Joint Stereo vs Stero [quality]

[aliumalik]

I have searched google for this but can't seem to find a final conclusion while some people say that Joint Stereo is better in quality others say it completely copies the other channel. Can anyone tell me whether it offers better or worse quality than stereo.
Thanks a lot!

[stephanV]

there is not one final answer to that i think. it will depend on the bitrate your using and codec as well i think (i know MP3 has JS but i dont know about vorbis, etc.)

[Slogra]

I'll give it a shot.

There are 2 types of joint-stereo:
1) intensity stereo
This is used in mpeg layer 2. This kind of stereo will actually join parts of the sound. The lower sounds will be joint as it's difficult to hear where lower frequencies come from (that's why you can put a subwoofer anywhere yoou want). BUT some information is lost forever. This is especially noticable with Dolby Prologic(2) encoded stuff. So it's not recommended to use intensity stereo with mp2.

2) Mid/side stereo. This is used in mp3s and most other formats, even lossless formats. It's kinda hard to explain how it works but i'll try anyway.

For example: you have a 128kbps mp3.
Normal stereo will have to distribute the 128kbps on each channel. So in most cases it'll give each channel 64kbps. Sometime a channel might get 96kbps and the other 32kbps if the first channel is a lot more difficult to encode.
Now imagine that the music is mono, but it's encoded to stereo nevertheless. Each channel will sound just the same so normal stereo will give each channel 64kbps.
This is not very efficent, it would be better if it gave one channel 128kbps and simply play this through both speakers. This is exactly what mid/side stereo does. You will gain a double bitrate compared to the 64kbps in normal stereo!
When the sound is not totally mono but only slighty different, the side channel channel will store what is different. So the Mid channel might get 96kbps and the side channel 32kbps.

In other words: Mid/side is much more efficient with mono-ish stuff. And stereo is better in stereo-ish stuff.
An mp3 can contain both joint-stereo and stereo frames. So what a good mp3 encoder like lame does is this: determine which stereo mode is best for each frame and encode the sound accordingly.

I hope this made things clearer :) cya

[Kurtnoise]

Some interesting links :
http://andreas.welcomes-you.com/projects/dv/index.html (mp2)
http://www.hydrogenaudio.org/index.php?act=ST&f=16&t=4678 (mp3)
http://www.hydrogenaudio.org/index.php?act=ST&f=15&t=995 (mp3)
http://www.hydrogenaudio.org/index.php?act=ST&f=15&t=683 (mp3)

[Bogalvator]

And also another one:
http://www.hydrogenaudio.org/index.php?showtopic=20007&hl=

[Fr4nz]

Originally posted by aliumalik
I have searched google for this but can't seem to find a final conclusion while some people say that Joint Stereo is better in quality others say it completely copies the other channel. Can anyone tell me whether it offers better or worse quality than stereo.
Thanks a lot!

Whit a well implemented Joint Stereo (like the LAME or Fraunhofer one), using JS is always better than using simple stereo.

[JohnMK]

I remember talking to some LAME developers about this a couple years ago, and the conclusion I reached was that even at 320kbps, joint stereo probably produces benefits that exceed the potential for error. LAME has well-implemented joint stereo, you really need not fear it.

[Fr4nz]

Originally posted by JohnMK
I remember talking to some LAME developers about this a couple years ago, and the conclusion I reached was that even at 320kbps, joint stereo probably produces benefits that exceed the potential for error. LAME has well-implemented joint stereo, you really need not fear it.

Absolutely :)

[Slogra]

LAME mp3s (aps, apx, api) use both joint-stereo and normal stereo frames, whatever is best for that frame. Note these "mixed" mp3s are recognized (by Winamp etc) as joint-stereo mp3s.