Ahhh the magic. The hours I have put forth on this subject.
Here is the skinny, surprisingly - it has not been addressed.
In a computer - a disk is read 010101001 blah blah blah. if it comes to an error- it stops says HEY I HAVE AN ERROR - BAD FILE, it might mark the file bad, or if its between the files in the "other " area between files - it says - disk bad.
In comes Red book audio, wearing its oversampling hat, looking all snazzy. It says I will keep on --keeping on even when I see a bump. I will have enough data from surrounding bits to either skip or guess (depending on quality of drive /firmware) the information that was there. If we were talking about a spreadsheet / database you would have lost data - and the file would be unreadable - you might run a tool on it to "recover" the file which would be guessing what was cutout -or adding a EOF .....
SINCE we are talking about music, specifically a series group of 0 and 1 sampled at VERY specific rate who is to say missing a 1 or a 0 is audible when you are listening to 200,000 of them per minute- through a speaker (either ribbon or cone) that is vibrating/moving as fast as it can. If we start missing 100 bits - or a 1000 bits - eventually it becomes noticable, on cheep speakers - they dont move fast enough for you to hear the difference - so there is none. If you take a CD recorded at 1400bits per second - and re-encode it at 128 bits per second - you have HACKED a whole lot of data out, good speakers - mean bigger diffence, played thru the pc speaker = no difference.
I hope that simplifes the Data extracton piece in relation to audible notes.
More magic, black magic
Given a perfect world, perfect cd recording, perfect drive, perfect temperature - once the file is extracted - it will be perfect reproduced from file to file to file, till you write it to an imperfect medium, going down to a floppy, and or CDrom - introduced medium flaws back into the loop, depending on the mood of your drive - you could have different results on the actual recording (and hence play back) - if you want to test this, use a cheep cd rom drive, rip a wav or other high quality file to a cd, take a skrew driver and run it over a different factory cd several times and really mar the surface, put the disk in the drive and extract the cd using ultra secure mode - let it whail for 2 hours or so, stop the process, insert a new blank cd and immediatly re-record the high quality file. play back of both disks will show the stress on the burner causing imperfect writes to the disk, if you were to re-rip the disks and compare the files - you would see a difference in file size. Better CDrom drives - we have all heard of plextor - will handle the abuse better. I have burned out cheep drives doing data extraction, and tested the process over and over. This type of stress happens to all the audio equipment once it heats up - or its pusshed hard - the quality changes. again you get what you pay for.
Once you have a Wav or cda file - its a pure as it can get (from that cd/drive combination). These files can move over a network - from hd to HD with out loosing a beat, the medium is not tolerent of errors (it doesnt use the redbook standard) . Hence - you can zipit - unzip it, ape it - unape it , mp3 it - ......wait a minute - there is no - un mp3 - once you hack it its gone..... and do a binary compare of the files is EXACTLY the same. As far as diffences heard - on a PC - there could be lots of things causing the problem - not the file manipulation. Once we discover latiencey we see that cpu speed, process running, HD performance, sound card quality, codecs, DSP all effect the speed and timing of the play back ( and the audio standard of keep on - keeping on comes back) ape files, expecially ones using the highest compressions settings on a slower computer (less then 2ghz) with not enough ram (less then 128meg), running lots of process (virus scan sw, Instant messinger,...) does cause a quality performance issue, sometime is causes pauses or pops, sometimes is subtle - like it just dont sound right. I have notice this on more faster complex music, lots of instruments playing at once, and have played slower beat less complex music and not seen the problem.
Ok - so I am a novelist wanna be, I will leave further discussions of how SW handle music and "interperts" it for a later thread.
hope this helps
Phatanhappy
20years in computers, 4 working for a "golden ear" audio file