公开(公告)号：US20210366497A1

公开(公告)日：2021-11-25

申请号：US17326035

申请日：2021-05-20

Applicant: Electronics and Telecommunications Research Institute ,  The Trustees of Indiana University

Inventor： Woo-taek LIM ,  Seung Kwon BEACK ,  Jongmo SUNG ,  Mi Suk LEE ,  Tae Jin LEE ,  Inseon JANG ,  Minje KIM ,  Haici YANG

IPC: G10L19/032

Abstract: Methods of encoding and decoding a speech signal using a neural network model that recognizes sound sources, and encoding and decoding apparatuses for performing the methods are provided. A method of encoding a speech signal includes identifying an input signal for a plurality of sound sources; generating a latent signal by encoding the input signal; obtaining a plurality of sound source signals by separating the latent signal for each of the plurality of sound sources; determining a number of bits used for quantization of each of the plurality of sound source signals according to a type of each of the plurality of sound sources; quantizing each of the plurality of sound source signals based on the determined number of bits; and generating a bitstream by combining the plurality of quantized sound source signals.

公开(公告)号：US20210233547A1

公开(公告)日：2021-07-29

申请号：US17156006

申请日：2021-01-22

Applicant: Electronics and Telecommunications Research Institute ,  The Trustees of Indiana University

Inventor： Mi Suk LEE ,  Seung Kwon BEACK ,  Jongmo SUNG ,  Tae Jin LEE ,  Jin Soo CHOI ,  Minje KIM ,  Kai ZHEN

IPC: G10L19/038 ,  G10L25/18 ,  G10L25/30 ,  G10L25/21 ,  G10L19/028 ,  G06N3/08

Abstract: A method and apparatus for processing an audio signal are disclosed. According to an example embodiment, a method of processing an audio signal may include acquiring a final audio signal for an initial audio signal using a plurality of neural network models generating output audio signals by encoding and decoding input audio signals, calculating a difference between the initial audio signal and the final audio signal in a time domain, converting the initial audio signal and the final audio signal into Mel-spectra, calculating a difference between the Mel-spectra of the initial audio signal and the final audio signal in a frequency domain, training the plurality of neural network models based on results calculated in the time domain and the frequency domain, and generating a new final audio signal distinguished from the final audio signal from the initial audio signal using the trained neural network models.

公开(公告)号：US20210142812A1

公开(公告)日：2021-05-13

申请号：US17098090

申请日：2020-11-13

Applicant: Electronics and Telecommunications Research Institute ,  The Trustees of Indiana University

Inventor： Minje KIM ,  Kai ZHEN ,  Mi Suk LEE ,  Seung Kwon BEACK ,  Jongmo SUNG ,  Tae Jin LEE ,  Jin Soo CHOI

IPC: G10L19/08 ,  G10L19/032 ,  G10L19/26 ,  G10L21/0208 ,  G10L25/30 ,  G10L13/02 ,  G06N3/08

Abstract: Disclosed are a method for coding a residual signal of LPC coefficients based on collaborative quantization and a computing device for performing the method. The residual signal coding method includes: generating encoded LPC coefficients and LPC residual signals by performing LPC analysis and quantization on an input speech; Determining a predicted LPC residual signal by applying the LPC residual signal to cross module residual learning; Performing LPC synthesis using the coded LPC coefficients and the predicted LPC residual signal; It may include the step of determining an output speech that is a synthesized output according to a result of performing the LPC synthesis.

公开(公告)号：US20210005209A1

公开(公告)日：2021-01-07

申请号：US16814103

申请日：2020-03-10

Applicant: Electronics and Telecommunications Research Institute ,  Kwangwoon University Industry-Academic Collaboration Foundation

Inventor： Seung Kwon BEACK ,  Jongmo SUNG ,  Mi Suk LEE ,  Tae Jin LEE ,  Hochong PARK

IPC: G10L19/02 ,  G10L19/032 ,  G10L21/038 ,  G06N3/04

Abstract: Disclosed are a method of encoding a high band of an audio, a method of decoding a high band of an audio, and an encoder and a decoder for performing the methods. The method of decoding a high band of an audio, the method performed by a decoder, includes identifying a parameter extracted through a first neural network, identifying side information extracted through a second neural network, and restoring a high band of an audio by applying the parameter and the side information to a third neural network.

公开(公告)号：US20190035412A1

公开(公告)日：2019-01-31

申请号：US16081169

申请日：2017-03-21

Applicant: Electronics And Telecommunications Research Institute

Inventor： Seung Kwon BEACK ,  Tae Jin LEE ,  Jongmo SUNG ,  Mi Suk LEE ,  Dae Young JANG ,  Jin Soo CHOI

IPC: G10L19/022 ,  G10L19/02 ,  G10L19/032 ,  G10L19/04

Abstract: Provided is an apparatus and method for encoding/decoding audio based on a block. A method of encoding an audio signal may include dividing each of frame of input signal that constitute an audio signal into a plurality of subframes; transforming the subframes to a frequency domain; determining a two-dimensional (2D) intra block using the subframes transformed to the frequency domain; and encoding the 2D intra block. The 2D intra block may be a block that two-dimensionally displays frequency coefficients of the subframes transformed to the frequency domain using a time and a frequency.

公开(公告)号：US20170076735A1

公开(公告)日：2017-03-16

申请号：US15260717

申请日：2016-09-09

Applicant: Electronics and Telecommunications Research Institute

Inventor： Seung Kwon BEACK ,  Tae Jin LEE ,  Jong Mo SUNG ,  Kyu Tae YANG ,  Bong Ho LEE ,  Mi Suk LEE ,  Hyoung Soo LIM ,  Jin Soo CHOI

IPC: G10L19/22 ,  G10L19/005 ,  G10L19/16 ,  G10L19/008

CPC classification number: G10L19/167 ,  G10L19/18

Abstract: Disclosed is a unified speech and audio coding (USAC) audio signal encoding/decoding apparatus and method for digital radio services. An audio signal encoding method may include receiving an audio signal, determining a coding method for the received audio signal, encoding the audio signal based on the determined coding method, and configuring, as an audio superframe of a fixed size, an audio stream generated as a result of encoding the audio signal, wherein the coding method may include a first coding method associated with extended high-efficiency advanced audio coding (xHE-AAC) and a second coding method associated with existing advanced audio coding (AAC).

Abstract translation: 公开了用于数字无线电业务的统一语音和音频编码（USAC）音频信号编码/解码装置和方法。 音频信号编码方法可以包括接收音频信号，确定所接收的音频信号的编码方法，根据所确定的编码方法对音频信号进行编码，并且将作为固定大小的音频超帧配置为以 编码音频信号的结果，其中编码方法可以包括与扩展的高效率高级音频编码（xHE-AAC）相关联的第一编码方法和与现有高级音频编码（AAC）相关联的第二编码方法。

公开(公告)号：US20160150182A1

公开(公告)日：2016-05-26

申请号：US14951005

申请日：2015-11-24

Applicant: Electronics and Telecommunications Research Institute

Inventor： In Ki HWANG ,  Mi Suk LEE

IPC: H04N7/14 ,  G06T3/00 ,  H04N13/00 ,  G06T7/00

CPC classification number: H04N7/144 ,  H04N13/111 ,  H04N13/128 ,  H04N13/239 ,  H04N13/271 ,  H04N13/383

Abstract: The present invention relates to a new eye-contact function providing method which provides a natural eye-contact function to attendances by using a stereo image and a depth image to estimate a precise depth value of the occlusion region and improve a quality of a composite eye-contact image when there are two or more remote attendances in one site at the time of a video conference using a video conference system and an apparatus therefor.

Abstract translation: 本发明涉及一种新的眼睛接触功能提供方法，其通过使用立体图像和深度图像来估计接近区域的精确深度值并提高复合眼睛的质量，为参加者提供自然的眼睛接触功能 当使用视频会议系统的视频会议系统及其装置在一个站点中存在两个或更多个远程出席者时，联系图像。