[openal] Small HRTF SOFA reader lib
christian.hoene at symonics.com
Mon Dec 5 03:14:02 EST 2016
Thank you Chris,
for your feedback. All your issues have been addressed and solved in the latest commits
My plan is it to add libmysofa to OpenAl-soft in addition to the tradition makehrtf-file-reader, if I may.
> currently, we are writing a small library to read HRTF SOFA files.
> It is intended for embedded devices or mobiles that have a small
> (code) memory footprint. Of course, it might be useful for OpenAL, too.
> Because you do not need the hdf5 and netcdf libs.
This looks interesting. Especially if it doesn't have any external dependencies.
> It is not yet fully finished but I though it is time to collect
> feedback so that I can enhance it.
> If you have same spare time, please let me know on what to change to
> make it useful for you guys…
I can't comment too much on the code, since I'm not familiar with the
HDF5 format or how SOFA files are actually put together. But I'll offer what I can.
First, in regards to mysofa.h, it's a good idea to prefix public functions and types to help avoid conflicts in other code. In particular, the HRTF_ATTRIBUTE and HRTF structs should have MYSOFA_ or something prepended to them. I also see that the __cplusplus/extern "C"
stuff is commented out, but that would be needed to include it in C++ code.
For CMakeLists.txt, I see it's forcing a set of CFLAGS. Forcing -O0 -g should be unnecessary since that's implicit with CMake's Debug build type. -fno-omit-frame-pointer shouldn't be needed since it's only ever implicitly enabled when it doesn't interfere with debugging, and requires an explicit -fomit-frame-pointer otherwise.
For the remaining flags, you should probably use CheckCCompilerFlag to ensure each individual one is available before adding it. That will help for compatibility with other compilers, like MSVC.
If I get some time, I'll try to look at it more.
openal mailing list
openal at openal.org
More information about the openal