[vorbis] Bitrate restrictions

Peter Schuller peter.schuller at infidyne.com
Fri Jul 26 04:36:23 PDT 2002



> if you want a certain bitrate, why are you using a qualitative setting
> and not a bitrate setting ? 

Why not?

I want a specific quality. But bitrate is also important in that I want
as little as possible. Now, if the minimum possible quality is *better*
than the quality I want, I am obviously interested in decreasing the
quality and saving bits. Yes, quality is the primary concern. But that
does not stop bitrate from being relevant. I don't care about the exact
bitrate. But the ballpark figure is important. At the moment I can't get
much below 50kbit for normal music. The bitrate drops some (but not
much) for voice data recorded with my mike (admittedly with background
noise). But it's still *waaaay* better than your typical phone or IP
Telephony software. I don't *need* that good quality, so it's a waste of
bits.

And there are reasons to use Ogg Vorbis instead of, say, GSM. Firstly,
O/V has quality based encoding. GSM and similar "voice codecs" tend to
work poorly in some instances, particularly with lots of noise.

I don't know how far down Ogg Vorbis could go without the hard coded
cut-off without producing artifacts, but I would like to *try*.

>  maybe for the audio you're encoding -1
> quality is up in the 50kbit range.  Have you tried doing -b 40 ?

Regardless of -b switch, I never get below 51kbit on the particular
piece I'm testing with.

> qualitative settings depend on the source, what you want is a specific
> range and there are options for it.  I assume -b is an option in the
> oggenc you're using.  

I know it depends on the source. I am just trying to illustrate that the
worst possible quality is still pretty good (as discussed in a previous
thread also). This is relevant even with music as an example *because*
quality based mode is supposed to yield a certain quality. In other
words, if I find -1 to be a lot better than I need for a piece of music,
it's going to be better than I need for voice comm too[1].

(I'm not *just* interested in voice comm, but that is a example where
low bitrates are desireable. Another is streaming at 28k8 modem speeds.
Perhaps it'll sound like crap without downsampling, but perhaps not.
Particularly if the stuff being streamed is non-complex. In this case it
would be good to be able to do -q -5 -M 28.)

[1] I'm not saying I think -q -1 is good enough to encode my music to.
But based on the quality I hear, it's obviously better than I need for
speech, assuming a fairly constant quality at -q -1.


-- 
/ Peter Schuller, InfiDyne Technologies HB

PGP userID: 0xE9758B7D or 'Peter Schuller <peter.schuller at infidyne.com>'
Key retrival: Send an E-Mail to getpgpkey at scode.org
E-Mail: peter.schuller at infidyne.com Web: http://www.scode.org

<p><p>--- >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