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

[Lorenzo Prince]

I see what you mean about the guitar.  However it seemed to me like the b45
sounded better than the b64.  I think this may be a bug in the encoder, since
b45 sounds better than b64 or q0.
I also couldn't get the last two WMA files to play for some reason.  Not sure
why, unless it has to do with the fact that I am using Winamp to play WMA,
although it played the first 64k file with no trouble.  I also liked the mp3 pro
much better than the WMA, as the WMA file put a lot of ringing in the guitar
both in the beginning and at the end of the clip where the guitar plays with the
drums.

Lorenzo

E Pluribus Unix
----- Original Message -----
From: "John E. Leon Guerrero" <jlg2003feb at live365.com>
To: <vorbis at xiph.org>
Sent: Wednesday, February 19, 2003 11:03 PM
Subject: [vorbis] trying to get better ogg quality for this clip

<p>> hi folks, in my (unlucky) first test of ogg vs other encoders, i found a
> case where wma and mp3pro sound much better than ogg at 64k.  can anyone
> suggest a setting that i haven't tried yet that can rival the wma and
> mp3pro samples at 64k?  it's the "gravel effect" that is troublesome.
>
> the part in question is the first 15 seconds of this wave file:
> http://qa01.live365.com/jlg/VH_simple.wav
>
> this piece starts with a stereo guitar that is rich in harmonic content
> (like chimes) with some flanging/chorus/delay type effects.  way in the
> background, there is some high frequency delay/reverb that goes easily
> unnoticed unless using head phones.  it sounds like the guitar pick
> scratching against the strings.
>
> using decent (but not top of the line) headphones for evaluation, i
> found that wma and mp3pro did a very nice job of encoding the file where
> the differences from the raw wav file were practically imperceptible to
> me.  however, when encoded with ogg q0-q2, the sounds is noticeably
> altered and the "high frequency guitar pick noise" sounds like someone
> walking through gravel.  at q3 it starts to sound better and at q4 it is
> about equal to the others.
>
> by my ears, the 64k wma file does the best job at avoiding the "gravel
> effect".  it also seems to add some presence & volume that doesn't exist
> in the original wave but *does* sound good...at least for this clip.
>
> focusing on streaming, i also encoded using the options:
> "-b 64"
> "-b 64 --managed"
> "-b 45"
> "-b 45 --managed"
>
> (dunno why, but i couldn't get it to do 32k nor could i get the
> --advanced-encoder-option bitrate_average_window=3 option to work either.)
>
> the directory isn't browsable but here are the links for my test cases:
>
> http://qa01.live365.com/jlg/VH_simple.wav
> http://qa01.live365.com/jlg/VH_simple_64k.wma
> http://qa01.live365.com/jlg/VH_simple_pro.mp3
> http://qa01.live365.com/jlg/VH_simple_q0.ogg
> http://qa01.live365.com/jlg/VH_simple_q1.ogg
> http://qa01.live365.com/jlg/VH_simple_q2.ogg
> http://qa01.live365.com/jlg/VH_simple_q3.ogg
> http://qa01.live365.com/jlg/VH_simple_q4.ogg
> http://qa01.live365.com/jlg/VH_simple_b64.ogg
> http://qa01.live365.com/jlg/VH_simple_b64managed.ogg
> http://qa01.live365.com/jlg/VH_simple_b45.ogg
> http://qa01.live365.com/jlg/VH_simple_b45managed.ogg
>
> dunno if these are really 40k and 48k since the encoder only processed 3
> seconds of the clip.  they sound good though.
> http://qa01.live365.com/jlg/VH_simple_48k.wma
> http://qa01.live365.com/jlg/VH_simple_40k.wma
>
> thanks,
> jlg
>
> --- >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.
>

<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.