Pydub

- Fix a bug in db_to_float() where all values were off by a factor of 2

- Allow users to set the location of their converter by setting `pydub.AudioSegment.converter = "/path/to/ffmpeg"` and added a shim to support the old method of assigning to `pydub.AudioSegment.ffmpeg` (which is deprecated now that we support avconv)

- Add support for Python 3.4
- Audio files opened with format "wave" are treated as "wav" and "m4a" are treated as "mp4"
- Add `pydub.silence` module with simple utilities for detecting and removing silence.
- Fix a bug affecting auto-detection of ffmpeg/avconv on windows.
- Fix a bug that caused pydub to only work when ffmpeg/avconv is present (it should be able to work with WAV data without any dependencies)

- Add a runtime warning when ffmpeg/avconv cannot be found to aid debugging

- Added support for pypy (by reimplementing audioop in python). Also, we contributed our implementation to the pypy project, so that's 💯
- Add support for avconv as an alternative to ffmpeg
- Add a new helper module `pydub.playback` which allows you to quickly listen to an audio segment using ffplay (or avplay)
- Add new function `pydub.utils.mediainfo('/path/to/audio/file.ext')` which reports back the results of ffprobe (or avprobe) including codec, bitrate, channels, etc