Gaudio Lab, Inc.

Gaudi Audio Lab, Inc.

Sangbae Chon

Jeonghun Seo

Hyunoh Oh

Taegyu Lee

Hyunjoo Chung

Kyutae Park

Sewoon Jeon

Younghoon Kwon

Yonghyun Baek

Jaesung Choi

Byoungjoon Ahn

Jiwon Oh

Cheolwoo Jeong

Soochul Park

Sangmoon Lee

Kyungyun Lee

Disclosed is an audio signal processing device that estimates a room impulse response (RIR) filter of a space. The audio signal processing device may include a processor. The processor receives an input signal, and estimates, from the received input signal by using a machine learning model, an RIR filter representing a characteristic of a response transferred to a listener from a sound source located in a predetermined location in a space where the input signal is collected, wherein the predetermined location is a relative location determined based on the listener, wherein the machine learning model is trained to estimate a reference RIR filter from signals response associated with a reference RIR filter corresponding to the predetermined location and a plurality of additional RIR filters that represent characteristics of responses transferred to the listener from sound sources located in a space represented by the reference RIR filter.

Audio signal processing device generating room impulse response filter using machine learning model

An audio signal processing device comprises: a receiver for receiving an input audio signal; a processor for generating loudness metadata corresponding to the input audio signal; and an outputter for transmitting the loudness metadata generated by the processor. The processor is configured to acquire loudness information analyzed from input content, acquires loudness information about the input audio signal by measuring the loudness of the input audio signal, generates the loudness metadata by converting the loudness information, and transmits, through the outputter, the generated loudness metadata to an output device for outputting the input audio signal.

Audio signal processing method and device for controlling loudness level

A method for performing loudness normalization by an audio signal processing device includes: receiving an audio signal; receiving information relating to integrated loudness of the audio signal; receiving information relating to target loudness of the audio signal; obtaining a corrected integrated loudness by correcting the integrated loudness based on one or more audio processing; and obtaining a normalized audio signal by normalizing the audio signal based on the corrected integrated loudness and the target loudness.

Method for performing normalization of audio signal and apparatus therefor

Disclosed is a device for processing an audio signal, which renders an audio signal. The device for processing an audio signal includes a processor. The processor receives metadata including an audio signal and first element reference distance information and renders a first element signal on the basis of the first element reference distance information, wherein the first element reference distance information indicates the reference distance of an element signal. The audio signal is capable of including a second element signal which may be simultaneously rendered with the first element signal, and the metadata is capable of including second element distance information indicating the distance of the second element signal. The number of bits required for representing the first element reference distance information is smaller than the number of bits required for representing the second element distance information.

Method and device for processing audio signal, using metadata

Disclosed is an audio signal processing method including: receiving a stereo signal; transforming the stereo signal into a frequency-domain signal; rendering the first signal based on a first ipsilateral filter coefficient; generating a frontal ipsilateral signal relating to the frequency-domain signal; rendering the second signal based on a second ipsilateral filter coefficient; generating a side ipsilateral signal relating to the frequency-domain signal; rendering the second signal based on a contralateral filter coefficient; generating a side contralateral signal relating to the frequency-domain signal; transforming an ipsilateral signal, generated by mixing the frontal ipsilateral signal and the side ipsilateral signal, and the side contralateral signal into a time-domain ipsilateral signal and a time-domain contralateral signal, which are time-domain signals, respectively; and generating a binaural signal by mixing the time-domain ipsilateral signal and the time-domain contralateral signal.

Method for generating binaural signals from stereo signals using upmixing binauralization, and apparatus therefor

Disclosed is an operation method of an audio signal processing device configured to process an audio signal including a first audio signal component and a second audio signal component. The operation method includes: receiving the audio signal; normalizing loudness of the audio signal, based on a pre-designated target loudness; acquiring the first audio signal component from the audio signal having the normalized loudness, by using a machine learning model; and de-normalizing loudness of the first audio signal component, based on the pre-designated target loudness.

Audio signal processing method and apparatus

Disclosed is a method of operating an audio signal processing apparatus playing content including an audio signal. The method includes receiving the audio signal, receiving metadata including information related to a loudness of the audio signal, the metadata including loudness distribution information indicating, for each of a plurality of steps separated according to a loudness magnitude, a ratio between an amount of the audio signal corresponding to each of the plurality of steps of the audio signal and a total amount of the audio signal, and adjusting the loudness of the audio signal based on the metadata.

Audio signal processing method and apparatus for controlling loudness level and dynamic range

An audio signal processing apparatus is provided. The audio signal processing apparatus includes an input terminal receiving an input audio signal, a processor obtaining a difference between a playback loudness of the input audio signal and a desired loudness thereof and correcting a frequency band spectrum of an output audio signal for each of a plurality of frequency bands based on the difference between the playback loudness and the desired loudness of the input audio signal and a relationship between a loudness and a sound pressure for each of the plurality of frequency bands, and an output terminal outputting the output audio signal. The playback loudness is a loudness of the output audio signal when the input audio signal is output without the correction.

Audio signal processing method and apparatus for frequency spectrum correction

A method of generating audio loudness performed by an audio loudness generation device may include: receiving loudness information on each of a plurality of audio tracks included in one group; predicting an intermediate loudness distribution, which is a loudness distribution for the one group, on the basis of the loudness information on each of the plurality of audio tracks; and generating an integrated loudness for the one group on the basis of the intermediate loudness distribution.

Method for generating audio loudness metadata and device therefor

Disclosed is an audio signal processing apparatus for rendering an input audio signal. The audio signal processing apparatus may include a processor configured to obtain an input audio signal including an ambisonics signal and a non-diegetic channel difference signal, render the ambisonics signal to generate a first output audio signal, mix the first output audio signal and the non-diegetic channel difference signal to generate a second output audio signal, and output the second output audio signal.

Growing community of inventors

Seoul, South Korea

Sangbae Chon