reencoding cds burnt from mp3s to mp3

Message Bookmarked
Bookmark Removed
how bad an idea is this? will it sound terrible? will i even notice?

toby (tsg20), Friday, 14 October 2005 11:52 (twenty years ago)

if you encode them to the same rate they were before you burnt them to CD, there is nothing left to lose, data wise, so they would sound the same - or am i completely wrong here?

Sociah T Azzahole (blueski), Friday, 14 October 2005 12:13 (twenty years ago)

I think you may be wrong, Steve.

You're essentially copying a copy - there is loss from original to copy, and there is further loss from copy of original to copy of copy.

The ripping programme does not know that it is not ripping from a full-audio source rather than a burnt MP3, so presumably it will compress it by as much again.

Sick Mouthy (Nick Southall), Friday, 14 October 2005 12:15 (twenty years ago)

Nick's got it -- basically, since mp3 is lossy, it's not reversible. mp3 gets decoded when burned to a cd so some audio data is just tossed out. When you reencode, you're once again turning a lossless audio source into a lossy one, so you're going to lose even more data.

mike h. (mike h.), Friday, 14 October 2005 12:21 (twenty years ago)

mp3 gets decoded when burned to a cd so some audio data is just tossed out.

I thought this only happened during the encoding process, not the burning process.

I wish it was like GIF and not JPEG.

Sociah T Azzahole (blueski), Friday, 14 October 2005 12:24 (twenty years ago)

I have done this and didn't notice any difference.

>>mp3 gets decoded when burned to a cd so some audio data is just tossed out.

I'm not sure this is true - why would audio data be lost when converting from mp3 to cda/wav? That doesn't make sense.

Colonel Poo (Colonel Poo), Friday, 14 October 2005 12:28 (twenty years ago)

AFAIK, data is only discarded when encoding to MP3. Decoding from MP3 uses the MP3 data to reconstruct an approximation of the original signal. If the reconstruction is subsequently encoded, more data would be thrown away. Presumably not as much as originally, as a lot of the psychoacoustic compression (i.e., removing subjectively imperceptible sounds) would have been done already. Though if there were artefacts introduced by the original MP3 encoding process, these would build up.

I have heard that after a few generations of MP3 encoding, one cannot tell the difference. Though I haven't seen any experiments performed to objectively quantify the amount of data lost or the subjective perception of sound quality.

acb (acb), Friday, 14 October 2005 13:43 (twenty years ago)

Someone should do a Steve Reich style loop where a piece of music is encoded then re-encoded until it's a piece of mush.

Billy Dods (Billy Dods), Friday, 14 October 2005 13:51 (twenty years ago)

i've definitely read that doing this like 100 times causes notceable sound quality loss - so

I have done this and didn't notice any difference.

is reassuring.

some results of a google search i just did (but haven't read yet):

http://www.hydrogenaudio.org/forums/index.php?act=ST&f=1&t=6410&hl=reencoding&s=


http://www.hydrogenaudio.org/forums/index.php?showtopic=20625

toby (tsg20), Friday, 14 October 2005 14:27 (twenty years ago)

I've done it, yes in theory you loose quality, but in reality it isn't really discernable.

kyle (akmonday), Friday, 14 October 2005 14:34 (twenty years ago)

Maybe if you do it 100 times, but I was just talking about doing it once! I used to burn compilation CDRs of mp3s from Napster, Audiogalaxy etc, then when I got an iPod I ripped them all back to mp3 so I could put them on that. Sound quality is the same. Of course that's listening on an iPod, which isn't exactly the best medium for audiophiles.

Colonel Poo (Colonel Poo), Friday, 14 October 2005 14:38 (twenty years ago)

One important thing: raw sound (ie eg .wav) -> mp3 is *not* in itself a well-defined transformation (which the opposite is) (wav->mp3 depends on selected psychoacoustic filter, bitrate etc); consequently, settings etc of the wav->mp3 machinery determine a lot.

More simply in practice: it's actually a bit unpredictable innit, but for a one-time transformation it'll probably be fine if the wav/cd thing sounds fine.

The Vintner's Lipogram (OleM), Friday, 14 October 2005 22:56 (twenty years ago)


You must be logged in to post. Please either login here, or if you are not registered, you may register here.