First off, I would like to thank this forum as a former lurker and newbie to this stuff, it is very much appreciated!

I am ripping all of my cd's to my computer to play through the stereo. A couple of months ago, I had a cd of mine break (luckilly, a friend had the same CD and burned me a copy). I figure that having lossless backups for these CD's would be a nice safety bonus as well.

In wavepack, you have the option of turning on or off joint stereo with the -j0 switch. If ripping to joint stereo, what would be the effects of ripping the backup back to a CD? Would the regular two-channel stero be restored, or would the joint stereo solution be burned on the CD? I have tried looking everywhere for an answer to this, but can't find any (maybe because it's too simple ). Thank You.

I suspect it won't matter at all. WavPack is lossless, after all.

Default behaviour is -j1 (non-adaptive j/s on) . There are no quality issues as output is lossless and compression is increased for samples with stereo correlation.

Use -x1 or -x for adaptive j/s

I wanted to ask Bryant about this for a few weeks now:

David,

Is it possible to add a smart j/s switch like -ja that will encode faster than -x1 ?

Simple stereo audio would be decoded from the joint stereo and written to the CD.

It is a lossless codec. If your output is bit identical to your input when you decode it, then you don't lost nothing. If the file using JS is smaler, and the previous condition is true, then it is better because it compress more.

See this for a explanation of how joint stereo works. It is for lossy Mp3, but the base is the same as for wavpack. See especialy the "Mid/side stereo vs Left/Right stereo".

Hi shadowking, sorry I missed this before...

At this time, all that -x1 does is determine the optimum joint-stereo setting. It does this simply by trying both (LR and MS) for each block and then using the best one for the "real" encode, so it takes from 2-3 times the standard encoding time. The reason it isn't fully 3 times slower is that the "trial" runs don't do everything required for an actual encode (which is important for higher x modes to save lots of time). I suppose I could do a special case for -x1 that would be exactly 2x slower than standard, but I'm not sure it would be worth it.

BTW, simply comparing -j0 and -j1 to get an idea of the extra compression that joint stereo provides is somewhat misleading. The reason is that in the default and high modes the decorrelation functions include cross-channel information, so the advantage of joint stereo is reduced. If you compare -j0 and -j1 in fast mode you will see that the joint stereo provides a bigger advantage for this reason. In fact, Ghido told me that OptimFROG has such good cross-channel decorrelation that he doesn't need to use joint stereo at all!