Patents by Inventor Zoran Fejzo

Zoran Fejzo has filed for patents to protect the following inventions. This listing includes patent applications that are pending as well as patents that have already been granted by the United States Patent and Trademark Office (USPTO).

  • Publication number: 20230197091
    Abstract: A method comprises receiving input audio and target audio having a target audio characteristic. The method includes estimating key parameters that represent the target audio characteristic based on one or more of the target audio and the input audio. The method further comprises configuring a neural network, trained to be configured by the key parameters, with the key parameters to cause the neural network to perform a signal transformation of the input audio, to produce output audio having an output audio characteristic corresponding to and that matches the target audio characteristic.
    Type: Application
    Filed: January 31, 2023
    Publication date: June 22, 2023
    Inventors: Atti Venkatraman, Zoran Fejzo, Antonius Kalker
  • Publication number: 20230186926
    Abstract: A method comprise: receiving input audio and target audio having a target audio characteristic; using a first neural network, trained to generate key parameters that represent the target audio characteristic based on one or more of the target audio and the input audio, generating the key parameters; and configuring a second neural network, trained to be configured by the key parameters, with the key parameters to cause the second neural network to perform a signal transformation of the input audio, to produce output audio having an output audio characteristic corresponding to and that matches the target audio characteristic.
    Type: Application
    Filed: January 31, 2023
    Publication date: June 15, 2023
    Inventors: Zoran Fejzo, Antonius Kalker, Atti Venkatraman
  • Patent number: 11380340
    Abstract: A frequency domain long-term prediction system and method for estimating and applying an optimum long term predictor. Embodiments of the system and method include determining parameters of a single-tap predictor using a frequency-domain analysis having an optimality criteria based on spectral flatness measure. Embodiments of the system and method also include determining parameters of the long-term predictor by accounting for the performance of the vector quantizer in quantizing the various subbands. In some embodiments other encoder metrics (such as signal tonality) are used as well. Other embodiments of the system and method include determining the optimal parameters of the long-term predictor by accounting for some of the decoder operation. Other embodiments of the system and method include extending a 1-tap predictor to a k-th order predictor by convolving the 1-tap predictor with a pre-set filter and selecting from a table of such pre-set filters based on a minimum energy criteria.
    Type: Grant
    Filed: September 8, 2017
    Date of Patent: July 5, 2022
    Assignee: DTS, Inc.
    Inventors: Elias Nemer, Jacek Stachurski, Zoran Fejzo, Antonius Kalker
  • Patent number: 11132984
    Abstract: There are disclosed automatic mixers and methods for creating a surround audio mix. A set of rules may be stored in a rule base. A rule engine may select a subset of the set of rules based, at least in part, on metadata associated with a plurality of stems. A mixing matrix may mix the plurality of stems in accordance with the selected subset of rules to provide three or more output channels.
    Type: Grant
    Filed: May 1, 2017
    Date of Patent: September 28, 2021
    Assignee: DTS, Inc.
    Inventors: Zoran Fejzo, Fred Maher
  • Patent number: 10950251
    Abstract: Systems and methods include audio encoders having improved coding of harmonic signals. The audio encoders can be implemented as transform-based codecs with frequency coefficients quantized using spectral weights. The frequency coefficients can be quantized by use of the generated spectral weights applied to the frequency coefficients prior to the quantization or by use of the generated spectral weights in computation of error within a vector quantization that performs the quantization. Additional apparatus, systems, and methods are disclosed.
    Type: Grant
    Filed: November 7, 2018
    Date of Patent: March 16, 2021
    Assignee: DTS, Inc.
    Inventors: Elias Nemer, Zoran Fejzo
  • Patent number: 10616587
    Abstract: Systems and methods are described for processing data from a sequential series of groups of frames to achieve a target average processing bit rate for a particular group of frames in the series. In an example, a look-ahead buffer circuit can be populated with a number of frames from a particular group of frames, and a bit allocation can be determined for a frame in the look-ahead buffer circuit using bit request information about all of the frames in the buffer. The look-ahead buffer circuit can be populated with streaming frame information in a first-in-first-out manner, and bit allocation processing can be performed for each frame, in a particular group of frames, based on a frame position in the look-ahead buffer circuit and further based on bit requests associated with other frames in the look-ahead buffer circuit.
    Type: Grant
    Filed: April 26, 2018
    Date of Patent: April 7, 2020
    Assignee: DTS, Inc.
    Inventors: Jacek Stachurski, Zoran Fejzo
  • Publication number: 20190272837
    Abstract: Systems and methods include audio encoders having improved coding of harmonic signals. The audio encoders can be implemented as transform-based codecs with frequency coefficients quantized using spectral weights. The frequency coefficients can be quantized by use of the generated spectral weights applied to the frequency coefficients prior to the quantization or by use of the generated spectral weights in computation of error within a vector quantization that performs the quantization. Additional apparatus, systems, and methods are disclosed.
    Type: Application
    Filed: November 7, 2018
    Publication date: September 5, 2019
    Inventors: Elias Nemer, Zoran Fejzo
  • Patent number: 10251016
    Abstract: Systems, devices, and methods are described herein for adjusting a relationship between dialog and non-dialog signals in an audio program. In an example, information about a long-term dialog balance for an audio program can be received. The long-term loudness dialog balance can indicate a dialog-to-non-dialog loudness relationship of the audio program. A dialog loudness preference can be received, such as from a user, from a database, or from another source. A desired long-term gain or attenuation can be determined according to a difference between the received long-term dialog balance for the audio program and the received dialog balance preference. The long-term gain or attenuation can be applied to at least one of the dialog signal and the non-dialog signal of the audio program to render an audio program that is enhanced according to the loudness preference.
    Type: Grant
    Filed: October 27, 2016
    Date of Patent: April 2, 2019
    Assignee: DTS, Inc.
    Inventors: Jean-Marc Jot, Brandon Smith, Jeffrey K. Thompson, Zoran Fejzo
  • Publication number: 20180316923
    Abstract: Systems and methods are described for processing data from a sequential series of groups of frames to achieve a target average processing bit rate for a particular group of frames in the series. In an example, a look-ahead buffer circuit can be populated with a number of frames from a particular group of frames, and a bit allocation can be determined for a frame in the look-ahead buffer circuit using bit request information about all of the frames in the buffer. The look-ahead buffer circuit can be populated with streaming frame information in a first-in-first-out manner, and bit allocation processing can be performed for each frame, in a particular group of frames, based on a frame position in the look-ahead buffer circuit and further based on bit requests associated with other frames in the look-ahead buffer circuit.
    Type: Application
    Filed: April 26, 2018
    Publication date: November 1, 2018
    Inventors: Jacek Stachurski, Zoran Fejzo
  • Patent number: 9984692
    Abstract: A post-encoding bitrate reduction system and method for generating one more scaled compressed bitstreams from a single encoded plenary file. The plenary file contains multiple audio object files that were encoded separately using a scalable encoding process having fine-grained scalability. Activity in the data frames of the encoded audio object files at a time period are compared with each other to obtain a data frame activity comparison. Bits from an available bitpool are assigned to all of the data frames based on the data frame activity comparison and corresponding hierarchical metadata. The plenary file is scaled down by truncating bits in the data frames to conform to the bit allocation. In some embodiments frame activity is compared to a silence threshold and the data frame contains silence if the frame activity is less than or equal to the threshold and minimal bits are used to represent the silent frame.
    Type: Grant
    Filed: December 15, 2015
    Date of Patent: May 29, 2018
    Assignee: DTS, INC.
    Inventor: Zoran Fejzo
  • Publication number: 20180075855
    Abstract: A frequency domain long-term prediction system and method for estimating and applying an optimum long term predictor. Embodiments of the system and method include determining parameters of a single-tap predictor using a frequency-domain analysis having an optimality criteria based on spectral flatness measure. Embodiments of the system and method also include determining parameters of the long-term predictor by accounting for the performance of the vector quantizer in quantizing the various subbands. In some embodiments other encoder metrics (such as signal tonality) are used as well. Other embodiments of the system and method include determining the optimal parameters of the long-term predictor by accounting for some of the decoder operation. Other embodiments of the system and method include extending a 1-tap predictor to a k-th order predictor by convolving the 1-tap predictor with a pre-set filter and selecting from a table of such pre-set filters based on a minimum energy criteria.
    Type: Application
    Filed: September 8, 2017
    Publication date: March 15, 2018
    Applicant: DTS, Inc.
    Inventors: Elias Nemer, Jacek Stachurski, Zoran Fejzo, Antonius Kalker
  • Publication number: 20170301330
    Abstract: There are disclosed automatic mixers and methods for creating a surround audio mix. A set of rules may be stored in a rule base. A rule engine may select a subset of the set of rules based, at least in part, on metadata associated with a plurality of stems. A mixing matrix may mix the plurality of stems in accordance with the selected subset of rules to provide three or more output channels.
    Type: Application
    Filed: May 1, 2017
    Publication date: October 19, 2017
    Applicant: DTS, Inc.
    Inventors: Zoran Fejzo, Fred Maher
  • Publication number: 20170127212
    Abstract: Systems, devices, and methods are described herein for adjusting a relationship between dialog and non-dialog signals in an audio program. In an example, information about a long-term dialog balance for an audio program can be received. The long-term loudness dialog balance can indicate a dialog-to-non-dialog loudness relationship of the audio program. A dialog loudness preference can be received, such as from a user, from a database, or from another source. A desired long-term gain or attenuation can be determined according to a difference between the received long-term dialog balance for the audio program and the received dialog balance preference. The long-term gain or attenuation can be applied to at least one of the dialog signal and the non-dialog signal of the audio program to render an audio program that is enhanced according to the loudness preference.
    Type: Application
    Filed: October 27, 2016
    Publication date: May 4, 2017
    Inventors: Jean-Marc Jot, Brandon Smith, Jeffrey K. Thompson, Zoran Fejzo
  • Patent number: 9641952
    Abstract: Devices and methods are adapted to characterize a multi-channel loudspeaker configuration, to correct loudspeaker/room delay, gain and frequency response or to configure sub-band domain correction filters.
    Type: Grant
    Filed: April 20, 2015
    Date of Patent: May 2, 2017
    Assignee: DTS, Inc.
    Inventors: Zoran Fejzo, James D. Johnston
  • Patent number: 9640163
    Abstract: There are disclosed automatic mixers and methods for creating a surround audio mix. A set of rules may be stored in a rule base. A rule engine may select a subset of the set of rules based, at least in part, on metadata associated with a plurality of stems. A mixing matrix may mix the plurality of stems in accordance with the selected subset of rules to provide three or more output channels.
    Type: Grant
    Filed: March 12, 2014
    Date of Patent: May 2, 2017
    Assignee: DTS, Inc.
    Inventors: Zoran Fejzo, Fred Maher
  • Patent number: 9564136
    Abstract: A post-encoding bitrate reduction system and method for generating one more scaled compressed bitstreams from a single encoded plenary file. The plenary file contains multiple audio object files that were encoded separately using a scalable encoding process having fine-grained scalability. Activity in the data frames of the encoded audio object files at a time period are compared with each other to obtain a data frame activity comparison. Bits from an available bitpool are assigned to all of the data frames based on the data frame activity comparison and corresponding hierarchical metadata. The plenary file is scaled down by truncating bits in the data frames to conform to the bit allocation. In some embodiments frame activity is compared to a silence threshold and the data frame contains silence if the frame activity is less than or equal to the threshold and minimal bits are used to represent the silent frame.
    Type: Grant
    Filed: March 6, 2014
    Date of Patent: February 7, 2017
    Assignee: DTS, Inc.
    Inventor: Zoran Fejzo
  • Patent number: 9552819
    Abstract: A multiplet-based spatial matrixing codec and method for reducing channel counts (and thus bitrates) of high-channel count (seven or more channels) multichannel audio, optimizing audio quality by enabling tradeoffs between spatial accuracy and basic audio quality, and converting audio signal formats to playback environment configurations. An initial N channel count is reduced to M channels by spatial matrix mixing to a lower number of channels using multiplet pan laws. The multiplet pan laws include doublet, triplet, and quadruplet pan laws. For example, using a quadruplet pan law one of the N channels can be downmixed to four of the M channels to create a quadruplet channel. Spatial information as well and audio content is contained in the multiplet channels. During upmixing the downmixed channel is extracted from the multiplet channels using the corresponding multiplet pan law. The extracted channel then is rendered at any location within a playback environment.
    Type: Grant
    Filed: November 26, 2014
    Date of Patent: January 24, 2017
    Assignee: DTS, Inc.
    Inventors: Jeffrey Kenneth Thompson, Zoran Fejzo
  • Patent number: 9530421
    Abstract: The present invention provides a novel end-to-end solution for creating, encoding, transmitting, decoding and reproducing spatial audio soundtracks. The provided soundtrack encoding format is compatible with legacy surround-sound encoding formats, so that soundtracks encoded in the new format may be decoded and reproduced on legacy playback equipment with no loss of quality compared to legacy formats.
    Type: Grant
    Filed: March 15, 2012
    Date of Patent: December 27, 2016
    Assignee: DTS, Inc.
    Inventors: Jean-Marc Jot, Zoran Fejzo
  • Publication number: 20160099000
    Abstract: A post-encoding bitrate reduction system and method for generating one more scaled compressed bitstreams from a single encoded plenary file. The plenary file contains multiple audio object files that were encoded separately using a scalable encoding process having fine-grained scalability. Activity in the data frames of the encoded audio object files at a time period are compared with each other to obtain a data frame activity comparison. Bits from an available bitpool are assigned to all of the data frames based on the data frame activity comparison and corresponding hierarchical metadata. The plenary file is scaled down by truncating bits in the data frames to conform to the bit allocation. In some embodiments frame activity is compared to a silence threshold and the data frame contains silence if the frame activity is less than or equal to the threshold and minimal bits are used to represent the silent frame.
    Type: Application
    Filed: December 15, 2015
    Publication date: April 7, 2016
    Applicant: DTS, Inc .
    Inventor: Zoran Fejzo
  • Patent number: 9253574
    Abstract: There is disclosed methods and apparatus for decomposing a signal having a plurality of channels into direct and diffuse components. The correlation coefficient between each pair of signals from the plurality of signals may be estimated. A linear system of equations relating the estimated correlation coefficients and direct energy fractions of each of the plurality of channels may be constructed. The linear system may be solved to estimate the direct energy fractions. A direct component output signal and a diffuse component output signal may be generated based in part on the direct energy fractions.
    Type: Grant
    Filed: September 12, 2012
    Date of Patent: February 2, 2016
    Assignee: DTS, Inc.
    Inventors: Jeff Thompson, Brandon Smith, Aaron Warner, Zoran Fejzo, Jean-Marc Jot