GithubHelp home page GithubHelp logo

Comments (8)

chanil1218 avatar chanil1218 commented on June 2, 2024
  • CHIME3(유료) 대체하는 Unseen noise dataset 공유 @chanil1218

from attention-se.pytorch.

chanil1218 avatar chanil1218 commented on June 2, 2024

Noisy dataset make reference(matlab version) - https://github.com/jtkim-kaist/Speech-enhancement/blob/master/Datamake/make_train_noisy.m
(python version) https://github.com/Sato-Kunihiko/audio-SNR

from attention-se.pytorch.

chanil1218 avatar chanil1218 commented on June 2, 2024

Python version noisy speech generation + DEMAND Dataset for Noise only unseen dataset candidate Tutorial https://engineering.linecorp.com/ko/blog/voice-waveform-arbitrary-signal-to-noise-ratio-python/

from attention-se.pytorch.

chanil1218 avatar chanil1218 commented on June 2, 2024

데이터셋 로딩할 때에 noisy speech와 clean speech 까지만 모델로 전달하고 STFT 적용해서 spectrogram으로 바꾸는 파트는 DCUnet.pytorch 구현의 경우에는 model layer로 들어가 있어요.
ISTFT도 model layer로 들어가있고요.

CC: @rownrrns

from attention-se.pytorch.

ddw02141 avatar ddw02141 commented on June 2, 2024

다른 분들이 진행시에 데이터 파일이 가장 먼저 필요할 것 같아서 필요한 데이터셋 먼저 공유해 드립니다

https://drive.google.com/file/d/1R76jEq-TYBK_VeUIjcHI1g2izpH9ek4E/view?usp=sharing (9.8 GB)

clean speech로 부터 noise를 mix하여 noisy 파일을 만들 때 사용한 snr 값은 아래와 같습니다

  • train, valid, test, test2 : 0 ~ 20 dB

  • test3 : -5 ~ 0 dB

다른 궁금한 점 있으시면 알려주세요 !

from attention-se.pytorch.

chanil1218 avatar chanil1218 commented on June 2, 2024

헉 용량;

이거 Noisy speech를 만들어서 올린거지?
그러면 Unseen noise에 대해서 Noisy Speech 만들 때에는 다시 만들어서 올려야 할 것 같은데,
Clean audio랑 Noise audio를 가지고 코드에서 만드는게 낫지 않으려나?

그리고 Train / Valid / Test 를 분리해서 각각 압축 파일로 만들어놔야 할 것 같아.
안그럼 테스트 할 때에도 Train셋까지 전체를 다운 받아야 해서.

from attention-se.pytorch.

ddw02141 avatar ddw02141 commented on June 2, 2024

코드에서 만들었고 해당 코드는 저희 프로젝트 내용에는 추가할 필요가 없는 것 같아서 업로드하지는 않았습니다! 필요하다면 업로드 할게요

from attention-se.pytorch.

ddw02141 avatar ddw02141 commented on June 2, 2024

train : https://drive.google.com/open?id=1CULVCAq0T3wqZTPGIqPja6OtwjYJkGAy (4.9GB)
valid : https://drive.google.com/open?id=1WE229Jt9WV2iZbxY7YjkYIfSZyCHz9Iq (1.6GB)
test : https://drive.google.com/open?id=1mERBbcwBgGjRpHeGf3d871DEW4EvI7fi (3.3GB)
gdrive 저장 용량이 부족해 기존 링크는 파기하였습니다

from attention-se.pytorch.

Related Issues (5)

Recommend Projects

  • React photo React

    A declarative, efficient, and flexible JavaScript library for building user interfaces.

  • Vue.js photo Vue.js

    🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.

  • Typescript photo Typescript

    TypeScript is a superset of JavaScript that compiles to clean JavaScript output.

  • TensorFlow photo TensorFlow

    An Open Source Machine Learning Framework for Everyone

  • Django photo Django

    The Web framework for perfectionists with deadlines.

  • D3 photo D3

    Bring data to life with SVG, Canvas and HTML. 📊📈🎉

Recommend Topics

  • javascript

    JavaScript (JS) is a lightweight interpreted programming language with first-class functions.

  • web

    Some thing interesting about web. New door for the world.

  • server

    A server is a program made to process requests and deliver data to clients.

  • Machine learning

    Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.

  • Game

    Some thing interesting about game, make everyone happy.

Recommend Org

  • Facebook photo Facebook

    We are working to build community through open source technology. NB: members must have two-factor auth.

  • Microsoft photo Microsoft

    Open source projects and samples from Microsoft.

  • Google photo Google

    Google ❤️ Open Source for everyone.

  • D3 photo D3

    Data-Driven Documents codes.