README.md

GStreamer DeepSpeech Plugin

DeepSpeech is a speech recognition project created by Mozilla.

This project provides a GStreamer element which can be placed into an audio pipeline, it will then report any recognised speech via bus messages. It automatically segments audio based on configurable silence thresholds making it suitable for continuous dictation.

Here’s a couple of example pipelines using gst-launch.

To perform speech recognition on a file, printing all bus messages to the terminal:

gst-launch-1.0 -m filesrc location=/path/to/file.ogg ! decodebin ! audioconvert ! audiorate ! audioresample ! deepspeech ! fakesink

To perform speech recognition on audio recorded from the default system microphone, with changes to the silence thresholds:

gst-launch-1.0 -m pulsesrc ! audioconvert ! audiorate ! audioresample ! deepspeech silence-threshold=0.3 silence-length=20 ! fakesink

Installation

If you have downloaded DeepSpeech native client and not installed it in some standard location, you can use the --with-deepspeech ./configure option to point to where it is located, for example:

./configure --prefix=/home/fran/local --with-deepspeech=/home/fran/source/DeepSpeech/native_client

If you want to link to your own version of Tensorflow instead of the one bundled with DeepSpeech then you can use the --enable-tensorflow option to the ./configure script.