diversity to each incoming signal. In an embodiment, the reflected sub-channel(s) are sent to upward-firing speakers or speakers pointed to a surface for indirect transmission of sound to the desired location.
It should be noted that, in the context of upmixing signals, the reflected acoustic waveform can optionally make no distinction between reflections off of a specific surface and reflections off of any arbitrary surfaces that result in general diffusion of the energy from the non-directed driver. In the latter case, the sound wave associated with this driver would in the ideal, be directionless (i.e., diffuse waveforms are those in which the sound comes from not one single direction).
With respect to the decomposition process 1700, it is important to note that energy preservation is preserved between the reflected sub-channel and the direct sub-channel at each stage in the process. For this calculation, the variable a is defined as that portion of the input channel that is associated with the direct sub-channel, and—is defined as that portion associated with the diffuse sub-channel. The relationship to determined energy preservation can then be expressed according to the following equations:
y(k)DIRECT=x(k)αk, ∀k
y(k)DIFFUSE=x(k)√{square root over (1−|αk|2)}, ∀k
where β=√{square root over (1−|αk|2)}
In the above equations, x is the input channel and k is the transform index. In an embodiment, the solution is computed on frequency domain quantities, either in the form of complex discrete Fourier transform coefficients, real-based MDCT transform coefficients, or QMF (quadrature mirror filter) sub-band coefficients (real or complex). Thus in the process, it is presumed that a forward transform is applied to the input channels, and the corresponding inverse transform is applied to the output sub-channels.
Where SDi are the frequency-domain coefficients for an input channel of index i, while SDj are the coefficients for the next spatially adjacent input audio channel, of index j. The E { } operator is the expectation operator, and can be implemented using fixed averaging over a set number of blocks of audio, or implemented as an smoothing algorithm in which the smoothing is conducted for each frequency domain coefficient, across blocks. This smoother can be implemented as an exponential smoother using an infinite impulse response (IIR) filter topology.
The geometric mean between the ICC of these two adjacent channels is computed and this value is a number between −1 and 1. The value for a is then set as the difference between 1.0 and this mean. The ICC broadly describes how much of the signal is common between two channels. Signals with high inter-channel correlation are routed to the reflected channels, whereas signals that are unique relative to their nearby channels are routed to the direct subchannels. This operation can be described according to the following example pseudocode:
Where pICC refers to the ICC of the i−I input channel spatially adjacent the current input channel i, and niCC refers to the ICC of the i+I indexed input channel spatially adjacent to the current input channel i. In step 1904, the system computes the transient scaling terms for each input channel. These scaling factors contribute to the reflected versus direct mix calculation, where the amount of scaling is proportional to the energy in the transient. In general, it is desired that transient signals be routed to the direct sub-channels. Thus a is compared against a scaling factor sf which is set to 1.0 (or near 1.0 for weaker transients) in the event of a positive transient detection
αi=max(αi, sfi)
Where the index i corresponds to the input channel i. Each transient scaling factor sf has a holdparameter as well as a decay parameter to control how the scaling factor evolves over time after the transient. These hold and decay parameters are generally on the order of milliseconds, but the decay back to the nominal value of a can extend to upwards of a full second. Using the a values computed in block 1902 and the transient scaling factors computed in 1904, the system splits each input channel into reflected and direct sub-channels such that sum energy between the sub-channels is preserved, step 1906.
As an optional step, the reflected channels can be further decomposed into reverberant and non-reverberant components, step 1908. The non-reverberant sub-channels could either be summed back into the direct sub-channel, or sent to dedicated drivers in the output. Since it may not be known which linear transformation was applied to reverberate the input signal, a blind deconvolution or related algorithm (such as blind source separation) is applied.
A second optional step is to further decorrelate the reflected channel from the direct channel, using a decorrelator that operates on each frequency domain transform across blocks, step 1910. In an embodiment the decorrelator is comprised of a number of delay elements (the delay in milliseconds corresponds to the block integer delay, multiplied by the length of the underlying time-to-frequency transform) and an all-pass IIR (infinite impulse response) filter with filter coefficients that can arbitrarily move within a constrained Z-domain circle as a function of time. In step 1912, the system performs equalization and delay functions to the reflected and direct channels. In a usual case, the direct sub-channels are delayed by an amount that would allow for the acoustic wavefront from the direct driver to be phase coherent with the principal reflected energy wavefront (in a mean squared energy error sense) at the listening position. Likewise, equalization is applied to the reflected channel to compensate for expected (or measured) diffuseness of the room in order to best match the timbre between the reflected and direct sub-channels.
Another option would be to control the algorithm through the use of an environmental sensing microphone that could be present in the room. This would allow for the calculation of the direct-to-reverberant ratio (DR-ratio) of the room. With the DR-ratio, final control would be possible in determining the optimal split between the diffuse and direct sub-channels. In particular, for highly reverberant rooms, it is reasonable to presume that the diffuse sub-channel will have more diffusion applied to the listener position, and as such the mix between the diffuse and direct sub-channels could be affected in the blind deconvolution and decorrelation steps. Specifically, for rooms with very little reflected acoustic energy, the amount of signal that is routed to the diffuse sub-channels, could be increased. Additionally, a microphone sensor in the acoustic environment could determine the optimal equalization to be applied to the diffuse subchannel. An adaptive equalizer could ensure that the diffuse sub-channel is optimally delayed and equalized such that the wavefronts from both sub-channels combine in a phase coherent manner at the listening position.
Virtualizer
In an embodiment, the adaptive audio processing system includes a component for virtual rendering of object-based audio over multiple pairs of loudspeakers, that may include one or more individually addressable drivers configured to reflect sound. This component performs virtual rendering of object-based audio through binaural rendering of each object followed by panning of the resulting stereo binaural signal between a multitude of cross-talk cancelation circuits feeding a corresponding multitude of speaker pairs. It improves the spatial impression for both listeners inside and outside of the cross-talk canceller sweet spot over prior virtualizers that simply use a single pair of speakers. In other words it overcomes the disadvantage that crosstalk cancelation is highly dependent on the listener sitting in the position with respect to the speakers that is assumed in the design of the crosstalk canceller. If the listener is not sitting in this so-called “sweet spot”, then the crosstalk cancellation effect may be compromised, either partially or totally, and the spatial impression intended by the binaural signal is not perceived by the listener. This is particularly problematic for multiple listeners in which case only one of the listeners can effectively occupy the sweet spot.
In spatial audio reproduction system, the sweet spot may be extended to more than one listener by utilizing more than two speakers. This is most often achieved by surrounding a larger sweet spot with more than two speakers, as with a 5.1 surround system. In such systems, sounds intended to be heard from behind, for example, are generated by speakers physically located behind all of the listeners, and as such, all of the listeners perceive these sounds as coming from behind. With virtual spatial rendering over stereo loudspeakers, on the other hand, perception of audio from behind is controlled by the HRTFs used to generated the binaural signal and will only be perceived properly by the listener in the sweet spot. Listeners outside of the sweet spot will likely perceive the audio as emanating from the stereo speakers in front of them. As described previously, however, installation of such surround systems is not practical for many consumers, or they simply may prefer to keep all speakers located at the front of the listening environment, oftentimes collocated with a television display. By using multiple speaker pairs in conjunction with virtual spatial rendering, a virtualizer under an embodiment combines the benefits of more than two speakers for listeners outside of the sweet spot and maintains or enhances the experience for listeners inside of the sweet spot in a manner that allows all utilized speaker pairs to be substantially collocated.
In an embodiment, virtual spatial rendering is extended to multiple pairs of loudspeakers by panning the binaural signal generated from each audio object between multiple crosstalk cancellers. The panning between crosstalk cancellers is controlled by the position associated with each audio object, the same position utilized for selecting the binaural filter pair associated with each object. The multiple crosstalk cancellers are designed for and feed into a corresponding multitude of speaker pairs, each with a different physical location and/or orientation with respect to the intended listening position. A multitude of objects at various positions in space may be simultaneously rendered. In this case, the binaural signal may expressed by a sum of object signals with their associated HRTFs applied. With a multi-object binaural signal, the entire rendering chain to generate the speaker signals, in a system with M pairs of speakers may be expressed in the following equation:
where
oi=audio signal for the ith object out of N
Bi=binaural filter pair for the ith object given by Bi=HRTF{pos(oi)}
αij=panning coefficient for the ith object into the jth crosstalk canceller
Cj=crosstalk canceller matrix for the jth speaker pair
sj=stereo speaker signal sent to the jth speaker pair
The M panning coefficients associated with each object i are computed using a panning function which takes as input the possibly time-varying position of the object:
In an embodiment, for each of the N object signals oi, a pair of binaural filters Bi, selected as a function of the object position pos(oi), is first applied to generate a binaural signal. Simultaneously, a panning function computes M panning coefficients, ail . . . aiM, based on the object position pos(oi). Each panning coefficient separately multiplies the binaural signal generating M scaled binaural signals. For each of the M crosstalk cancellers, Cj, the jth scaled binaural signals from all N objects are summed This summed signal is then processed by the crosstalk canceller to generate the jth speaker signal pair sj which is played back through the jth speaker pair.
In order to extend the benefits of the multiple loudspeaker pairs to listeners outside of the sweet spot, the panning function is configured to distribute the object signals to speaker pairs in a manner that helps convey the object's desired physical position to these listeners. For example, if the object is meant to be heard from overhead, then the panner should pan the object to the speaker pair that most effectively reproduces a sense of height for all listeners. If the object is meant to be heard to the side, the panner should pan the object to the pair of speakers than most effectively reproduces a sense of width for all listeners. More generally, the panning function should compare the desired spatial position of each object with the spatial reproduction capabilities of each loudspeaker pair in order to compute an optimal set of panning coefficients.
In one embodiment, three speaker pairs are utilized, and all are collocated in front of the listener.
For both the generation of the cross-talk cancellers associated with each of the speaker pairs as well as the binaural filters for each audio object, parametric spherical head model HRTFs are utilized. These HRTFs are dependent only on the angle of an object with respect to the median plane of the listener. As shown in
Associated with each audio object signal oi is a possibly time-varying position given in Cartesian coordinates {xi yi zi}. Since the parametric HRTFs employed in the preferred embodiment do not contain any elevation cues, only the x and y coordinates of the object position are utilized in computing the binaural filter pair from the HRTFs function. These {xi yi} coordinates are transformed into equivalent radius and angle {ri θi}, where the radius is normalized to lie between zero and one. The parametric does not depend on distance from the listener, and therefore the radius is incorporated into computation of the left and right binaural filters as follows:
BL=(1−√{square root over (ri)})+r√{square root over (ri)}HRTFL{θi}
BR=(1−√{square root over (ri)})+r√{square root over (ri)}HRTFR{θi}
When the radius is zero, the binaural filters are simply unity across all frequency, and the listener hears the object signal equally at both ears. This corresponds to the case when the object position is located exactly within the listener's head. When the radius is one, the filters are equal to the parametric HRTFs defined at angle θi. Taking the square root of the radius term biases this interpolation of the filters toward the HRTF, which better preserves spatial information. Note that this computation is needed because the parametric HRTF model does not incorporate distance cues. A different HRTF set might incorporate such cues in which case the interpolation described by the equation above would not be necessary.
For each object, the panning coefficients for each of the three crosstalk cancellers are computed from the object position {xi yi zi}, relative to the orientation of each canceller. The upward-firing driver pair 2004 is meant to convey sounds from above by reflecting sound off of the ceiling. As such, its associated panning coefficient is proportional to the elevation coordinate zi. The panning coefficients of the front and side-firing driver pairs 2006, 2008 are governed by the object angle θi, derived from the {xi yi} coordinates. When the absolute value of θi is less that 30 degrees, object is panned entirely to the front pair 2008. When the absolute value of θi is between 30 and 90 degrees, the object is panned between the front and side pairs. And when the absolute value of 8; is greater than 90 degrees, the object is panned entirely to the side pair 2006. With this panning algorithm, a listener in the sweet spot receives the benefits of all three cross-talk cancellers. In addition, the perception of elevation is added with the upward firing pair, and the side firing pair adds an element of diffuseness for objects mixed to the side and back which can enhance perceived envelopment. For listeners outside of the sweet-spot, the cancellers lose much of their effectiveness, but the listener can still appreciate the perception of elevation from the upward-firing driver pair 2004 and the variation between direct and diffuse sound from the front to side panning.
In an embodiment, the virtualization technique described above is applied to an adaptive audio format that contains a mixture of dynamic object signals along with fixed channel signals, as described above. The fixed channels signals may be processed by assigning a fixed spatial position to each channel.
As shown in
Features and Capabilities
As stated above, the adaptive audio ecosystem allows the content creator to embed the spatial intent of the mix (position, size, velocity, etc.) within the bitstream via metadata. This allows an incredible amount of flexibility in the spatial reproduction of audio. From a spatial rendering standpoint, the adaptive audio format enables the content creator to adapt the mix to the exact position of the speakers in the room to avoid spatial distortion caused by the geometry of the playback system not being identical to the authoring system. In current consumer audio reproduction where only audio for a speaker channel is sent, the intent of the content creator is unknown for locations in the room other than fixed speaker locations. Under the current channel/speaker paradigm the only information that is known is that a specific audio channel should be sent to a specific speaker that has a predefined location in a room. In the adaptive audio system, using metadata conveyed through the creation and distribution pipeline, the reproduction system can use this information to reproduce the content in a manner that matches the original intent of the content creator. For example, the relationship between speakers is known for different audio objects. By providing the spatial location for an audio object, the intention of the content creator is known and this can be “mapped” onto the user's speaker configuration, including their location. With a dynamic rendering audio rendering system, this rendering can be updated and improved by adding additional speakers.
The system also enables adding guided, three-dimensional spatial rendering. There have been many attempts to create a more immersive audio rendering experience through the use of new speaker designs and configurations. These include the use of bi-pole and di-pole speakers, side-firing, rear-firing and upward-firing drivers. With previous channel and fixed speaker location systems, determining which elements of audio should be sent to these modified speakers has been guesswork at best. Using an adaptive audio format, a rendering system has detailed and useful information of which elements of the audio (objects or otherwise) are suitable to be sent to new speaker configurations. That is, the system allows for control over which audio signals are sent to the front-firing drivers and which are sent to the upward-firing drivers. For example, the adaptive audio cinema content relies heavily on the use of overhead speakers to provide a greater sense of envelopment. These audio objects and information may be sent to upward-firing drivers to provide reflected audio in the listening environment to create a similar effect.
The system also allows for adapting the mix to the exact hardware configuration of the reproduction system. There exist many different possible speaker types and configurations in consumer rendering equipment such as televisions, home theaters, soundbars, portable music player docks, and so on. When these systems are sent channel specific audio information (i.e. left and right channel or standard multichannel audio) the system must process the audio to appropriately match the capabilities of the rendering equipment. A typical example is when standard stereo (left, right) audio is sent to a soundbar, which has more than two speakers. In current systems where only audio for a speaker channel is sent, the intent of the content creator is unknown and a more immersive audio experience made possible by the enhanced equipment must be created by algorithms that make assumptions of how to modify the audio for reproduction on the hardware. An example of this is the use of PLII, PLII-z, or Next Generation Surround to “up-mix” channel-based audio to more speakers than the original number of channel feeds. With the adaptive audio system, using metadata conveyed throughout the creation and distribution pipeline, a reproduction system can use this information to reproduce the content in a manner that more closely matches the original intent of the content creator. For example, some soundbars have side-firing speakers to create a sense of envelopment. With adaptive audio, the spatial information and the content type information (i.e., dialog, music, ambient effects, etc.) can be used by the soundbar when controlled by a rendering system such as a TV or A/V receiver to send only the appropriate audio to these side-firing speakers.
The spatial information conveyed by adaptive audio allows the dynamic rendering of content with an awareness of the location and type of speakers present. In addition information on the relationship of the listener or listeners to the audio reproduction equipment is now potentially available and may be used in rendering. Most gaming consoles include a camera accessory and intelligent image processing that can determine the position and identity of a person in the room. This information may be used by an adaptive audio system to alter the rendering to more accurately convey the creative intent of the content creator based on the listener's position. For example, in nearly all cases, audio rendered for playback assumes the listener is located in an ideal “sweet spot” which is often equidistant from each speaker and the same position the sound mixer was located during content creation. However, many times people are not in this ideal position and their experience does not match the creative intent of the mixer. A typical example is when a listener is seated on the left side of the room on a chair or couch in a living room. For this case, sound being reproduced from the nearer speakers on the left will be perceived as being louder and skewing the spatial perception of the audio mix to the left. By understanding the position of the listener, the system could adjust the rendering of the audio to lower the level of sound on the left speakers and raise the level of the right speakers to rebalance the audio mix and make it perceptually correct. Delaying the audio to compensate for the distance of the listener from the sweet spot is also possible. Listener position could be detected either through the use of a camera or a modified remote control with some built-in signaling that would signal listener position to the rendering system.
In addition to using standard speakers and speaker locations to address listening position it is also possible to use beam steering technologies to create sound field “zones” that vary depending on listener position and content. Audio beam forming uses an array of speakers (typically 8 to 16 horizontally spaced speakers) and use phase manipulation and processing to create a steerable sound beam. The beam forming speaker array allows the creation of audio zones where the audio is primarily audible that can be used to direct specific sounds or objects with selective processing to a specific spatial location. An obvious use case is to process the dialog in a soundtrack using a dialog enhancement post-processing algorithm and beam that audio object directly to a user that is hearing impaired.
Matrix Encoding
In some cases audio objects may be a desired component of adaptive audio content; however, based on bandwidth limitations, it may not be possible to send both channel/speaker audio and audio objects. In the past matrix encoding has been used to convey more audio information than is possible for a given distribution system. For example, this was the case in the early days of cinema where multi-channel audio was created by the sound mixers but the film formats only provided stereo audio. Matrix encoding was used to intelligently downmix the multi-channel audio to two stereo channels, which were then processed with certain algorithms to recreate a close approximation of the multi-channel mix from the stereo audio. Similarly, it is possible to intelligently downmix audio objects into the base speaker channels and through the use of adaptive audio metadata and sophisticated time and frequency sensitive next generation surround algorithms to extract the objects and correctly spatially render them with an adaptive audio rendering system.
Additionally, when there are bandwidth limitations of the transmission system for the audio (3G and 4G wireless applications for example) there is also benefit from transmitting spatially diverse multi-channel beds that are matrix encoded along with individual audio objects. One use case of such a transmission methodology would be for the transmission of a sports broadcast with two distinct audio beds and multiple audio objects. The audio beds could represent the multi-channel audio captured in two different teams' bleacher sections and the audio objects could represent different announcers who may be sympathetic to one team or the other. Using standard coding a 5.1 representation of each bed along with two or more objects could exceed the bandwidth constraints of the transmission system. In this case, if each of the 5.1 beds were matrix encoded to a stereo signal, then two beds that were originally captured as 5.1 channels could be transmitted as two-channel bed 1, two-channel bed 2, object 1, and object 2 as only four channels of audio instead of 5.1+5.1+2 or 12.1 channels.
Position and Content Dependent Processing
The adaptive audio ecosystem allows the content creator to create individual audio objects and add information about the content that can be conveyed to the reproduction system. This allows a large amount of flexibility in the processing of audio prior to reproduction. Processing can be adapted to the position and type of object through dynamic control of speaker virtualization based on object position and size. Speaker virtualization refers to a method of processing audio such that a virtual speaker is perceived by a listener.
This method is often used for stereo speaker reproduction when the source audio is multi-channel audio that includes surround speaker channel feeds. The virtual speaker processing modifies the surround speaker channel audio in such a way that when it is played back on stereo speakers, the surround audio elements are virtualized to the side and back of the listener as if there was a virtual speaker located there. Currently the location attributes of the virtual speaker location are static because the intended location of the surround speakers was fixed. However, with adaptive audio content, the spatial locations of different audio objects are dynamic and distinct (i.e. unique to each object). It is possible that post processing such as virtual speaker virtualization can now be controlled in a more informed way by dynamically controlling parameters such as speaker positional angle for each object and then combining the rendered outputs of several virtualized objects to create a more immersive audio experience that more closely represents the intent of the sound mixer.
In addition to the standard horizontal virtualization of audio objects, it is possible to use perceptual height cues that process fixed channel and dynamic object audio and get the perception of height reproduction of audio from a standard pair of stereo speakers in the normal, horizontal plane, location.
Certain effects or enhancement processes can be judiciously applied to appropriate types of audio content. For example, dialog enhancement may be applied to dialog objects only. Dialog enhancement refers to a method of processing audio that contains dialog such that the audibility and/or intelligibility of the dialog is increased and or improved. In many cases the audio processing that is applied to dialog is inappropriate for non-dialog audio content (i.e. music, ambient effects, etc.) and can result is an objectionable audible artifact. With adaptive audio, an audio object could contain only the dialog in a piece of content and can be labeled accordingly so that a rendering solution would selectively apply dialog enhancement to only the dialog content. In addition, if the audio object is only dialog (and not a mixture of dialog and other content, which is often the case) then the dialog enhancement processing can process dialog exclusively (thereby limiting any processing being performed on any other content).
Similarly audio response or equalization management can also be tailored to specific audio characteristics. For example, bass management (filtering, attenuation, gain) targeted at specific object based on their type. Bass management refers to selectively isolating and processing only the bass (or lower) frequencies in a particular piece of content. With current audio systems and delivery mechanisms this is a “blind” process that is applied to all of the audio. With adaptive audio, specific audio objects in which bass management is appropriate can be identified by metadata and the rendering processing applied appropriately.
The adaptive audio system also facilitates object-based dynamic range compression. Traditional audio tracks have the same duration as the content itself, while an audio object might occur for a limited amount of time in the content. The metadata associated with an object may contain level-related information about its average and peak signal amplitude, as well as its onset or attack time (particularly for transient material). This information would allow a compressor to better adapt its compression and time constants (attack, release, etc.) to better suit the content.
The system also facilitates automatic loudspeaker-room equalization. Loudspeaker and room acoustics play a significant role in introducing audible coloration to the sound thereby impacting timbre of the reproduced sound. Furthermore, the acoustics are position-dependent due to room reflections and loudspeaker-directivity variations and because of this variation the perceived timbre will vary significantly for different listening positions. An AutoEQ (automatic room equalization) function provided in the system helps mitigate some of these issues through automatic loudspeaker-room spectral measurement and equalization, automated time-delay compensation (which provides proper imaging and possibly least-squares based relative speaker location detection) and level setting, bass-redirection based on loudspeaker headroom capability, as well as optimal splicing of the main loudspeakers with the subwoofer(s). In a home theater or other listening environment, the adaptive audio system includes certain additional functions, such as: (1) automated target curve computation based on playback room-acoustics (which is considered an open-problem in research for equalization in domestic listening rooms), (2) the influence of modal decay control using time-frequency analysis, (3) understanding the parameters derived from measurements that govern envelopment/spaciousness/source-width/intelligibility and controlling these to provide the best possible listening experience, (4) directional filtering incorporating head-models for matching timbre between front and “other” loudspeakers, and (5) detecting spatial positions of the loudspeakers in a discrete setup relative to the listener and spatial re-mapping (e.g., Summit wireless would be an example). The mismatch in timbre between loudspeakers is especially revealed on certain panned content between a front-anchor loudspeaker (e.g., center) and surround/back/wide/height loudspeakers.
Overall, the adaptive audio system also enables a compelling audio/video reproduction experience, particularly with larger screen sizes in a home environment, if the reproduced spatial location of some audio elements match image elements on the screen. An example is having the dialog in a film or television program spatially coincide with a person or character that is speaking on the screen. With normal speaker channel-based audio there is no easy method to determine where the dialog should be spatially positioned to match the location of the person or character on-screen. With the audio information available in an adaptive audio system, this type of audio/visual alignment could be easily achieved, even in home theater systems that are featuring ever larger size screens. The visual positional and audio spatial alignment could also be used for non-character/dialog objects such as cars, trucks, animation, and so on.
The adaptive audio ecosystem also allows for enhanced content management, by allowing a content creator to create individual audio objects and add information about the content that can be conveyed to the reproduction system. This allows a large amount of flexibility in the content management of audio. From a content management standpoint, adaptive audio enables various things such as changing the language of audio content by only replacing a dialog object to reduce content file size and/or reduce download time. Film, television and other entertainment programs are typically distributed internationally. This often requires that the language in the piece of content be changed depending on where it will be reproduced (French for films being shown in France, German for TV programs being shown in Germany, etc.). Today this often requires a completely independent audio soundtrack to be created, packaged, and distributed for each language. With the adaptive audio system and the inherent concept of audio objects, the dialog for a piece of content could an independent audio object. This allows the language of the content to be easily changed without updating or altering other elements of the audio soundtrack such as music, effects, etc. This would not only apply to foreign languages but also inappropriate language for certain audience, targeted advertising, etc.
Embodiments are also directed to a system for rendering object-based sound in a pair of headphones, comprising: an input stage receiving an input signal comprising a first plurality of input channels and a second plurality of audio objects, a first processor computing left and right headphone channel signals for each of the first plurality of input channels, and a second processor applying a time-invariant binaural room impulse response (BRIR) filter to each signal of the first plurality of input channels, and a time-varying BRIR filter to each object of the second plurality of objects to generate a set of left ear signals and right ear signals. This system may further comprise a left channel mixer mixing together the left ear signals to form an overall left ear signal, a right channel mixer mixing together the right ear signals to form an overall right ear signal; a left side equalizer equalizing the overall left ear signal to compensate for an acoustic transfer function from a left transducer of the headphone to the entrance of a listener's left ear; and a right side equalizer equalizing the overall right ear signal to compensate for an acoustic transfer function from a right transducer of the headphone to the entrance of the listener's right ear. In such a system, the BRIR filter may comprise a summer circuit configured to sum together a direct path response and one or more reflected path responses, wherein the one or more reflected path responses includes a specular effect and a diffraction effect of a listening environment in which the listener is located. The direct path and the one or more reflected paths may each comprise a source transfer function, a distance response, and a head related transfer function (HRTF), and wherein the one or more reflected paths each additionally comprise a surface response for one or more surfaces disposed in the listening environment; and the BRIR filter may be configured to produce a correct response at the left and right ears of the listener for a source location, source directivity, and source orientation for the listener at a particular location within the listening environment.
Aspects of the audio environment of described herein represents the playback of the audio or audio/visual content through appropriate speakers and playback devices, and may represent any environment in which a listener is experiencing playback of the captured content, such as a cinema, concert hall, outdoor theater, a home or room, listening booth, car, game console, headphone or headset system, public address (PA) system, or any other playback environment. Although embodiments have been described primarily with respect to examples and implementations in a home theater environment in which the spatial audio content is associated with television content, it should be noted that embodiments may also be implemented in environments. The spatial audio content comprising object-based audio and channel-based audio may be used in conjunction with any related content (associated audio, video, graphic, etc.), or it may constitute standalone audio content. The playback environment may be any appropriate listening environment from headphones or near field monitors to small or large rooms, cars, open air arenas, concert halls, and so on.
Aspects of the systems described herein may be implemented in an appropriate computer-based sound processing network environment for processing digital or digitized audio files. Portions of the adaptive audio system may include one or more networks that comprise any desired number of individual machines, including one or more routers (not shown) that serve to buffer and route the data transmitted among the computers. Such a network may be built on various different network protocols, and may be the Internet, a Wide Area Network (WAN), a Local Area Network (LAN), or any combination thereof. In an embodiment in which the network comprises the Internet, one or more machines may be configured to access the Internet through web browser programs.
One or more of the components, blocks, processes or other functional components may be implemented through a computer program that controls execution of a processor-based computing device of the system. It should also be noted that the various functions disclosed herein may be described using any number of combinations of hardware, firmware, and/or as data and/or instructions embodied in various machine-readable or computer-readable media, in terms of their behavioral, register transfer, logic component, and/or other characteristics. Computer-readable media in which such formatted data and/or instructions may be embodied include, but are not limited to, physical (non-transitory), non-volatile storage media in various forms, such as optical, magnetic or semiconductor storage media.
Unless the context clearly requires otherwise, throughout the description and the claims, the words “comprise,” “comprising,” and the like are to be construed in an inclusive sense as opposed to an exclusive or exhaustive sense; that is to say, in a sense of “including, but not limited to.” Words using the singular or plural number also include the plural or singular number respectively. Additionally, the words “herein,” “hereunder,” “above,” “below,” and words of similar import refer to this application as a whole and not to any particular portions of this application. When the word “or” is used in reference to a list of two or more items, that word covers all of the following interpretations of the word: any of the items in the list, all of the items in the list and any combination of the items in the list.
While one or more implementations have been described by way of example and in terms of the specific embodiments, it is to be understood that one or more implementations are not limited to the disclosed embodiments. To the contrary, it is intended to cover various modifications and similar arrangements as would be apparent to those skilled in the art. Therefore, the scope of the appended claims should be accorded the broadest interpretation so as to encompass all such modifications and similar arrangements.
This application claims the benefit of priority to U.S. Provisional Patent Application No. 61/696,056 filed on 31 Aug. 2012, hereby incorporated by reference in its entirety.
| Filing Document | Filing Date | Country | Kind |
|---|---|---|---|
| PCT/US2013/057052 | 8/28/2013 | WO | 00 |
| Publishing Document | Publishing Date | Country | Kind |
|---|---|---|---|
| WO2014/036121 | 3/6/2014 | WO | A |
| Number | Name | Date | Kind |
|---|---|---|---|
| 6839438 | Riegelsberger | Jan 2005 | B1 |
| 20050177256 | Shintani | Aug 2005 | A1 |
| 20070263888 | Melanson | Nov 2007 | A1 |
| 20100014692 | Schreiner | Jan 2010 | A1 |
| 20110150228 | Yoon | Jun 2011 | A1 |
| 20120183162 | Chabanne | Jul 2012 | A1 |
| 20140133683 | Robinson | May 2014 | A1 |
| Number | Date | Country |
|---|---|---|
| 2941692 | Apr 1981 | DE |
| 3201455 | Jul 1983 | DE |
| 1416769 | May 2004 | EP |
| 1971187 | Sep 2008 | EP |
| 60-079900 | May 1985 | JP |
| 06-153290 | May 1994 | JP |
| 2002-199487 | Jul 2002 | JP |
| 2007-288405 | Nov 2007 | JP |
| 2009-520419 | May 2009 | JP |
| 2010-258653 | Nov 2010 | JP |
| 2010-538572 | Dec 2010 | JP |
| 1332 | Aug 2013 | RS |
| 2007127781 | Nov 2007 | WO |
| 2010076850 | Jul 2010 | WO |
| Entry |
|---|
| Stanojevic, T. et al “The Total Surround Sound System”, 86th AES Convention, Hamburg, Mar. 7-10, 1989. |
| Stanojevic, T. et al “Designing of TSS Halls” 13th International Congress on Acoustics, Yugoslavia, 1989. |
| Stanojevic, T. et al “TSS System and Live Performance Sound” 88th AES Convention, Montreux, Mar. 13-16, 1990. |
| Stanojevic, Tomislav “3-D Sound in Future HDTV Projection Systems” presented at the 132nd SMPTE Technical Conference, Jacob K. Javits Convention Center, New York City, Oct. 13-17, 1990. |
| Stanojevic, T. “Some Technical Possibilities of Using the Total Surround Sound Concept in the Motion Picture Technology”, 133rd SMPTE Technical Conference and Equipment Exhibit, Los Angeles Convention Center, Los Angeles, California, Oct. 26-29, 1991. |
| Stanojevic, T. et al. “TSS Processor” 135th SMPTE Technical Conference, Oct. 29-Nov. 2, 1993, Los Angeles Convention Center, Los Angeles, California, Society of Motion Picture and Television Engineers. |
| Stanojevic, Tomislav, “Virtual Sound Sources in the Total Surround Sound System” Proc. 137th SMPTE Technical Conference and World Media Expo, Sep. 6-9, 1995, New Orleans Convention Center, New Orleans, Louisiana. |
| Stanojevic, T. et al “The Total Surround Sound (TSS) Processor” SMPTE Journal, Nov. 1994. |
| Stanojevic, Tomislav “Surround Sound for a New Generation of Theaters, Sound and Video Contractor” Dec. 20, 1995. |
| Brown, P. et al “A Structural Model for Binaural Sound Synthesis” IEEE Transactions on Speech and Audio Processing, vol. 6, No. 5, Sep. 1998, pp. 476-488. |
| Avendano, C. et al “A Head-and-Torso Model for Low-Frequency Binaural Elevation Effects” Proc. 1999 IEEE Workshop on Applications of Signal Processing to Audio and Acoustics, New Paltz, New York, Oct. 17-20, 1999, pp. 179-182. |
| Blauert, Jens “Spatial Hearing: The Psychophysics of Human Sound Localization” MIT Press, Dec. 1983. |
| Number | Date | Country | |
|---|---|---|---|
| 20150223002 A1 | Aug 2015 | US |
| Number | Date | Country | |
|---|---|---|---|
| 61696056 | Aug 2012 | US |
