Ripoff: Python tools for CD ripping

1. a highly configurable CD ripper written in Python and designed for high-throughput ripping
2. a collection of Python modules for the various tasks common to CD rippers: CD-ROM drive control, CDDB/CDIndex disc ID computations, CDDB lookup/caching, ID3 tagging, etc.

Currently, Ripoff is only available through CVS. Use the links on the right.

Principles

Extraction is the process of extracting digital audio from a CD. Ripoff relies on an external program, such as cdparanoia or cdda2wav for extraction. The output of audio extraction is very large uncompressed audio files (about 10 MB per minute of audio, or around 600-700 MB for a full CD).

Encoding is the process of encoding digital audio to a compressed, usually lossy, format, such as MP3 or Ogg Vorbis. Ripoff relies on an external program such as lame or bladeenc for encoding.

To maximize throughput, it's best if extraction and encoding run simultaneously, so this is what Ripoff tries to do by default.

The script

If you're an adventurous Linux/Unix user not afraid of checking code out of CVS and playing around, Ripoff might be useful for you -- especially if you're more interested in getting fiddly details (like filenames, ID3v2 tags, and the command-line arguments to your encoder of choice) "just right" than in a fancy GUI.

Ripoff is also very keen on maximizing throughput, ie. the number of CDs you can rip per hour. It has two basic modes of operation:

• synchronous: extraction and encoding are connected via a pipe, and start and stop simultaneously. This minimizes temporary disk usage (no enourmous .WAV files ever), and is the best way to go if encoding is faster than extraction. (This is probably the case on most modern CPUs, but it depends on the exact methods used for extraction and encoding.) Because the two processes are connected by a pipe, the encoder cannot "run ahead" of the extractor; it is I/O-blocked by the kernel when it tries to read (encode) data that hasn't been extracted yet.
• asynchronous: extraction writes to a regular file, which is then read by the encoder. In the worst case (one long track), this will require ~700 MB of temporary disk space. But it's the only way to go if you want to keep unencoded audio data around (eg. for burning to a CD, or to experiment with different encoders).

Ripoff's user interface is rather unusual. Like many Unix programs, it's controlled by command-line options and an editable configuration file; use

  ripoff --help

And like most Unix programs, Ripoff runs in a text console. But unlike most other Unix programs, Ripoff opens up other consoles where its "slaves" -- namely, audio extraction and encoding programs -- do all the hard work.

The rationale for this unusual interface is that the audio extraction and encoding tools I've played with under Linux have fairly noisy console output: they print a running progress report to their terminal. This is interesting and useful information, and I didn't want to throw it away or compress it to a little progress bar in a GUI. But have you ever tried to run, say, cdparanoia and lame simultaneously in the same window? It isn't pretty. I toyed for a while with running Ripoff in a big window, and doing some curses and pseudo-tty magic to make it all work. But then I remembered: that's what windowing systems are for! So now Ripoff just fires off a couple of terminal emulators (xterm, rxvt, gnome-terminal, whatever you like) and runs its slaves inside of them. It's weird, it's wonky, and it takes up a lot of screen real estate. But it works!

If you're a Python programmer looking to write your own CD ripper (and, after re-reading the above explanation, I'm not sure I blame you...), then Ripoff is most likely useful to you as a library. The ripoff package includes the following modules:

ripoff.cdrom	C extension for controlling a CD-ROM drive (useful for reading the TOC to compute disc IDs) (stolen from Ben Gertzfield)
ripoff.disc_id	code for computing CDDB and CDIndex disc IDs. CDDB code stolen from Ben Gertzfield, CDIndex code from Guido van Rossum.
ripoff.disc_info	provides the DiscInfo class, which abstracts the information provided by both CDDB and CDIndex. Based on code by Guido van Rossum.
ripoff.cddb	provides an implementation of the CDDB protocol (stolen from Guido van Rossum) and a local CDDB cache
ripoff.id3v1	ID3v1 (including v1.1) implementation (stolen from Ryan Finnie)
ripoff.id3v2	partial ID3v2 implementation (based on Ryan Finnie's code)

(The real reason it's called "ripoff" should be clear from the above list of credits.)