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: Ogg Quality (VBR) (Read 8013 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

Ogg Quality (VBR)

Hi, Hope i can get a Pro advise:

In Mp3 i use

Razorlame
EAC
MP3Gain
MP3TRIM

And my definition in Razor are from r3mix.net (VBR) Wich i achieve good results.


But i want to change to OGG,

I can still use EAC to ripp to wave
I can use OggGain equal to Mp3Gain

But my problem is i have check OggDrop and he have CBR, ABR And quality manager the 1 to 10, what i want to know is, where is VBR settings, how can configure oggDrop to use from 0kb to 500kb (VBR) Or i have to stick to quality manager and choose a Constant Bitrate (i see q6 is equal to mp3 192kb).

Another problem, is there any soft equal to MP3 Trim????


I want quality with the best size possible.....I don't like Wma because of the Vbr, i want to change to Ogg but..........  VBR.


Thanks for all your reply's. B)

Ps: Grip for linux (i use debian and XP) is better then Eac, can i get better wave quality with Grip.

Ogg Quality (VBR)

Reply #1
Ogg Vorbis is only VBR codec to begin with?  You have quality settings instead of bitrate. Because some samples are more difficult to encode than others. When you use quality 6 for instance you may get quite low or high bitrate depending on the type of music (high for harpsichord for instance, and low for mono recordings). So I'd suggest you stick to Vorbis' quality scheme that is provided, it is a superior method to achieve results similar in quality.

But I guess you should be able to play with the switches (then you need to use the command line version) to set the highest and lowest possible bitrate and maybe even the average bitrate.

Quote
I want quality with the best size possible.....


Umm but how much quality you want determines the size. If you want transparent quality quality 6 should be fine with Ogg. However if you want the highest possible quality at high bitrates your best bet is MPC.

Quote
Grip for linux (i use debian and XP) is better then Eac, can i get better wave quality with Grip.


I bet this claim is very dubious. cdparanoia (which grip uses, right?) is a very good ripper though. But all rippers achieve the same quality (=CD quality) but they can skip (causing glitches, pops, clicks)... I imagine you're using grip on linux side and then use OggDrop under Windows to rip your music. If you use EAC on windows side with ogg encoder that would save you a lot of trouble (also in tagging) and would be at least as accurate.
The object of mankind lies in its highest individuals.
One must have chaos in oneself to be able to give birth to a dancing star.

Ogg Quality (VBR)

Reply #2
Thanks For the Reply.

Ogg is Vbr, thanks for solve my doubt, but for instance if i choose q6 wich is 192kb,so i will have Vbr from 0kb to 192kb but i'm sure any of my music (linkinpark and others) will need higher bitrates at some point, 256kb or even 320kb.

See my point, so with q6 i will only have to 192kb.


From command line i have to check it out if i can set minimum 0kb to maximum 500kb.


One more question: in ogg can i use join stereo for instance, in Mp3 i have read and listen that produce better results.


Thanks (Sorry i make a lot of questions,but when i'm not in it.......you know what i mean.)

Ogg Quality (VBR)

Reply #3
Quote
Thanks For the Reply.

Ogg is Vbr, thanks for solve my doubt, but for instance if i choose q6 wich is 192kb,so i will have Vbr from 0kb to 192kb but i'm sure any of my music (linkinpark and others) will need higher bitrates at some point, 256kb or even 320kb.

See my point, so with q6 i will only have to 192kb.


From command line i have to check it out if i can set minimum 0kb to maximum 500kb.


One more question: in ogg can i use join stereo for instance, in Mp3 i have read and listen that produce better results.


Thanks (Sorry i make a lot of questions,but when i'm not in it.......you know what i mean.)

Yuo got it all wrong
it's no going from 0 to 192kbits

i tries to stick the hole file on a average of 192kbits.
It can go up to 256 (or hgigher) if its need so.

The 192kbits is a guide...rule of thumb.
kinda like when doing ABR with lame buts it IS VBR

-- EDIT-
misspellings
Sven Bent - Denmark

Ogg Quality (VBR)

Reply #4
Ogg Vorbis q6 is not 192kbps. You are talking about the average bitrate I assume. Average bitrate (filesize in kb/duration in seconds) usually is around that bitrate range. However it would go much higher than that when it is necessary as well as it would go lower when quality is achieved with lower bitrates. So I don't see your point? I guess you don't know how Vorbis works that's why. I'd suggest you read the previous threads...

Quote
i'm sure any of my music (linkinpark and others) will need higher bitrates at some point


The encoder is meticulously tuned, it usually knows very well where higher bitrate is necessary. Much better than one could estimate. So don't be surprised if any of your music that you'd consider to require high bitrate, actually does not.

I don't know what your reasons were in picking Ogg Vorbis. And Ogg is the name of the container, so Vorbis is a better name. But you have other alternatives like AAC and MPC. I'd suggest you compare them too.

Quote
One more question: in ogg can i use join stereo for instance, in Mp3 i have read and listen that produce better results.


Vorbis uses joint stereo by default. But passes to lossless stereo for >q6 AFAIK.
The object of mankind lies in its highest individuals.
One must have chaos in oneself to be able to give birth to a dancing star.

Ogg Quality (VBR)

Reply #5
Quote
Vorbis uses joint stereo by default. But passes to lossless stereo for >q6 AFAIK.

So q5.99 uses joint stereo and q6 "pure" stereo... then q5.99 could be better in SOME situations... ???

Ogg Quality (VBR)

Reply #6
First, I am speculating based on what I read before on HA when I was using Vorbis. Now I use MPC and am not sure at what quality Vorbis uses lossless stereo.

However q6 quality should be better than q5.99 because lossless stereo is theoretically better than joint stereo although comes with higher bitrate expense. If you observe the quality/bitrate graph of Vorbis (you can find that somewhere here), there's a jump at 5.99->6 because of this reason.
The object of mankind lies in its highest individuals.
One must have chaos in oneself to be able to give birth to a dancing star.

Ogg Quality (VBR)

Reply #7
First, Vorbis uses no 'joint stereo', only normal, 8-, 4- and 2-phase and point stereo.
I don't know how Vorbis uses its stereo modes in ABR, but it uses all lossless stereo at quality >=6.
In lower quality levels it mixes stereo modes.
ruxvilti'a

Ogg Quality (VBR)

Reply #8
Uhm... i thought vorbis uses lossless stereo (which is some joint stereo implementation AFAIK) for q >= 6.0 and lossy stereo (something like mp3's intensity stereo?) for the q levels below...

Ogg Quality (VBR)

Reply #9
The losless stereo is like a implemetation of joint stereo but it guarantee a perfect chanel separation like a normal stereo encoding and offer more compresion. The lossy stereo is implemet using de point stero( a Phase stereo mode, the 8 & 4 pashes modes arent implemented in v1.0 but they are in rc2, 2 phases is a point stereo mode). This is realy quite confusing but in low bitartes uses a special mix of the two modes but i don't now the determination of this and how it work.

Ogg Quality (VBR)

Reply #10
You know, I keep reading these threads, and keep wondering if I should switch to OGG...

I've been a dedicated MPC user since Buchmann was working on it, and my entire collection (minus a few albums downloaded) is in that format.  I believe that it is a superb music archiving format, and I'm sure many here will agree with me.

However, recently I decided to test out OGG...and I must admit I was really impressed with the quality it has attained.  I'm also totally amped about the fact that it is now getting portable support, as this is a big issue with me.  I listen to music at work, where I normally have a simple portable CD player, but I'd love to be able to fit several albums on a portable, without transcoding the music off my computer.

I wanted to ask, just how active is OGG development?  And at what settings was it designed to encode at?  For instance, MPC can achieve great results at roughly 150-180kbps (average) on most music, going up towards 200+ on a few tracks.  Is OGG similar?  Or was it designed to average out at a more conservative number, say 128, or even lower bitrates?  I've searched around on the forum, but there just doesn't seem to be anyone asking what OGG was designed to run at, just discussions about it's relative quality at various bitrates.

I don't want anyone posting about how "1337" OGG is, I just want some objective answers (I don't mean to sound snappy, but these boards do get a lot of that).  Thanks so much!

Ogg Quality (VBR)

Reply #11
Quote
I don't know what your reasons were in picking Ogg Vorbis. And Ogg is the name of the container, so Vorbis is a better name. But you have other alternatives like AAC and MPC. I'd suggest you compare them too.



I must confess that i'm not a pro in Ogg and i want to get some info, because in first is open source like Xvid for instance (mpeg4 codec that i like much) and what i have read is that it's superior to mp3, i think that i come to the right place to gathering info about it 


Ok, i get it now how it works in quality manager, i get it how it use JS OR Lossless. Thanks.


MPC (I have heard that only works good in HighBitrates), but i have to inform myself better about it.


Thanks for all your time.

Ogg Quality (VBR)

Reply #12
If you are using Ogg Vorbis to get near transparency, I'd suggest using GT3b1.  Our wonderful member here, Garf, has done his own tuning from q's of 5 to 10 and implemented a very effective transient detector which is very liberal at upping the bitrate in complex parts of music, esp. parts where pre-echo is common in Vorbis 1.0.

For example, I have a track which has lots of guitaring and cymbals in the background.  Vorbis 1.0 at q5 compressed it at an average of 165 kbps while Vorbis 1.0 GT3b1 was at around 200 kbps.  This shows that GT3b1 knew that it had to use more bits to encode those pre-echo-prone parts of music.  Second example is classical music which tends to have more silence and less of these sudden 'attacks'.  GT3b1 at q5 compressed it at around an average 146 kbps.  So you can see just how well it does in order to maintain constant quality.

Vorbis 1.0 from Xiph.org is more predictable in terms of quality-bitrate so it tends to be more conservative at upping or downing (is that a word? ) the number of bits.  So if you are concerned about predictable filesize, I'd suggest using Xiph.org's Vorbis 1.0 CVS.  Otherwise, if you are more concerned about constant quality and reducing pre-echo, I strongly recommend Vorbis GT3b1.

EDIT:  I'll speculate a bit here based in theory.  If you are looking for the best quality achieved for a certain bitrate for music which is rich in sharp transients, then MPC is the way to go I guess.  The subband paradigm of MPC allows good localisation in time as well as sparse representation of sharp transients and singularities.  On the other hand, pure transform coders like Ogg Vorbis require a considerably larger bitrate (also accompanying the reduction of block size to deal with pre-echo) to represent those sharp transients at a certain fidelity compared with subband coders (mpc).  Conversely, transform coders have excellent frequency localisation so music which is very tonal and slow moving I assume would be compressed with lesser bits using Ogg Vorbis.  Again, I'm just speculating here based on theory and intuition.  I haven't done any tests to verify this.

Ogg Quality (VBR)

Reply #13
@silver_cpu

I agree with QuantumKnot that Garf's tuned Vorbis encoder is certainly superior to straight Vorbis in many situations.  B)

You really need to do some checks yourself though. When I dabbled in various forms of Vorbis, I found it had a noise problem which bugged me. I don't have the threads readily at hand, but I know that Guruboolez posted some tests regarding it, and one or two other members have as well here in HA. In my own case, the noise was apparent over a wide range of bitrates, so I could not rate Vorbis as being truly transparent at the time I tested, even at high bit rates. This was all with Vorbis I and the GT versions, but there may be some other compiles/release candidates that do not have this problem. 

Some people don't seem bothered by it, and it may vary across the different versions of the encoder available, but I've moved on now to other formats, so you need to check this for yourself.

As you know already though, despite it having no portable support, MPC is a very, very hard act to follow quality wise. Vorbis is also slower to encode/decode if this bugs you.

Den.

Ogg Quality (VBR)

Reply #14
Quote
MPC (I have heard that only works good in HighBitrates), but i have to inform myself better about it.

it really depends on your definition of high bitrates.  I view high bitrates as above 192kbps.  I think generally high bitrate is percieved as above 128kbps.

If 192kbps isn't a problem than neither should be mpc -q6 (standard) which seems to average to me somewhere around 170-180kbps.

In my opinion if you are not concerned about open source or future portable/hardware compatibility and you are just going to listen to music on your pc or burn it to cd then mpc is the best choice for transparent audio.

If you are into open source software and/or future portable/hardware player support then ogg will do you good.

If you need maximum compatibility and don't care about gapless playback then you can always stick with mp3 and encode using lame --alt-preset standard

Ogg Quality (VBR)

Reply #15
Edit: what you said, QuantumKnot

Well, I guess that since OGG Vorbis tends to give me the same bitrates as MPC, and because of the way it's designed, it's not truely what I'm looking for in an archival format.  If the files weren't so large, I'd just use OptimFrog (probably dual stream).  I can say for certain now that, without more development and tuning, I will not use OGG Vorbis for archiving, as it doesn't seem to be completely suitable for this task.  Maybe I'll just stick to transcoding MPC -> MP3 for my cheap-o portable player, and see where MPC/Vorbis goes in the future.

Ogg Quality (VBR)

Reply #16
Well, it's all up to the ears.  For me, Ogg Vorbis is good enough for my crude hearing so MPC is sort of like overkill for me.

Ogg Quality (VBR)

Reply #17
For the same audio quality at high bitrates, an audio file is smaller with mpc than with ogg vorbis. In this case, vorbis, not mpc, is overkill for you.

Quote
windoze9x: If 192kbps isn't a problem than neither should be mpc -q6 (standard) which seems to average to me somewhere around 170-180kbps.


Erm, the standard preset in mpc is not -q6. It is -q5. And as we all know, -q5 is THE setting to start using on the vast majority of music when encoding in mpc, reserving higher settings only for troublesome files.

Note: Sorry for being quite off-topic here, just needed to clarify some important things.

Ogg Quality (VBR)

Reply #18
Quote
For the same audio quality at high bitrates, an audio file is smaller with mpc than with ogg vorbis. In this case, vorbis, not mpc, is overkill for you.

That is a very general statement since quality is subjective.  To my ears, q4 Vorbis sounds no different to q5 mpc.  My point is that I already started off archiving my music in Vorbis so there isn't much point redoing them all in mpc for a negligible increase in quality that I cannot notice.

Ogg Quality (VBR)

Reply #19
Quote
To my ears, q4 Vorbis sounds no different to q5 mpc.

Then try q3 MPC

Quote
there isn't much point redoing them all in mpc for a negligible increase in quality that I cannot notice


Yeah reencoding takes time. But the difference in space consumption is immediately noticeable.
The object of mankind lies in its highest individuals.
One must have chaos in oneself to be able to give birth to a dancing star.