probs with mp3s - EAC does not produce standard wave files?
Hey guys... I'm yet another one out there on the quest for perfect digital copies etc =D. After finally getting my read/write offsets sorted, just about everything is okay.
When I <USE EAC> to rip a wave from a pressed CD, then take this wave, record it to a blank CDR in DAO mode, extract the wave off the CDR, then do a bit-for-bit comparism, everything is fine. NO differences what-so-ever. However :
When I use winamp to decode an mp3 to a .WAV file (yes im turning EQ off), Write this wave to a CDR, extract it off the CDR and then do a bit-for-bit comparism, the freshly extracted file is ALWAYS about 0.006 seconds longer ???
Just to keep things systematic, I took the wave that I extracted off the CDR, burnt it to ANOTHER CDR, then from the second CDR I ripped it again. did a bit-for-bit compare, and low and behold, the files are identical. The only expanation is this :
EAC and Winamp must be storing the wave file <slightly> differently? perhaps some different header information? Because the .WAV from winamp is always just a fraction shorter than the one from EAC. I have also used other software (Mp3 To Wave) to do the task, thinking perhaps winamp was at fault. Same Problem.
Any ideas anyone (Andrae especially?)
When you convert to mp3 you are compressing and loosing information. When you decompress again to wav you will never
get the same wav you began from.
Endless discussions are about different encoders and settings
to get the best results.
Using big bitrates you will get mp3 that you can reconvert to wav
and you will not be able to ear the difference with the original, but the difference exists and the PC can read it.
When you compress to mp3 and when you decompress to wav
the algorithm of the programs introduce little differences in the
lenght of the song.
When you burn a wav to CDDA the lenght of the file must fill
an integral number of frames (this means that the lenght must be multiple of 1/75 seconds). If it is not, the burning program
(or the burner?) will complete the missing samples
Ahhh thanks for that... I think I understand now.
So basically what it boils down to, is that .WAV files written to CDDA format must have length which is multiple of 1/75 of a second, which is the smallest unit (1 'sample' if you will) ?
And extra samples are added (presumably, 'null' samples?) to fix the length?
Is there a way to force Winamp etc to decode its mp3s to waves which have a length which is a multiple of 1/75 second?
I don't think you can decode with Winamp to a wav multiple of 1/75 sec, because for sure the mp3 you are coming from it isn't.
There is a program called CDWAV that split wav's in an integer number of frames and I think you can correct with this program the lenght. Is easier than to use a wav editor.
There is a good discussion of why mp3 files, when decoded back to wav files have a different length.
Read about it here if you are interested
In addition, although you can split mp3 files (using something like mp3 cutter) there is often a 'glitch' introduced as a result... it varies depending on the cut point. On one mp3 file where I knew the exact break point (from EAC .cue sheet) it sounded perfect but other times the results have not been good.