Hi, I have few MPC in BrainDead or Insane profile and I would like to decode them to waves, and then encode to MPC Extreme Profile to save some space on HD. What do you think about this? Will I loose on quality?

Yes : every transscoding operation from lossy to lossy --> lower quality.
But difference may be inaudible.

Is your hard drive so limited to such operation ? I suggest you to delete some useless files, and to keep your nice mpc -q8

Keep your original files!!

You could always archive the -q8 to cd, then make --standard encodes for your hard drive.

You never know you might not even hear the difference

First, it's called reencoding because it's encoding again from the same format.
Second, i think reencoding it to extreme from insane (or even braindead) is a really bad idea, since the gain in HD space you'll get won't be so significient, plus you'll just be losing more quality.
Also, it's not necessary to decode to wav and then reencode to MPC, you can use MPC Batch Encoder.

Whatever happened to bit-stripping? Is that not a feature of mpc that allows stripping away data from an mpc to lower birate without reencoding? Useful.... but i don't know if any current mpc software does it. Oh well.

You're talking about 'bitrate peeling', and i don't think there are plans for this in musepack (someone correct me i'm wrong here). In vorbis there were, but it seems they've been pushed back

Yes, it´s really a sad thing, that more and more ppl think they MUST choose the highest quality-levels possible in MPC, not seeing the limited upstream-bandwith they can offer only for spreading their stuff. But, ok, this topic we already discussed till having been "braindead". I´m still have to do one session per week with my psychiater because of this.

But back to your question: Yes, re-encoding is indeed the most terrible thing you do (nearly as terrible as a napalm-bomb). As far as I know, re-encoding from one quality-level in MPC to another leads to more quality-losses than the so-called transcoding (e.g. MPC -> MP3). But that is an unconfirmed information now. I still didn´t check it out by myself, as I saw no use to do it.

And yes, disc-space gets cheaper and cheaper. So, this shouldn´t be the problem anymore. But the limited upstream-bandwiths still is, was and will be for a longer time (at least the next 5 years) a big prob.

bollox!
If you don't know for certain, don't post it. this is a great way of spreading false information - assuming.
Ofcourse you get a loss from re-encoding, but it's not even close to what you describe.
You should be able to survive one generation of re-encoding with MPC without such a loss it would yield in a disturbingly artefacting second generation MPC.

Just did a test: transcoded (or re-encoded, the two terms are interchangeable) MPC insane v0.90o to MPC v1.14 standard (no xlevel either time, but no clipping warnings either). Both times I decoded with mppdec v1.93j, default parameters.

I picked a completely random song (The Cardigans, "Lovefool") and just listened for a few seconds to the transcoded version until something sounded wrong (didn't take long, about 8 seconds in, where Nina sings "you"). Her voice has some sort of flange to it. I did a lightning ABX test, one MPC versus the other, 10/10. In other words, it sounded pretty crappy for MPC.

Perhaps using a better source MPC file (one encoded more recently) would improve the score, but this test supports my long-time opinion on the matter of MPC -> MPC transcoding.

try a more recent version off mppenc and see if the resulting MPC has a similar artefacting issue

@ MRTH: With "unconfirmed information" I meant, that I can´t remember the article here in HA where it was written, but I was sure I have read this before (but I couldn´t citate it). But now I can remeber: It was this article of "Sometimes Warrior" from April, 4th, 2002.