Plugin idea: Silence finder about wavesurfer.js HOT 11 CLOSED

Hey Patrick,

Yes, great idea! I'll be glad to help should you decide to do it yourself.

Large files are a problem we couldn't solve (it seems decodeAudioData has limits on audio buffer size, and it's crashing when the file is over 60 MB). However, speech is easily compressed due to narrower frequency ranges, plus high quality usually isn't required.

from wavesurfer.js.

Cool! Is there any documentation on writing WaveSurfer plugins? Otherwise I guess I can just try to model it from the interface in the ELAN plugin. My Javascript skills are nowhere near as good as yours, but I want to try!

Have you looked into the “streaming” capability of createMediaElementSource for the possibility of addressing the file size issues you mention?

In my line of work we do use huge .wav files, often (gulp) half a gig, because sometimes we do acoustic analysis on them—we use nice mics, etc.

Incidentally, have you seen this blog post? He mentions WaveSurfer and claims to have come up with a way to handle long files, but I can’t seem to find the Github link he mentions.

from wavesurfer.js.

ELAN "plugin" is just a number of functions in WaveSurfer namespace, it's not integrated in any way with the main program. Whereas you'll have to access WaveSurfer instance to analyze the peaks array and mark silent sections, so yours will be the first real plugin. Just pass a WaveSurfer instance as a parameter, and we'll figure out how to enable plugins by names via main init options later.

Chunking is a way to go, but it will be 1) slow (the browser will have to download the half gig of chunks anyway), 2) you'll have to prepare the chunks somehow (outside wavesurfer). I think the better similar option is to setup a server and decode audio/find peaks there, then just download peaks and use Audio element/createMediaElementSource for playing (SoundCloud's waveform.js does exactly that).

I'll see this weekend how Aurora.js decodes MP3. They must somehow preprocess audio stream before playing it (I don't think they can be fast enough to decode in realtime). If I'm right, we could use that to draw waveforms from audio streams.

from wavesurfer.js.

I've started working on this for a transcription tool: http://www.wavesurfer.fm/example/annotation/

from wavesurfer.js.

kats, you don't even know how grateful I am for that demo. :') I was designing a music annotator with wavesurfer.js and fabric.js and, seriously, that demo will help me to organize some ideas. 💃

Thanks!!!

from wavesurfer.js.

@echo66, you're welcome, man!

from wavesurfer.js.

As I mentioned before I’m a big fan of this plugin too!

It seems that the audio isn’t currently loading?

from wavesurfer.js.

Eh? Loading for me.

from wavesurfer.js.

+1 for the Silence Finder

from wavesurfer.js.

+1 for the Silence Finder

Anyone made an extension for silence finding?

from wavesurfer.js.

Anyone made an extension for silence finding?

Doesn't the demo annotation plugin do this? Demo.
Relevant source: http://wavesurfer-js.org/example/annotation/app.js

from wavesurfer.js.