204f7d8811
Rename From2D to FromACN2D
2020-12-25 06:40:13 -08:00
0d3b041aa2
Avoid AL types and enums in the effect processors
2020-12-24 22:49:55 -08:00
d578bc6cb1
Move logging to core
2020-12-17 02:47:03 -08:00
56af137ba0
Move fmt_traits to core
2020-12-13 21:12:03 -08:00
191fe888b4
Move ambidefs.h to core
2020-12-12 10:38:24 -08:00
4ee95dc296
Use a smaller FFT for convolution
2020-12-05 14:18:19 -08:00
50e33ce8f4
Change some macros into constexpr variables
2020-12-04 13:53:56 -08:00
69d55d7e03
Move the filters to core
2020-12-04 11:15:50 -08:00
32b9a46b39
Move AL EffectProp handling to separate sources
2020-11-25 13:55:29 -08:00
4e760bbecc
Use a separate structure for the active effect slot properties
2020-11-07 08:36:49 -08:00
d912b92a60
Support all buffer layouts for convolution
2020-09-22 08:09:42 -07:00
8ef242bce0
Don't use direct channels with convolution
...
It may come back as an effect property, but it's probably better to not try
forcing it by default.
2020-09-18 10:12:58 -07:00
07989a34dc
Allow higher-order B-Format buffers for convolution
...
Only the first-order channels will be used, but higher-order buffers will be
allowed.
2020-09-18 00:06:10 -07:00
e7d8e7eb3f
Avoid extraneous buffers and buffer clears
2020-09-17 21:34:26 -07:00
0974b6b47c
Use inline wrappers to clarify forward/inverse FFTs
2020-09-13 04:18:40 -07:00
ce0a45cbc7
Don't assume iterators are pointers
2020-09-09 14:02:02 -07:00
68a099ba35
Apply the first convolution segment in the time domain
...
This avoids an inherent delay from the effect, at the cost of higher CPU use.
Having a customizable user-specified delay (with said user ensuring a properly
trimmed impulse response) could help alleviate the cost since once the delay
exceeds the segment size, the initial FIR filter could be skipped.
2020-09-09 02:23:38 -07:00
f4a55cc8c2
Don't leave the negative frequencies as 0 for inverse FFT
2020-09-08 23:20:06 -07:00
29566b995c
Avoid an extraneous ConvolutionFilter sub-object
2020-09-06 00:16:50 -07:00
c52bf8c401
Rework effect slot buffer setting
...
Rather than creating an effect-specific buffer that gets passed along as a
property, the buffer is set the effect state when the effect state is created,
the device is updated, or the buffer is changed. The buffer can only be set
while the effect slot isn't playing, so it won't be changed or updated while
the mixer is processing the effect state.
2020-09-05 20:48:56 -07:00
af97a92952
Move more processing to the ConvolutionFilter struct
2020-08-31 06:43:42 -07:00
88425becb2
Remove a couple unnecessary pointers
2020-08-30 04:41:19 -07:00
82873486b7
Dynamically allocate the convolution channel mixing data
2020-08-30 04:28:01 -07:00
ebe0765ce3
Reorganize some convolution fields
2020-08-30 03:42:44 -07:00
3fdee6c814
Calculate the correct number of convolution segments
2020-08-28 23:08:44 -07:00
ecfb0d4f5b
Limit convolution processing to the output ambisonic order
2020-08-28 06:47:21 -07:00
7266cd8b6c
Don't dereference a null convolution filter
2020-08-28 06:44:05 -07:00
52531d8b72
Support B-Format impulse responses for convolution
2020-08-28 05:48:26 -07:00
986a58d5b4
Pass a BufferStorage to EffectState::createBuffer
2020-08-28 00:44:55 -07:00
a9ebcdcb7f
Handle 0-length convolution buffers
2020-08-27 21:32:25 -07:00
2a01940041
De-duplicate LoadSampleArray and FmtTypeTraits
2020-08-26 21:29:16 -07:00
9dab2db33b
Avoid copying the convolution filter history to shift it
...
Instead, the current/head/newest segment is tracked, and decremented with each
process to overwrite the oldest.
2020-08-25 15:43:48 -07:00
801c7a9226
Initial implementation of the convolution effect
...
Currently limited to mono and stereo impulse responses, and stereo IRs try to
use direct/real output rather than panning.
2020-08-25 04:21:10 -07:00
1a9fbc1b2f
Stub out a convolution effect state
2020-08-24 20:04:16 -07:00
Chris Robinson