I just talked with one of my friends who encodes his music with MPC 1.1 and I pointed to the fact, that there's version 1.15r out already... As far as I know, versions 1.12, 1.14 and maybe 1.16 in the near future should be "stable" although they are marked as beta; versions 1.11, 1.13, 1.15 are alphas and should not be used? Or am I wrong? Just asking as I read in some thread JohnV said that due to the Frank Klemm version 1.14 isn't interesting anymore... It's a bit confusing. I didn't find any explanation so I ask the right way here...

the 1.15x binaries are meant for testing; they could screw up some encodes, or totally bloat the bitrate. Stick with 1.14 until Frank makes another stable release. Or test the 1.15x versions with problem samples - in comparison to 1.14 - to help Frank tune.

Most people interested in the best quality have been using 1.15 for awhile now. 1.15r should be safe for use, and is what I'd probably use myself if I were encoding right now. At the very least, I'd use 1.15f.

1.14 is still recommended in the "Recommended Settings" thread because of the 'unstable/experimental' profile issue in the tags, and "just to be safe". This is mostly geared towards newbies and so recommending 1.14 can possibly prevent some issues for them if they are not aware of possible risks.

1.15r is superior to 1.14 with at least 3-4 samples out of 50 I tried. It doesn't fail on any known test sample and it's better at handling transients. There is no need to be afraid. Staying with older versions of mppenc does not make sense.

Thanx guys for very prompt answers, it helped us a lot.

Seed:

Is it really a way better or just a little bit? As I did no comparable listening test, does it mean, should I advise him to better re-encode all the albums he's done so far? When you take into account he's encoded it with 1.1...

It's only useful if he could hear a difference in the two, which I believe he most unlikely can

cheers

---
David

So can we still use --quality x --xlevel with mppenc 1.15r?

definately ;o)

Allthough I'm a newbie I took the advice and switched to 1.15r
Now all my newly encoded files are marked as 'Unstable/Experimental' in the profile tag, is this normal?

Notice the comment setting as shown in my signature below. Anyone using 1.15r should comment the file to log the encoder version used and the quality setting.

As far as I know, versions 1.12, 1.14 and so on are betas (considered stable), versions 1.13, 1.15 are alphas; due to this fact they have this tag marker (Unstable/Experimental).

EDIT: See this thread from the beginning...

lol ok i've been using version 1.1 all along...

this thread makes it sound like there isn't any real reason to upgrade... are there any advantages at all? my encodes all sound transparent, so i'm not going to go and re-rip my entire collection (64 albums, will be 65 next tuesday when that new I Mother Earth album comes out... goodie!) But perhaps I should consider switching to a newer version for any following rips? Any advice?

I still have files made with 0.90s. I don't have those CDs anymore to re-rip/re-encode but the difference in quality between 0.90s and today's 1.15r is not major. It's less than say, Vorbis RC2 and 1.0 release or LAME 3.87 192kbps VBR and LAME 3.90.2 aps.

MPC has been "transparent" for a while now. The newer releases just keep shrinking the pool of known problems but even an older encoder version will generally offer excellent results (assuming you've been using --standard or better).

MPC can only be said to be transparent when there are no known problem samples.

Hence the quotes.

I doubt a codec will be able to be transparent on everything. Somewhere, there's a sound that will screw it up. It only took a hyper active ping pong game to screw up MP3 and Vorbis(I'm referring to c44).

I was wondering if someone could point me in the direction of the source code for one of these releases if it is public? I've been fooling around with Musepack on Mac OS X, and it would be very nice to know that I was using the "proper" version.

Thanks in advance.

- Dan

What's c44?

http://static.hydrogenaudio.org/extra/samp...amples/c44.flac

A test sample

If 1.15r is better than 1.14 (fixes to problem samples) then why is it not renamed as a beta instead of an alpha?

OMG I'm such an audiophile . ABXed this as mp3 insane ang ogg q9
This is the Killer Sample.
Ok I confess than I can't ABX mpc standard and ogg q10.
Is this some magickal sample or true music?

p.s. sorry for quite OT

Everything can be considered as true music. Even silence...
A sample is considered as valid when it came from a 'music' audio disc. Generated samples are more problematics from an epistemological point of vue : they are very useful for testing some critical point, but it's difficult to make any audioformat choice with them.

Musical neighbour of c44 are castanets passage, like this one :
http://static.hydrogenaudio.org/extra/samp...castanets2.flac



Valid/Invalid duality is less important than the Frequent/Rare one. For an exemple, if you listen a lot of eighteenth century music (in Europa), you must take care from harpsichord problems, because you will probably find this instrument on 80% of all the baroque recordings. If you listen to heavy metal, harpsichord issue can't be considered as a major problem (but they are maybe some similar problems).

Short samples are only illustrating some problems you may find by encoding a disc. They must be considered carefully, but with some intelligence too.

btw, harpsichord is not only occuring on baroque music,
Alan Parsons Project (Eric Woolfson himself, in another song another musician) played harpsichord on their album:

The Alan Parsons Project (1976) Tales Of Mystery And Imagination (Edgar Allan Poe)