int WDL_ConvolutionEngine_Div::Avail(int wantSamples)
{
timingEnter(1);
int wso=wantSamples;
int x;
#ifdef WDLCONVO_ZL_ACCOUNTING
int cnt=0;
static int maxcnt=-1;
int h=0;
#endif
for (x = 0; x < m_engines.GetSize(); x ++)
{
WDL_ConvolutionEngine *eng=m_engines.Get(x);
#ifdef WDLCONVO_ZL_ACCOUNTING
eng->m_zl_fftcnt=0;
#endif
int a=eng->Avail(wso+eng->m_zl_dumpage) - eng->m_zl_dumpage;
#ifdef WDLCONVO_ZL_ACCOUNTING
cnt += !!eng->m_zl_fftcnt;
#if 0
if (eng->m_zl_fftcnt)
h|=1<<x;
if (eng->m_zl_fftcnt && x==m_engines.GetSize()-1 && cnt>1)
{
char buf[512];
wsprintf(buf,"fft flags=%08x (%08x=max)\n",h,1<<x);
OutputDebugString(buf);
}
#endif
#endif
if (a < wantSamples) wantSamples=a;
}
#ifdef WDLCONVO_ZL_ACCOUNTING
static DWORD lastt=0;
if (cnt>maxcnt)maxcnt=cnt;
if (GetTickCount()>lastt+1000)
{
lastt=GetTickCount();
char buf[512];
wsprintf(buf,"maxcnt=%d\n",maxcnt);
OutputDebugString(buf);
maxcnt=-1;
}
#endif
if (wantSamples>0)
{
WDL_FFT_REAL *tp[WDL_CONVO_MAX_PROC_NCH];
for (x =0; x < m_proc_nch; x ++)
{
memset(tp[x]=(WDL_FFT_REAL*)m_samplesout[x].Add(NULL,wantSamples*sizeof(WDL_FFT_REAL)),0,wantSamples*sizeof(WDL_FFT_REAL));
}
for (x = 0; x < m_engines.GetSize(); x ++)
{
WDL_ConvolutionEngine *eng=m_engines.Get(x);
if (eng->m_zl_dumpage>0) { eng->Advance(eng->m_zl_dumpage); eng->m_zl_dumpage=0; }
WDL_FFT_REAL **p=eng->Get();
if (p)
{
int i;
for (i =0; i < m_proc_nch; i ++)
{
WDL_FFT_REAL *o=tp[i];
WDL_FFT_REAL *in=p[i];
int j=wantSamples;
while (j-->0) *o++ += *in++;
}
}
eng->Advance(wantSamples);
}
}
timingLeave(1);
int av=m_samplesout[0].Available()/sizeof(WDL_FFT_REAL);
return av>wso ? wso : av;
}
void C_THISCLASS::smp_render(int this_thread, int max_threads, char visdata[2][2][576], int isBeat, int *framebuffer, int *fbout, int w, int h)
{
if (!enabled) return;
unsigned int *f = (unsigned int *) framebuffer;
unsigned int *of = (unsigned int *) fbout;
unsigned int *lfo = (unsigned int *) lastframe;
int start_l = ( this_thread * h ) / max_threads;
int end_l;
if (this_thread >= max_threads - 1) end_l = h;
else end_l = ( (this_thread+1) * h ) / max_threads;
int outh=end_l-start_l;
if (outh<1) return;
int skip_pix=start_l*w;
f += skip_pix;
of+= skip_pix;
lfo += skip_pix;
int at_top=0, at_bottom=0;
if (!this_thread) at_top=1;
if (this_thread >= max_threads - 1) at_bottom=1;
timingEnter(0);
{
if (at_top)
// top line
{
int x;
// left edge
{
int r=_R(f[1]); int g=_G(f[1]); int b=_B(f[1]);
r += _R(f[w]); g += _G(f[w]); b += _B(f[w]);
f++;
r-=_R(lfo[0]); g-=_G(lfo[0]); b-=_B(lfo[0]);
lfo++;
if (r < 0) r=0;
else if (r > 255) r=255;
if (g < 0) g=0;
else if (g > 255*256) g=255*256;
if (b < 0) b=0;
else if (b > 255*65536) b=255*65536;
*of++=_RGB(r,g,b);
}
// middle of line
x=(w-2);
while (x--)
{
int r=_R(f[1]); int g=_G(f[1]); int b=_B(f[1]);
r += _R(f[-1]); g += _G(f[-1]); b += _B(f[-1]);
r += _R(f[w]); g += _G(f[w]); b += _B(f[w]);
f++;
r/=2; g/=2; b/=2;
r-=_R(lfo[0]); g-=_G(lfo[0]); b-=_B(lfo[0]);
lfo++;
if (r < 0) r=0;
else if (r > 255) r=255;
if (g < 0) g=0;
else if (g > 255*256) g=255*256;
if (b < 0) b=0;
else if (b > 255*65536) b=255*65536;
*of++=_RGB(r,g,b);
}
// right block
{
int r=_R(f[-1]); int g=_G(f[-1]); int b=_B(f[-1]);
r += _R(f[w]); g += _G(f[w]); b += _B(f[w]);
f++;
r-=_R(lfo[0]); g-=_G(lfo[0]); b-=_B(lfo[0]);
lfo++;
if (r < 0) r=0;
else if (r > 255) r=255;
if (g < 0) g=0;
else if (g > 255*256) g=255*256;
if (b < 0) b=0;
else if (b > 255*65536) b=255*65536;
*of++=_RGB(r,g,b);
}
}
// middle block
{
int y=outh-at_top-at_bottom;
while (y--)
{
int x;
// left edge
{
int r=_R(f[1]); int g=_G(f[1]); int b=_B(f[1]);
r += _R(f[w]); g += _G(f[w]); b += _B(f[w]);
r += _R(f[-w]); g += _G(f[-w]); b += _B(f[-w]);
f++;
r/=2; g/=2; b/=2;
r-=_R(lfo[0]); g-=_G(lfo[0]); b-=_B(lfo[0]);
lfo++;
if (r < 0) r=0;
else if (r > 255) r=255;
if (g < 0) g=0;
else if (g > 255*256) g=255*256;
if (b < 0) b=0;
else if (b > 255*65536) b=255*65536;
*of++=_RGB(r,g,b);
}
// middle of line
x=(w-2);
#ifdef NO_MMX
while (x--)
{
int r=_R(f[1]); int g=_G(f[1]); int b=_B(f[1]);
r += _R(f[-1]); g += _G(f[-1]); b += _B(f[-1]);
r += _R(f[w]); g += _G(f[w]); b += _B(f[w]);
r += _R(f[-w]); g += _G(f[-w]); b += _B(f[-w]);
f++;
r/=2; g/=2; b/=2;
r-=_R(lfo[0]); g-=_G(lfo[0]); b-=_B(lfo[0]);
lfo++;
if (r < 0) r=0;
else if (r > 255) r=255;
if (g < 0) g=0;
else if (g > 255*256) g=255*256;
if (b < 0) b=0;
else if (b > 255*65536) b=255*65536;
*of++=_RGB(r,g,b);
}
#else
__asm
{
mov esi, f
mov edi, of
mov edx, lfo
mov ecx, x
mov ebx, w
shl ebx, 2
shr ecx, 1
sub esi, ebx
align 16
mmx_water_loop1:
movd mm0, [esi+ebx+4]
movd mm1, [esi+ebx-4]
punpcklbw mm0, [zero]
movd mm2, [esi+ebx*2]
punpcklbw mm1, [zero]
movd mm3, [esi]
punpcklbw mm2, [zero]
movd mm4, [edx]
paddw mm0, mm1
punpcklbw mm3, [zero]
movd mm7, [esi+ebx+8]
punpcklbw mm4, [zero]
paddw mm2, mm3
movd mm6, [esi+ebx]
paddw mm0, mm2
psrlw mm0, 1
punpcklbw mm7, [zero]
movd mm2, [esi+ebx*2+4]
psubw mm0, mm4
movd mm3, [esi+4]
packuswb mm0, mm0
movd [edi], mm0
punpcklbw mm6, [zero]
movd mm4, [edx+4]
punpcklbw mm2, [zero]
paddw mm7, mm6
punpcklbw mm3, [zero]
punpcklbw mm4, [zero]
paddw mm2, mm3
paddw mm7, mm2
add edx, 8
psrlw mm7, 1
add esi, 8
psubw mm7, mm4
packuswb mm7, mm7
movd [edi+4], mm7
add edi, 8
dec ecx
jnz mmx_water_loop1
add esi, ebx
mov f, esi
mov of, edi
mov lfo, edx
};
#endif
// right block
{
int r=_R(f[-1]); int g=_G(f[-1]); int b=_B(f[-1]);
r += _R(f[w]); g += _G(f[w]); b += _B(f[w]);
r += _R(f[-w]); g += _G(f[-w]); b += _B(f[-w]);
f++;
r/=2; g/=2; b/=2;
r-=_R(lfo[0]); g-=_G(lfo[0]); b-=_B(lfo[0]);
lfo++;
if (r < 0) r=0;
else if (r > 255) r=255;
if (g < 0) g=0;
else if (g > 255*256) g=255*256;
if (b < 0) b=0;
else if (b > 255*65536) b=255*65536;
*of++=_RGB(r,g,b);
}
}
}
// bottom line
if (at_bottom)
{
int x;
// left edge
{
int r=_R(f[1]); int g=_G(f[1]); int b=_B(f[1]);
r += _R(f[-w]); g += _G(f[-w]); b += _B(f[-w]);
f++;
r-=_R(lfo[0]); g-=_G(lfo[0]); b-=_B(lfo[0]);
lfo++;
if (r < 0) r=0;
else if (r > 255) r=255;
if (g < 0) g=0;
else if (g > 255*256) g=255*256;
if (b < 0) b=0;
else if (b > 255*65536) b=255*65536;
*of++=_RGB(r,g,b);
}
// middle of line
x=(w-2);
while (x--)
{
int r=_R(f[1]); int g=_G(f[1]); int b=_B(f[1]);
r += _R(f[-1]); g += _G(f[-1]); b += _B(f[-1]);
r += _R(f[-w]); g += _G(f[-w]); b += _B(f[-w]);
f++;
r/=2; g/=2; b/=2;
r-=_R(lfo[0]); g-=_G(lfo[0]); b-=_B(lfo[0]);
lfo++;
if (r < 0) r=0;
else if (r > 255) r=255;
if (g < 0) g=0;
else if (g > 255*256) g=255*256;
if (b < 0) b=0;
else if (b > 255*65536) b=255*65536;
*of++=_RGB(r,g,b);
}
// right block
{
int r=_R(f[-1]); int g=_G(f[-1]); int b=_B(f[-1]);
r += _R(f[-w]); g += _G(f[-w]); b += _B(f[-w]);
f++;
r-=_R(lfo[0]); g-=_G(lfo[0]); b-=_B(lfo[0]);
lfo++;
if (r < 0) r=0;
else if (r > 255) r=255;
if (g < 0) g=0;
else if (g > 255*256) g=255*256;
if (b < 0) b=0;
else if (b > 255*65536) b=255*65536;
*of++=_RGB(r,g,b);
}
}
}
memcpy(lastframe+skip_pix,framebuffer+skip_pix,w*outh*sizeof(int));
#ifndef NO_MMX
__asm emms;
#endif
timingLeave(0);
}
C_RBASE *R_Water(char *desc)
{
if (desc) { strcpy(desc,MOD_NAME); return NULL; }
return (C_RBASE *) new C_THISCLASS();
}
static C_THISCLASS *g_this;
static BOOL CALLBACK g_DlgProc(HWND hwndDlg, UINT uMsg, WPARAM wParam,LPARAM lParam)
{
switch (uMsg)
{
case WM_INITDIALOG:
if (g_this->enabled) CheckDlgButton(hwndDlg,IDC_CHECK1,BST_CHECKED);
return 1;
case WM_COMMAND:
if (LOWORD(wParam) == IDC_CHECK1)
{
if (IsDlgButtonChecked(hwndDlg,IDC_CHECK1))
g_this->enabled=1;
else
g_this->enabled=0;
}
return 0;
}
return 0;
}
HWND C_THISCLASS::conf(HINSTANCE hInstance, HWND hwndParent)
{
g_this = this;
return CreateDialog(hInstance,MAKEINTRESOURCE(IDD_CFG_WATER),hwndParent,g_DlgProc);
}
