HydrogenAudio

Lossy Audio Compression => MP3 => MP3 - Tech => Topic started by: Benjamin Lebsanft on 2006-04-02 14:39:25

Title: Mp3 Decoders
Post by: Benjamin Lebsanft on 2006-04-02 14:39:25
I just stubled upon a post on computerbase.de in which a forum user claims the MP3 decoders are just good for mp3s encoded with the corresponding encoder e.g. lame for lame encoded files or fraunhofer for fraunhofer encoded files. If you mix them it's bad. Is this true or are all mp3 decoders decoding the bitstream exactly the same?
Title: Mp3 Decoders
Post by: guruboolez on 2006-04-02 14:48:37
MP3 is a standard. A compliant decoder shouldn't advantage any specific encoded file (lame-lame, fhg-fhg, etc...). A good decoder (i.e. with advanced dithering and noise shaping) would be good for all MP3 files, whatever the original encoder name.
Moreover, there are so many Fhg encoders and decoders that an ideal combination would be hard to find. I guess this person is inventing stories about audio compression (such as joint stereo spoil stereo image, VBR is making quality variable, ISO decoders have a better tonal purity, etc...). If doubts, ask for ABX results.


EDIT: LAME - LAME is nevertheless good to prevent encoder&decoder offset.
Bitstream aren't necessary the same between two decoders, but difference should only concern LSB (excepted rare case, like old and famous Nitrate decoder maybe).
Title: Mp3 Decoders
Post by: Benjamin Lebsanft on 2006-04-02 14:57:23
Quote
MP3 is a standard. A compliant decoder shouldn't advantage any specific encoded file (lame-lame, fhg-fhg, etc...). A good decoder (i.e. with advanced dithering and noise shaping) would be good for all MP3 files, whatever the original encoder name.
Moreover, there are so many Fhg encoders and decoders that an ideal combination would be hard to find. I guess this person is inventing stories about audio compression (such as joint stereo spoil stereo image, VBR is making quality variable, ISO decoders have a better tonal purity, etc...). If doubts, ask for ABX results.


EDIT: LAME - LAME is nevertheless good to prevent encoder&decoder offset.
Bitstream aren't necessary the same between two decoders, but difference should only concern LSB (excepted rare case, like old and famous Nitrate decoder maybe).
[a href="index.php?act=findpost&pid=378176"][{POST_SNAPBACK}][/a]

Would be nice if you could go a bit into detail here as I want to be the winner on the argumentation side
Title: Mp3 Decoders
Post by: guruboolez on 2006-04-02 15:02:06
Does it mean that he have arguments to defend his idea?
Title: Mp3 Decoders
Post by: Benjamin Lebsanft on 2006-04-02 15:04:33
Quote
Does it mean that he have arguments to defend his idea?
[a href="index.php?act=findpost&pid=378180"][{POST_SNAPBACK}][/a]

no, thats not the problem  But I don't want anyone there to believe him.
Title: Mp3 Decoders
Post by: guruboolez on 2006-04-02 15:24:11
Then ask him to defend his idea. Tell him that decoders don't embedd post-processing filters which could lower th audibility of artefact produced by a typical encoder.

Or use the same kind of plausible fantasie with exaggeration to show people how theorys could quickly be invented:
A MP3 sound better if you can play it on the same computer used for the encoder.
Always use the same OS for encoding and for decoding.
A MP3 sound better if you play it on the same hygrometric conditions than those encountered during the encoding stage. Subtle variation in humidity are changing the sound quality.
Always defragment your MP3 to avoid subtle lack in harmonics.
Title: Mp3 Decoders
Post by: smack on 2006-04-03 08:47:53
Quote
I just stubled upon a post on computerbase.de in which a forum user claims the MP3 decoders are just good for mp3s encoded with the corresponding encoder e.g. lame for lame encoded files or fraunhofer for fraunhofer encoded files. If you mix them it's bad. Is this true or are all mp3 decoders decoding the bitstream exactly the same?
[a href="index.php?act=findpost&pid=378174"][{POST_SNAPBACK}][/a]

Well, you can't prove that something like that does not exist. A believer of some "MP3 conspiracy theory" could always claim that there exists a certain version of a decoder that applies some "magical post-processing" to mp3 files that were encoded using the encoder made by the "evil empire". 
Title: Mp3 Decoders
Post by: Andavari on 2006-04-03 17:37:44
Quote
Or use the same kind of plausible fantasie with exaggeration to show people how theorys could quickly be invented
[a href="index.php?act=findpost&pid=378186"][{POST_SNAPBACK}][/a]

LMAO, thanks for giving me the first laugh of the day.
...and of course if the mp3 was encoded on a rainy day, it should logically be decoded on a rainy day as well.