Just wondering. Ogg at ???kbps will be equal to MP3 at 128kbps?(I mean nominal bit rates.)
And it would be great if anyone can provide the same information for WMA as well.

Thank you!

Depends on the MP3 encoder. For LAME, I'd estimate 90-100kbps.

Depends on your definition of equal. Vorbis, to me, sounds better at -q0 than MP3 at 128 does due to lack of irritating artifacting, although it does sound like it has less fidelity. They're very different codecs, so there's no tried and true answer. Garf knows what he's talking about though; these are my opinions.

Ha, maybe, but I try to not to be too optimistic about Vorbis' performance, to avoid being accused of format zealotery

Best way to avoid that would be to be optimistic about Vorbis's performance, then recommend he encode using MPC instead.

I don't know which mp3 encoder you are using or which settings, but vorbis -q0 definitely creates lots more irritating artifacting than a good quality 128kbps mp3 encoding. Not to mention -q0 completely demolishes the stereo separation.

Just tested with LAME and Vorbis, transcoded from MPC -q6 on Galaxy Bounce by the Chemical Brothers. I was too lazy to rip for a nice scientific test.

I can clearly hear "wateryness" in the MP3. I hear the oft-heard HF-boosting sound of Vorbis, with less stereo-sep if I were trying to seriously distinguish the two, as well as a little less "fidelity". These are the artifacts that jump out at me most, and I find the "wateryness" far more annoying than Vorbis's artifacts. Maybe I'm just less trained in Vorbis artifacts (I don't use it that much), but the Vorbis encode sounds better to me than the MP3. Neither is transparent.

It's probably just that I haven't trained myself to pick out Vorbis artifacts yet, and I trained myself very early on in my MP3 encoding career to find that watery sound and re-encode in a higher bitrate when/if I found it. As a result, maybe, I loathe that sensation.

I'd say they're being a tad modest when they say -q3(112k VBR) is a bit better than 128k MP3. I'd say it's around -q2(96k VBR) or so. Many say -q0 sounds like a 128k MP3, yet I guess a 112k MP3 would be good for them too.

why not just use -q4 (around 128kbits) and enjoy the improved audio quality. Rather then try to cut the last bits.

i only use -q2 (96kbits) for movies encoding of tv-series (mostly talking)


--edit
corrected -q3 to -q4 as noted by AstralStorm

Comparing VBR Vorbis encoding to CBR MP3 encoding is useless.
Apples and oranges.

But most likely Vorbis at -q2.5 is much better than any 128kbps CBR MP3.
Might be comparable to good 128kbps ABR MP3.

-q0 sounds like a badly encoded 128kbps MP3 - has slightly less annoying artifacts, but is very monoish.

/EDIT\ -q4 is around 128kbps, for reference \EDIT/

-q0 is pretty good for around 64kbps, yes. But I don't think many would choose it over *properly* encoded mp3 around 128kbps. If lame alt-preset 128 (abr) doesn't sound a lot better, then some latest FhG FastEnc-128 definitely should sound better.
-q0's causes quite perceptible stereo separation reduction. This saves bits and makes artifacts harder to hear. It works pretty well often, and people endure the stereo separation reduction much better than bad artifacting. Though -q0 can produce significant artifacting regardless of that, but it's relatively good anyway.

I've actually encountered a few cases where Vorbis 1.0 at -q0 surpassed lame at 128kbps ABR/CBR due to the heavy artifacting introduced by lame.

I will have to "more than" agree with dev0 here!

With lame 4, encoding with alt preset 128 is always a disaster for me? However, I am quiet happy with alt preset medium though

*sigh* I guess we can thank John33 about this..
Lame4 shouldn't even be in public testing yet, and as far as I know nobody has even touched the presets in it which will have to be completely retweaked.

Of course there are cases when something sounds better than something else, but it is imo prettty far fetched to claim that generally -q0 would be better than what the best of mp3 can do around 128kbps.
In essence, mp3-128 can sound decent, but -q0 has pretty much always something which exposes it as very low bitrate encoding.

I just encoded the same audio file using lame 3.90. It sounds exactly the same as the file encoded with lame 4 @ 128!

Its not a music file, its a speech file ???

Did you use --alt-preset 128, --alt-preset cbr 128 or just plain 128kbps?

Anyway, please post a sample.

/EDIT\ Was that vanilla 3.90 or 3.90.3? \EDIT/

(I miss one feature of Winamp forum - the ability to attach files of limited size to posts)

/EDIT\
A great win for Vorbis (GT3b1) - it encodes electronic music very well even at -q 2,5.
Of course it isn't transparent, but nice anyway. (91 kbps)
LAME 3.90.3 AP128 (132) is just much worse.
LAME 3.90.3 APS (218) isn't transparent on the sample (slightly more metallic percussion & preecho) (14/16), APE is (238).
-q 5 (208) isn't transparent too (same artifact, but w/o preecho) ABX 15/16, but is better than APS.
It is transparent for me at -q 6 (234).

At q0 (63), Vorbis performs way better if the sample is lowpassed to 14kHz (SoX) than (otherwise similarly to)
--alt-preset 80 (83) pre-lowpassed.

I think I have tested this sample previously...
\EDIT/