Search

JP-7857392-B2 - Systems and methods for covariance smoothing

JP7857392B2JP 7857392 B2JP7857392 B2JP 7857392B2JP-7857392-B2

Inventors

  • マグラス，デイヴィッド エス．
  • ブラウン，ステファニー
  • トレス，ジュアン フェリックス

Assignees

  • ドルビー ラボラトリーズ ライセンシング コーポレイション

Dates

Publication Date
20260512
Application Date
20241227
Priority Date
20190801

Claims (12)

  1. A method of providing audio services, The steps include comparing the effective count of bins within a frequency band with the minimum count of bins for the frequency band for the time-domain sequence of the signal frame, The steps include: calculating a forgetting coefficient for the frequency band as the ratio of the effective count to the minimum count, wherein the forgetting coefficient is limited to the maximum allowable forgetting coefficient; For each frequency band of a plurality of frequency bands, the steps of generating each smoothed band-specific covariance matrix for the current signal frame across a window , using the forgetting coefficient and previously generated values of the respective smoothed band-specific covariance matrices for previous signal frames relative to the current signal frame; The steps include generating an audio signal for the audio service using a frequency-domain representation of the time-domain sequence of the signal frame and further using the corresponding set of the respective smoothed band-specific covariance matrices, The minimum count of bins for the frequency band corresponds to the minimum number of bins determined to give a good statistical estimate for the window, A method wherein the comparison step, the calculation step, and the generation step are performed by a system including one or more computer processors.
  2. The method according to claim 1, wherein the effective number of bins for each frequency band of the plurality of frequency bands is calculated as the sum of the filter bank response values for the frequency bands.
  3. The method according to claim 1, wherein the step of generating the respective smoothed band-specific covariance matrices is performed using a first-order autoregressive low-pass filter.
  4. The method according to claim 3, wherein the use of the first-order autoregressive low-pass filter comprises calculating the difference between the input covariance matrix for the current frame and the previously generated values of the respective smoothed band-specific covariance matrices for the previous frames, the difference being weighted by the forgetting coefficient.
  5. A step to detect whether or not a transient event occurred within the monitored frame, The method according to claim 3, further comprising the step of resetting a low-pass filter in response to a detected transient phenomenon.
  6. The method according to claim 5, wherein the detection step is performed by using an embodiment of a ducking inverse correlator.
  7. The method according to claim 6, wherein the detection step is performed on all channels.
  8. The method according to claim 7, wherein the detection step involves detecting that a transient phenomenon has occurred if any of the transient phenomena are detected in any channel for the monitored frame.
  9. The method according to claim 8, wherein the detection step detects the occurrence of a transient phenomenon only when the transient phenomenon occurs on a specific channel.
  10. The steps include storing the respective smoothed band-specific covariance matrices, The method according to claim 1, further comprising the step of resampling each of the smoothed band-specific covariance matrices stored over the plurality of frequency bands.
  11. One or more computer processors, A system comprising: a non-temporary computer-readable medium storing instructions that, when executed by the one or more computer processors, cause the one or more computer processors to perform the operation according to claim 1; and a system comprising:
  12. A non-temporary computer-readable medium that, when executed by one or more computer processors, stores instructions causing those one or more computer processors to perform the operation described in claim 1.

Description

[Cross-reference to related applications] This application claims priority by reference to U.S. Provisional Patent Application No. 62/881,825, filed on 1 August 2019, and U.S. Provisional Patent Application No. 63/057,533, filed on 28 July 2020. [Technical field] This disclosure relates to improvements for signal processing. In particular, this disclosure relates to processing audio signals to improve covariance smoothing for improved processing. One aspect of audio signal processing involves presenting multi-channel audio to a listener in a way that allows the listener to determine the virtual spatial location of the audio, thereby providing the listener with an immersive experience. Early implementations of this were stereo, where the listener could spatially determine the "direction" from which the sound source was coming. Recent developments in this technology utilize inter-channel dependencies in multi-channel systems to present a more immersive sound experience. This may involve the use of audio channel covariance matrices. Various signal processing systems and methods are disclosed herein. Some of these systems and methods may include smoothing the covariance values of bands across consecutive frames. In some examples, a system and method for smoothing an estimate of a covariance matrix for a sequence of signal frames within a frequency band includes the steps of: comparing the effective count of a bin within a frequency band with the count of a desired bin for the frequency band; calculating a forgetting coefficient for the band as the ratio of the effective count to the desired count; and, if the effective count of a bin within a frequency band is less than the desired count, generating a current estimate of the covariance matrix value for the current frame using the forgetting coefficient and a previously generated estimate of the covariance matrix value for a previous frame relative to the current frame. The comparing, calculating, and generating steps are performed by a system including one or more computer processors. The smoothed covariance matrix can be used to further improve signal processing by reducing artifacts caused by rough transitions in the matrix. In some such examples, the system and method may include calculating the number of effective bins as the sum of the filter bank response values for a frequency band. In some of these examples, the generation step uses a primary filter. In some of these examples, the first-order filter includes the difference between the value for the current frame and the previously generated estimate for previous frames, and this difference is weighted by a forgetting coefficient. In some such examples, the system and method also include the steps of comparing the forgetting coefficient to the maximum forgetting coefficient and setting the forgetting coefficient to the minimum of the calculated forgetting coefficient and the maximum allowable forgetting coefficient, the comparison step and the setting step being performed before generating the current estimate of the covariance matrix value for the current frame. In some such examples, the system and method also include the steps of detecting whether a transient has occurred within the monitored frame, and, in response to the detection of a transient, resetting the smoothing by setting the current estimate of the covariance matrix value for the monitored frame to the original value of the covariance matrix value for the monitored frame, without using a forgetting coefficient for the monitored frame. The detection step can be performed by using an embodiment of a ducking decorrelator. In some such examples, the system and method include the steps of storing a current estimate of the covariance matrix of the current frame, and resampling the stored current estimates across frequency bands to transform one banding structure into another. This system and method can be incorporated as part of an audio signal decoder. Some or all of the methods described herein may be executed by one or more devices according to instructions (e.g., software) stored on one or more non-temporary media. Such non-temporary media may include, but are not limited to, random access memory (RAM) devices, read-only memory (ROM) devices, and other memory devices as described herein. Therefore, various aspects of the inventions of the subject matter described herein may be implemented on non-temporary media storing software. The software may be executable, for example, by one or more components of a control system such as those disclosed herein. The software may include, for example, instructions for executing one or more of the methods disclosed herein. At least some aspects of this disclosure may be implemented via an apparatus or a set of apparatuses. For example, one or more devices may be configured to perform at least partially the methods disclosed herein. In some implementations, the apparatus may include an interface system and