[Openal-devel] Cone angles
ewmailing at gmail.com
Fri Apr 22 00:32:30 PDT 2011
On 4/21/11, Chris Robinson <chris.kcat at gmail.com> wrote:
> It seems there's a problem with the way OpenAL Soft handles the cone angles.
> believe it's been brought up before, and I think I should do something about
> The way OpenAL Soft interprets AL_INNER_CONE_ANGLE and AL_OUTER_CONE_ANGLE
> that it specifies half of the entire engle. That is, it covers the area as
> specified from the center point outward -- the alpha in the folowing image:
> This is how it's been done all along, ever since starting with the old
> software code, which treats it the same way.
> However, it's come to my attention that the Windows hardware drivers treat
> as the full angle (ie, theta in the aforementioned image). It also seems
> OSX treats it this way, too. Additionally, the spec does seem to imply,
> not out-right say, this is the correct behavior.
My interpretation of the spec was indeed that it was the full angle
and led me to draw the following figure for my book (which I spent way
too long drawing):
As far as I know, Mac, iOS, and Creative implementations do it this way.
> So I'm seriously considering fixing this bug, even though some apps may be
> affected. I can always make a config option or environment variable to
> the buggy behavior if any applications rely on it.
> If there's any objections, feel free to speak up.
I am on the side of seeing this fixed in OpenAL Soft. I would rather
not see another fragmentation issue like we had with Doppler in 1.0.
Beginning iPhone Games Development
More information about the Openal-devel