If John already has compiles of the CVS code for win32 and supposedly, they are identical to Vorbis 1.0.1, why haven't Xiph.org officially released Vorbis 1.0.1 yet?

I also wonder about that.

And to think I postponed my 64kbps test one week to wait for the relase... Thanks God I later decided it's not a good idea to depend on Xiph's schedule

There was a delay sorting out how to package the win32 SDK stuff. (DLLs, Libs., Visual Studio Projects, etc.) It looks like all that was fixed last week. I would guess we should see something soon, but my information only comes from reading the Vorbis developer mailing list not from anyone official.

edit: fixed bad grammar

This is pretty much the case, I think. I did a number of SDK builds for Xiph, but all seems to have gone quiet again ATM!!!

Actually, I've been wanting to ask that as well.

Its been over a month since the last Ogg traffic, and I've been holding off re-ripping all my stuff into Vorbis ( iRiver support coming soon! B) ). I'm getting kind of impatient.. hope some positive news will be released soon!

AFAIK, the biggest quality improvement in 1.0.1 will be at low bitrates (that's why I postponed my test). It's some bug Guruboolez found on very quiet music, IIRC.

Are Garf's GT3 tunings going to be included even? I don't think so.

I've been wondering the same thing about the "official" release of 1.0.1 by Xiph.

However, as of about 15 minutes ago, I just finished re-ripping and encoding my whole music collection in "what we have so far" of Vorbis 1.0.1 (the 2.3-1.0.1-P4 package from RareWares, dated internally 09-09-2003), hoping this will be at least close if not identical to the official release of the 1.0.1 encoder. But as Roberto says, the fixes are probably going to be all minor anyway, with about the only quality improvement affecting just the low bitrates (not affecting my -q 5 encodings).

At -q 5 I guess I could just as well have used Post 1.0 CVS, huh?

They've always been kinda slow

http://www.xiph.org/minutes/2003/november/raw/

My understanding is that the code in the CVS is complete as far as 1.0.1 is concerned. The delays in the release centered around the precise composition of the published SDK(s), particularly whether the dlls should be statically or dynamically linked and which of the MSVC complier(s) should be supported in regard to the provision of project files. I still haven't actually seen a statement that finally resolves these issues.

In the meantime, the compiles at Rarewares are built from what I believe are the 1.0.1 libs. At the last count, Garf's tuning efforts (GT3b1) were stated as going to be incorporated in 1.1, as and when that arrives.

Well, according to this post by Carsten Haese on Sept 4,

http://www.hydrogenaudio.org/forums/index....pic=7197&st=25&

the bugfix that addresses the bug found by guruboolez was committed to the CVS. Hence, John's 20030909 compile would include this bugfix.

But Carsten mentioned there were other bugfixes that were to be addressed in 1.0.1 and the CVS state at that time (Sept 4) was still 'work in progress'. So am I right in assuming there will be other bugfixes that were committed to the CVS after Sept 9?

You can see for yourself on the xiph.org cvs mailing list archives. Nothing was changed that could affect sound quality.

Nice. Most of the changes afterwards seem to be documentation and build changes. So it does seem the 20030909 build should sound the same as 1.0.1 then.

Where can I get this build then? And do I have to compile oggenc myself?

Rarewares.

RareWares

If you have a Pentium 4, download Oggenc2.3 - 1.0.1 P4-only (259Kb). If not, download Oggenc2.3 - 1.0.1 (271Kb).

Thanks, found it. I notice the files are about 10-20kb bigger than with 1.0, using -q8.

Hmmm... I thought the 1.0.1. version contained a bugfix for low bitrates, but according to your findings it would also affect higher bitrates.

But I still think you should use Garf's GT3b1 version for high (> -q5) bitrates.

Hmm.. so does 1.01 affect encodings at q >5.0 or not? From CVS commits, it would seem not, however, there seem to be (relatively) minor changes as indicated by mad_arab.

I guess I'm in no position to demand anything, but I wish Xiph would just release the darn encoder and tools, and resolve whatever issues they have with the SDK at a later date, or heck, at least an official Ogg Traffic stating the issues... It's killing me waiting for this

yeah i think it does more bad than good when they release so slowly

Just use the 1.0.1 compiles on Rarewares. There will be very little to no difference between this encoder and the official one in terms of quality. I've checked the CVS logs and after the CVS state on Sep 9, there has been no changes to things that affect quality. The other major difference I remember Carsten Haese mentioning is that the new oggenc in 1.0.1 will update the progress % less often.

If you are using q > 5.0, then GT3b1 would be the best option. Or I guess you could wait for Garf to merge his tunings into 1.0.1.

I downloaded the nightly cvs tarball from Xiph.org and compiled it for linux. It definitely says 1.0.1 when you run oggenc and lo and behold, the percentage updates less frequently.