zdoom/raze
0
0
Fork 0
mirror of https://github.com/ZDoom/Raze.git synced 2025-05-30 08:51:08 +00:00
Code Issues Projects Releases Packages Wiki Activity

Patch from Nuke.YKT to replace audiolib's volume table with floating point calculations

Browse source 
This fixes an issue where rapidly changing the player's angle could result in the audio "popping" due to the sudden change to a new position in the volume table, and the resulting difference in volume between one sample and the next. This also fixes a DC offset issue when applying the volume level to certain sample types.

git-svn-id: https://svn.eduke32.com/eduke32@7773 1a8010ca-5511-0410-912e-c29ae57300e0
This commit is contained in:
terminx 2019-07-08 00:41:08 +00:00 • committed by Christoph Oelckers
parent cf7ac8f16e
commit 027afc2347
4 changed files with 94 additions and 148 deletions
Download patch file Download diff file Expand all files Collapse all files

65
source/audiolib/src/mix.cpp
View file

@ -28,23 +28,25 @@
*/
// 8-bit mono source, 16-bit mono output
uint32_t MV_Mix16BitMono(struct VoiceNode const * const voice, uint32_t length)
uint32_t MV_Mix16BitMono(struct VoiceNode * const voice, uint32_t length)
{
auto const source = (uint8_t const *)voice->sound;
auto dest = (int16_t *)MV_MixDestination;
uint32_t position = voice->position;
uint32_t const rate = voice->RateScale;
float const volume = voice->volume;
float const volume = voice->volume*MV_GlobalVolume;
do
{
uint8_t const usample0 = SCALE_SAMPLE(source[position >> 16], volume);
int32_t const isample0 = FLIP_SIGN(source[position >> 16]) << 8;
position += rate;
*dest = (int16_t)clamp(SCALE_SAMPLE(MV_LeftVolume[usample0], MV_GlobalVolume) + *dest, INT16_MIN, INT16_MAX);
*dest = (int16_t)clamp(SCALE_SAMPLE(isample0, volume*voice->LeftVolume) + *dest, INT16_MIN, INT16_MAX);
dest += MV_SampleSize >> 1;
voice->LeftVolume = SMOOTH_VOLUME(voice->LeftVolume, voice->LeftVolumeDest);
}
while (--length);
@ -54,25 +56,28 @@ uint32_t MV_Mix16BitMono(struct VoiceNode const * const voice, uint32_t length)
}
// 8-bit mono source, 16-bit stereo output
uint32_t MV_Mix16BitStereo(struct VoiceNode const * const voice, uint32_t length)
uint32_t MV_Mix16BitStereo(struct VoiceNode * const voice, uint32_t length)
{
auto const source = (uint8_t const *)voice->sound;
auto dest = (int16_t *)MV_MixDestination;
uint32_t position = voice->position;
uint32_t const rate = voice->RateScale;
float const volume = voice->volume;
float const volume = voice->volume*MV_GlobalVolume;
do
{
uint8_t const usample0 = SCALE_SAMPLE(source[position >> 16], volume);
int32_t const isample0 = FLIP_SIGN(source[position >> 16]) << 8;
position += rate;
*dest = (int16_t)clamp(SCALE_SAMPLE(MV_LeftVolume[usample0], MV_GlobalVolume) + *dest, INT16_MIN, INT16_MAX);
*dest = (int16_t)clamp(SCALE_SAMPLE(isample0, volume*voice->LeftVolume) + *dest, INT16_MIN, INT16_MAX);
*(dest + (MV_RightChannelOffset >> 1))
= (int16_t)clamp(SCALE_SAMPLE(MV_RightVolume[usample0], MV_GlobalVolume) + *(dest + (MV_RightChannelOffset >> 1)), INT16_MIN, INT16_MAX);
= (int16_t)clamp(SCALE_SAMPLE(isample0, volume*voice->RightVolume) + *(dest + (MV_RightChannelOffset >> 1)), INT16_MIN, INT16_MAX);
dest += MV_SampleSize >> 1;
voice->LeftVolume = SMOOTH_VOLUME(voice->LeftVolume, voice->LeftVolumeDest);
voice->RightVolume = SMOOTH_VOLUME(voice->RightVolume, voice->RightVolumeDest);
}
while (--length);
@ -82,27 +87,27 @@ uint32_t MV_Mix16BitStereo(struct VoiceNode const * const voice, uint32_t length
}
// 16-bit mono source, 16-bit mono output
uint32_t MV_Mix16BitMono16(struct VoiceNode const * const voice, uint32_t length)
uint32_t MV_Mix16BitMono16(struct VoiceNode * const voice, uint32_t length)
{
auto const source = (int16_t const *)voice->sound;
auto dest = (int16_t *)MV_MixDestination;
uint32_t position = voice->position;
uint32_t const rate = voice->RateScale;
float const volume = voice->volume;
float const volume = voice->volume*MV_GlobalVolume;
do
{
int16_t const isample0 = B_LITTLE16(source[position >> 16]);
split16_t const usample0{FLIP_SIGN(SCALE_SAMPLE(isample0, volume))};
int32_t const isample0 = B_LITTLE16(source[position >> 16]);
position += rate;
int32_t const sample0 = (SCALE_SAMPLE(MV_LeftVolume[usample0.l()], MV_GlobalVolume) >> 8) +
SCALE_SAMPLE(MV_LeftVolume[usample0.h()], MV_GlobalVolume) + 128;
int32_t const sample0 = SCALE_SAMPLE(isample0, volume*voice->LeftVolume);
*dest = (int16_t)clamp(sample0 + *dest, INT16_MIN, INT16_MAX);
dest += MV_SampleSize >> 1;
voice->LeftVolume = SMOOTH_VOLUME(voice->LeftVolume, voice->LeftVolumeDest);
}
while (--length);
@ -112,7 +117,7 @@ uint32_t MV_Mix16BitMono16(struct VoiceNode const * const voice, uint32_t length
}
// 16-bit mono source, 16-bit stereo output
uint32_t MV_Mix16BitStereo16(struct VoiceNode const * const voice, uint32_t length)
uint32_t MV_Mix16BitStereo16(struct VoiceNode * const voice, uint32_t length)
{
auto const source = (int16_t const *)voice->sound;
auto dest = (int16_t *)MV_MixDestination;
@ -123,20 +128,20 @@ uint32_t MV_Mix16BitStereo16(struct VoiceNode const * const voice, uint32_t leng
do
{
int16_t const isample0 = B_LITTLE16(source[position >> 16]);
split16_t const usample0{FLIP_SIGN(SCALE_SAMPLE(isample0, volume))};
int32_t const isample0 = B_LITTLE16(source[position >> 16]);
position += rate;
int32_t const sample0 = (SCALE_SAMPLE(MV_LeftVolume[usample0.l()], MV_GlobalVolume) >> 8) +
SCALE_SAMPLE(MV_LeftVolume[usample0.h()], MV_GlobalVolume) + 128;
int32_t const sample1 = (SCALE_SAMPLE(MV_RightVolume[usample0.l()], MV_GlobalVolume) >> 8) +
SCALE_SAMPLE(MV_RightVolume[usample0.h()], MV_GlobalVolume) + 128;
int32_t const sample0 = SCALE_SAMPLE(isample0, volume*voice->LeftVolume);
int32_t const sample1 = SCALE_SAMPLE(isample0, volume*voice->RightVolume);
*dest = (int16_t)clamp(sample0 + *dest, INT16_MIN, INT16_MAX);
*(dest + (MV_RightChannelOffset >> 1))
= (int16_t)clamp(sample1 + *(dest + (MV_RightChannelOffset >> 1)), INT16_MIN, INT16_MAX);
dest += MV_SampleSize >> 1;
voice->LeftVolume = SMOOTH_VOLUME(voice->LeftVolume, voice->LeftVolumeDest);
voice->RightVolume = SMOOTH_VOLUME(voice->RightVolume, voice->RightVolumeDest);
}
while (--length);
@ -145,25 +150,15 @@ uint32_t MV_Mix16BitStereo16(struct VoiceNode const * const voice, uint32_t leng
return position;
}
void MV_16BitReverb(char const *src, char *dest, const int16_t *volume, int32_t count)
void MV_16BitReverb(char const *src, char *dest, const float volume, int32_t count)
{
auto input = (uint16_t const *)src;
auto output = (int16_t *)dest;
do
{
int const sample0 = *input++;
#if 0 // def BIGENDIAN
int sample0l = sample0 >> 8;
int sample0h = (sample0 & 255) ^ 128;
#else
int sample0l = sample0 & 255;
int sample0h = (sample0 >> 8) ^ 128;
#endif
sample0l = (volume)[sample0l] >> 8;
sample0h = (volume)[sample0h];
*output++ = (int16_t)(sample0l + sample0h + 128);
int16_t const isample0 = (int16_t)*input++;
*output++ = SCALE_SAMPLE(isample0, volume);
}
while (--count > 0);
}