Michael Matz explains how audioCD ripping works:

Audiocd:/ creates a virtual view (or better multiple views) of the content of audio CDs. These multiple views are collected in virtual directories which all (besides one exception) are subdirectories of the root directory. The actual set of views available to the user depends on how kio_audiocd was compiled, and if a network connection was possible (for the CDDB request). Most of these views consist of one file for each audio track. The difference inthe views is only the actual name of these files, and the content, when those (up to then virtual) files are read:

1. There is one view listing the tracks by number in "By Track/"; the files are named "Track XX.wav" with XX being a number. When read they are normal .wav files.
2. One view lists the track by track-name; in "By Name/". The files are called "XX .wav", where 'name' is the track title as it comes from CDDB. The content of these files are identical to those from (1), just the name is different. This view only exists if a network connection exists for CDDB lookups.
3. Another view has the same content as (2) (same name and content of the .wav files) but is located in the directory "", where ToD is the Title of the Disc coming from CDDB. If there was no CDDB query this view nevertheless exists, in which case the title is "No Title" and the track names are "Track XX.wav" (so the content is identical to (1)).
4. If libmp3lame was available when compiling kio_audiocd, there is a view in the directory "MP3/" (screenshot). The names of the files there are the same as used by (3), except the extension is ".mp3" instead of ".wav". When read, those files are normal mp3 files (they are coded on the fly, while ripping the track/reading the file).
5. If libvorbis was available when compiling kio_audiocd, there is a view in the directory "Ogg Vorbis/" (screenshot). The names are the same as in (3), except now the extension is ".ogg". When read those files are Vorbis Ogg files (created the same way as (4).
6. The root directory itself has a view. The files are in the form "trackXX.cda" (screenshot), which contain the raw headerless data of this track (the exact same info as the .wav files, without the 44 Bytes header).
7. The are two other directories which should be ignored for now: "Information" and "dev").

It might be good to know, that some of the strings above are i18n()'d (internationalized). These include the "Track XX", "By Name", "By Track", "No Title" and "Information" strings. The files in the "MP3", "Ogg Vorbis" and "dev" directories and the "trackXX.cda" files are not i18n()'d.

As should now be clear, the existance of certain directories depends on compile time environment and runtime environment and is not changeable by the user (except the CDDB information depends on a network connection). The files in most subdirectories are virtual, because they don't exist, similar to e.g. the pop3 slave which also looks as if it provides "files" although these are messages on some mail-server. But the virtuality makes no difference, as no user of audiocd:/ could tell if this or that particular URL now is a file, or not, which of course is true for all URLs, not only the audiocd:/ ones).

When a URL from one of the "MP3" or "Ogg Vorbis" directories is dropped, its content is copied to the destination. For (1)-(3) this results in .WAV files, for (4) and (5) in some compressed audio files (corresponding to .MP3 and .OGG, respectively), which additionally have encoded the CD-title, author and trackname if CDDB was available at the time (those file formats provide a means to encode such auxilliary information).





<  |  >