I noticed that some of the popular encoders, such as LamedropXP encode mp3's as joint stereo
instead of pure stereo - even at the higher bitrates.
I always believed JS was a compromise intended only for the lower bitrates, ie. 160 kbps and below.
JS is preferred now?
+ Reply to Thread
Results 1 to 8 of 8
Your impression agreed with mine, afa compromise @ lower bitrates. I'm guessing the programmer didn't know the differences and just guessed.
I suppose anything is possible, perhaps somebody "in the know" could explain it, if there is actually
a reason for it or whether it's a choice of convenience for the programmer.
LAMEdropXP produces a very nice encode, vbr average ~200 kbps. Sounds great. I'd like to be able
to compare it with the straight stereo, but this it's not possible with this tool (as far as I can see).
Joint stereo encodes the stereo as a sum channel and a difference channel as I recall. For a given bitrate this allows more bits to the common sounds than to the difference channel, which should result in a better sounding encode. Regular stereo will have a lot of information repeated in both channels, thus 'wasting' bits. I use joint stereo for that reason. There is usually more in common between the channels than there is a difference. Some exceptions include 'Space Oddity' by David Bowie and 'Up From the Skies' by Hendrix. With the extreme panning and ping-ponging joint stereo probably uses more bits...
I understand the theory re: the bitrate allocation, but somewhere in the back of my mind is this thought that
JS was primarily intended for the lower bitrates. Perhaps I'm wrong. I wonder if LAME has been optimized
so that JS is now the preferred mode?
No, you're not wrong, because at the higher bitrates (256, 320) the effect of JS bit allocation vs. plain Stereo allocation is extremely negligible (for lots of program material, but not all), unlike lower bitrates.
BTW, if you ever work with binaural, Dolby Surround, or other material that makes much use of HRTF, Phase, or Delay, you will find that JS is actually detrimental to the sound quality.
I made a similar post at Hydrogenaudio. The gist of the answer is that with recent LAME innovations,
the encoder will pick, on a frame-by-frame basis, whether to use LR or MS stereo resulting in the best optimization
at any given bitrate. Here's the thread:
That was very helpful!