Re: [css3-speech] Comments: 'voice-balance'

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On Wed, 28 Jul 2004, csant wrote:

> 
> * 'voice-balance'
> "Note that unlike voice-volume, voice-balance does apply to cues."
> 
> Does this actually imply that I can have my TTS (text-to-speech) string  
> and the preceding cue on one channel only? And that if I would want the  
> cue to be placed on one side and the speech on the other side I would need  
> to use generated content ::before or ::after an element and voice-balance  
> that?

voice-balance applies to both cues and synthetic speech. The reason 
that voice-volume applies to synthetic speech but not to cues is
that the CSS3 Speech module is tied to SSML. In SSML, the prosody
element allows you to set the volume for speech but doesn't apply
to audio cues. So that's the way it is in CSS3 speech, since we
want to be able to take advantage of the speech engines developed
by vendors for the much larger market for SSML. The remaining
point is that SSML doesn't support a stereo audio model, but a
CSS3 speech processor can apply this in a post synthesis audio
mixing step as mentioned at the end of the section on voice
properties.

The work around is to use the CSS3 speech module's cue properties
to set the volume of the cue independently or relative to that of
synthetic speech.

> 
> My take would rather be to apply voice-balance to voice only (why
> would it be *voice-*balance? Would simply 'balance' not be more
> appropriate if refering to all sounds?) and add more control over
> where to place the cues (maybe in the form of a new property
> 'cue-balance'). CSS 2.0/2.1 'azimuth' was not nominally linked to
> 'voice': I don't understand why the 'voice-' prefix has been added
> if balance is not to refer to voice only.
> 
> Regards,
> /c
> 
> 

- -- 
 Dave Raggett <dsr@w3.org>  W3C lead for voice and multimodal.
 http://www.w3.org/People/Raggett +44 1225 866240 (or 867351)
 
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.4 (GNU/Linux)

iD8DBQFBCkHTb3AdEmxAsUsRAnT8AKDWjElN7VXmQYrx1nU301oyIA1RsgCgzhxb
qZQlBOBfqii1DagtfgunUQQ=
=HAcj
-----END PGP SIGNATURE-----

Received on Friday, 30 July 2004 08:41:31 UTC