[X3D-Public] improved validation tests for sound nodes; add ability avoiding audio attenuation altogether

[Don Brutzman]

On 1/4/2014 8:29 AM, Michalis Kamburelis wrote:
> Don Brutzman wrote:
>> I've often thought we might benefit by having a way to just make a
>> sound audible without having to worry about all this non-visible
>> geometry (see attached slideset excerpt for some sense of how
>> involved it can be).
>>
>> Why don't we add an initializeOnly field that allows an author to
>> disable all attenuation?  This would seem to be essential for
>> user-interface responses.  Putting some set of 4 extremely large
>> values in for min/max front/back is error prone and not very
>> intuitive.  How about the following backwards-compatible addition to
>> Sound node:
>>
>> SFBool  []  attenuate TRUE
>>
>> If author sets attenuate="false" then all of the computational sound
>> features are disabled and the clip simply plays as recorded.
> 
> I agree, such field would be useful.
> 
> As for view3dscene, we actually treat "spatialize" field just like you 
> described above. So "spatialize FALSE" disables all 3D effects and the 
> sound will play as-is.
> 
> Which is not exactly correct per the specification (that says that 
> ellipsoid dimensions should still have an effect even spatialize=FALSE).

yes, last paragraph of 16.4.2 Sound
http://www.web3d.org/files/specifications/19775-1/V3.3/Part01/components/sound.html#Sound

> But it's useful, just you describe --- a simple way to disable 3D sounds 
> is often needed, e.g. for UI sounds. And it was natural to implement it 
> that way too :)
> 
> Also, stereo sound files are never attenuated. They always behave like 
> spatialize=FALSE for view3dscene. Because that's how OpenAL behaves, and 
> it just makes sense (it's not obvious how you would attenuate 
> multi-channel sound).

Michalis, thanks for the description of your current implementation approach.

Your handling of stereo sound files is curious... might make sense for a Beatles song, but might not be good for an audio track that has left/right channels that communicating sound effects or cues.

Wondering:  does the proposed combination of both fields, attenuate and spatialize, give the author sufficient control over the aural rendering to match the purposes of whatever type of sound signal might be provided?

It currently appears to me that the combinations are pretty rich, and the model is simple to understand/implement.  If there are any counterexample use cases, they would be interesting to consider.

all the best, Don
-- 
Don Brutzman  Naval Postgraduate School, Code USW/Br       brutzman at nps.edu
Watkins 270,  MOVES Institute, Monterey CA 93943-5000 USA   +1.831.656.2149
X3D graphics, virtual worlds, navy robotics http://faculty.nps.edu/brutzman