HRTF support

It's been often asked about, so here's an announcement for the interested. I 
just pushed some changes to Git/SVN that adds proper HRTF support to OpenAL 
Soft. It uses data based on the KEMAR HRTF measurements (compact version), 
found here:
http://sound.media.mit.edu/resources/KEMAR.html

The data was processed so it could be compiled directly into the lib[1]. It's 
handled in such a way that (hopefully) allows the data to be placed in read-
only, shareable memory, so the physical memory used by it should be mininal.

HRTF must be enabled in the config file, and only works with 44100hz stereo 
output (it automatically disables if this is not the case). It currently only 
works on mono sources[2], and the cf_level and head_dampen config options are 
ignored while it's active[3].

The future may see improvements to it, such as detecting headphone output vs 
normal stereo, and enabling it by default when appropriate. It can also be 
improved by merging the direct path low-pass filter, so both of them can apply 
at once.

Try it out and let me know how it works. If you have any recommendations for 
improvemnts, I'd like to hear them, too.


[1] For the curious, it was done using this rather hacky tool:
http://kcat.strangesoft.net/dumper.c
Compile it, change to the directory where compact.zip was extracted, and run 
it. It should then silently create hrtf_tables.inc, which is included by 
hrtf.c.

[2] This may be expanded in the future, however the actual filter isn't 
exactly cheap in either memory and CPU use... approx. 1536 bytes * max input 
channels * allocated sources, and it uses a 128-tap FIR filter per output 
channel, per input channel. It's doable, it just may have a cost.

If you want, you can check out mplayer (the Linux one). It's OpenAL output 
uses separate sources for each channel, so a multi-channel audio file gets 
separated into multiple mono sources placed around the listener, which are 
each filtered accordingly. It works, and gives a neat effect for multi-channel 
audio.

[3] Because of the cf_level option being ignored, multi-channel sounds may 
again exhibit a super-stereo effect when using headphones. This should be 
dealt with once HRTF is applied to multi-channel sources.