Skip to main content ↑ ↓

Notice

Please note that most of the software linked on this forum is likely to be safe to use. If you are unsure, feel free to ask in the relevant topics, or send a private message to an administrator or moderator. To help curb the problems of false positives, or in the event that you do find actual malware, you can contribute through the article linked here.
Topic: Vorbis 1.1 RC1 in SVN (Read 66042 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

Vorbis 1.1 RC1 in SVN

Reply #75 –
The bitrate is extraordinarily high because of the way I've tuned it to reduce pre-echo in samples like castanets.  For q 4, trying to get castanets to sound transparent was perhaps overkill so I'm in the process of down-tuning it to reduce the bitrate (so far, down 10 kbps).

Vorbis 1.1 RC1 in SVN

Reply #76 –
I look forward to trying it out.    Although even with that inflated bitrate it still isnt right (at 2.5 secs) and i prefer the encode made from a managed version of megamix 2 (ive abxd successfully).    Can anyone explain why this would be the case?  Even just using gt3b2 with managed limits is pretty good - but at under q5.    And definitely better than either gt3b2/megamix2 at q4.99. 

Does engaging the managed mode (without the -q mode) have an effect on the way the gt3b2 tunings (im assuming a lot of the tunings are for pre/post echo) are used?  Or maybe the usage of short vs long blocks?

Vorbis 1.1 RC1 in SVN

Reply #77 –
Quote
I look forward to trying it out.    Although even with that inflated bitrate it still isnt right (at 2.5 secs) and i prefer the encode made from a managed version of megamix 2 (ive abxd successfully).    Can anyone explain why this would be the case?  Even just using gt3b2 with managed limits is pretty good - but at under q5.    And definitely better than either gt3b2/megamix2 at q4.99. 

Does engaging the managed mode (without the -q mode) have an effect on the way the gt3b2 tunings (im assuming a lot of the tunings are for pre/post echo) are used?  Or maybe the usage of short vs long blocks?
[a href="index.php?act=findpost&pid=226205"][{POST_SNAPBACK}][/a]


Unfortunately, I don't have an answer for that but I'm certain someone else like nyaochi will since he knows the internals better than me.  This block switching thing is quite interesting and I wonder whether all this pre-echo tuning I do is really worth the effort.  I'll stop flooding this thread with binaries since I don't want to cause confusion with these different experimental versions floating around.

There will always be killer samples for all codecs, much like how amnesia, waiting, and fsol kill mpc --standard.  Good thing is that with these killer samples, you can devise fixes for them.

Vorbis 1.1 RC1 in SVN

Reply #78 –
Hi.  This is my first (successful) foray into ABX testing.  I have never been able to successfully ABX anything unless the samples were encoded to super low bitrates (e.g. 64...).  I decided to embark on an experiment with the different Vorbis tunings available because I own a Rio Karma.  I am deciding if I should use either the new Xiph Vorbis 1.1 RC1, aoTuV b2, or the new megamix v2.  The current encoder/settings I use for my Karma are aoTuV b2 -q4.  While experimenting, I found a sample that I can very easily ABX from the original.

The sample is from Images and Words by Dream Theater.  The song is called “Surrounded” and I found I could ABX the solo keyboard introduction of the song.  I am very shocked at how easy it is to ABX this track with all three Vorbis tunings I tested.  All three tunings exhibit a louder sounding background hiss (this could be some sort of pre-echo or some other artifact that I have not the experience to name correctly) vs the original file.  Below are the ABX results.  I used foobar2000's ABX program. 

Code: [Select]
foo_abx v1.2 report
foobar2000 v0.8.2
2004/07/18 12:17:18

File A: file://E:\Stuff\My Documents\My Music\Karma Music\Dream Theater - Images And Words - 04 - Surrounded.ogg
File B: file://E:\Stuff\My Documents\My Music\Karma Music\Dream Theater - Images And Words - 04 - Surrounded.flac
12:17:45 : Test started.
12:18:31 : 01/01  50.0%
12:18:39 : 02/02  25.0%
12:18:52 : 03/03  12.5%
12:19:02 : 04/04  6.3%
12:20:08 : 05/05  3.1%
12:20:20 : 06/06  1.6%
12:20:35 : 07/07  0.8%
12:20:57 : 08/08  0.4%
12:21:11 : Test finished.

 ---------- 
Total: 8/8 (0.4%)


oggenc-megamix2 -q3.5 - -o %d

bitrate = 127
channels = 2
samplerate = 44100
bitrate_nominal = 120
codec = Vorbis
vorbis_vendor = Vorbis Megamix II: Xiph.Org 1.1 RC1 [20040629] + QKTune beta 3.2 + GTune 3 beta 2


Code: [Select]
foo_abx v1.2 report
foobar2000 v0.8.2
2004/07/18 12:37:08

File A: file://E:\Stuff\My Documents\My Music\Karma Music\Dream Theater - Images And Words - 04 - Surrounded.flac
File B: file://E:\Stuff\My Documents\My Music\Karma Music\Dream Theater - Images And Words - 04 - Surrounded.ogg
12:40:07 : Test started.
12:40:21 : 01/01  50.0%
12:40:35 : 02/02  25.0%
12:40:42 : 03/03  12.5%
12:40:58 : 04/04  6.3%
12:41:09 : 05/05  3.1%
12:41:21 : 06/06  1.6%
12:41:29 : 07/07  0.8%
12:41:42 : 08/08  0.4%
12:41:45 : Test finished.

 ---------- 
Total: 8/8 (0.4%)


oggenc-1.1rc1 -q4 - -o %d

bitrate = 125
channels = 2
samplerate = 44100
bitrate_nominal = 128
codec = Vorbis
vorbis_vendor = Xiph.Org libVorbis I 20040629


Code: [Select]
foo_abx v1.2 report
foobar2000 v0.8.2
2004/07/18 12:51:07

File A: file://E:\Stuff\My Documents\My Music\Karma Music\Dream Theater - Images And Words - 04 - Surrounded.flac
File B: file://E:\Stuff\My Documents\My Music\Karma Music\Dream Theater - Images And Words - 04 - Surrounded aotuv.ogg

12:51:28 : Test started.
12:51:51 : 01/01  50.0%
12:51:59 : 02/02  25.0%
12:52:08 : 03/03  12.5%
12:52:15 : 04/04  6.3%
12:52:27 : 05/05  3.1%
12:52:36 : 06/06  1.6%
12:52:44 : 07/07  0.8%
12:52:53 : 08/08  0.4%
12:53:05 : Test finished.

 ---------- 
Total: 8/8 (0.4%)


oggenc-aoTuV b2 -q4 - -o %d

bitrate = 125
channels = 2
samplerate = 44100
bitrate_nominal = 128
codec = Vorbis
vorbis_vendor = AO; aoTuV b2 [20040420]  (based on Xiph.Org's 1.0.1)


I lowered the quality setting on megamix v2 because I wanted to test at a similar bitrate (128) and the setting of 3.5 gave me closest to that bitrate with this tuning. 

I don't know if this helps anyone out much, but I hope it does.

Sample uploaded here:
http://www.hydrogenaudio.org/forums/index....ndpost&p=227000

Vorbis 1.1 RC1 in SVN

Reply #79 –
Dont know if this will help, best results i could get with my problem sample (without resorting to using the managed mode) was from garfs GT2 tuning of the old RC2 encoder:

http://sjeng.org/vorbisgt2.html

Vorbis 1.1 RC1 in SVN

Reply #80 –
Quote
Quote
That's an good suggestion. I guess oggenc does have much less tweaking options than mppenc.


Or even better a few added command-line switches to the advanced-encode-option on the encoder that allow you to tweak noise biasing and mapping values, etc? Would be quite convenient ;-D.
[{POST_SNAPBACK}][/a]


There is already a switch for tweaking noise bias.  Check [a href="http://www.hydrogenaudio.org/forums/index.php?showtopic=23466&view=findpost&p=227974]this[/url] out. 

Vorbis 1.1 RC1 in SVN

Reply #81 –
Quote
Given different time zones, I'll have compiles at Rarewares later in the day (my time).
[a href="index.php?act=findpost&pid=224046"][{POST_SNAPBACK}][/a]


John : could you make a compile of libvorbis.dll using this 1.1 RC1 ?? This library is not on rarewares.org now...
 

Vorbis 1.1 RC1 in SVN

Reply #82 –
Quote
Quote

Given different time zones, I'll have compiles at Rarewares later in the day (my time).
[a href="index.php?act=findpost&pid=224046"][{POST_SNAPBACK}][/a]


John : could you make a compile of libvorbis.dll using this 1.1 RC1 ?? This library is not on rarewares.org now...
[a href="index.php?act=findpost&pid=237138"][{POST_SNAPBACK}][/a]

Standard and P4 compiles now at Rarewares.