Recommended Encoder Versions and Settings Options

[QuantumKnot]

From the Vorbis listening test and the recent 128 kbps multiformat test, Aoyumi's aoTuV Vorbis tuning was determined to be the best Vorbis encoder and hence it is now the recommended encoder.

Maybe I'm misunderstanding things a bit but shouldn't it state that aoTuV is recommended for lower bitrates and GT3b2 is recommended for higher bitrates?

I thought that would be implied in the "128 kbps multiformat test'. But it was only a very quick change I made to get it up there. I'll definitely rework it, now that I have some time.

[bani]

[Linux:

oggenc aoTuV beta 2 (static gcc compile)

not very static at all

$ ./oggenc-aotuv
./oggenc-aotuv: /lib/libc.so.6: version `GLIBC_2.3' not found (required by ./oggenc-aotuv)

how about 100% static version for those of us without glibc2.3?

[QuantumKnot]

not very static at all

$ ./oggenc-aotuv
./oggenc-aotuv: /lib/libc.so.6: version `GLIBC_2.3' not found (required by ./oggenc-aotuv)

how about 100% static version for those of us without glibc2.3?

Oh. I'll look into it then. Judging from the filesize, I was confident that it included everything. Maybe not

[de Mon]

I also think GT3b2 can be newbie confusing. Actualy most of them don't know of New/Old versions. And somebody can accidentally use GT3b2-Old.
I propose to name merged version as GT3F (Final) or GT3S (Stable)

[xmixahlx]

well, theres nothing "stable" about it

it doesn't at all address the problems <q4, and tests have shown that >q4 needs better/more tuning to be (more) effective than aotuv is

at this point it should be considered to tune aotuv for higher bitrates and drop gt3 all together... garf doesn't tune it any more, and all later versions from gt3b1 are just hacks anyways...


later

[Tang]

well, theres nothing "stable" about it

it doesn't at all address the problems <q4, and tests have shown that >q4 needs better/more tuning to be (more) effective than aotuv is

at this point it should be considered to tune aotuv for higher bitrates and drop gt3 all together... garf doesn't tune it any more, and all later versions from gt3b1 are just hacks anyways...


later

Seems to be a good choice to avoid any confusion indeed...

[kidi]

Hi! Could somebody help me? :
I`d like to use a lowpass filter when encoding with low bitrates (EAC). What is the additional parameter for the command line? Thanks in advance.

[QuantumKnot]

Hi! Could somebody help me? :
I`d like to use a lowpass filter when encoding with low bitrates (EAC). What is the additional parameter for the command line? Thanks in advance.

CODE

--advanced-encode-option lowpass_frequency=NN

for NN kHz cutoff.

This post has been edited by QuantumKnot: Sep 26 2004, 01:58

[NoXFeR]

I think it would be nice if some information about possible candidates for next recommended versions is on the first post of the thread. That way people can get some information about them without searching wild. Like information about the Megamix merge and testing of it.

This post has been edited by NoXFeR: Aug 3 2004, 14:55

[QuantumKnot]

I think it would be nice if some information about possible candidates for next recommended versions is on the first post of the thread. That way people can get some information about them without searching wild. Like information about the Megamix merge and testing of it.

There is already a thread that lists out all the different versions.

http://www.hydrogenaudio.org/forums/index....showtopic=21126

As for candidates for the next recommended versions, I can only decide via listening tests from others as well as their opinions. Vorbis development is in a state of flux at the moment so one version chosen as recommended may soon be outdated by another (eg. Vorbis 1.1 may be coming out soon)

[esa372]

How do I know which encoder was used to make this Ogg Vorbis file?
Using either the ogginfo program or file info in your player, you can tell from the vendor tag:

Xiph.Org libVorbis I 20020717 => 1.0

Xiph.Org libVorbis I 20030909 (1.0.1)  => 1.0.1

I've got a question regarding this:
I'm using dBpowerAmp music converter (Release 10.1) to convert from FLAC to Ogg (-q6). According to dBpowerAmp, it's using 'Release 11.1 (Ogg V1.0.1)' of Ogg Vorbis. But if I look at the tag of any given converted file more than once, it displays different information. One time it may read 'Xiph.Org libVorbis I 20030909 (1.0.1)' and the next time I open it, it may read 'Xiph.Org libVorbis I 20020717'.

Any idea what's going on? (or what version I'm actually using?)

Thanks in advance for the help!

~esa372

[Peter Harris]

But if I look at the tag of any given converted file more than once, it displays different information.  One time it may read 'Xiph.Org libVorbis I 20030909 (1.0.1)' and the next time I open it, it may read 'Xiph.Org libVorbis I 20020717'.

Any idea what's going on?  (or what version I'm actually using?)

The old Winamp Vorbis plugin would overwrite the actual encoder version with its version whenever you changed any of the tags. This doesn't change the quality of the encode, it only makes the encoder string incorrect.

Maybe whatever you're using to tag your Vorbis files has the same bug?

[QuantumKnot]

IIRC, the old Vorbis plugin bug caused the vendor string to be changed to RC3.

esa372:

What program are you using the view the tag information?

[DreamTactix291]

Just would like to add that the compile of 1.1 with the impulse_trigger_profile switch available from rarewares has a vendor tag of Xiph.Org libVorbis I 20040920 and not 20040629 like 1.1 RC1.

[QuantumKnot]

Just would like to add that the compile of 1.1 with the impulse_trigger_profile switch available from rarewares has a vendor tag of Xiph.Org libVorbis I 20040920 and not 20040629 like 1.1 RC1.

Thanks. Change made. Actually, this is a bit confusing since the vendor tag hasnt changed in the source code. Hence the linux build with ITP is probably using the old vendor tag while the win32 builds use this new one.

[esa372]

esa372:
What program are you using the view the tag information?

I found out what the problem was... Peter Harris was right...

The old Winamp Vorbis plugin would overwrite the actual encoder version with its version whenever you changed any of the tags. This doesn't change the quality of the encode, it only makes the encoder string incorrect.

Maybe whatever you're using to tag your Vorbis files has the same bug?

Apparently, this happens with several tag editors - JetAudio and Rio Music Manager both change the encoder version in the tag... as well as dbPowerAmp(!). Gotta be careful with these things!

Now I'm using Foobar2000 without any such trouble. But, what a hassle it was to trace the problem!

[Garf]

Prolly interesting to add:

Xiphophorus libVorbis I 20010816 (gtune 1)
1.0 RC2 GT1

Xiphophorus libVorbis I 20011014 (GTune 2)
1.0 RC2 GT2

This post has been edited by Garf: Sep 25 2004, 19:20

[Tang]

Thanks for your EDIT of the 1st post QK... Very nice job...

[jwu42]

Nice revision there QK - much appreciated

[sergelac]

i'd like to use the new 1.1 but i hate these switches (i use oggdropXPd)
can someone make a new compile of the new 1.1 with GT3b2 tunings ?

This post has been edited by sergelac: Sep 27 2004, 03:36

[esa372]

Thanks for the update, QuantumKnot... excellent work..!

Just one question:
I see that the default setting for the 'microattack' is 2.7 - what (if any) is the default setting for the 'pre-echo'?

Anybody know? Is there a way to check this?

Thanks!

~esa

[QuantumKnot]

Thanks for the update, QuantumKnot...  excellent work..!

Just one question:
I see that the default setting for the 'microattack' is 2.7 - what (if any) is the default setting for the 'pre-echo'?

Anybody know?  Is there a way to check this?

Thanks!

~esa

The answer is simple. The default value for impulse_noisetune is 0 impulse_noisetune is a relative setting. It will add to whatever is the default for that quality level. To force more aggressive coding of short blocks, lower the impulse_noisetune to negative values (max negative value is -15).

[esa372]

The default value for impulse_noisetune is 0   impulse_noisetune is a relative setting.  It will add to whatever is the default for that quality level.  To force more aggressive coding of short blocks, lower the impulse_noisetune to negative values (max negative value is -15).

Thanks, QK!


Here's another question for you:
You said that...

...the notion that you have to sacrifice something for a gain in speed may not be untrue here and some people prefer the safety and stability of the MSVC builds. If you want to be doubly sure that your new and shiny Ogg Vorbis files are 'stable' and 'safe', go for the MSVC builds.

Can you explain what this "something" is that we might be loosing by using the new ICL compiles rather than the MSVC? What are the potential drawbacks?

Thanks (again)!

~esa

This post has been edited by esa372: Sep 27 2004, 15:56

[QuantumKnot]

Here's another question for you:
You said that...

...the notion that you have to sacrifice something for a gain in speed may not be untrue here and some people prefer the safety and stability of the MSVC builds. If you want to be doubly sure that your new and shiny Ogg Vorbis files are 'stable' and 'safe', go for the MSVC builds.

Can you explain what this "something" is that we might be loosing by using the new ICL compiles rather than the MSVC? What are the potential drawbacks?

Thanks (again)!

~esa

I'm not familiar with the problems with ICL and in most of the cases where I've used it, it seems great. However, perhaps that something is 'certainty' or 'the warm and fuzzy feeling' But yeah, I think people have pointed out some issues with (particularly older versions of) ICL cutting corners in some ways that aren't good. I'm afraid that's the most detailed answer I can give. I'm sure there are plenty of others who know more about this than me.

[esa372]

Thanks for all your help, QK! It really helps us newbies out!

One more question:
Regarding the 'impulse_trigger_profile' setting, you said:

Note 1: There is the possibility that relaxed block switching may cause other quality problems and create suboptimal Vorbis files.  So use sparingly and with caution.

I'm not sure what you mean by 'relaxed block switching'. Are you referring to a lower profile number in the 'impulse trigger profile' string (ie, --advanced-encode-option impulse_trigger_profile=0 (or 1))? In other words, are you recommending the use of higher ITP values?

I was intending to use an ITP value of 3 for most (if not all) of my files, but your warning to 'use sparingly and with caution' caused me to balk. Any clarification would be great!


Also, I noticed that in the 'Vorbis History' section, it still reads:

...John33 merged the sources to give us GT3b2, which is now the recommended Vorbis encoder.

Does this still hold true in light of the release of 1.1?



Thanks again. I hope all these questions aren't a bother...

~esa

This post has been edited by esa372: Sep 28 2004, 15:23