Will "--xlevel" improve the audio quality?

Null vote for me.

I suggest you try encoding the same file twice; one time without, and one time with --xlevel. Then, decode them to WAV and compare the files..

--xlevel only sets in if there is internal clipping. Otherwise the file remains untouched, CRCs will be the same.

My point exactly.

As long as no clipping occured:
- the decoded WAV's will be identical
- I also suppose that the encoded MPC's are always bit-identical, but that shouldn't matter.

So, for kotrtim, once and for all:
- the --xlevel never decreases quality
- the --xlevel helps to avoid most clipping situations
- the --xlevel does not increase filesize

Therefore, including --xlevel in your command line is wise thing to do, with no drawbacks except for the older (obsolete) decoders which don't handle it.

Everybody says that using --xlevel is a good choice.
I think that it would be more comfortable if xlevel protection were turned on by default. It would be not necessary to write "--xlevel" each time...

I agree, especially since every CD is compressed. It's one of those why isn't it questions. Not every person will know that implimenting --xlevel is wise if they don't frequent the HA forum.

You can't force --xlevel if you want backwards compatibility. MPCs created with --xlevel sound absolutely horrible on the old decoders. You don't want to confuse users.

And mppenc's output will say "use --xlevel" when it generates clipped MPC files. It's not kept as a big secret.