PDA

View Full Version : 80 vbr joint stereo or 80vbr stereo?


azmoth
21st October 2010, 16:18
Just a quick question which is better using 80vbr joint stereo or stereo 80vbr. I would not expect any surounds to be kept in a bitrate as low as this, but would stereo retain the signal better at a low bitrate?

nurbs
21st October 2010, 19:06
Assuming you are talking about encoding MP3 with LAME use Joint Stereo regardless of bitrate. There is a reason why it's on by default. With other MP3 encoders you might want to do some research, but I think every encoder that isn't completely outdated should handle it fine. If you look at newer formats like AAC encoders often won't let you turn it off.
By the way using Joint Stereo doesn't mean that he whole track will be encoded as M/S. The encoder will switch between M/S and L/R depending on the correlation between the channels in different parts of the track. Further information available here (http://wiki.hydrogenaudio.org/index.php?title=Joint_stereo).

azmoth
21st October 2010, 19:59
You got it-LAME encoding-Thankyou your help and input is appreciated.

shon3i
21st October 2010, 21:39
Joint Stereo M/S is completly lossless process (simmilar to CABAC in H264), and it's used also in lossless encoders like FLAC and others. JS/MS is highly recommend always, if encoder support it. Some implementations like nurbs says have even better algo, that can switch per frame, newer compression algo like AAC can do switch per subband, and this give much better results. JS/MS may not always help, there is some cases where JS/MS uses more bits than classical stereo, that's why is better to turn it on and leave to encoder determine is better to use it or not.

azmoth
21st October 2010, 21:48
It is the best option, but doesn't it kill suround signals below 192 bitrate?

shon3i
21st October 2010, 22:13
again JS/MS is lossless, that mean it will not change or destroy stereo image and it will be same as orignial but with some bit savings.

azmoth
22nd October 2010, 13:49
Shon3i js/ms is good, but say you had a bitrate at 128kb/s in stereo or lower would any surround infomation be lost if you encoded it to 128js/ms instead? Many say it does destroy surround information but LAME preserves signal. I'm on fence at moment. Your view took me by complete surprise since joint stereo stirs up so many negative views.

Thanks.

nurbs
22nd October 2010, 14:41
Joint Stereo doesn't destroy surround information. Transformation between M/S and L/R is lossless and perfectly reversible as illustrated in the link I posted.
People might say that it hurts surround, but people say a lot of things that are flat out wrong. I've never heard anyone with a clue about audio processing complain about Joint Stereo in general. Sure there were complaints about specific implementations in old mp3 encoders, but that was a decade ago. These days the only ones telling people not to use Joint Stereo are Audiophiles and, as usual, they have no idea what they are talking about.

azmoth
22nd October 2010, 17:15
I was on the fence regarding this, but will say joint stereo produces high quality and if it kills surround below 192? then maybe a prologic decoder will fill in the missing audio info in anycase. The article(if relevant now, probably not as it refers to layer II) below shows to some extent the Audiophile argument which is quite a good read with graphs showing sound data .

http://forum.videohelp.com/threads/80513-Does-joint-stereo-destroy-Dolby-Surround-Sound-(rewritten)

Advocates of joint stereo read this:

http://harmsy.freeuk.com/mostync/

shon3i
22nd October 2010, 17:27
Ok, you completlly misunderstood this let me explain.

First there is two types of Joint Stereo.

1. Mid/Side (MS) which we already say is completly lossless, so will not damage stereo or surround image in case prologic, DPL etc. If you get different results, your encoder or decoder is definitly broken, and it's not by standard
2. Intesivity Stereo (IS) which is completly different than MS. This mode of Joint Stereo will destroy stereo image, and will encode practicly as mono in some cases, to save bits, process is completly lossy, and is not possible to revert stereo image to original state.

In your first link, guy talking about MP2 encoding which by now, i know all MP2 implementations use JS/IS. And that's explain why Single Stereo is Better. And if you look date of this post (2002) is time where implementations are been poor.

nurbs
22nd October 2010, 17:34
http://forum.videohelp.com/threads/80513-Does-joint-stereo-destroy-Dolby-Surround-Sound-(rewritten
This article is about MP2. You asked about MP3 in this thread. There is a world of difference between those two formats. MP2 uses an older less efficient joint stereo method and AFAIK doesn't allow switching between joint and normal stereo on a per frame basis. This has been greatly improved upon in MP3 and newer formats like AAC.

edit: shon3i was faster :)

azmoth
22nd October 2010, 18:30
... The article(if relevant now, probably not as it refers to layer II)

Guys I knew old article was MPEG1 layer II as above shows, I wanted your valued opinions anyway, so put link in to get feedback and to contrast how things have changed from then till now.