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 absolute killer (Read 23251 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

Vorbis absolute killer

Reply #25
A tuning correction that drastically improves this behavior has been committed to CVS. A bugfix release that includes this fix (and many others) is just around the corner.

Best regards,

Carsten Haese
Ogg Traffic Editor, Xiph.org Foundation

Vorbis absolute killer

Reply #26
Cool, I find it nice to see the HA community having a direct influence on improving audio codecs  I couldn't pick any major problems out above -q1.5..
< w o g o n e . c o m / l o l >

Vorbis absolute killer

Reply #27
Quote
A tuning correction that drastically improves this behavior has been committed to CVS. A bugfix release that includes this fix (and many others) is just around the corner.

Best regards,

Carsten Haese
Ogg Traffic Editor, Xiph.org Foundation

Good news

Vorbis absolute killer

Reply #28
Just a tongue in cheek observation,

Mr. Guruboolez does so much excellent work finding problem samples of music
for all of us at HA.  When does this overworked individual ever have time to just sit back
and listen and enjoy his music? With his sensitivity to artifacts of all kinds, can he ever
use compressed music on his HD other than lossless files?

Oh the bane of extraordinary hearing   
you will make mp3's for compatibility reasons.

Vorbis absolute killer

Reply #29
Doesn't he use Musepack still?
I recall a thread in which he was agonising over placebo in mpc artifacts. .

Vorbis absolute killer

Reply #30
This might be the thread.  Here's hoping I never hear that well. 
I am *expanding!*  It is so much *squishy* to *smell* you!  *Campers* are the best!  I have *anticipation* and then what?  Better parties in *the middle* for sure.
http://www.phong.org/

 

Vorbis absolute killer

Reply #31
Geez!  And Vorbis (Post1.0CVS) -q 4.25 is transparent to my ears on all samples I've tested so far (>20).  But I have heard some of the known problems people have pointed out in the past with various codecs/settings/samples, so I'm pretty sure my hearing isn't too far from "average".

As odd as this may sound: Thank goodness my hearing's no better than it is!  Using a quality setting with a 136kbps nominal bitrate, all 5194 of the songs in my collection *just* fit onto my 20GB disk volume.  *whew*  If I had better hearing I'd have to go hard disk shopping.  There's some money I can put off spending!

Vorbis absolute killer

Reply #32
Quote
As odd as this may sound: Thank goodness my hearing's no better than it is!  Using a quality setting with a 136kbps nominal bitrate, all 5194 of the songs in my collection *just* fit onto my 20GB disk volume.  *whew*  If I had better hearing I'd have to go hard disk shopping.  There's some money I can put off spending!

Still, you have to stop buying (or getting) more music, cause in that case you would have to lower the bitrate in order to fit everything in...eventually you might have to deal with listenable artifacts.

Vorbis absolute killer

Reply #33
Quote
Still, you have to stop buying (or getting) more music

Stop getting music?      Blasphemy!   

Quote
cause in that case you would have to lower the bitrate in order to fit everything in...eventually you might have to deal with listenable artifacts.

I'm not re-encoding anything for a while, as I've spent the past week-and-a-half running EAC, WaveGain and Oggifier behind all my normal daily activities.

Boss: "Why's your PC so slow?"
Me: "Dunno.  Network traffic?"
 

Plus the fact that I actually will be buying another HDD soon anyway (external USB drive, since it's a notebook PC).  But you're right, I'll be pushing my capacity limit the next time I try to add music.  In the meantime, I guess I'll have to start deleting my ABBA and Bangles collections to make room. 

Vorbis absolute killer

Reply #34
Quote
In the meantime, I guess I'll have to start deleting my ABBA and Bangles collections to make room.  

DELETE ABBA???  sounds like a good idea to me. 
gentoo ~amd64 + layman | ncmpcpp/mpd | wavpack + vorbis + lame

Vorbis absolute killer

Reply #35
Quote
A tuning correction that drastically improves this behavior has been committed to CVS. A bugfix release that includes this fix (and many others) is just around the corner.

Best regards,

Carsten Haese
Ogg Traffic Editor, Xiph.org Foundation

This is great news.  Is this bugfix release the Vorbis 1.0.1 Monty was supposedly working on now?

Anymore hints on what other tunings are planned for this release?

Vorbis absolute killer

Reply #36
Quote
This is great news.  Is this bugfix release the Vorbis 1.0.1 Monty was supposedly working on now?

Anymore hints on what other tunings are planned for this release?

I'm especially curious about this too, considering that I just finished spending most of my leisure time using Vorbis Post1.0CVS to encode everything on my shelves.

Then again, I couldn't foresee any changes that would benefit me personally in terms of sound quality.  -q 4.25 in this version is transparent for me (so far), and I can't hear any of the "high frequency noise" not present with other encoders that many people seem to be able to hear with Vorbis.  Unless they're referring to cymbals...  (J/K!)  Unless they can make, perhaps, -q 2 100% transparent to me?  That would be nice...the filesizes of 96kbps nominal.

I would think that maybe more of the tuning would be in lower bitrates (-q 0 range) to fix the problem guruboolez pointed out in this thread, plus maybe Garf's GT3b1 upper bitrate enhancements could be merged into the main development path as well with this release.  That would be fantastic, IMO.

Quote
DELETE ABBA??? sounds like a good idea to me.

Hehe...if only I really had any ABBA to delete, it'd be done by now.  On the other hand, I wouldn't have needed to delete it for capacity gain, as it could be encoded in 8kbps and sound just as good to me...or better still...0kbps.   

But seriously, I do have several Neil Young and Bob Dylan albums that I *never* listen to.  They'll be the first to go.

Vorbis absolute killer

Reply #37
Quote
Hehe...if only I really had any ABBA to delete, it'd be done by now.  On the other hand, I wouldn't have needed to delete it for capacity gain, as it could be encoded in 8kbps and sound just as good to me...or better still...0kbps.  

Stop it! I know you love Chiquitita.


BTW: Your avatar creeps me out.

Vorbis absolute killer

Reply #38
Quote
Stop it! I know you love Chiquitita.

  *LOL*  Actually, weren't they over six feet?  Kinda tall for "chiquititas".  Maybe it was the platform shoes.

Quote
BTW: Your avatar creeps me out.

That's why I like it...the controversy of it.  I found it on the Storm Thorgerson website, the John Peel 30th Anniversary cover drawn by Sam Brooks.  Anyway, it kinda goes with my current obsession with enraged revolt against the RIAA in the US and the European recording industry, both trying to steal back music they sell to us (via copy-protection).  I saw it and thought, "Yeah, that's how I feel!"  Anyhow, at the moment it was either that or the AHM cow, and someone here's already using that.

Vorbis absolute killer

Reply #39
Quote
Quote
A tuning correction that drastically improves this behavior has been committed to CVS. A bugfix release that includes this fix (and many others) is just around the corner.

Best regards,

Carsten Haese
Ogg Traffic Editor, Xiph.org Foundation

This is great news.  Is this bugfix release the Vorbis 1.0.1 Monty was supposedly working on now?

Anymore hints on what other tunings are planned for this release?

Yes, I am talking about the Vorbis 1.0.1 bugfix release. The work-in-progress Release Notes can be found at http://wiki.xiph.org/Release101.

Since 1.0.1 is a bugfix release, not a lot of tuning is going in. Apart from the tuning to correct the problem with quiet signals, there is a "midrange noise floor adjustment" (Monty's words) going in, but I am not privy to which problems are corrected with that adjustment.

Major tuning and quality enhancements are planned for Vorbis 1.1, but that's somewhere in the future.

Best Regards,

Carsten Haese
Ogg Traffic Editor, Xiph.org Foundation

Vorbis absolute killer

Reply #40
Hey Now lets stop all the ABBA bashing!!!

I am proud to say that ABBA has such a special place in my heart, I refuse to despoil them by even storing their greatest hits cd as lossless files. I store this cd as a cue sheeted wav. And I have 200 backups of the cd in case I lose the original.                 
you will make mp3's for compatibility reasons.

Vorbis absolute killer

Reply #41
Quote
I am proud to say that ABBA has such a special place in my heart, I refuse to despoil them by even storing their greatest hits cd as lossless files. I store this cd as a cue sheeted wav. And I have 200 backups of the cd in case I lose the original.


Unfortunately I really do have some ABBA cds backed up as lossless files, but my wife made me do it...Honest... 

She is an ABBA nut, and we lost a couple of her cds to oxidation, so after much footwork, we found new ones, and these have now developed cd cancer as well, so yep, we got 'em secure ripped and archived as Wavpack lossless... 

There must be something about these ABBA cds. "Your mission Den, should you decide to accept it is to listen to these ABBA tunes for the lifetime of this CD. This message, CD and its contents will slowly self distruct in 10 years..." 

Vorbis absolute killer

Reply #42
Quote
Quote
Quote
A tuning correction that drastically improves this behavior has been committed to CVS. A bugfix release that includes this fix (and many others) is just around the corner.

Best regards,

Carsten Haese
Ogg Traffic Editor, Xiph.org Foundation

This is great news.  Is this bugfix release the Vorbis 1.0.1 Monty was supposedly working on now?

Anymore hints on what other tunings are planned for this release?

Yes, I am talking about the Vorbis 1.0.1 bugfix release. The work-in-progress Release Notes can be found at http://wiki.xiph.org/Release101.

Since 1.0.1 is a bugfix release, not a lot of tuning is going in. Apart from the tuning to correct the problem with quiet signals, there is a "midrange noise floor adjustment" (Monty's words) going in, but I am not privy to which problems are corrected with that adjustment.

Major tuning and quality enhancements are planned for Vorbis 1.1, but that's somewhere in the future.

Best Regards,

Carsten Haese
Ogg Traffic Editor, Xiph.org Foundation

Nice.  The less frequent updating in oggenc would be good.  I've always wondered the overhead of updating that decimal number so quickly.

So is this release only available for CVS or will it be made official on the vorbis.com site to download?

Many thanks.

Vorbis absolute killer

Reply #43
Quote
So is this release only available for CVS or will it be made official on the vorbis.com site to download?

Many thanks.

The "release" hasn't happened yet. What's in CVS is work in progress. When the dust settles, everything will be packaged into an official release that will be made available for download.

Best regards,

Carsten Haese
Ogg Traffic Editor, Xiph.org Foundation

Vorbis absolute killer

Reply #44
Though CVS work aren't finished, I must admit than the biggest flaws are now removed. No more electric noise ; no loud chirping anymore on Stimmung sample. Bitrate, at -q0, is now 52 kbps (against ridiculous 27 kbps with previous encodeur). Nice work.

But with Benediction piano sample, things are not so nice. Latest CVS file seemed to be more noisy (slight increase - not annoying), and introduce an obvious and not welcome reverberation (more audible on sec. 07:00 - 11:00). I heard the same kind of strange problem during Roberto 64 listening test, using the same encoder (september, O7 - rarewares) on Sample 09, piano, and another one. It's not very pleasing (it sound like a piano recorded on an airport), and I never heard this before.

Vorbis absolute killer

Reply #45
Quote
Though CVS work aren't finished, I must admit than the biggest flaws are now removed. No more electric noise ; no loud chirping anymore on Stimmung sample. Bitrate, at -q0, is now 52 kbps (against ridiculous 27 kbps with previous encodeur). Nice work.

But with Benediction piano sample, things are not so nice. Latest CVS file seemed to be more noisy (slight increase - not annoying), and introduce an obvious and not welcome reverberation (more audible on sec. 07:00 - 11:00). I heard the same kind of strange problem during Roberto 64 listening test, using the same encoder (september, O7 - rarewares) on Sample 09, piano, and another one. It's not very pleasing (it sound like a piano recorded on an airport), and I never heard this before.

I think Xiph.org should employ our member, guruboolez, as one of their 'resident ears'. hehehe

Vorbis absolute killer

Reply #46
Quote
I think Xiph.org should employ our member, guruboolez, as one of their 'resident ears'. hehehe

Isn't Garf a resident vorbis tester, and developer too, with a nice hearing, and using sometimes abc/hr unlike most people in this board (no offense: simple report of a common, sadly but daily reality) :/

Vorbis absolute killer

Reply #47
Quote
Isn't Garf a resident vorbis tester, and developer too, with a nice hearing, and using sometimes abc/hr unlike most people in this board (no offense: simple report of a common, sadly but daily reality) :/

As far as I know, Garf is in no way linked to Xiph.org. I think all the work he has done on Vorbis was voluntary (which explains why a good deal of it hasn't been fold back into official Vorbis).

Vorbis absolute killer

Reply #48
About reverberation behaviour of Vorbis low bitrate encoding :

I searched in user comments of Roberto's test, and found that some users noticed it :
- schnofler on samples 1, 8, 9, 11 ("hollow sound")
- phong on samples 7, 8, 10 ("hollow sound")
- garf, on sample 9 ("hollow sound")
- mac (sample 7 : "reverberated guitar")
- anon1, on sample 06, 09, 10 (reverberation)
- anon11, on sample 09 (reverberation)
- myself (3, 5, 9 & 12) ("reverberation")

Other reference to the problem are about other encoders, but 80% of the phenomenon are related to Vorbis. Sample 09 (piano music) is the more often designed.
It seems that the latest CVS used in this test introduce an artifact that deosn't appear on previous version. I hope that this will be corrected on fixed 1.01