Re: voice-volume, absolute and relative volume units [css3-speech][css3-values][CSS-ISSUE-184]

On 29 Jun 2011, at 23:43, L. David Baron wrote:

> http://dev.w3.org/csswg/css3-speech/#voice-volume defines a relative
> 'dB' unit (which does make sense as a relative unit), which
> CSS-ISSUE-184 proposes to move to css3-values.
>
> However, the 'voice-volume' property then says its computed value is
> "specified value", which doesn't make sense for relative values.  In
> particular, an element with 'voice-volume: +5dB' should have a
> different computed value depending on what its inherited value was.
>
> We've had lots of problems in the past when we had properties whose
> computed values couldn't be represented as valid syntax for the
> property.  I think adding another such case (I think we've fixed the
> existing ones, e.g., with the 'font-weight' changes in CSS 2.1) is a
> bad idea.  So I think if you want relative units in this manner, you
> should also have a syntax for combining them with the possible
> absolute values.

I understand the problem you are describing, but I can't figure-out a  
way to solve it. Absolute values for the audio amplitude level are  
materialized via the x-soft, soft, medium, loud, x-loud keywords, but  
this is still dependent on the selected voice instance (and it is  
linked to the perceived loudness, i.e. to the user).

By the way, I've looked at CSS2.1 font-weight and I can see how the  
characteristics of a particular fonts "instance" are mapped to the  
property values (the "see text" for "Computed value:" really means  
"read the whole damn thing" ;o) ... and that's the only way to really  
get a grasp of the implementation requirements):

http://www.w3.org/TR/CSS21/fonts.html#font-boldness

> Furthermore, the definition of <decibel> values for 'voice-volume'
> says:
>  # This represents a change (positive or negative) relative to the
>  # default or inherited volume level.
> which seems unnecessarily vague.  It should specify that it is
> relative to the default for the root element and the inherited level
> otherwise.

Good point.

Thanks a lot!
Dan

Received on Wednesday, 6 July 2011 21:35:23 UTC