DeepSpeech-examples/vad_transcriber

Transcribing longer audio clips

The Command and GUI tools perform transcription on long wav files. They take in a wav file of any duration, use the WebRTC Voice Activity Detector (VAD) to split it into smaller chunks and finally save a consolidated transcript.

0. Prerequisites

0.1 Install requiered packages

Install the package which contains rec on the machine:

Fedora:

sudo dnf install sox

Tested on: 29

Ubuntu/Debian

sudo apt install sox

A list of distributions where the package is available can be found at: https://pkgs.org/download/sox

0.1 Download Deepspeech

Either clone from git via git clone, or Download a version from the release page

For the next steps we assume you have extracted the files to ~/Deepspeech

0.2 Setup your environment

Ubuntu/Debian:

~/Deepspeech$ sudo apt install virtualenv
~/Deepspeech$ cd examples/vad_transcriber
~/Deepspeech/examples/vad_transcriber$ virtualenv -p python3 venv
~/Deepspeech/examples/vad_transcriber$ source venv/bin/activate
(venv) ~/Deepspeech/examples/vad_transcriber$ pip3 install -r requirements.txt

Fedora

~/Deepspeech$ sudo dnf install python-virtualenv
~/Deepspeech$ cd examples/vad_transcriber
~/Deepspeech/examples/vad_transcriber$ virtualenv -p python3 venv
~/Deepspeech/examples/vad_transcriber$ source venv/bin/activate
(venv) ~/Deepspeech/examples/vad_transcriber$ pip3 install -r requirements.txt

Tested on: 29

1. Command line tool

The command line tool processes a wav file of any duration and returns a trancript which will the saved in the same directory as the input audio file.

The command line tool gives you control over the aggressiveness of the VAD. Set the aggressiveness mode, to an integer between 0 and 3. 0 being the least aggressive about filtering out non-speech, 3 is the most aggressive.

(venv) ~/Deepspeech/examples/vad_transcriber
$ python3 audioTranscript_cmd.py --aggressive 1 --audio ./audio/guido-van-rossum.wav --model ./models/0.4.1/


Filename                       Duration(s)          Inference Time(s)    Model Load Time(s)   Scorer Load Time(s)
sample_rec.wav                 13.710               20.797               5.593                17.742

Note: Only wav files with a 16kHz sample rate are supported for now, you can convert your files to the appropriate format with ffmpeg if available on your system.

ffmpeg -i infile.mp3  -ar 16000 -ac 1  outfile.wav

2. Minimalistic GUI

The GUI tool does the same job as the CLI tool. The VAD is fixed at an aggressiveness of 1. The output is displayed in the transcription window and saved into the directory as the input audio file as well.

(venv) ~/Deepspeech/examples/vad_transcriber
$ python3 audioTranscript_gui.py

2.1. Sporadic failures in pyqt

Some systems have encountered Cannot mix incompatible Qt library with this with this library issue. In such a scenario, the GUI tool will not work. The following steps is known to have solved the issue in most cases

(venv) ~/Deepspeech/examples/vad_transcriber$ pip3 uninstall pyqt5
(venv) ~/Deepspeech/examples/vad_transcriber$ sudo apt install python3-pyqt5 canberra-gtk-module
(venv) ~/Deepspeech/examples/vad_transcriber$ export PYTHONPATH=/usr/lib/python3/dist-packages/
(venv) ~/Deepspeech/examples/vad_transcriber$ python3 audioTranscript_gui.py

2.2 Useful Tips

The GUI programm immediately crashes when you press start recording

This happens when you don't load the models via the "Browse Models" button, before pressing the "Start recording" button.

What does error XYZ mean?

You can find a list of error codes and what they mean at https://mozilla-voice-stt.readthedocs.io/en/latest/Error-Codes.html