No, not quite.

What comes off the disc is fed to a decoder. Small blocks of data are used to represent data, called symbols. Symbols are easier to fix than pure binary, so you can read off the disk with high error rates and yet still get all your original data back.

The symbol decoder corrects any busted symbols where possible (note!) and then decodes them into pure binary data (the original data generated during the recording process in the studio). This lot then enters the output chain of your CD player, and what part of it is actually used to create sound is up to the designer of the output chain. In two examples I know of, one had the bottom three bits chopped off each sample for each stereo channel (it's not as simple as this, by the way ) due to tests that showed this would improve the quality of the output. The other used single bit output, where the sample streams were manipulated on the fly to improve noise figures (a very expensive DSP {for 1990 that is} was needed for that one, I can tell you). Note that if a symbol is irretrievably busted, this is where the player has to act smart by either guessing what the correct value would be (it makes up a value based on trends) or simply causing the output to mute before the erroneous value is presented to the DACs. You can see that all sorts of stuff can take place in your CD's output stage before you hear it. Each player is different, as well, 'cos playback strategies on damaged symbols are formulated differently by different companies.

The difference with the data used by a RIPper is that when you RIP using a CD-ROM drive, at the point where you actually get the binary channel samples, the MP3 encoder is given data with all of the bits of each sample available. They are unmodified, and can contain uncorrected samples with error markers. With a CD player, the output stage designer may choose to correct (described above), or even modify the data word size by choice while trying to avoid the effects of signal spectrum modification in the time domain (ie. after the signal has been reconstructed). In the early days this was to compensate for non-linearity in the DACs or to cover noise. With the MP3 encoder, who knows what choices the designers have made about the content of the data stream and their strategies on damaged symbol data?

One of the few remaining Mk1 owners... #00015