Well i've got over 300 cd's i want to rip

I've tried 192kbps mp3 to no avail, and 128kbps AAC, after that i decided to switch back to ogg

I've kept all my cd's ripped in 64kbps vorbis to save as much space as possible, but until a few weeks ago i never realized all the quality i was missing out on

My dilemma is I have a 40gb hard drive (only around 35 gigs after formatting and the os) I want to fit all my CD's on there and have room for many more later on

-q 0 is barely acceptable for me now ...i'd rather try a higher quality....-q 4 offers me a huge diffrence and would only take up apporx 20 gigs for 5000 songs...but i can tell a slight difference between -q 4 and -q 6 (significant enough that i'd prefer -q 6), only problem is at 192kbps 5000 songs would use almost all my space (hence my dilemma)

I'm using the "Oggenc2.3 using libVorbis 1.1.0 with IMPULSE_TRIGGER_PROFILE Option" encoder

any suggestions or counseling you can give me to help me overcome my audio pickiness?

If quality 0 is "barely acceptable" to you, I find it odd that you consider quality 6 significant enough of an improvement over 4 to prefer the former. Sounds like the placebo effect rearing its head again.

Quality 0 is rather degraded compared to the original WAV source, but 4 is much much closer to the original...to the point where if you have space concerns you shouldn't have to bother with 6.

And it seems weird that you've been listening to quality 0 all this time (presumably with an encoder version that doesn't have aoTuV's latest low bitrate tweaks) and yet you are worried about having to resort to 128kbps instead of 192kbps.

Vorbis suffers from fairly steep diminishing returns (IMO). Once you get to quality 4.5+, you really don't get much more performance for the extra space needs.

If you want the best bang for the buck, you should look at Musepack and encode at quality 5 (175kbps average). But if you want to stick with Vorbis (nuthin' wrong there) I recommend you stay in the 4 range, given you want to limit space consumption. You can always use 4.99 or 5.

There is no dilemma,
Use Vorbis 1.1 Case Build at Q5,
This is what is actually the quality vs size optimized setting with vorbis, for your use.

People that rip at Q4 are ripping CD as if it were DVD, they will realize their mistake later.

People that rip at Q6 are ripping OGG as if it were MP3, they will realize their mistake later.

Ripping Music at Q1-2-3-4 is a major mistake as you wasted quality & you are in the bad range for music, the +16kbps by Step Range ==> you will have to re-rip.

Ripping music at Q6-7-8 is a minor mistake as you only wasted space, but you are in the good range for music, the +32kbps by Step Range ==> you will not need to re-rip.

People using Q4 & lower for music are deaf or noobs,
people using Q6 & upper for music are paranoid or golden eared,
People Using Q0-Q9 for music are stupid.

Don't use impulse_trigger_build ... there is no gain if you care for quality vs size.

It's for paranoid or golden eared people that would use Q6 & upper anyway just to get the "warm & fuzzy felling" that their rips are so much better than official Q5 ...

People that use Impulse_Trigger are people that use Lossy as if they would use Lossless, they will realize their misstake later.

Ripping with Impulse_Trigger is like ripping at Q9 for me,
if you are ready to sacrifice space for a virtual quality gain,
use Flac 1.11 Q8 instead of Impulse_Trigger ...

You can't be wrong if you use official between Q5 & Q8, now the choice in between relies on your earing skills & your HD size.

Hope it helped,
Cya.

i tried musepack earlier today too...pretty impressed

but i want something compatible w/ a portable player....i'm looking at getting a iRiver H340

I think i'm gonna stick with -q 4, since you're probably right about the placebo effect...hopefully i won't change my mind and delete all my songs again

well i tried a couple ABX tests w/ foobar2000

i flunked them all, even comparing as low as -q 4 to an original wav

i obiviously don't have great hearing

I'm going to try MusePack --quality 5 for awhile since it uses less cpu/ram to decode (i'm big on saving resources) and since a lot of people consider it to be transparent

if i run out of space i'll have to figure something else out

You are very selfish with this statement.
I mean most of HA and all over there world accept q3&q4 as very listenable and so do I.
You should differentiat the statements: listenable and transparent.
EG if I'm going to rip my CDs in q2 I don't want to transport them and I don't want to archive them.
I my opinion it's just different… There's a big thing I learned at HA.org:
There's no "one for all" solution if you know what I mean.

An icepick in the ear?

For all the time you will spend reripping, I suggest popping for a bigger hard drive. Shop the sales and HD space is typically US$0.50/GB for 3.5 inch (ie desktop) drives.

Edit: Things destined for a portable player I generally have in q2 through 4. For flash (tight space and non-critical listening) generally q0 is fine.

Firstly, can I say that without using an explicit switch, ITP is not activated during normal operation, so the ITP build is safe to use and you will get exactly the same encoding experience as a non-ITP build.

Secondly, I don't think I have ever suggested that ITP gives a quality gain for golden-eared people. It was introduced to alleviate smearing with microattack samples. Vorbis has trouble with the block switching with these samples. So it was designed to fix problems rather than to improve quality. I don't have golden-ears yet I was able to easily abx the smearing on some microattack samples such as this one at q 4, unless you are someone who has some sort of bias against third-party vorbis development or belong to the same category as those listeners of vorbis at q 4

Not only is it selfish. I find that statement very insulting as well. Quality and transparency is subjective and differs for certain people. You cannot call people who listen to q 4 as being deaf.

Quantumknot:
I know ITP is not activated by default, but I don't know how to easyly detect for sure if ITP is used or not except by checking if size is bigger or quality better, so I regret the times when people where using GTB, which I could discard easier.
Don't have anything about aotuv, just don't like any GTB like tuning, incl. QK MegaMess ... this is not new ... been arguing real interests of minor tunings with Garf before QK ...

"Not only is it selfish. I find that statement very insulting as well. Quality and transparency is subjective and differs for certain people. You cannot call people who listen to q 4 as being deaf."
I started ripping at Q4 myself & I was very happy till I realize 128kbps was not in the +32Kbps range & re-ripped all ...

Q4 is very listenable, & there are clever use for this setting:
Use Q4 to rip with basic DVD with 1024kbps for video
Or Use Q4 to rip with portable device in noisy environment with limited HD
(& keep an ogg Q5 or flac Q8 version for PC backup)

I have nothing against people that willingly sacrifice quality for space in these two case ... still I have problem when these 2 kind of people claim Q4 is a transparent bitrate for PC backup ... it is not ... & will most likely never be ... at last with vorbis I ...

as I already explained there is not a single perfect bitrate for me ... but only a good range of bitrates for a specific use ...
"most likely transparent" for PC backup Use
==> +32Kbps by step range
"most likely not transparent" for Portable Device Use + Basic DVD ripping Use
==> +16Kbps by step range

... just don't mix the two ...

indeed transparency of Q4 is subjective ... but non-transparency of 80kbps is much less arguable ... & 80Kbps is in the same +16kbps range as 128Kbps ...

so I ain't gonna use 160-256kbps because I always like these settings & add 128kbps which is an orpheline bitrate from +16kbps range just because sometimes 128kbps sounds like 160kbps ... I sacrifice 128kbps on the shrine of logic ...

All my apologizes if the intelligence of any Q4 users as been insulted by my stupid statements

That's all folks ...

Well I for one find -q4 transparent for me. Tell me why it makes more sense to waste space with a higher quality level when it is just as transparent to the user as a lower quality level? The whole reason for lossy codecs is to save space by compressing to the lowest file size while maintaining transparency for the user. You defeat the purpose of a lossy codec with your logic.



WTF?

All this is about choice & priority:

------------------------------------------------------------------------

Quality Oriented People: (Me)

1st Priority Quality:
080-096-112-128Kbps = +16kbps by step range = most likely not transparent on most samples for most people
160-192-224-256Kbps = +32Kbps by step range = most likely transparent on most samples for most people

If you care for quality mainly, you will use 160-192-224-256Kbps,
as it's always better than 080-096-112-128Kbps

2nd Priority Size:
Chose the lowest in 160-192-224-256Kbps
It's 160Kbps

-----------------------------------------------------------------------

Size Oriented People: (You)

1st Priority Size:
080-096-112-128Kbps = +16kbps by step range
160-192-224-256Kbps = +32Kbps by step range

If you care for size mainly, you will use 080-096-112-128Kbps,
as it's always smaller than 160-192-224-256Kbps

2nd Priority Quality:
Chose the biggest in 080-096-112-128Kbps
It's 128Kbps

------------------------------------------------------------------------

So some people will favor 128kbps as it's smaller & some people will favor 160kbps as it's better ... no one is right no one is wrong ...

What is important is to know why you use this specific bitrate for this specific use,
people using 128Kbps & justifying it by quality argument are wrong ...

you can use 128Kbps but plz justify it by saving size, not by quality claims,
the only thing you can claim about 128Kbps is that it's enought for you ...

the transparency range for most people on most sample is 160 to 256Kbps,
don't claim 128Kbps to be in this range ... it is not & will never be ...

now feel free to use 128Kbps if you think the difference in size between 128 & 160 is too big ...

but my personnal choice is that the difference in size between 128 & 160 is not worth switching to a lower range ... specially as even if I would use 128kbps I would not use 80-96-112Kbps for CD backup ...

so I favor staying in the +32Kbps range ... specially as I use 192-224-256Kbps & I sacrifice 128Kbps ....

switching from the +32Kbps range to the +16Kbps range for CD backup use is a major switch IMHO ... actual codecs are not ready for CD backup in the +16Kbps range ... & 128Kbps alone is not enought to justify using the +16Kbps for CD backup.

now if quality isn't a priority, yes 080-096-112-128Kbps can be used ... specially for basic DVD & portable device ... but for CD backup let me disagree ...

I have to agree that the perfect range for Vorbis is between Q4 and Q5. Going above Q5 doesn't make much sense IMO. You won't get transparent quality even with Q6 (it's much worse than mpc standard, yet has bitrates comparable to mpc xtreme). And anything above which might give you minor quality improvements is useless because it's just a huge waste of space.

I disagree with you in all.

I encoded many files in q6 [1.1RC1] and mpc extreme (1.14).

To the same sources wav files (music in general):

In my own experience, filesize of mpc extreme files and ogg [1.1RC1] in the quality range of q6,7 to q6,85 are very similar; Of course, depending of type of music.

On the other hand 1.1RC1 filesizes in q6 is very similar to mpc standard; in the worst case, ogg bitrates is aprox. 6% more big than mpc standard.

Sound quality??? very interesting point...

I compared some music files (not killer samples) with mpc standard and 1.1RC1 q6 in ABX tests (jazz and rock material)

[1.1RC1 (-q6) vs. mpc standard (1.14)]

RESULTS: I don't hear differences

Other people with better ears probably will hear differences, but for me no.

As a reference, my ears have tuned and checked audio amplifiers succesfully for more than 10 years...

Q6 is undoubtedly worse than mpc standard. And on my music, mpc standard bitrates are comparable to Vorbis Q5 bitrates.

And if Q6 sounds fine to you, I'd suggest you try if you can ABX Q5.

ITP is an advanced encoder option. There is a one in a few billion chance that anyone with butter fingers will accidentally enable it as they are typing into oggenc.



The purpose of a true VBR codec is maintain constant quality, not bitrate. GT3b1 produces sharper transients than the official builds at q 5 and above. You may not like that sort of tuning, but why assume that everyone else will be interested in constant average bitrate rather than constant quality? Why do you assume that others will like to listen to smearing on microattacks and tell them to avoid ITP?



To you, it may not be transparent. But to some others, it may be. Its possible that qualities which you find transparent, may not be transparent for people like guruboolez. So there are no absolutes here. It's relative. Hence I don't see any justification with having a problem with other people's hearing abilities.



I dont think we are on the same wavelength here. What is a "step range"?

I think he's talking about the fact that the with Vorbis, q-2 through q4 have a nominal bitrate increase of 16kbps with each step up through the scale. While q5 through q9 have a nominal bitrate increase of 32kbps.

I have to agree with everyone else though, I don't see the logic in OggZealot's argument at all. Almost all of the scenarios you are describing apply to you. Everyone has their own hearing capabilities and equipment. The way you are describing Vorbis is irrational. The quality settings are there for a reason, the nominal bitrate is just a target.

Vorbis is still a VBR perceptual codec unless I missed the boat on this one. I personally perceive that q4 is fine for me in most situations. You might think otherwise, but this doesn't make me deaf, it probably means I don't thouroghly examine my music with microscope, I listen and enjoy.

Now to reply to the OP:



See if you can't abx q4 why would you worry about possibly wasting more bits? Down the road you might find a problem song/sample on a certain album, couldn't you just retest/rerip that album?

I don't know I saw myself going through something similar a while ago. Now for transporting music to my other computers (work etc,) I transcode from FLAC to vorbis q4.

well on some songs i can't even ABX q0 but there are are a few that sound like crap at q0

i can't abx q4, but i feel like i'm wasting space

i wanna try q1-3 but i need a problem song that is really sensitive to the quality changes, any suggestions?

I'm ripping these cds so i can stream them of a samba/http server so the lower the bitrate the better

EDIT: Think i found the perfect song to ABX

i can easily tell the difference between the wav a q 0-2 so far

Best "Range of Bitrates" for Quality:
The +32Kbps Range = 160-256Kbps

Best "Quality Vs. Size" Step in this Range:
160Kbps

indeed in you don't perceive the big holes between Q0 & 1, Q4 & Q5, & between Q8 & 9 each corresponding to a switch of range, yes all downward steps are equal & all I explain is meaningless ... but that just mean you don't understand the strategic steps of bitrate ...

what are we arguing about ? 128 Vs. 160 as the best quality VS size ratio ...
we will not solve the dilemma ... & that for a very long time because

128Kbps is the "Highest Quality" Step of the "Lowest Size" Range (+16Kbps Range: 80-128Kbps)
160Kbps is the "Lowest Quality" Step of the "Highest Size" Range (+32Kbps Range: 160-256Kbps)

so the difference between the 2 is very slim ....
if the question is "What is the optimized Quality Vs. Size bitrate for Vorbis?"

then the answer is BOTH 128Kbps & 160Kbps BUT:
128Kbps is the optimized Quality Vs. Size Ratio if YOU care for Size Mainly
160Kbps is the optimized Quality Vs. Size Ratio if YOU care for Quality Mainly

Both 128Kbps & 160KBps are strategic point in the bitrate scale ... & both can be defended cleverly ... but the use of 128Kbps is explained mainly by saving space not by quality ...

if 128Kbps is sometimes as transparent as 160Kbps it is normal as it is just the bitrate "next door" ... but if you take the ranges as a whole "80-128Kbps" can't be said to be as transparent as "160-256Kbps" ... & 128Kbps belongs to "80-128Kbps" aka +16Kbps range ... so I sacrifice it ...

all is about choice ... yours are not mine ...
but the difference between you & me is that I know why & can explain exactly why I use 160-256Kbps ... & not 128Kbps

I am not sure the reason why your favorite bitrate is your favorite is as clear as mine in your head ... & here is a very good hint of it :
Quote Zoom:
"While q5 through q9 have a nominal bitrate increase of 32kbps."

Q9 is NOT in the 32Kbps Range, your misstake shows you don't perceive the logic switch ... there is an architectura behind bitrates, you can't claim to understand bitrates if you don't understand the structure behind & just say "well, today 128kbps is transparent for me"

both listening tests & bitrate structure understanding are complementary,
it is NOT random if we are all so attracted by 128 or 160Kbps

these are both "near transparency point" for most people AND "strategic Quality Vs Size ratio" point

if you understand the logic hole between 128 & 160, & decide to that's it's ok, then it's ok for me ... you chosed size over quality, & you chosed that quality of 128kbps was worth a switch of range.

now let me ask 2 questions:
1: What is the min bitrate under witch you would delete a rip because it's crap ?
2: What is the max bitrate above witch you would delete a rip because it's overkill ?

My personnal answer are:
1: 160
2: 256

which by a wonderfull randomness (which is not random AT ALL if you understand what I explain) correspond to the +32Kbps Range of Bitrate ...

128 is the upper edge of 80-128 ... using 160-256Kbps + 128Kbps ... you use 128Kbps as an orpheline bitrate ...

I just want to be sure you are aware of it ... I started ripping at Q4 myself we all make the same error ... because we think that because Vorbis is superior to MP3 we can switch from 192-APS to XXXKbps much lower ...

& as the big standards for MP3 were 192 for CD & 128 for DVD,
when we look downward we all think of 128Kbps first because it's a familiar bitrate for us ... but it's an error Vorbis is far superior to MP3 in the way that it produces "much more listenable than mp3 but still not transparent" quality at lower bitrate ... but being listenable doesn't mean it's transparent ... vorbis has just slightly moved downward the transparency point even if it has highly moved downward the global listenability at lower bitrate ...

you were using lame APS APE before vorbis don't you ? so you were in the +32Kbps range, vorbis just brings you the perfect APM setting lame only added recently (recent considerig the long history of lame ...) the APM setting that no one use, just by habbit of using APS ...

so Vorbis is the perfect lame ... now you can use 128 for DVD & 160 for CD ...
but don't ask Vorbis to divide all bitrates by 2 & start using 128 for CD ...

80Kbps is not transparent & for a long time & this is why I don't use 128Kbps for CD ...

indeed explaining the fact that I don't use 128Kbps by the fact that 80Kbps is not transparent may sound like an heresy for you all used to Guru & Roberto listening tests ... & I understand it very well ... but I started thinking "by range of bitrate" because I realized that there is no perfect bitrate ... all the flaming wars about this bitrate being better than this bitrate are boring & are the result of the fact that we all don't think

1:by use (CD Vs DVD ...)
2:by priority (Size Vs Quality, PC Vs Portable ...)
3:by range
4:by bitrate

many many people only think by Use & by Bitrate (myself incl. in the past), & that is where the miss-understanding lies between us ...

When I started using vorbis, I had to argue with people using Q6 because they were using lame 192 for CD ... so "yesterday 192 was transparent for them"

Now the same kind of people have realized vorbis is superior & they start using Q4 because they were using lame 128 for DVD ... so "today 128 is transparent for them"

this is funny & pathetic at the same time ... but I don't blame them ... we all follow the same path of learning ... & the truth lies in between ...

128 lame for DVD was listanable, & it is much more with vorbis
160 lame for CD should have been a standard with APM but APM came too late

Vorbis is a perfect lame successor, prize it for what it is, but don't ask miracle ...
Monty is not Jesus ... not yet at last

The original poster is concerned about space so you have to give bitrate a lot of consideration.

Lossy codecs are all about compromise. That's why we use them: give me 95+% of the performance at 1/6th of the size of lossless. The relationship between size and performance is not linear, however. It's more like this:

Performance - Size
30% - 5%
50% - 10%
75% - 12%
90% - 14%
95% - 16%
98% - 20%
99% - 30%
99.5% - 50%
99.9% - >99%

Diminishing returns. Therefore if space is a fairly important constraint in the equation it makes no sense aiming for 99+% performance. While lower bitrates will provide reduced performance that doesn't necessarily mean reduced utility. Just because you can ABX between an original and lossy sample doesn't mean the lossy sample is unpleasant. It's just different/less accurate. But you are getting big space savings by giving up transparency...and if you don't have limitless space, the compromise is often worth it.

For Q0, try to resample (SSRC, of course) @ 32000 Hz... IS the best quality/size choice !!!

Okay, this thread has gotten a bit out of control.

This is a highly subjective topic, but the general rules of HA.org still apply, so if you are making any statements about quality please make sure those are backed up by blind tests.

OggZealot seems to have serious problems understanding Vorbis and makes up his arguments using some sort of weird pseudo-logic, so I'd advise you to just ignore whatever he posts.

Yeah but you gotta give him that: He's definetely got passion for it!

Come on ... your way of thinking will lead you to lower your favorite setting each time a new vorbis version will get out ... so you started ripping Q6 with vorbis 1.0, you tested Q5 with 1.01 & now you think Q4 is the best with vorbis 1.1 ...

the above way is of thinking is "logic at first look" but in the end it is absurd ...

indeed codec efficiency increase ... that doesn't mean you have to change the way you rip every six months ...
what are you gonna use with next vorbis version ? 112Kbps Q3 ?

just look at lame history ... once APS has become a standard, no one have switched from APS or CBR 192 to a lower setting over the time ... IRC teams channels leaders have chosen APS & now it's APS forever ...

I am not locked to +32kbps range forever ... but I will consider changing my way of thinking only when a major increase at low bitrate will happen ...

& that will most likely happen with vorbis II now ...

it seems IRC teams channels leaders have chosen Q5 for vorbis I CD ripping & it's gonna be Q5 for a long time ...

your logic of always going down as codecs efficiency increase is right, I am not telling you that you are wrong, I am just trying to explain to you that you must put a limit somewhere as always lowering transparency point just for minor codec updates is a myth & this limit is actually the frontier between bitrate ranges ...

anyway ... topic closed ... use whatever you want ...

Hey OggZealot, where does -q4.5 fit in? The lowly +16kbps range "never transparent" or the magical +32kbps range "wohoo transparent audio!"?

Q4.5 aka 144Kbps is in the +32Kbps range so it's ok but unusual, ... I never said the +16Kbps range was never transparent did I ? if I did it was either a misstake from me or a miss-understanding from you ...

I said when the upper bitrates of the +16Kbps Range are "as transparent" as the lower bitrates of the +32Kbps Range that doesn't mean you have to use the upper edge of the 16Kbps range ... because doing so you are pushing the codec to its limit, & you may realize later that there are clear artefacts in your "power-user 128kbps rip" ...

so if you want a little safety it's better to use 128kbps for case when you know there will be loss anyway: like lossy source=DVD or like noisy environment=Portable Device ...

if your source is lossless, & your listening environment/equipment is ok then you may (like me) want a little safety & favor 160 to 256kbps ...

As I don't think keeping both 128kbps for portable & 160kbps for pc is clever ...

Personnaly I use
- 128+1024Kbps for lame+xvid (maybe for vorbis+theora later but I may use 1280+160Kbps if HD-DVD source worths it)
- 160kbps for portable
- 160 to 256Kbps for pc
- flac Q8 (wv, ape, shn ...) for Hi-Fi ...

I don't have a 40gigs HD ... I have more than 500gigs of HDs ... so what is 10meg more or less for me ...

using 128kbps as orpheline to save a few megs just don't mean sense for ME ...
but that's just me ...

... arg I answered too quick & realized I get catched in my own game LOL
it's
] 40-48-56-64 ] 80-96-112-128 ] 160-192-224-256 ]

so Q4.5 is in the +32Kbps range [Edit: It's Corrected Now]

I have Q5.5 , Q6.5, Q7.5 on my HD ...
I don't have any problem with 144-176-208-240Kbps

I should have some Q4.5 too ... but it just seems no-one use it ...

you're right, following my logic 144Kbps is a very good bitrate
so I recall now shortly asked myself if I should use it or not ...

I just don't use it because no-one use it ... using lame for so long has just made everybody use 128, 160 or 192 ...
... you're right if you "think vorbis" instead of "thinking lame" then 144kbps is not the "horrible bitrate" it seems instinctivly ... even for me ...

once the world will have understund my crazy theory we'll all use 144Kbps ! rotfl ...

nice to see at last one guy understand what I say, congratulations breez !!!

Edit: In fact I favor rounded settings because if you start using half-setting ... you start deconstructing the exponential scale to a linear scale ... & to push further you could use quarter-setting as well ... doing so you destroy the meaning of the exponential scale ... & you destroy all my theory of ranges as well ... that's why I use 160kbps over 144Kbps

... but your trap-post lead me to this conclusion for this topic that may be more understandable for many people:

128Kbps is in the +16Kbps Range, 129kbps is in the +32Kbps Range,
transparency point "for most people on most samples" is somewhere in this area

for minimum safety "most people on most music" should use 160Kbps because it's the first rounded bitrate above this 128-129kbps invisible line which also often represents the codec limits for transparency ...

& indeed 128 & 129Kbps sounds the same ... but still there is an invisible artificial theoric difference in the exponential bitrate architectura at this point ...

personnaly, I use this line as a point of no-return ... as with actual codecs efficiency it works fine ...

feel free not to use it, but be aware it exists ...

now enought bitrates ... got a headache myself