pyspeech (python) - Transcribe mp3 files?

Question

I'd like to transcribe mp3 (speech-to-text) using the pyspeech API. I don't know if this is possible, though.

Is it? How?

Preset

Answer 1

pyspeech seems to be merely a python interface to the regular Windows speech APIs. Most likely you'd create some method of treating mp3 playback as an audio source for that speech API to listen to.

Korney

Could you please explain how to do this in fuller detail? – Preset

Also, I have a Mac. Will this prevent me from using pyspeech? – Preset

You'd have to run Windows on your mac. EIther directly on the hardware or in a virtual machine (Parallels, VMWare, etc..). MacOS doesn't support Windows APIs (not should it). As for how to do it, I don't know. I'm not a Win32 programmer. But since the Windows apis are intended to "listen" to an audio source, you'd have to programatically turn your mp3 into an audio device that the APIs would listen to. – Korney

Answer 2

I don't know about pyspeech, but if it is a Python wrapper around the Microsoft speech APIs, then some other posts may be helpful.

Microsoft Speech engines do not require microphone input. They can accept audio files.

If you are doing transcription, you will need a dictation grammar. Dictation grammars are included with the client versions of Microsoft speech engines that ship with Windows 7 and Vista. Dictation grammars are not provided on the Server engines Microsoft provides.

A simple example of using a dictation grammar in C# with the System.Speech namespace is in the StackOverflow question SAPI and Windows 7 Problem

Almira