Voice2Text

Voice2Text is transcript media file to txt file to use Google Speach API &

Installation

Voice2Text need GOOGLE_APPLICATION_CREDENTIALS files. if you don't have this, please build google cloud projects and get from it.

Gcloud Project build

Google Cloud SDK Install

brew cask install google-cloud-sdk

Setting Gcloud Projects

gcloud auth login
gcloud alpha projects create voicetotext-123456 --name voice2text

Go to Projects URL and enable Google Speech API.
Please Enable (Billing)[https://support.google.com/cloud/answer/6293499?hl=en].
Create Service Key and Downlaod (Ref:Service Acount.)
set GOOGLE_APPLICATION_CREDENTIALS

export GOOGLE_APPLICATION_CREDENTIALS='/your/service/acount/key/xxx.json'

Install

pip install voicetotext

Usage

This application has two commands. splitvoice is convert the voice diving. voicetotext is voice existing in the folder into a text through google api. (See help command)

splitvoice --help
voicetotext --help

Sample

Split Audio Files

Sample Japanese voices from here

$ splitvoice voices/hana_1.mp3 --relative
spliting /57
spliting Done!
File was separete 57 filesOutput Separeted files? [Y/n]:y
separeted done! Have a nice Day!⏎

Transript Japanese audio files

$ voicetotext results/ -s 22050 -l "ja_JP"
芥川龍之介
花
line
朗読池田秀雄
禅智内供の鼻といえば池で知らないものはない
長澤語録すがって上唇の上から顎の下まで下がっている

Error Handling

"Sample rate in request does not match FLAC header."

You need to examine the sample rate. I recommend ffprove to examine.

$ ffmprove results/000.flac
Input #0, flac, from 'results/000.flac':
  Metadata:
    ENCODER         : Lavf57.56.101
  Duration: 00:00:01.87, start: 0.000000, bitrate: 184 kb/s
    Stream #0:0: Audio: flac, 22050 Hz, mono, s16

You can get framerate. In this case, frame rate is 22050. So, your commands is this.

$ voicetotext results -s 22050

Contributing

Fork it!
Create your feature branch: git checkout -b my-new-feature
Commit your changes: git commit -am 'Add some feature'
Push to the branch: git push origin my-new-feature
Submit a pull request :D

Debugging

# virtualenv
python3 -m venv env
source ./env/bin/activate

# python packages install
pip install -r requirements.txt

History

License

This software is released under the MIT License, see LICENSE.txt.

mkazutaka/voicetotext