[vorbis] trying to get better ogg quality for this clip

ndrw mchl grnbrg agreenbu at nyx.net
Tue Feb 25 11:10:35 PST 2003



John E. Leon Guerrero <jlg2003feb at live365.com> wrote:
: Yes, that sample sounds much better in the beginning since the "gravel 
: effect" is gone and there is not much "swishing" in the guitar sound. 
: However, IMHO, the rest of the piece sounds "flat" most likely due to 
: the frequency limit.  For instance, with good headphones, you can hear 
: the snare springs from under the snare drum in the other pieces when you 
: A-B compare the samples.  Interestingly, the ogg samples that I provided 
: earlier all sound wonderful in full 44khz after the initial 15 second 
: intro.  For comparision sake, I encoded the sample with LAME at 64kbps, 
: 22khz, JS and found similar results.  Ie, the intro sounds good (though 
: it doesn't have the same "bell-like" chiming quality that mp3 PRO and 
: WMA have) due to the lack of the gravel effect and the swishing 
: sound...but it too sounds flat for the rest of the sample which you can 
: notice with the snare drum springs.
: 
: http://qa01.live365.com/jlg/VH_simple_64kb_22kh.mp3
: 
: I don't have a clue how hard this would be, but it would be wonderful if 
: the encoder could tell that the first 15 seconds were going to be 
: difficult to encode at 44khz and automatically downsample at 22khz 
: (where it at least sounds normal despite lacking a little high end 
: presence), and then resume the normal 44khz encoding once the "high 
: demand" is over.  I noticed in winamp that the higher q settings 
: consumed more bits-per-second for the opening piece, which could be an 
: indicator.

I think it would be quite a feat to create a file format with a variable
sample rate. A variable lowpass filter would make a little bit more sense,
and would probably provide similar results, although probably with less
file size savings.


-- 
 agreenbu @ nyx . net                             andrew michael greenburg
                     http://www.nyx.net/~agreenbu/
--- >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