So which version of Ogg Vorbis, since the beginning of the beta releases, do you think is the best, in terms of encoding the musical genre that you listen to the most and why (ie. why is it better than the latest release, if your favourite isnt 1.1)?

Oggenc2.3 using MegaMix2 2004-07-21

I read a post which said some tunings of aotuv beta 2 could have been broken in the move to 1.1. I am unable to ABX killer samples with Aotuv beta 2 at Q7, so this is my encoder and Q level of choice.

I'm also very interested in the opinions expressed here as the number of different encoder versions currently on Rarewares is somewhat excessive and needs rationalising.

I use Vorbis aoTuVb2 for bitrate below 128kbps
It's transparent at -q 4 for Rock music to me

The official releases has obvious (1.01) hiss even at 128kbps
I haven't try 1.1 yet, i'm too lazy so i still stick with aoTuVb2
and for bitrates higher than -q5 I use Megamix2.

I would use megamix1. megamix2 produces higher bitrate, and I didn't noticed improvements (in pre-echo area) justifying this inflation.
aoTuV beta 3 is a potential favorite, but I had to test it (some issues were noticed, but some progress too).

I currently use mostly Musepack or lossless for archiving. Vorbis I use primarily for streaming. I stream with foobar2000 using the oddcast plugin. I use aoTuVb2 (dll's from rarwares.org) with this, because 1) it's claimed to be good at low bitrates and 2) the oddcast plugin doesn't work with Xiph.org 1.1 and derivates (including aoTuVb3). I've posted on Oddsocks forum on this topic, but he seems reluctant updating his software, claiming he can hardly find the time. Considering Oddsock.org is a one-man-operation this is understandable. I just hope he will update his tools to work with 1.1 some day.

[EDIT] Ooops. Seems I was a bit fast there. Seems this issue have just been fixed. http://www.oddsock.org/tools/oddcastv2_foo/
All this was maybe a bit off-topic anyways. Sorry [/EDIT]

I use aoTuV b2 and b3 @ q-4. I can't really ABX anything -q4 and above, so -q4 was my rational decision. I chose aoTuV based on it's performance in the last 128k listening test. My music collection consists mainly of rock/metal. Vorbis is the only codec I use and it's used on my DAP. I use lossless on my computer.

I tested for many of my personal music (jazz and rock in general)
with aoTuV b2; aoTuV b3 and 1.1...

I can't really ABX in q6 for any of this versions of vorbis...
For killer samples (for example badvilbel and others..) i can't really ABX in q7 in my headphones...

I believe, for my ears and music in general,
q6 is sufficient for total transparency... , but, is my personal opinion...

from a poll i started not so long ago...

it is clear that majority of the ppl are using 1.1 / aoTuV b2 while megamixII came a distant second. i would assume that many 1.1/ab2 usrs would switch to ab3 after some rigorous testing and tuning is done.

aoumi indicated that he would be realeasing another beta to improve on the flaws pointed out by guru recently.

qk has hinted a mega modified encoder based on some wavelet stuff

so there is a lot in store for vorbis in the comming weeks/months!!!

maybe i start one more poll after these developments take place

btw, i still prefer to use megamix 2. i gives me (garf's + qk's + aoumi's) tunings. i think it is a good allround encoder for all bitrates

I'm currently toying with aoTuV b3 and vanilla 1.1. I might get my lazy ass to do a comparison at -q 4.25 later today.

Vorbis 1.1 with merged tunings for most part I think ;-d. Pre-echo doesn't really bother me (it did with FhG algorithm though ;-d) with Vorbis and I will encode as low as -q 2 for music that does not have a lot of transients, hence some solo acoustic guitar track's etc, but I have used as high as -q 5 with an impulse_noisetune=5 on some of my electronic material due to agressive percussion, which was causing noise to be rotated towards the middle on lower -q values. If I listened to metal/hardcore oriented material I might encode with -q 5 and above too due to cymbals being all of over the place and agressive drumming. Anything -q 6 and above to me is overall kill with hardware I have ;-D as well.

The best so far for me is Vorbis 1.1. BUT, my demands are probably quite different to many of HA's users, since I'm encoding music routinely at -q -1 / 22.05KHz [ssrc'd] since size (without too bad quality) is critical to me. At this quality level Vorbis 1.1 seems (untrained ear, non-scientificaly-tested disclaimers apply) marginally the best of all the Vorbis variants so far. aotuv b3 seems slightly better for some tracks but generally a little worse, particularly a little more destructive to cymbals at this quality level.

My favourite is aoTuVb3. Using q 6. This gives always transparency for my ears.

jergar

I encode with OggVorbis when I need something at very low bitrate - I use Archer sped up version of aoTuV beta3 -q-2 then.

I'm generally using 1.1, but I also really dig that MegaMix2..!

I dont use Ogg Vorbis myself because of the problems i encountered with pre-echo and impulse trigger etc. QK has come up with some good solutions and the 2 additional switches are extremely useful, but these usually increase the bitrate noticeably and this solution would either mean encoding everything with the optional switches turned on or going back and re-encoding problem tracks.

Interestingly the old GT2 encoder at vbr 160 exhibited none of these problems to my ears and whilst not perfect exhibited less objectional results and performed more predictably.

oggvorbis 1.1 does really good at q 2
I remember when doing a abx test that megamix 1 and 2 both had a higher bitrate and both echoed so i could abx it easily.When i encode in to aoTuV b3
I hear smearing so i stick with 1.1 aka aoTuV b2.

I usually rip into .wav then transcode to .flac for the comp. Then I transcode that into -q6 ogg using the fb ogg converter component for my Karma. Or if its an album I don't think I'm gonna keep (eg one copied from a friend..) I'll go straight to a -q6.

Me thinks that ripping to wav and then flac on the PC is bit of an overkill. EAC rips to a wav file anyway and then does the encoding.

Why don't you do this:
Use EAC and then call the mareo.exe that you can configure to create you your flac files and ogg vorbis (or whatever else you want) at the same time. That'll save you a lot of time.

An example from my mareo setup (ini) file may help:

; ---------------------------------------------------------------------------------------------------------------------------
; OGG VORBIS
; ---------------------------------------------------------------------------------------------------------------------------
EXT = ogg
PATH = C:\Audio\lossy\ogg_aotuvb3\@artist@\@album@
ENCODER = "C:\My Downloads\Encoders\ogg\oggenc2.3aoTuVb3P4\oggenc2.exe"
PARAMETERS = -q 4 "@source@" -o "@dest@" -a "@artist@" -l "@album@" -t "@title@" -N "@track@" -d @year@ -G "@genre@" -c "COMMENT=VORBIS+MAREO"

; ---------------------------------------------------------------------------------------------------------------------------
; OGG VORBIS - Various Artists
; ---------------------------------------------------------------------------------------------------------------------------

; EXT = ogg
; PATH = C:\Audio\lossy\ogg_aotuvb3\Various Artists\@album@
; ENCODER = "C:\My Downloads\Encoders\ogg\oggenc2.3aoTuVb3P4\oggenc2.exe"
; PARAMETERS = -q 4 "@source@" -o "@dest@" -a "@artist@" -l "@album@" -t "@title@" -N "@track@" -d @year@ -G "@genre@" -c "COMMENT=VORBIS+MAREO"

; ---------------------------------------------------------------------------------------------------------------------------
; FLAC
; ---------------------------------------------------------------------------------------------------------------------------
EXT = flac
PATH = C:\Audio\lossless\flac\@artist@\@album@
ENCODER = "C:\My Downloads\Encoders\flac\flac 1.1.1a\flac.exe"
PARAMETERS = -5 "@source@" -o "@dest@" -T artist="@artist@" -T album="@album@" -T title="@title@" -T tracknumber="@track@" -T date=@year@ -T genre="@genre@" -T comment="FLAC+MAREO"

; ---------------------------------------------------------------------------------------------------------------------------
; FLAC - Various Artists
; ---------------------------------------------------------------------------------------------------------------------------
; EXT = flac
; PATH = C:\Audio\lossless\flac\Various Artists\@album@
; ENCODER = "C:\My Downloads\Encoders\flac\flac 1.1.1a\flac.exe"
; PARAMETERS = -5 "@source@" -o "@dest@" -T artist="@artist@" -T album="@album@" -T title="@title@" -T tracknumber="@track@" -T date=@year@ -T genre="@genre@" -T comment="FLAC+MAREO"

; EOF

I use 1.1RC1 as my personal encoder. This have a very good sound and good speed encode. My settings are: Q5, with ITP=4

Yeah, I know its overkill. :\

Thanks for that link, I'll check it out!

shameless bump! ~ and 1st post (long time reader)

I've got a Rio Karma coming my way, prob tomorrow, and have spent the weekend ripping CDs to OGG. I'm using EAC and oggenc from vorbis-tools 1.0.1. I think I'm doing something right so far (not having the player yet to check) since they're tagged well, in proper directories, and play gaplessly in foobar2k (and WA.)

Thought I'd ask here, since this seems like the most pertinent thread, though perhaps dated.

Seems like there are WAY too many options of encoders to choose... as much as I love open-source and all, this just seems silly. Looks like there's a lot of development going on currently, so that's cool. It's just a bit confusing, just starting out.

I'm now compressing to -q 7.5, a bit high from what most people seem to use here. Files are ~20% the original WAV size. Too high? Pointlessly so?

For such high bit rates, what is the currently preferred / best version of the encoder?

My settings in EAC for oggenc are as follows:
-q 7.5 -a "%a" -t "%t" -l "%g" -d "%y" -N "%n" -G "%m" %s

Anything to change there? And if a different encoder is recommended, any additions to the above?

Thanks for any guidance, all!

( yeah, I've read many posts, and used search... but seems this is a rapidly changing scene, and I'm not sure what the current concensus is. )

Vorbis 1.01 is outdated. You can safely use 1.1. If you really want a very high quality (it's probably the case if you're using -q7,5), you can also use the version called Megamix, merging different kind of tuning. Megamix 2.1 have lower pre-echo than official release (but bitrate is as consequence higher for the same -q setting). Also take into consideration than the same results as megamix are possible when using vorbis with some advanced options (impulse trigger).

To make things clear:

- simplest and safest solution: use vorbis 1.1

- for higher quality, you can use:
o vorbis megamix
o vorbis aoTuV beta 3
o any vorbis encoder offering impulse trigger profile

The recommended encoder is
http://www.hydrogenaudio.org/forums/index....showtopic=15049



Maybe... depends on your ears only. Me personally got tired of doing listening tests for each and every album, wherefore I keep lossless copies of them. They are then transcoded into Q5+ Vorbis, depending on genre and instruments used (well, that's my method).




For portable use, Q5 (maybe even less) should be all you need.

PS: I knew, someone would be replying faster

ed: quotes ^^

Thanks for the info/advice you two.

I'd read that thread, and was still unsure about the overall concensus, so I bumped this one.

I'd got the feeling that the 1.1 is the "offically" recommended one, and that all the others were kind of optimized for Q <= 5. From people's comments elsewhere, it kind of led me to believe that with my relatively high Q setting, I didn't need to worry about the pre-echo/impulse settings, and that's why I asked about the switches I've used.

I'm ripping these not only for the portable but for this PC as well, which is pumped through my HiFi, so I figured a higher Q would be better. Generally, I already have lossless copies - my CDs! (~400GB in FLAC/SHN of live shows though.)

I use Vorbis 1.1RC1, It's my personal encoder; and for me the sound is very clear and natural. In my ABX tests I don't hear differences in comparison with the original wav file (with exception of a few killer samples).

My setting is Q6.