Develop Reference

How to convert wav with custom codec to mp3?

I've received a set of .wav files which are encoded with a custom codec. I wish to convert these files to .mp3 (or really any standard audio format that other software will recognize).
The codec has been provided to me as a .msi file, which allows me to install the codec in Windows. Running the installer creates a .acm file (in its own folder, not in the standard codec location which seems to be c:\windows\system32 ). Once I do so, Windows Media Player will play the audio in the files. No other software will play them (I've tried Audacity, VLC, QuickTime, ...).
I've tried to convert the files using ffmpeg, but it doesn't recognize the format either.
I've attempted the technique of burning the audio to a CD (with the intent of ripping the CD as mp3), but WMP gives an error when it attempts to burn them.
I'm proficient in Python and Javascript so I don't mind coding a solution if that's possible--I'm not really familiar with the theory of audio codecs. I looked at a couple python libraries but they didn't seem able to do anything with the .msi file I've been given.
Can anyone suggest a solution? Solutions might include:
How to point ffmpeg or Audacity or VLC to the installed codec so that it is able to play/convert the files?
Point me to some resources/libraries/code snippets for coding a converter based on the supplied .acm file?
How to convice WMP to save these files in some other format?

Detecting audio file integrity with python

If I download an audio file from the web and something bad happens to the download process, how does one efficiently detect that the audio file is incomplete with python?
There are some ideas, such as using the file command in linux:
file audio.mp4
But it recognizes that it's mp4:
audio.mp4: ISO Media, MPEG v4 system, version 2
Even mplayer detects the mp4 audio type, but fails when trying to play. I don't think launching mplayerfrom python and checking if it failed is a scalable solution though.
Here is a sample of broken file:
https://www.dropbox.com/s/5rpscb9r1xrrx4t/They
The sample above fails with mutagen and mp4file, causing them to hang indefinitely. It has to do with fileObject.tell().

There are many different audio file formats, and container formats for things that that may or may not be audio files.
Fortunately, there are libraries that can a wide variety of different kinds of files. And there are Python wrappers for:
Portable command-line tools like ffmpeg and mplayer.
Portable libraries like libavcodec (what ffmpeg uses).
Platform-specific libraries like Core Audio or QuickTime or Windows Media.
If you're willing to use separate wrappers for separate file types, there are even more choices (e.g., libmp4v2 is great for MP4 files, but useless for anything else).
Of course there are huge tradeoffs—the more powerful libraries are often going to be more complex, or have more prerequisites. Do some searching at http://pypi.python.org/ to see what turns up; you should be able to find something that does everything you want.
For one really simple example, mp4file will attempt to parse any MPEG4 container. If it's incomplete, or has any invalid atoms, you'll get an exception. So, the check is just one line, mp4file.Mp4File(path). If it succeeds, it's complete; if it throws an exception, it's incomplete or invalid. But of course this will accept a complete MPEG4 video file, or MPEG4 with no audio or video in it, and it will reject a complete MP3, or even a complete M4A with one broken metadata tag.

Best python solution to play ALL kinds of audio on Windows (and Linux)?

I'm trying to write some scripts to play some of my music collection with python. Finding python modules that will play ogg and mp3 is not a problem. However, I'm having repeated failures with aac-encoded m4a files from iTunes (not DRM). pygame's audio machinery doesn't support them, so I tried pymedia:
a = pymedia.player.Player()
a.start()
a.startPlayback("myM4a.m4a", format='aac')
I've tried several versions of the last line of code, including omitting the format argument, changing the files to mp4, etc. mp3's work fine, however.
pymedia even claims to support aac encoded files, but the project appears to have been abandoned anyway.
Is there a good, up to date, solution for playing ALL types of audio in python? What is used by existing python media centers/players?
I should add that I intend to use this primarily on windows, so windows support for the library is a must, but cross-platform would obviously be preferable.

You should look at the gStreamer API. It has plugins for many major audio types, is used by many audio players including Banshee and Rhythmbox and it can run on Linux, Windows and Mac. It has Python bindings as well as bindings for many other languages:
http://gstreamer.freedesktop.org/bindings/

MPlayer plays most known audio formats, and there's Python wrapper for it:
http://code.google.com/p/python-mplayer/
And a list of audio codecs supported by MPlayer:
http://www.mplayerhq.hu/DOCS/codecs-status.html#ac

Python: decode mp3

I need to decode an MP3 file with Python. What is a good library which can do that?
I looked at PyMedia but the project seems dead and it didn't worked on MacOSX. Then I found out about pyffmpeg but I didn't got it working on MacOSX so far.
Any suggestion?

I did try an easy_install of PyMedia on OS X / Fink, and it did not work because it could not find the source. This module does look quite dead…
One way of decoding MP3 is to call ffmpeg without going through pyffmpeg, but by calling ffmpeg using the standard subprocess module instead.

You really need an external library. It'd be very difficult to do in Python with any sort of speed - see How to convert MP3 to WAV in Python for some discussion.
How about python-mad? MAD being the 'mpeg audio decoder'; there's a python library. It'd give you the audio data. Never used it myself...

I decided to code this myself based on subprocess and ffmpeg.
Some code can be found here:
https://github.com/albertz/learn-midi/blob/master/decode.py

Please try https://github.com/sampsyo/audioread
It's fast, installs from pypi and works well

Python: Detect codecs used in a video container

I have a couple of video container files which contain audio and video in various codecs. Now I'd like to inspect the container from a Python script to know which codec is used for audio+video. This is on a linux box so I have all the tools available if necessary.
I thought that maybe gstreamer could help me here but I was unable to find an API which could help me here.
Any ideas? I'm also open for any suggestion, doesn't need to be gstreamer as long as it is free software :-)
fs

ffprobe -show_format -show_streams -loglevel quiet -print_format json YOUR_FILE
Just invoke this with subprocess.check_output and you'll get a beautiful JSON description of your media file. If you need it to grab the data from stdin, replace YOUR_FILE with pipe:0.
ffprobe comes with ffmpeg.

Try downloading the ffmpeg source and look at the source for their command line programs. I've hacked up similar utilities in the past. I'm not posting my solution because ffmpeg likes to change their API, so my old code is unlikely to compile with the current version. You'll want to do enough work to create codec context, which you can inspect to get what you need.
Some other alternatives:
MediaInfo: http://mediainfo.sourceforge.net/en
GSpot (Windows only): http://www.headbands.com/gspot/
EDIT:
http://code.google.com/p/pyffmpeg/ might have what you want (I haven't used it myself).

You can use decodebin2 in Gstreamer. Take a look at TAE for code examples.