search

Kaljurand / dictate.js

A small Javascript library for browser-based real-time speech recognition, which uses Recorderjs for audio capture, and a WebSocket connection to the Kaldi GStreamer server for speech recognition.
https://kaljurand.github.io/dictate.js/
BSD 3-Clause "New" or "Revised" License
216 stars 62 forks source link

Using resampler redundant? #16

Closed psukys closed 6 years ago

psukys commented 6 years ago

As observed recorderWorker.js [1] mainly does resampling, yet gstreamer server does it as well [2]. Is there a specific reason to do so, if yes - why?

[1] - https://github.com/Kaljurand/dictate.js/blob/master/lib/recorderWorker.js#L2 [2] - https://github.com/alumae/kaldi-gstreamer-server/blob/master/kaldigstserver/decoder.py#L44

Kaljurand commented 6 years ago

@alumae can you comment on this?

alumae commented 6 years ago

JavaScript does this in order to save bandwidth.

psukys commented 6 years ago

Understood, thanks for the answer.