Hi everyone,
I'm pretty sure this as already been discussed but I'm unable to find any thread about that. There's surely a keyword about that but I don't have it.

Do I really lose quality when recompressing ?
Of course I don't mean starting with a 128CBR and ending with 256CBR with the hope that the quality will be improved.

But what if I take a 320CBR sample and recompressing at alt-preset-standard (or 256CBR) ?

Is there really a difference between these two ?
01. original wav > MP3 alt preset standard
02. original wav > MP3 320 > MP3 alt preset standard

I know that 02. should be avoided, I can understand why, but in this case, will I get an audible difference between the 01 and 02 final mp3s ?

Any answers would be appreciated.

Yes. What you're doing is decompressing the mp3 to uncompressed PCM and then recompressing it to MP3, including all the existing artifacts. If you're asking whether it's noticable, that's a matter of opinion. I'm going to try some samples now.

Edit:
ABX removed, see below.

Thanks for your answer Defsac, but actually that was not the point.

You made :
01. Original > MP3 320CBR > APS
02. Original > MP3 320CBR
And thus compared two files of different bitrates.

Here's my point :
01. Original > MP3 320CBR > APS
03. Original > APS
The two files should have almost the same size and almost the same bitrate.

Maybe I wasn't clear before, I'm not a native english speaker.

http://www.hydrogenaudio.org/forums/index....ndpost&p=308932

the key word you are looking for is "transcode"

Sorry, that was a minunderstanding on my part. I'm currently testing another group of samples with the correct method.

Exactly ! Thanks.

The goal of the topic starter was to compress from 320k down to preset standard, which would be 200k.
IMo, such compression isn't efficient, the introduction of doubled mp3 artefacts (transcoding) vs. the little bitrate gain isn't worth.
If you would go from 320k or 256k down to 128-140k, then you would get additional transcoding artefacts, too, but if you weigh the bitrate gain in this case vs. the qualities of a wav->128-140k vs. wav->256-320k->128-140k encoding, you could accept the way of wav->256-320k->128-140k.
As listening isn't a linear procedure, (simple example: sound volume vs. required power isn't linear, it is exponential, see db scale) some small differences in bitrate don't change the quality of sound much, but if you double bitrates (always with same/similar optimized models, eg. psychoacoustics), you can perceive clear improvements of sound.
Though I wanna add, that 128-140k mp3 is only acceptable sound quality for portable, eg. in car, not that somebody thinks, 128k would be Hi-Fi..
preset standard from ca. 192-220k offers already good sound quality, and the image of this preset should not be deteriorated by transcodings with additional artefacts.

This is the corrected ABX (Sample A: APS, Sample B: Source > 320kbps CBR > APS).



Sample A
Sample B
Source

MP3 -> MP3 doesn't work very well.
http://www.hydrogenaudio.org/forums/index....showtopic=32440

Even from 320 to VBR at high bitrate.

Thanks for the replies.
In fact, I said 320 to APS so as to get the A case instead of the B case on this draw.

A draw is so much better than lines and lines of text.
Am I understood with this draw ?
I mean, I could have said 256CBR to 128CBR. The goal was to always and inevitably obtain a lower bitrate for each sample after transcoding.

If you go to half of bitrate, eg. from 256-320 down to 128, it will be acceptable, valuing the effort and the win (of filesize without too much quality loss compared to direct encode)
It is all relative.
As I wrote above, 128k is only acceptable for noisy environments, where quality does not count too much, like portable usage, cars, running etc.

-Will there be a difference? Yes.
-Will it be audible? Depends on what you can hear.

You might not be able to hear the difference, but there is a difference. #2 will always be of a lower quality than #1, but you may not be able to hear it.

The short answer is to always do #1 if possible and avoid #2 when you can. There are times when transcoding is perfectly acceptable, such as when you transcode to a lower bitrate for a portable player or what have you. When you're not looking for the absolute best quality, in other words.

Ok, thanks for your answers, that's right for me, all is clear and logical.