lame 3.97 alpha11 is now available in CVS

you can checkout lame using cvs yourself:
cvs -d:pserver:anonymous@cvs.sourceforge.net:/cvsroot/lame co lame

or get the source at RareWares:
http://www.rarewares.org/files/mp3/lame_cv...50710.14.tar.gz

binaries at RareWares/Debian are updated.


later

I dont see an update at rarewares.

Why do you belive, that this is alpha11 and not a daily snapshot? The filename does not indicat this.

Win32 bundle now at Rarewares.

It's named alpha11 because 'version.h' has been amended to that effect.

It features some changes for abr/cbr in the 160-320kbps range.

Thanks.
To clarify: the changes you've done are to those areas mentioned above only, and thus if I were to only ever use the vbr settings ('-Vx' etc), then there would be no reason to use this new alpha over a10?
... or I suppose if I did then it would result in bit identical results since nothing has been changed.

try it and see

There are times when 'try it and see' is a relevant reply, e.g. 'I've never tried Foobar or Winamp: which one would you recommend?'.
But in this instance it isn't. Cos I'm not about to do some listening tests with vbr settings that I'll probably not here any difference between even if they had changed (which looks like they haven't). Instead, I was really wanting a reply off one of the LAME devs (most likely Gabriel) just to clarify.
However, I probably answered myself anyway in my first post.

Oh yeah, I don't want to hijack this thread with a 'thread of wars' about what constitutes a relevant/adequate reply or not.

fair enough

still, if a song encoded once with each version matches bit for bit...

Only 160-320kbps abr/cbr should have changed.

Ooops, sorry, I see your point.
I don't think I'd remembered what I actually said in my post. Sheesh, Monday morning an'all!

Thanks Gabriel for the clarification.

*gets coat*

Just fyi:
Yep, bit-identical between 3.97a10 & 3.97a11 with the '-V 4 --vbr-new' &'-V 2 --vbr-new' switch. I'm not gonna test anymore 'V' switchs, as I'm satisfied with that test.

BTW I couldn't test earlier as I was at work.

i referred to RareWares/Debian (binary) and RareWares (source) - i didn't think john had noticed the upstream release, so i thought i would bring it to his attention ^^

apologies, my cvs checkout script packages it in this way and i just moved it over to the server

eheh, thanx john


also, the package at RareWares/Debian is compiled with the new gcc4 in unstable.


later

You were right!! I check the CVS probably 2 - 3 times a week, so I may have caught it tomorrow.

I have just added a compile of the LAME DirectShow filter to the Rarewares 'mp3' page.

This is a 3.97a11/ICL8.1 compile of the recently updated CVS. This is untested by me, but certainly registers OK and brings up Property pages in Graphedit OK.

You should read the text files included, and there are batch files for registering/unregistering the filter.

Lame.exe shows up as:

LAME version 3.97 (alpha 10, Mar 21 2005 09:50:39)

!

Which one? If the win32 compile you have shows that, you've got them confused. The one on Rarewares shows:

Thanks John, it was my fault. I forgot to put the new exe into my eac dir. Which I did lately for a shorter command line...

Hey does someone use wapet.exe for ape tags? Does this kind of command line (in eac encoder settings) run? --> wapet.exe ... ... ... ..\lame\lame.exe ... ...
(I mean the ..\lame\lame.exe)

I need to test. The thing is, with wapet.exe the command line gets close to the limit and needs to be short...

If you're using wapet with LAME, you might want to consider using MAREO or REACT.

For jaybeee:
The LAME tag will say alpha11 and not alpha10, eh?

Eh??
Is this reply really meant to be to me or Squeller?
I never asked about LAME tags or how to ID the exe etc.

EDIT: gramma

is it safe to use it for 128kbps CBR or -V2 --vbr-new ?
(as it is stated that only 160-320 cbr/abr should have changed)

is --vbr-new going to be default for -V2 since it is prooven to be better?
(guruoolez test)

I don't know, but cbr and abr around 160 - 192 has been nicely improved, as guru says, there's less ringing or warbling in the background of some samples.

I would say yes. CBR/ABR at less than 160 kbps could sometime offer some warbling effect (noise in low frequencies: try this organ sample for an example), especially on tonal moments, and therefore could offer worse quality than older version. But on average, 3.97 is clearly better to my ears, and is safe to use. Just keep in mind that possible undetected problems might be revealed with time (but same goes for all previous LAME encoders, even older and tested ones ).