doem97 / audio_to_SRT

Generate .srt subtitle file automatically from audio file or video file. Features in Indian accent or Singalish recognition.

This tool generates .srt format subtitle from audios.

Usage

STEP 1 Clone this script to your local folder

apt-get -y update
apt-get -y install git
git clone https://github.com/doem97/audio_to_SRT.git
cd ./audio_to_SRT

STEP 2 Get XUNFEI cloud certification

This script calls XUNFEI api. You should first get a XUNFEI cloud account at https://console.xfyun.cn/services/lfasr. Go to 语音识别 - 语音转写 to get your APPID and SecretKey.

STEP 3 Get sound track from video

You can get .wav audio from video file using ffmpeg.

E.g. below shows how to get the sound track from sample.mp4 using sampling rate 16K.

ffmpeg -i sample.mp4 -ar 16000 audio.wav

*if you want to use other format rather than .wav such as .mp3 or .prm, you should change code block in get_srt.py

STEP 4 Put sound track

Put audio.wav file to ./work_dir/

STEP 5 Run and get subtitle

Run auto-environment build script config_env.sh and then run main script get_srt.py to get the srt. As follow:

chmod +x config_env.sh
./config_env.sh
source ./xunfei_venv/bin/activate
python get_srt.py APPID SECRET-KEY

in which the APPID and SECRET-KEY are gotten in STEP 2. E.g. python get_srt.py 1j3i3g fjiwgo23jio23t8932tj

and it will generate subtitle at ./work_dir/audio.srt