forked from len0rd/rockbox
MIDI: At long last, though quick and dirty, pitch bend depth! Or, I think it works. Tested on two
files. Let me know if anyone discovers any problems with this. This commit also includes Nils's synth loop optimization patch. I hope committing it does not cause problems. git-svn-id: svn://svn.rockbox.org/rockbox/trunk@15112 a1c6a512-1295-4272-9138-f99709370657
This commit is contained in:
Stepan Moskovchenko
parent
99f9550881
commit
1515ff8522
5 changed files with 240 additions and 162 deletions
|
|
@ -65,6 +65,7 @@ int initSynth(struct MIDIfile * mf, char * filename, char * drumConfig)
|
|||
chPan[a]=64; /* Center */
|
||||
chPat[a]=0; /* Ac Gr Piano */
|
||||
chPW[a]=256; /* .. not .. bent ? */
|
||||
chPBDepth[a]=2; /* Default bend value is 2 */
|
||||
}
|
||||
for(a=0; a<128; a++)
|
||||
{
|
||||
|
|
@ -255,191 +256,195 @@ inline void stopVoice(struct SynthObject * so)
|
|||
so->decay = 0;
|
||||
}
|
||||
|
||||
static inline int synthVoice(struct SynthObject * so)
|
||||
static inline void synthVoice(struct SynthObject * so, int32_t * out, unsigned int samples)
|
||||
{
|
||||
struct GWaveform * wf;
|
||||
register int s;
|
||||
register unsigned int cpShifted;
|
||||
register short s1;
|
||||
register short s2;
|
||||
register int s1;
|
||||
register int s2;
|
||||
|
||||
register unsigned int cp_temp = so->cp;
|
||||
|
||||
wf = so->wf;
|
||||
|
||||
const int mode_mask24 = wf->mode&24;
|
||||
const int mode_mask28 = wf->mode&28;
|
||||
const int mode_mask_looprev = wf->mode&LOOP_REVERSE;
|
||||
|
||||
/* Is voice being ramped? */
|
||||
if(so->state == STATE_RAMPDOWN)
|
||||
const unsigned int num_samples = (wf->numSamples-1) << FRACTSIZE;
|
||||
|
||||
const unsigned int end_loop = wf->endLoop << FRACTSIZE;
|
||||
const unsigned int start_loop = wf->startLoop << FRACTSIZE;
|
||||
const int diff_loop = end_loop-start_loop;
|
||||
|
||||
while(samples > 0)
|
||||
{
|
||||
if(so->decay != 0) /* Ramp has been started */
|
||||
samples--;
|
||||
/* Is voice being ramped? */
|
||||
if(so->state == STATE_RAMPDOWN)
|
||||
{
|
||||
so->decay = so->decay / 2;
|
||||
if(so->decay != 0) /* Ramp has been started */
|
||||
{
|
||||
so->decay = so->decay / 2;
|
||||
|
||||
if(so->decay < 10 && so->decay > -10)
|
||||
so->isUsed = 0;
|
||||
if(so->decay < 10 && so->decay > -10)
|
||||
so->isUsed = 0;
|
||||
|
||||
return so->decay;
|
||||
s1=so->decay;
|
||||
s2 = s1*chPan[so->ch];
|
||||
s1 = (s1<<7) -s2;
|
||||
*(out++)+=(((s1&0x7FFF80) << 9) | ((s2&0x7FFF80) >> 7));
|
||||
continue;
|
||||
}
|
||||
} else /* OK to advance voice */
|
||||
{
|
||||
cp_temp += so->delta;
|
||||
}
|
||||
} else /* OK to advance voice */
|
||||
{
|
||||
so->cp += so->delta;
|
||||
}
|
||||
|
||||
|
||||
cpShifted = so->cp >> FRACTSIZE;
|
||||
|
||||
|
||||
|
||||
s2 = getSample((cpShifted)+1, wf);
|
||||
s2 = getSample((cp_temp >> FRACTSIZE)+1, wf);
|
||||
|
||||
/* LOOP_REVERSE|LOOP_PINGPONG = 24 */
|
||||
if((wf->mode & (24)) && so->loopState == STATE_LOOPING && (cpShifted < (wf->startLoop)))
|
||||
{
|
||||
if(wf->mode & LOOP_REVERSE)
|
||||
if(mode_mask24 && so->loopState == STATE_LOOPING && (cp_temp < start_loop))
|
||||
{
|
||||
cpShifted = wf->endLoop-(wf->startLoop-cpShifted);
|
||||
so->cp = (cpShifted)<<FRACTSIZE;
|
||||
s2=getSample((cpShifted), wf);
|
||||
}
|
||||
else
|
||||
{
|
||||
so->delta = -so->delta; /* At this point cpShifted is wrong. We need to take a step */
|
||||
so->loopDir = LOOPDIR_FORWARD;
|
||||
}
|
||||
}
|
||||
|
||||
if((wf->mode & 28) && (cpShifted >= wf->endLoop))
|
||||
{
|
||||
so->loopState = STATE_LOOPING;
|
||||
if((wf->mode & (24)) == 0)
|
||||
{
|
||||
cpShifted = wf->startLoop + (cpShifted-wf->endLoop);
|
||||
so->cp = (cpShifted)<<FRACTSIZE;
|
||||
s2=getSample((cpShifted), wf);
|
||||
}
|
||||
else
|
||||
{
|
||||
so->delta = -so->delta;
|
||||
so->loopDir = LOOPDIR_REVERSE;
|
||||
}
|
||||
}
|
||||
|
||||
/* Have we overrun? */
|
||||
if( (cpShifted >= (wf->numSamples-1)))
|
||||
{
|
||||
so->cp -= so->delta;
|
||||
cpShifted = so->cp >> FRACTSIZE;
|
||||
s2 = getSample((cpShifted)+1, wf);
|
||||
stopVoice(so);
|
||||
}
|
||||
|
||||
|
||||
/* Better, working, linear interpolation */
|
||||
s1=getSample((cpShifted), wf);
|
||||
|
||||
s = s1 + ((signed)((s2 - s1) * (so->cp & ((1<<FRACTSIZE)-1)))>>FRACTSIZE);
|
||||
|
||||
|
||||
if(so->curRate == 0)
|
||||
{
|
||||
stopVoice(so);
|
||||
// so->isUsed = 0;
|
||||
|
||||
}
|
||||
|
||||
if(so->ch != 9 && so->state != STATE_RAMPDOWN) /* Stupid ADSR code... and don't do ADSR for drums */
|
||||
{
|
||||
if(so->curOffset < so->targetOffset)
|
||||
{
|
||||
so->curOffset += (so->curRate);
|
||||
if(so -> curOffset > so->targetOffset && so->curPoint != 2)
|
||||
if(mode_mask_looprev)
|
||||
{
|
||||
if(so->curPoint != 5)
|
||||
{
|
||||
setPoint(so, so->curPoint+1);
|
||||
}
|
||||
else
|
||||
{
|
||||
stopVoice(so);
|
||||
}
|
||||
cp_temp += diff_loop;
|
||||
s2=getSample((cp_temp >> FRACTSIZE), wf);
|
||||
}
|
||||
} else
|
||||
{
|
||||
so->curOffset -= (so->curRate);
|
||||
if(so -> curOffset < so->targetOffset && so->curPoint != 2)
|
||||
else
|
||||
{
|
||||
|
||||
if(so->curPoint != 5)
|
||||
{
|
||||
setPoint(so, so->curPoint+1);
|
||||
}
|
||||
else
|
||||
{
|
||||
stopVoice(so);
|
||||
}
|
||||
|
||||
so->delta = -so->delta; /* At this point cp_temp is wrong. We need to take a step */
|
||||
so->loopDir = LOOPDIR_FORWARD;
|
||||
}
|
||||
}
|
||||
|
||||
if(mode_mask28 && (cp_temp >= end_loop))
|
||||
{
|
||||
so->loopState = STATE_LOOPING;
|
||||
if(!mode_mask24)
|
||||
{
|
||||
cp_temp -= diff_loop;
|
||||
s2=getSample((cp_temp >> FRACTSIZE), wf);
|
||||
}
|
||||
else
|
||||
{
|
||||
so->delta = -so->delta;
|
||||
so->loopDir = LOOPDIR_REVERSE;
|
||||
}
|
||||
}
|
||||
|
||||
/* Have we overrun? */
|
||||
if(cp_temp >= num_samples)
|
||||
{
|
||||
cp_temp -= so->delta;
|
||||
s2 = getSample((cp_temp >> FRACTSIZE)+1, wf);
|
||||
stopVoice(so);
|
||||
}
|
||||
|
||||
/* Better, working, linear interpolation */
|
||||
s1=getSample((cp_temp >> FRACTSIZE), wf);
|
||||
|
||||
s = s1 + ((signed)((s2 - s1) * (cp_temp & ((1<<FRACTSIZE)-1)))>>FRACTSIZE);
|
||||
|
||||
if(so->curRate == 0)
|
||||
{
|
||||
stopVoice(so);
|
||||
// so->isUsed = 0;
|
||||
|
||||
}
|
||||
|
||||
if(so->ch != 9 && so->state != STATE_RAMPDOWN) /* Stupid ADSR code... and don't do ADSR for drums */
|
||||
{
|
||||
if(so->curOffset < so->targetOffset)
|
||||
{
|
||||
so->curOffset += (so->curRate);
|
||||
if(so -> curOffset > so->targetOffset && so->curPoint != 2)
|
||||
{
|
||||
if(so->curPoint != 5)
|
||||
{
|
||||
setPoint(so, so->curPoint+1);
|
||||
}
|
||||
else
|
||||
{
|
||||
stopVoice(so);
|
||||
}
|
||||
}
|
||||
} else
|
||||
{
|
||||
so->curOffset -= (so->curRate);
|
||||
if(so -> curOffset < so->targetOffset && so->curPoint != 2)
|
||||
{
|
||||
|
||||
if(so->curPoint != 5)
|
||||
{
|
||||
setPoint(so, so->curPoint+1);
|
||||
}
|
||||
else
|
||||
{
|
||||
stopVoice(so);
|
||||
}
|
||||
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
if(so->curOffset < 0)
|
||||
{
|
||||
so->curOffset = so->targetOffset;
|
||||
stopVoice(so);
|
||||
}
|
||||
|
||||
s = (s * (so->curOffset >> 22) >> 8);
|
||||
|
||||
/* need to set ramp beginning */
|
||||
if(so->state == STATE_RAMPDOWN && so->decay == 0)
|
||||
{
|
||||
so->decay = s*so->volscale>>14;
|
||||
if(so->decay == 0)
|
||||
so->decay = 1; /* stupid junk.. */
|
||||
}
|
||||
|
||||
|
||||
/* Scaling by channel volume and note volume is done in sequencer.c */
|
||||
/* That saves us some multiplication and pointer operations */
|
||||
s1=s*so->volscale>>14;
|
||||
|
||||
s2 = s1*chPan[so->ch];
|
||||
s1 = (s1<<7) - s2;
|
||||
*(out++)+=(((s1&0x7FFF80) << 9) | ((s2&0x7FFF80) >> 7));
|
||||
|
||||
}
|
||||
|
||||
if(so->curOffset < 0)
|
||||
{
|
||||
so->curOffset = so->targetOffset;
|
||||
stopVoice(so);
|
||||
}
|
||||
|
||||
s = (s * (so->curOffset >> 22) >> 8);
|
||||
|
||||
|
||||
/* need to set ramp beginning */
|
||||
if(so->state == STATE_RAMPDOWN && so->decay == 0)
|
||||
{
|
||||
so->decay = s*so->volscale>>14;
|
||||
if(so->decay == 0)
|
||||
so->decay = 1; /* stupid junk.. */
|
||||
}
|
||||
|
||||
|
||||
/* Scaling by channel volume and note volume is done in sequencer.c */
|
||||
/* That saves us some multiplication and pointer operations */
|
||||
return s*so->volscale>>14;
|
||||
so->cp=cp_temp; /* store this again */
|
||||
return;
|
||||
}
|
||||
|
||||
/* buffer to hold all the samples for the current tick, this is a hack
|
||||
neccesary for coldfire targets as pcm_play_data uses the dma which cannot
|
||||
access iram */
|
||||
int32_t samp_buf[256] IBSS_ATTR;
|
||||
|
||||
/* synth num_samples samples and write them to the */
|
||||
/* buffer pointed to by buf_ptr */
|
||||
void synthSamples(int32_t *buf_ptr, unsigned int num_samples) ICODE_ATTR;
|
||||
void synthSamples(int32_t *buf_ptr, unsigned int num_samples)
|
||||
{
|
||||
int i;
|
||||
register int dL;
|
||||
register int dR;
|
||||
register int sample;
|
||||
register struct SynthObject *voicept;
|
||||
while(num_samples>0)
|
||||
struct SynthObject *voicept;
|
||||
|
||||
rb->memset(samp_buf, 0, num_samples*4);
|
||||
|
||||
for(i=0; i < MAX_VOICES; i++)
|
||||
{
|
||||
dL=0;
|
||||
dR=0;
|
||||
voicept=&voices[0];
|
||||
|
||||
for(i=MAX_VOICES; i > 0; i--)
|
||||
voicept=&voices[i];
|
||||
if(voicept->isUsed==1)
|
||||
{
|
||||
if(voicept->isUsed==1)
|
||||
{
|
||||
sample = synthVoice(voicept);
|
||||
dL += sample;
|
||||
sample *= chPan[voicept->ch];
|
||||
dR += sample;
|
||||
}
|
||||
voicept++;
|
||||
synthVoice(voicept, samp_buf, num_samples);
|
||||
}
|
||||
|
||||
dL = (dL << 7) - dR;
|
||||
|
||||
/* combine the left and right 16 bit samples into 32 bits and write */
|
||||
/* to the buffer, left sample in the high word and right in the low word */
|
||||
*buf_ptr=(((dL&0x7FFF80) << 9) | ((dR&0x7FFF80) >> 7));
|
||||
|
||||
buf_ptr++;
|
||||
num_samples--;
|
||||
}
|
||||
|
||||
rb->memcpy(buf_ptr, samp_buf, num_samples*4);
|
||||
|
||||
/* TODO: Automatic Gain Control, anyone? */
|
||||
/* Or, should this be implemented on the DSP's output volume instead? */
|
||||
|
||||
|
|
|
|||
Loading…
Add table
Add a link
Reference in a new issue