[vorbis] applaud.wav sounds worse in 1.0?

John Morton jwm at plain.co.nz
Thu Jul 25 19:26:07 PDT 2002



On Fri, 26 Jul 2002 10:22, Nicola Larosa wrote:
>  > (my ears aren't that well trained, and I don't want them to be :-)
>
> Surely better trained than mine, then. :^)

Different folks have different sensitivities. I gather applaud.wav is a good 
pre-echo test case, but I don't think that's what I'm hearing, as I'm not 
particularly sensitive to that (or the warbling in 60.wav or stereo image 
stuff in general). 

>  > but this time I noticed that applaud.wav[1] seems to have a lot of
>  > conspicuous static-like noise in it, all the way up to -q 6 or there
>  > abouts - I could ABX out to 5, before I gave up and went to bed.
>
> I can't do that. applaud.wav was one of the very few samples where I could
> recognize artifacts in RC3; they were still barely noticeable at -q 4.99,
> but went away at -q 5.
>
> On 1.0 I have a hard time hearing that noise at -q 2!

There's definitely a crackling sort of noise that ghosts the clapping, and 
it's more pronounced in 1.0. Curiously, lame doesn't produce the same sort 
of artifact at -V 0 through to 5 (I didn't bother testing lower), which is 
why I raised the issue (that, and the artifact is really annoying).

> So, I've changed my chosen encoding level, from -q 4.99 to -q 3, the
> default, for a gain of 20-30% in size, and I'm confident I won't regret
> this decision.

I can detect warbling on 60.wav, and spot that a light piece of precussion 
(or echo) drops out on spahm.wav below 4.0. At 4.0, I think I can here 
something, but the ABX is borderline. Beyond that, it's fine. Which means 
that I'd probably stick to using 4.99 if it wasn't for the applaud.wav thing. 
I might test some other applause samples from my live albums, first, though.

> Vorbis is a-m-a-z-i-n-g! Thanks again, Monty and crew!

I'm impressed by q-1 - it's obviously not "cd" quality, but it's a lot more 
listenable than any other format at 'streaming to modems' level.

John

--- >8 ----
List archives:  http://www.xiph.org/archives/
Ogg project homepage: http://www.xiph.org/ogg/
To unsubscribe from this list, send a message to 'vorbis-request at xiph.org'
containing only the word 'unsubscribe' in the body.  No subject is needed.
Unsubscribe messages sent to the list will be ignored/filtered.



More information about the Vorbis mailing list