MPC --insane or --braindead
WMA 9 pro quality 98, 2channel, 24bit

Which one has better quality?
Anyone has the detail test report?

I'm not an expert. I compared them using Adobe Audition and found that WMA's spectrum was closer to the origional one, seems better than MPC. However, most people say that MPC does best on high bitrates. Now I'm totally confused. Maybe the means I used is improper?

Images don't mean anything because you are not going to listen to images.

At those high bitrates, any good codec is going to be transparent on most samples.

One can't say that one result yields more transparency than another. It's either transparent or it's not.

Hello, Incognito, welcome in HydrogenAudio.

First of all, read why graphs can't tell anything about the compression quality : http://www.hydrogenaudio.org/forums/index....indpost&p=77265 (this link is taken from the FAQ).
They can give some informations about the quality of a non compressed signal, but as soon as psychoacoustic compression is involved, like MPC or WMA, they are meaningless, since compression is nothing else than removing as much as possible (=introducing as much distortion as possible compared to the original) without audible effect.

Second, and before a discussion about sound quality arises, I recall, since you are still new and maybe not aware of the Terms of Service, that the rules about sound quality discussions are very strict here (read TOS 8 here : http://www.hydrogenaudio.org/forums/index.php?showtopic=3974 )

This said, I recently compared MP3, MPC, and WMA 9 at near transparent bitrates. It was not WMA9 pro, and it was not as high bitrates as you are targetting, but here are my results : http://www.musicplayer.com/ubb/ultimatebb....t=006691#000024

My understanding is that those with trained ears (ie not myself) do not believe WMA ever achieves transparency, while MPC does it with the --standard profile and up.

The whole idea behind lossy codecs is to throw out information that you can't hear. In theory, a good lossy codec will sound good even after throwing away tons of things which will make a spectrum look remarkably different than the source. This is a good thing - we don't hear a lot of the stuff that we see on those graphs!

Thanks for all of your replies. I made sense of it eventually.