aboutsummaryrefslogtreecommitdiffstats
Commit message (Collapse)AuthorAgeFilesLines
* GetProcBinary implementation for Haiku.David Carlier2020-07-211-0/+10
|
* Revert "Don't call find_package(WindowsSDK) if the platform ver is already set"Chris Robinson2020-07-092-13/+10
| | | | This reverts commit c1383e3a48c6b882adb97df2255a0f2938cb2887.
* Don't call find_package(WindowsSDK) if the platform ver is already setChris Robinson2020-07-082-10/+13
|
* Remove dead codeChris Robinson2020-06-301-92/+13
|
* Don't round up the dst size when the src size was saturatedChris Robinson2020-06-291-1/+1
|
* Fix SrcBufferSize calculation when downsamplingChris Robinson2020-06-291-20/+36
|
* Reduce the MAX_PITCH to 10Chris Robinson2020-06-291-1/+1
| | | | | 255 is insanely overkill. The bsinc resamplers fade to silence after a few octaves, and the others would be a mess.
* Avoid including windows.h in threads.hChris Robinson2020-06-283-9/+14
|
* Workaround a MinGW quirkChris Robinson2020-06-261-0/+1
| | | | | Sometimes MinGW defines strncasecmp to _strnicmp in string.h, which can cause problems with al::strncasecmp if it's not consistently replaced.
* Update default HRTFChris Robinson2020-06-211-0/+0
|
* Apply simulated HRIR occlusion in the frequency domainChris Robinson2020-06-211-50/+59
|
* Update the default HRTFChris Robinson2020-06-211-0/+0
|
* Synthesize missing elevations in the frequency domainChris Robinson2020-06-211-141/+147
| | | | | | This should help avoid destructive phase interference. The occlusion low-pass filter is still applied in the time domain due to no clear topology (cutoff frequency, slope, bandwidth, etc).
* Use multiple threads for magnitude calculationsChris Robinson2020-06-193-46/+107
| | | | Also allow user-declarable thread count.
* Update the default HRTFChris Robinson2020-06-191-0/+0
|
* Avoid blending multiple IRs for the -90 elevation IRChris Robinson2020-06-191-33/+27
|
* Calculate the HRIR onsets and magnitudes separately from loadingChris Robinson2020-06-171-21/+82
| | | | | This should help improve memory use a bit since the SOFA file can be unloaded before allocating some temp buffers for onset detection and FFT calculation.
* Add a makemhr option to use the farthest field onlyChris Robinson2020-06-171-8/+30
| | | | | All fields are used for equalization, but only the farthest field is resampled, reconstructed, normalized, and written to the mhr.
* Resample HRIRs after equalizationChris Robinson2020-06-171-93/+93
|
* Avoid reinitializing the resampler with each HRIRChris Robinson2020-06-171-10/+11
|
* Resample HRTFs in the frequency domain in makemhrChris Robinson2020-06-171-29/+93
| | | | | | | | | | | | This should produce far better results given it works directly on the frequency response magnitudes prior to phase reconstruction, as it doesn't deal with a linear phase filter on a truncated time-domain response (with the result also getting truncated in both direction). The in-library on-load HRTF resampler still uses the linear filter due to its relative performance and simplicity benefits. It's good enough as a backup, though users with custom HRTFs would benefit from resampling when creating the mhr (adjusting its window size as appropriate/desired).
* Print the synthesized elevations' azimuth countsChris Robinson2020-06-161-2/+4
|
* Avoid an unnecessary local variableChris Robinson2020-06-161-6/+4
|
* Set channel labels from WFX channel masksChris Robinson2020-06-155-118/+129
|
* Print the name of unhandled PulseAudio channelsChris Robinson2020-06-151-1/+3
|
* Remove the Lower channel labels and rename the Upper labelsChris Robinson2020-06-153-23/+21
|
* Don't bother with otherwise unused Aux channelsChris Robinson2020-06-154-103/+39
|
* Handle the front stablizer with the B-Format decoderChris Robinson2020-06-157-165/+254
|
* Fix a typoChris Robinson2020-06-121-1/+1
|
* Move a couple related functions to the backend baseChris Robinson2020-06-1220-145/+145
|
* Link with the CoreFoundation framework on iOSChris Robinson2020-06-111-2/+7
|
* More capitalization fixesChris Robinson2020-06-111-451/+451
|
* Fix some capitalizationChris Robinson2020-06-111-217/+217
|
* Merge pull request #439 from halx99/patch-1kcat2020-06-111-0/+1
|\ | | | | Enable to detect Windows SDK v2004
| * Enable to detect Windows SDK v2004HALX992020-06-111-0/+1
|/
* Don't export functions when static linkingChris Robinson2020-06-101-22/+28
|
* Use a range-for loop instead of for_eachChris Robinson2020-06-091-5/+4
|
* Use better types for array lengthsChris Robinson2020-06-091-17/+15
|
* Update changelogChris Robinson2020-06-071-1/+3
|
* Simplify the aligned over-allocation strategyChris Robinson2020-06-071-15/+14
|
* Avoid explicit checks for _BitScanForward[64]Chris Robinson2020-06-073-29/+10
|
* Check the returned sample endianess with sndioChris Robinson2020-06-071-0/+6
|
* Fix a return valueChris Robinson2020-06-071-1/+1
|
* Use a safer layout if sndio changes the channel countChris Robinson2020-06-071-61/+78
|
* Use a more appropriate size value for copyingChris Robinson2020-06-071-1/+1
|
* Use a better type for array accessChris Robinson2020-06-071-3/+3
|
* Merge pull request #435 from DanielGibson/sndio-more-channelskcat2020-06-071-2/+32
|\ | | | | sndio: Support more than 2 channels
| * sndio: Support more than 2 channelsDaniel Gibson2020-06-071-2/+32
|/ | | | | | | So far the sndio backend only supports Mono and Stereo, but sndio supports up to 16 channels (64 in their current git code). This adds support for all the openal-soft formats (DevFmt*). I tested this on Linux with 5.1 surround speakers.
* Avoid convolving an always-0 sampleChris Robinson2020-05-301-4/+4
|
* Clarify some changelog entriesChris Robinson2020-05-271-3/+4
|
generated by cgit v1.2.3 (git 2.39.1) at 2024-12-13 21:58:59 +0000