How to detect silence in audio files?

I'm working on a tool to edit .srt (subtitle) files within the browser (the tool is to be used for linguistic annotation). In desktop tools that are used for similar purposes, the user has access to the waveform, and can "see" where silences are in the signal, and thus select a particular phrase for transcr开发者_Python百科iption.

Such a tool might be buildable in-browser down the road (using Web Workers and Canvas, say), but for now, it's not feasible to do the sort of signal processing it would take to find those silences.

So, I'm thinking about the next-best approach: what free tool could I use to produce a list of timestamps of where silences (below some given threshold) start and stop? If I produce such a list offline and upload it with the audio file, then I can at least make it possible to navigate through the "phrases" (defined as periods of non-silence). I think that would still be a win for in productivity for doing the transcription.

Audacity can sort of doing this, but AFAICT, only if you install Nyquist, which seems to have some patent issues.

Are there any alternatives?

It would be nice if the tool could handle as many as possible of ogg, mp3, and wav files.