Re: Changing default panner model to equalpower instead of HRTF?

Hello Ray,

I had the impression that the reason behind this decision was "to support
awesome 3D sound that plays nicely with WebGL game demo." In other words, a
3D coordinate from WebGL code can be passed to the panner node directly.

I believe this decision rather caused another problem: the panner node does
not support parameter automation. There is no standard mapping mechanism
between the motion in 3D space and the parameter automation in audio
application. So the panner node dropped the support on AudioParam and chose
to be WebGL-friendly. (Not sure if it is a good call though...)

I know this is a long shot - but how about separating the panner node to 2D
version and 3D version? HRTF and equal-power panning are quite different in
terms of the signal processing. The latter can be easily automated with
AudioParam and developers can decide to save processing power from the
unnecessary HRTF load.

Best,
Hongchan




On Thu, Jul 17, 2014 at 10:07 AM, Raymond Toy <rtoy@google.com> wrote:

> The HRTF panner takes significant resources to load impulse responses as
> well as in processing power. For low-end mobile devices this represents
> significant resources that we may not want to use.
>
> Since the current default panner model is HRTF, once a panner is created,
> the HRTF database must be loaded, even if the user immediately switches to
> the equalpower panner.
>
> If we make the default be the equalpower model, we won't have this
> problem. Only if the application wants to use the HRTF panner will the
> extra memory be used.
>
> Can we change the default to equalpower? This would, obviously, be a
> breaking change in the spec.
>
> Ray
>
>
>
>


-- 
Hongchan Choi

PhD Candidate, Research Assistant
Center for Computer Research in Music and Acoustics (CCRMA)
Stanford University

http://ccrma.stanford.edu/~hongchan