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: Is it safe to start archiving music with vorbis 1.0 @ -q 6?? (Read 15747 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

Is it safe to start archiving music with vorbis 1.0 @ -q 6??

I think that since this codec is improving somehow like 30% better quality/compression than mp3, my goal is to know if vorbis works ok at 192kbps, rather than using mp3 at 256 to archive music

Ill compare vorbis 1.0 to my current archive codec: AAC. i'll post results later...

Is it safe to start archiving music with vorbis 1.0 @ -q 6??

Reply #1 –
OKOKOKOKOK...

Sorry, i have to modify this cuz it wasnt true

Well vorbis at 224 kb/s performs well but extreme AAC wich is somehow 224 has a very subtle advantage. I have to point out that AAC files where bigger for 700kb so i guess its good, i just have to make up my mind since AAC will get very little hardware support and vorbis will have a wave of this...

Is it safe to start archiving music with vorbis 1.0 @ -q 6??

Reply #2 –
You got to love Vorbis. Being different is good. But not sucking at the same time is hard. Somehow they have done a miracle! They have taken the cast off scrap of other codecs and built a monster of a codec. Monty and the gang should take a nice long vacation after this. OMG! What am I saying?! They must stay inside and work feverishly towards version 2! I can't wait!!!!!!!

Is it safe to start archiving music with vorbis 1.0 @ -q 6??

Reply #3 –
But can it beat MPC at high bitrates? It's hard to find faults with mppenc when you start pushing 192kbps+ VBR.

Is it safe to start archiving music with vorbis 1.0 @ -q 6??

Reply #4 –
Quote
Originally posted by mithrandir
But can it beat MPC at high bitrates? It's hard to find faults with mppenc when you start pushing 192kbps+ VBR.

Nope. There are quite a few known examples of clips which MPC copes with, but Vorbis doesn't yet. MPC is still the king for Mr. L33t Audiophile.

Vorbis is 'good enough' for me, though .

Is it safe to start archiving music with vorbis 1.0 @ -q 6??

Reply #5 –
are you all talking about _blind tests_ or _groupthinking_?
personally i don't believe to people claiming "X is better than Y", if they didn't do a blind test.
if you are tooo lazy you can at least do a b.t. with winamp.

...i even talked to a guy claiming that copied CDs sound with smaller volume than original, and that he can feel this every time he listened to his stereo.

Byebye ~Mr.Sax

Is it safe to start archiving music with vorbis 1.0 @ -q 6??

Reply #6 –
i forgot to mention that using very similar dipswitches the aac filezise was about 700kb higuer everytime, this leaves no doubt vorbis is better

on the other hand, i have to reencode all my albums, so lazy :mad:

Is it safe to start archiving music with vorbis 1.0 @ -q 6??

Reply #7 –
I hesitate using Ogg for archiving. I'm a fan of electronic music, and on very electric sounds, almost all codecs fail by producing weird artifacts, which AAC doesn't. I can't await these wavelets in Ogg tho. If they do 'fix' that, I will switch to Ogg for archiving.

Is it safe to start archiving music with vorbis 1.0 @ -q 6??

Reply #8 –
A bit off-topic here, but jkeating, wasn't there something about CD-Rs giving "softer" volume than originals a while back? I know I've had problems with some of my older, cheaper CD-Rs (didn't sound as loud as the originals). Anyway, I haven't experienced this problem these last couple of years, so either the quality of CD-Rs have gotten better, or I used some crappy programs/settings back then...

Is it safe to start archiving music with vorbis 1.0 @ -q 6??

Reply #9 –
digital is just zeros and ones...

Is it safe to start archiving music with vorbis 1.0 @ -q 6??

Reply #10 –
Quote
Originally posted by fileman
digital is just zeros and ones...


This is true, but it's not the only thing that influences the sound of an audio CD. Other factors include: Block error rate (BLER), C11, C21, C31, C12, C22 and C32 errors (and the correction mechanism of the CD player), jitter, ...

Is it safe to start archiving music with vorbis 1.0 @ -q 6??

Reply #11 –
Quote
Originally posted by Tom Servo
I hesitate using Ogg for archiving. I'm a fan of electronic music, and on very electric sounds, almost all codecs fail by producing weird artifacts, which AAC doesn't. I can't await these wavelets in Ogg tho. If they do 'fix' that, I will switch to Ogg for archiving.


After doing some experiments in image compression in 1997 I'm nearly
absolutely sure than wavelets are a pure buzzword in Audio compression.

This has nothing to do with a statement pro-Vorbis or cons-Vorbis.

The same occures with Fractal compression.
--  Frank Klemm

Is it safe to start archiving music with vorbis 1.0 @ -q 6??

Reply #12 –
omg frank klemm himself!!!

seems agter hearing him ill stick to AAC for archiving...

Is it safe to start archiving music with vorbis 1.0 @ -q 6??

Reply #13 –
Quote
Originally posted by Frank Klemm


After doing some experiments in image compression in 1997 I'm nearly
absolutely sure than wavelets are a pure buzzword in Audio compression.

This has nothing to do with a statement pro-Vorbis or cons-Vorbis.

The same occures with Fractal compression.


They have been treated that way, yes.

However, ignoring buzzwords entirely and looking at them from a mathematical standpoint, they offer a grey area in the mapping rotation bteween time and frequency, allowing multiple levels of tradeoff within.  This is useful.

MPC is limited by its design because it is a time domain representation and as such, frequency is not something it truly encodes except by brute force.

Vorbis is limited by its design because it is a frequency domain representation and as such, can encode impulses and short time events only by similar brute force.

Wavelets offer a transform that allows compromize between the two.  This is a useful property with a specific application, not a magic bullet.  Most of the compressions that used wavelets as such a 'magic bullet' tried to replace the filterbank/transform entirely with a wavelet transform.  Wavelets proved to be no more useful than MDCT or subbanding when used alone and as both subbanding and transform codecs are far more mature, wavelets failed to impress.  This result should have been entirely expected.

In the context of Vorbis, wavelets potentially offer good compaction of time events when used in conjunction with the normal MDCT.  The human ear has seperate physical and neural paths for processing time and frequency, and the goal is to analyze, divide and represent the original audio component with two codec pathways, each mapped to the workings of the ear itself.

Is it safe to start archiving music with vorbis 1.0 @ -q 6??

Reply #14 –
Quote
Originally posted by Phobos
omg frank klemm himself!!!

seems agter hearing him ill stick to AAC for archiving...


In each blind test I've participated in, on this site, on r3mix, and in privately run tests, I've not yet heard an AAC encoder that managed to impress me even over rugged old mp3 (and by mp3, I mean LAME).  AAC has generally delivered the sharp attacks it promises, but the tuning of the QT MP2/MP4 and the Psytel encoder let artifacts creep into practically every other aspect of the encoding.  I'm not talking about slight problems; I'm talking about swirling, flanging and fluttering on the order of annoyingness with WMA's creeping lowpass and laser war artifacts.  In ff123's tests, I put AAC at the bottom with Xing, below even WMA. 

In fariness to AAC, these tests have all be mid-to-low bitrate; 64kbps, 80kbps and 128kbps.  It could be that so far, the AAC encoders I mention are all tuned specifically for very high bitrates, in the same way I expend most of my effort on Vorbis at ~128ish and below.  The last full test I'm thinking of was also ~8 months ago, and I know how far Vorbis has come in that time (however, in the informal 64kbps testing going on in one of the General threads, the current AAC offerings are still unimpressive).

However, the results at these bitrates have been consistently utterly unimpressive.  AAC has been around for years now, and I still pick LAME over it every time.  MPC [for high bitrate] and Vorbis [for low] make the decision a no-brainer.  Very odd.

Is it safe to start archiving music with vorbis 1.0 @ -q 6??

Reply #15 –
Quote
Originally posted by Phobos
seems agter hearing him ill stick to AAC for archiving...


Huh?  He said 'Wavelets are a buzzword'.  I said 'Historically, I agree'.  What does that have to do with Vorbis or AAC?  Neither Vorbis nor AAC use wavelets.

Is it safe to start archiving music with vorbis 1.0 @ -q 6??

Reply #16 –
Quote
Originally posted by xiphmont
Huh?  He said 'Wavelets are a buzzword'.  I said 'Historically, I agree'.  What does that have to do with Vorbis or AAC?  Neither Vorbis nor AAC use wavelets.


He said he's not satisfied with Vorbis ATM, Mr. Montgomery.

He was hoping that implementation of Wavelets in Vorbis could improve it in a way to make him get satisfied with it. After Mr. klemm pointed out that wavelets aren't all that people have been rumoring, he gave up waiting for them, and decided to stick with AAC.

Regards;

R.

Is it safe to start archiving music with vorbis 1.0 @ -q 6??

Reply #17 –
Quote
Originally posted by xiphmont
(however, in the informal 64kbps testing going on in one of the General threads, the current AAC offerings are still unimpressive).


I didn't expect anything better than that. It's a well known fact that both best AAC implementations "available" (Psytel and FhG) suck at low bitrates due to lack of IS implementation.

Quote
However, the results at these bitrates have been consistently utterly unimpressive.  AAC has been around for years now, and I still pick LAME over it every time.  MPC [for high bitrate] and Vorbis [for low] make the decision a no-brainer.  Very odd.


Well, I won't post the reasons I use AAC for the Nth time, people are surely getting tired of that. If you want to check them, you can use the search function.

But I'm pretty sure this decision isn't as much "no-brainer" as you think it is. (or you expected it to be)

Regards;

Roberto.

Is it safe to start archiving music with vorbis 1.0 @ -q 6??

Reply #18 –
Quote
Originally posted by jkeating
are you all talking about _blind tests_ or _groupthinking_?


personally i don't believe to people claiming "X is better than Y", if they didn't do a blind test.


I'd have to agree with this.  Phobos, you say there is a very "subtle" advantage, but you haven't given any more information which would go towards validating this claim.

Once you start comparing these codecs at bitrates above 160kbps you need to be very specific and use a wide variety of difficult test samples and blind testing to really come to some sort of valid conclusion.

Is it safe to start archiving music with vorbis 1.0 @ -q 6??

Reply #19 –
Quote
Originally posted by Tom Servo
I hesitate using Ogg for archiving. I'm a fan of electronic music, and on very electric sounds, almost all codecs fail by producing weird artifacts, which AAC doesn't. I can't await these wavelets in Ogg tho. If they do 'fix' that, I will switch to Ogg for archiving.


This is not true.  I listen to quite a bit of very complex and unusual electronic music, and AAC does have problems with this.  I've even pointed out a few flaws I've heard in PsyTEL on some of the samples I use to Ivan in the past.  That's not to say that AAC does really bad on this type of stuff (I've only tested moderate to high bitrates), but it's certainly not perfect.  I certainly find it to be better than LAME (again at moderate to high bitrates).

If you want the best quality on this type of music, I suggest you use MPC.  MPC has a pretty large advantage on this type of music, partially due to it's very time critical nature and the fact that MPC being a subband codec handles this particularly well.

For that matter though, back with Garf's tuned modes, Vorbis was handling quite a bit of this pretty well.  I haven't done a whole lot of extensive testing with Vorbis since then though, so I can't really say how well it does now.  However, I can only imagine that things have probably improved even more since then.

Is it safe to start archiving music with vorbis 1.0 @ -q 6??

Reply #20 –
Quote
Originally posted by CiTay


This is true, but it's not the only thing that influences the sound of an audio CD. Other factors include: Block error rate (BLER), C11, C21, C31, C12, C22 and C32 errors (and the correction mechanism of the CD player), jitter, ...


If you are a real anorak, you can read all about it here in full technicolor:

http://www.ee.washington.edu/conselec/CE/k...audio2/95x7.htm

PS - I am only read it 'cos I am forced to

Is it safe to start archiving music with vorbis 1.0 @ -q 6??

Reply #21 –
Copied from

http://www.hydrogenaudio.org/forums/newrep...ly&postid=24336

to

http://www.hydrogenaudio.org/forums/newrep...ly&postid=24336
 

Quote
Originally posted by xiphmont

However, ignoring buzzwords entirely and looking at them from a mathematical standpoint, they offer a grey area in the mapping rotation between time and frequency, allowing multiple levels of tradeoff within.  This is useful.

MPC is limited by its design because it is a time domain representation and as such, frequency is not something it truly encodes except by brute force.

Vorbis is limited by its design because it is a frequency domain representation and as such, can encode impulses and short time events only by similar brute force.


MPC has problems with pure tones to encode. It can encode it, but needs more bits than
good designed Transform encoders. (Average) Tonality of older music seems to be
higher than the tonality of modern very busy music you hear if you switch on the
amplifier and not switch off fast enough the FM tuner ;-)

So the advantage of transform encoders is not so high for transform encoders.
For medium (120 kbps and below) bitrates transform encoder has a smoother degrade in quality. When I would design a 128 kbps encoder I would use a transform, for 192 kbps
a subband encoder (2 ch stereo).

Quote
Wavelets offer a transform that allows compromize between the two.  This is a useful property with a specific application, not a magic bullet.  Most of the compressions that used wavelets as such a 'magic bullet' tried to replace the filterbank/transform entirely with a wavelet transform.  Wavelets proved to be no more useful than MDCT or subbanding when used alone and as both subbanding and transform codecs are far more mature, wavelets failed to impress.  This result should have been entirely expected.


What we call "transients" are usually normal sinusoid signals with a medium stationarity
(compared to the long stationarity of other signals). The are 98% sinusoid waves,
you can easily describe it by a sine with a envelope.

(all) wavelet (I know) are much more different, so an approach using
MDCT+some modifications is much better. Wavelet also have some problems in
image compression, but these problems become very dominant in the world of
mass and springs.

Quote
[/b]

In the context of Vorbis, wavelets potentially offer good compaction of time events when used in conjunction with the normal MDCT.  The human ear has seperate physical and neural paths for processing time and frequency, and the goal is to analyze, divide and represent the original audio component with two codec pathways, each mapped to the workings of the ear itself.


I would always use a MDCT based transform. May be some additional window techniques
and a more flexible window handling or really strange transform method (end-to-front
encoding). This I would use to handle transient signals.

A second possiblity is to divide the signal into tonal and noise signals and
use short cuts for the noise. Not in the way of MP3Pro, not in the way of AAC V2, not
in the way of AAC+SBR, but in a way that noise and tonal components are equal.


Info for Dibrom:

- HA is buggy. When sending a posting not fast enough (wait 30 min), I always get the
  information that the thread I want to post is invalid.

- Please remove the army of meta information avoid caching even when it's
  only a second old. This is a pain for non-ADSL users (you must have
  ASDL/direct Ethernet and please use a normal modem on one day per week).
 
- I'm missing a https version. Everyone can read my password and
  post using my name if he can catch a packet.
--  Frank Klemm

Is it safe to start archiving music with vorbis 1.0 @ -q 6??

Reply #22 –
Thank you spoon. This site is interesting.

Is it safe to start archiving music with vorbis 1.0 @ -q 6??

Reply #23 –
Quote
Originally posted by Frank Klemm

Info for Dibrom:

- HA is buggy. When sending a posting not fast enough (wait 30 min), I always get the
  information that the thread I want to post is invalid.

- Please remove the army of meta information avoid caching even when it's
  only a second old. This is a pain for non-ADSL users (you must have
  ASDL/direct Ethernet and please use a normal modem on one day per week).
 
- I'm missing a https version. Everyone can read my password and
  post using my name if he can catch a packet.


I'm in the middle of moving again right now (for the Nth time.. :/)

In a week or so I should have some more time so I'll be able to get back to work on the site update which features completely different software.  Hopefully that will work better for you.

I'll also be able to possibly start working on the mailing list integration again.  Unfortunately I have to start from scratch due to a problem with the person previously working on this for HA.

I'm trying to improve things here, I just have so little time to focus on HA right now.  You'll just have to bear with me I'm afraid.

Is it safe to start archiving music with vorbis 1.0 @ -q 6??

Reply #24 –
Quote
Originally posted by rjamorim


He said he's not satisfied with Vorbis ATM, Mr. Montgomery.


Sure, but his explanations are vague to the point that I have to wonder what he's really thinking/hearing.  Stating something carefully and methodically, with explanation and reasoning, generally persuades me to take the assertion seriously even should I later chose to disagree.  "OMG!  U R dissed!  I'll stick to AAC!" contains nothing of substance on its own merits.  Even if the reasoning behind that statement is sound, there's nothing in the statement to suggest any reasoning actually exists.

Quote
Originally posted by rjamorim

He was hoping that implementation of Wavelets in Vorbis could improve it in a way to make him get satisfied with it. After Mr. klemm pointed out that wavelets aren't all that people have been rumoring, he gave up waiting for them, and decided to stick with AAC.

Regards;

R.


Had you posted such a thing yourself (and you'd not have; you'd go into detail), I'd possibly have read some substance into it.  As it was, it looked to me much more like "I like to assert my favorite opinion!  Again!"  I've not seen anything to suggest he knows what a wavelet *is*. 

Mmm.  Social dynamics.  Mmmm, sleep-deprived, release-frazzled Monty,