PC Welt 2004 March

home *** CD-ROM | disk | FTP | other *** search

/ PC Welt 2004 March / PCWELT_3_2004.ISO / pcwsoft / flaskmpeg_078_39_src.z.exe / flaskmpeg / Audio / MPEG / polyphase.cpp < prev next >

Wrap

C/C++ Source or Header | 2002-10-28 | 62.4 KB | 2,097 lines

#include <math.h> #include <string.h> #include <crtdbg.h> #include <stdio.h> #include "AMPDecoder.h" #define USE_ASM static float wincoef[16][32] = { /* window coefs */ { 0.000000000f, 0.000442505f, -0.003250122f, 0.007003784f, -0.031082151f, 0.078628540f, -0.100311279f, 0.572036743f, -1.144989014f, -0.572036743f, -0.100311279f, -0.078628540f, -0.031082151f, -0.007003784f, -0.003250122f, -0.000442505f, },{ 0.000015259f, 0.000473022f, -0.003326416f, 0.007919312f, -0.030517576f, 0.084182739f, -0.090927124f, 0.600219727f, -1.144287109f, -0.543823242f, -0.108856201f, -0.073059082f, -0.031478882f, -0.006118774f, -0.003173828f, -0.000396729f, },{ 0.000015259f, 0.000534058f, -0.003387451f, 0.008865356f, -0.029785154f, 0.089706421f, -0.080688477f, 0.628295898f, -1.142211914f, -0.515609741f, -0.116577141f, -0.067520142f, -0.031738281f, -0.005294800f, -0.003082275f, -0.000366211f, },{ 0.000015259f, 0.000579834f, -0.003433228f, 0.009841919f, -0.028884888f, 0.095169067f, -0.069595337f, 0.656219482f, -1.138763428f, -0.487472534f, -0.123474121f, -0.061996460f, -0.031845093f, -0.004486084f, -0.002990723f, -0.000320435f, },{ 0.000015259f, 0.000625610f, -0.003463745f, 0.010848999f, -0.027801514f, 0.100540161f, -0.057617184f, 0.683914185f, -1.133926392f, -0.459472656f, -0.129577637f, -0.056533810f, -0.031814575f, -0.003723145f, -0.002899170f, -0.000289917f, },{ 0.000015259f, 0.000686646f, -0.003479004f, 0.011886597f, -0.026535034f, 0.105819702f, -0.044784546f, 0.711318970f, -1.127746582f, -0.431655884f, -0.134887695f, -0.051132202f, -0.031661987f, -0.003005981f, -0.002792358f, -0.000259399f, },{ 0.000015259f, 0.000747681f, -0.003479004f, 0.012939452f, -0.025085449f, 0.110946655f, -0.031082151f, 0.738372803f, -1.120223999f, -0.404083252f, -0.139450073f, -0.045837402f, -0.031387329f, -0.002334595f, -0.002685547f, -0.000244141f, },{ 0.000030518f, 0.000808716f, -0.003463745f, 0.014022826f, -0.023422241f, 0.115921021f, -0.016510010f, 0.765029907f, -1.111373901f, -0.376800537f, -0.143264771f, -0.040634155f, -0.031005858f, -0.001693726f, -0.002578735f, -0.000213623f, },{ 0.000030518f, 0.000885010f, -0.003417969f, 0.015121460f, -0.021575928f, 0.120697014f, -0.001068115f, 0.791213989f, -1.101211548f, -0.349868774f, -0.146362305f, -0.035552979f, -0.030532837f, -0.001098633f, -0.002456665f, -0.000198364f, },{ 0.000030518f, 0.000961304f, -0.003372192f, 0.016235352f, -0.019531250f, 0.125259399f, 0.015228271f, 0.816864014f, -1.089782715f, -0.323318481f, -0.148773193f, -0.030609131f, -0.029937742f, -0.000549316f, -0.002349854f, -0.000167847f, },{ 0.000030518f, 0.001037598f, -0.003280640f, 0.017349243f, -0.017257690f, 0.129562378f, 0.032379150f, 0.841949463f, -1.077117920f, -0.297210693f, -0.150497437f, -0.025817871f, -0.029281614f, -0.000030518f, -0.002243042f, -0.000152588f, },{ 0.000045776f, 0.001113892f, -0.003173828f, 0.018463135f, -0.014801024f, 0.133590698f, 0.050354004f, 0.866363525f, -1.063217163f, -0.271591187f, -0.151596069f, -0.021179199f, -0.028533936f, 0.000442505f, -0.002120972f, -0.000137329f, },{ 0.000045776f, 0.001205444f, -0.003051758f, 0.019577026f, -0.012115479f, 0.137298584f, 0.069168091f, 0.890090942f, -1.048156738f, -0.246505737f, -0.152069092f, -0.016708374f, -0.027725220f, 0.000869751f, -0.002014160f, -0.000122070f, },{ 0.000061035f, 0.001296997f, -0.002883911f, 0.020690918f, -0.009231566f, 0.140670776f, 0.088775635f, 0.913055420f, -1.031936646f, -0.221984863f, -0.151962280f, -0.012420653f, -0.026840210f, 0.001266479f, -0.001907349f, -0.000106812f, },{ 0.000061035f, 0.001388550f, -0.002700806f, 0.021789551f, -0.006134033f, 0.143676758f, 0.109161377f, 0.935195923f, -1.014617920f, -0.198059082f, -0.151306152f, -0.008316040f, -0.025909424f, 0.001617432f, -0.001785278f, -0.000106812f, },{ 0.000076294f, 0.001480103f, -0.002487183f, 0.022857666f, -0.002822876f, 0.146255493f, 0.130310059f, 0.956481934f, -0.996246338f, -0.174789429f, -0.150115967f, -0.004394531f, -0.024932859f, 0.001937866f, -0.001693726f, -0.000091553f, } }; static const float wincoef2[]={ -0.001586914f, -0.023910521f, -0.148422241f, -0.976852417f, 0.152206421f, 0.000686646f, -0.002227783f, 0.000076294f, }; static const float coef32[31]={ /* 32 pt dct coefs */ .5006029983604062f, .5054709610427671f, .5154473132093669f, .5310425978615345f, .553103908021273f, .5829349877738026f, .6225041535926976f, .674808388163145f, .7445363421017894f, .8393497548926808f, .9725684113484575f, 1.16944022296938f, 1.484165144274385f, 2.057782139240278f, 3.407611810249888f, 10.19004086100494f, .5024192866920285f, .5224986197814337f, .5669440502448403f, .6468218211965951f, .7881547114622637f, 1.060677908257709f, 1.722447849872342f, 5.101156529442028f, .5097955811692735f, .6013449114834756f, .8999763602860835f, 2.562917284521227f, .5411961092767109f, 1.306563350304716f, .7071068387877132f, }; static void __inline forward_bf(int m, int n, const float x[], float f[], const float coef[]) { int j, n2; int p, q, p0; n2 = n >> 1; for (p0 = 0; p0 < 32; p0 += n) { p = p0; q = p + n - 1; for (j = 0; j < n2; j++, p++, q--) { f[p] = x[p] + x[q]; f[n2 + p] = coef[j] * (x[p] - x[q]); } } } /*------------------------------------------------------------*/ static void __inline back_bf(int m, int n, const float x[], float f[]) { int j, n2, n21; int p, q, p0; n2 = n >> 1; n21 = n2 - 1; for (p0 = 0; p0 < 32; p0 += n) { p = p0; q = p0; for (j = 0; j < n2; j++, p += 2, q++) f[p] = x[q]; p = p0 + 1; for (j = 0; j < n21; j++, p += 2, q++) f[p] = x[q] + x[q + 1]; f[p] = x[q]; } } /*------------------------------------------------------------*/ void fdct32(float x[], float c[], bool odd, bool invert_odd) { float a[32]; /* ping pong buffers */ float b[32]; int p, q; // special first stage // forward_bf(2, 16, a, b, coef32 + 16); if (invert_odd) { q = 15; for(p = 0; p < 8; p+=2, q-=2) { double a1, a2, a3, a4; // p: normal, q:inverted. a1 = x[p] - x[16+q]; a2 = -x[q] + x[16+p]; a3 = (x[p] + x[16+q]) * coef32[p]; a4 = (-x[q] - x[16+p]) * coef32[q]; b[p ] = a1 + a2; b[p+ 8] = (a1 - a2) * coef32[16+p]; b[p+16] = a3 + a4; b[p+24] = (a3 - a4) * coef32[16+p]; // p: inverted, q: normal. a1 = -x[p+1] + x[15+q]; a2 = x[q-1] - x[17+p]; a3 = (-x[p+1] - x[15+q]) * coef32[p+1]; a4 = (x[q-1] + x[17+p]) * coef32[q-1]; b[p+ 1] = a1 + a2; b[p+ 9] = (a1 - a2) * coef32[17+p]; b[p+17] = a3 + a4; b[p+25] = (a3 - a4) * coef32[17+p]; } } else { q = 15; for(p = 0; p < 8; p++, q--) { double a1 = x[p] + x[16+q]; double a2 = x[q] + x[16+p]; double a3 = (x[p] - x[16+q]) * coef32[p]; double a4 = (x[q] - x[16+p]) * coef32[q]; b[p ] = a1 + a2; b[p+ 8] = (a1 - a2) * coef32[16+p]; b[p+16] = a3 + a4; b[p+24] = (a3 - a4) * coef32[16+p]; } } forward_bf(4, 8, b, a, coef32 + 16 + 8); // forward_bf(8, 4, a, b, coef32 + 16 + 8 + 4); // forward_bf(16, 2, b, a, coef32 + 16 + 8 + 4 + 2); // back_bf(8, 4, a, b); // back_bf(4, 8, b, a); // back_bf(2, 16, a, b); for (p = 0; p < 32; p += 16) { double a0, a1, a2, a3; double b0, b1, b2, b3; double d0, d1, d2, d3, d4, d5, d6, d7; double d8, d9, dA, dB, dC, dD, dE, dF; b0 = a[p ] + a[p+3]; b1 = a[p+1] + a[p+2]; b2 = coef32[28] * (a[p ] - a[p+3]); b3 = coef32[29] * (a[p+1] - a[p+2]); a0 = b0 + b1; a1 = coef32[30] * (b0 - b1); a2 = b2 + b3; a3 = coef32[30] * (b2 - b3); d0 = a0; d1 = a2 + a3; d2 = a1; d3 = a3; b0 = a[p+4] + a[p+7]; b1 = a[p+5] + a[p+6]; b2 = coef32[28] * (a[p+4] - a[p+7]); b3 = coef32[29] * (a[p+5] - a[p+6]); a0 = b0 + b1; a1 = coef32[30] * (b0 - b1); a2 = b2 + b3; a3 = coef32[30] * (b2 - b3); d4 = a0; d5 = a2 + a3; d6 = a1; d7 = a3; b0 = a[p+8] + a[p+11]; b1 = a[p+9] + a[p+10]; b2 = coef32[28] * (a[p+8] - a[p+11]); b3 = coef32[29] * (a[p+9] - a[p+10]); a0 = b0 + b1; a1 = coef32[30] * (b0 - b1); a2 = b2 + b3; a3 = coef32[30] * (b2 - b3); d8 = a0; d9 = a2 + a3; dA = a1; dB = a3; b0 = a[p+12] + a[p+15]; b1 = a[p+13] + a[p+14]; b2 = coef32[28] * (a[p+12] - a[p+15]); b3 = coef32[29] * (a[p+13] - a[p+14]); a0 = b0 + b1; a1 = coef32[30] * (b0 - b1); a2 = b2 + b3; a3 = coef32[30] * (b2 - b3); dC = a0; dD = a2 + a3; dE = a1; dF = a3; ////////////////////////////////////////////////// double s1 = dC + dD; double s2 = dD + dE; double s3 = dE + dF; a[p ] = d0; a[p+ 2] = d4 + d5; a[p+ 4] = d1; a[p+ 6] = d5 + d6; a[p+ 8] = d2; a[p+10] = d6 + d7; a[p+12] = d3; a[p+14] = d7; a[p+ 1] = d8 + s1; a[p+ 3] = d9 + s1; a[p+ 5] = d9 + s2; a[p+ 7] = dA + s2; a[p+ 9] = dA + s3; a[p+11] = dB + s3; a[p+13] = dB + dF; a[p+15] = dF; } // back_bf(1, 32, b, c); // // reordering for windowing: // flip around 256 if odd if (odd) { c[0] = a[0]; c[512 - 16] = a[16] + a[17]; q = 1; for (p = 2; p < 30; p+=2, q++) { c[512 - p*16 ] = a[q]; c[512 - p*16 - 16] = a[q+16] + a[q+17]; } c[ 2*16] = a[15]; c[ 1*16] = a[31]; } else { q = 0; for (p = 0; p < 30; p+=2, q++) { c[p*16+0 ] = a[q]; c[p*16+16] = a[q+16] + a[q+17]; } c[30*16] = a[15]; c[31*16] = a[31]; } } #define float_to_long(x) ((converter = (double)(x)+6755399441055744.0),*(long *)&converter) void window_dual(const float vbuf[2][512], int vb_ptr, short *pcm) { double converter; int i, j; int si, bx; int coef; float *window, *window2; float sum1, sum2, sum3, sum4; long tmp; coef = 0; /* first */ i = 0; if (vb_ptr & 1) { sum1 = 0.0F; sum2 = 0.0F; window = &wincoef[i][(-vb_ptr) & 15]; si = 256-i*16; sum1 -= window[ 0] * vbuf[0][si+ 0]; sum2 -= window[ 0] * vbuf[1][si+ 0]; sum1 += window[ 1] * vbuf[0][si+ 1]; sum2 += window[ 1] * vbuf[1][si+ 1]; sum1 -= window[ 2] * vbuf[0][si+ 2]; sum2 -= window[ 2] * vbuf[1][si+ 2]; sum1 += window[ 3] * vbuf[0][si+ 3]; sum2 += window[ 3] * vbuf[1][si+ 3]; sum1 -= window[ 4] * vbuf[0][si+ 4]; sum2 -= window[ 4] * vbuf[1][si+ 4]; sum1 += window[ 5] * vbuf[0][si+ 5]; sum2 += window[ 5] * vbuf[1][si+ 5]; sum1 -= window[ 6] * vbuf[0][si+ 6]; sum2 -= window[ 6] * vbuf[1][si+ 6]; sum1 += window[ 7] * vbuf[0][si+ 7]; sum2 += window[ 7] * vbuf[1][si+ 7]; sum1 -= window[ 8] * vbuf[0][si+ 8]; sum2 -= window[ 8] * vbuf[1][si+ 8]; sum1 += window[ 9] * vbuf[0][si+ 9]; sum2 += window[ 9] * vbuf[1][si+ 9]; sum1 -= window[10] * vbuf[0][si+10]; sum2 -= window[10] * vbuf[1][si+10]; sum1 += window[11] * vbuf[0][si+11]; sum2 += window[11] * vbuf[1][si+11]; sum1 -= window[12] * vbuf[0][si+12]; sum2 -= window[12] * vbuf[1][si+12]; sum1 += window[13] * vbuf[0][si+13]; sum2 += window[13] * vbuf[1][si+13]; sum1 -= window[14] * vbuf[0][si+14]; sum2 -= window[14] * vbuf[1][si+14]; sum1 += window[15] * vbuf[0][si+15]; sum2 += window[15] * vbuf[1][si+15]; tmp = float_to_long(sum1); if (tmp > 32767) tmp = 32767; else if (tmp < -32768) tmp = -32768; pcm[i*2] = tmp; tmp = float_to_long(sum2); if (tmp > 32767) tmp = 32767; else if (tmp < -32768) tmp = -32768; pcm[i*2+1] = tmp; } else { sum1 = 0.0F; sum2 = 0.0F; window = &wincoef[i][(-vb_ptr) & 15]; si = 256+i*16; sum1 += window[ 0] * vbuf[0][si+ 0]; sum2 += window[ 0] * vbuf[1][si+ 0]; sum1 -= window[ 1] * vbuf[0][si+ 1]; sum2 -= window[ 1] * vbuf[1][si+ 1]; sum1 += window[ 2] * vbuf[0][si+ 2]; sum2 += window[ 2] * vbuf[1][si+ 2]; sum1 -= window[ 3] * vbuf[0][si+ 3]; sum2 -= window[ 3] * vbuf[1][si+ 3]; sum1 += window[ 4] * vbuf[0][si+ 4]; sum2 += window[ 4] * vbuf[1][si+ 4]; sum1 -= window[ 5] * vbuf[0][si+ 5]; sum2 -= window[ 5] * vbuf[1][si+ 5]; sum1 += window[ 6] * vbuf[0][si+ 6]; sum2 += window[ 6] * vbuf[1][si+ 6]; sum1 -= window[ 7] * vbuf[0][si+ 7]; sum2 -= window[ 7] * vbuf[1][si+ 7]; sum1 += window[ 8] * vbuf[0][si+ 8]; sum2 += window[ 8] * vbuf[1][si+ 8]; sum1 -= window[ 9] * vbuf[0][si+ 9]; sum2 -= window[ 9] * vbuf[1][si+ 9]; sum1 += window[10] * vbuf[0][si+10]; sum2 += window[10] * vbuf[1][si+10]; sum1 -= window[11] * vbuf[0][si+11]; sum2 -= window[11] * vbuf[1][si+11]; sum1 += window[12] * vbuf[0][si+12]; sum2 += window[12] * vbuf[1][si+12]; sum1 -= window[13] * vbuf[0][si+13]; sum2 -= window[13] * vbuf[1][si+13]; sum1 += window[14] * vbuf[0][si+14]; sum2 += window[14] * vbuf[1][si+14]; sum1 -= window[15] * vbuf[0][si+15]; sum2 -= window[15] * vbuf[1][si+15]; tmp = float_to_long(sum1); if (tmp > 32767) tmp = 32767; else if (tmp < -32768) tmp = -32768; pcm[i*2] = tmp; tmp = float_to_long(sum2); if (tmp > 32767) tmp = 32767; else if (tmp < -32768) tmp = -32768; pcm[i*2+1] = tmp; } /*-- 1-16, last 15 --*/ if (vb_ptr & 1) { #ifdef USE_ASM window = &wincoef[1][16-vb_ptr]; window2 = &wincoef[1][vb_ptr]; const void *vbuf2 = (void *)&vbuf[0][240]; __asm { push esi push vbuf2 mov ecx,window mov edx,window2 mov ebx,pcm pop esi push edi mov edi,ebx add ebx,2*2 add edi,62*2 push ebp push 0 push 0 push 0 push 0 mov ebp,15 ;ESI = FP input ;EBX = PCM output (low) ;ECX = window ;EDX = window2 ;EDI = PCM output (high) ;EBP = counter ; ;A sum = window[0...15] * vbuf[0] ;B sum = window2[15...0] * vbuf[0] ;C sum = window[0...15] * vbuf[1] ;D sum = window2[15...0] * vbuf[1] div0_loop: fld dword ptr [esi+15*4+0*2048] ;Lx fmul dword ptr [ecx+15*4] ;Ax fld dword ptr [esi+15*4+0*2048] ;Lx Ax fmul dword ptr [edx+ 0*4] ;Bx Ax fld dword ptr [esi+15*4+1*2048] ;Rx Bx Ax fmul dword ptr [ecx+15*4] ;Cx Bx Ax fld dword ptr [esi+15*4+1*2048] ;Rx Cx Bx Ax fmul dword ptr [edx+ 0*4] ;Dx Cx Bx Ax fld dword ptr [esi+14*4+0*2048] ;Lx Dt Ct Bt At fmul dword ptr [ecx+14*4] ;Ax Dt Ct Bt At fld dword ptr [esi+14*4+0*2048] ;Lx Ax Dt Ct Bt At fmul dword ptr [edx+ 1*4] ;Bx Ax Dt Ct Bt At fld dword ptr [esi+14*4+1*2048] ;Rx Bx Ax Dt Ct Bt At fmul dword ptr [ecx+14*4] ;Cx Bx Ax Dt Ct Bt At fld dword ptr [esi+14*4+1*2048] ;Rx Cx Bx Ax Dt Ct Bt At fmul dword ptr [edx+ 1*4] ;Dx Cx Bx Ax Dt Ct Bt At fxch st(3) ;Ax Cx Bx Dx Dt Ct Bt At fsubp st(7),st ;Cx Bx Dx Dt Ct Bt At fxch st(1) ;Bx Cx Dx Dt Ct Bt At faddp st(5),st ;Cx Dx Dt Ct Bt At fsubp st(3),st ;Dx Dt Ct Bt At faddp st(1),st ;Dt Ct Bt At fld dword ptr [esi+13*4+0*2048] ;Lx Dt Ct Bt At fmul dword ptr [ecx+13*4] ;Ax Dt Ct Bt At fld dword ptr [esi+13*4+0*2048] ;Lx Ax Dt Ct Bt At fmul dword ptr [edx+ 2*4] ;Bx Ax Dt Ct Bt At fld dword ptr [esi+13*4+1*2048] ;Rx Bx Ax Dt Ct Bt At fmul dword ptr [ecx+13*4] ;Cx Bx Ax Dt Ct Bt At fld dword ptr [esi+13*4+1*2048] ;Rx Cx Bx Ax Dt Ct Bt At fmul dword ptr [edx+ 2*4] ;Dx Cx Bx Ax Dt Ct Bt At fxch st(3) ;Ax Cx Bx Dx Dt Ct Bt At faddp st(7),st ;Cx Bx Dx Dt Ct Bt At fxch st(1) ;Bx Cx Dx Dt Ct Bt At faddp st(5),st ;Cx Dx Dt Ct Bt At faddp st(3),st ;Dx Dt Ct Bt At faddp st(1),st ;Dt Ct Bt At fld dword ptr [esi+12*4+0*2048] ;Lx Dt Ct Bt At fmul dword ptr [ecx+12*4] ;Ax Dt Ct Bt At fld dword ptr [esi+12*4+0*2048] ;Lx Ax Dt Ct Bt At fmul dword ptr [edx+ 3*4] ;Bx Ax Dt Ct Bt At fld dword ptr [esi+12*4+1*2048] ;Rx Bx Ax Dt Ct Bt At fmul dword ptr [ecx+12*4] ;Cx Bx Ax Dt Ct Bt At fld dword ptr [esi+12*4+1*2048] ;Rx Cx Bx Ax Dt Ct Bt At fmul dword ptr [edx+ 3*4] ;Dx Cx Bx Ax Dt Ct Bt At fxch st(3) ;Ax Cx Bx Dx Dt Ct Bt At fsubp st(7),st ;Cx Bx Dx Dt Ct Bt At fxch st(1) ;Bx Cx Dx Dt Ct Bt At faddp st(5),st ;Cx Dx Dt Ct Bt At fsubp st(3),st ;Dx Dt Ct Bt At faddp st(1),st ;Dt Ct Bt At fld dword ptr [esi+11*4+0*2048] ;Lx Dt Ct Bt At fmul dword ptr [ecx+11*4] ;Ax Dt Ct Bt At fld dword ptr [esi+11*4+0*2048] ;Lx Ax Dt Ct Bt At fmul dword ptr [edx+ 4*4] ;Bx Ax Dt Ct Bt At fld dword ptr [esi+11*4+1*2048] ;Rx Bx Ax Dt Ct Bt At fmul dword ptr [ecx+11*4] ;Cx Bx Ax Dt Ct Bt At fld dword ptr [esi+11*4+1*2048] ;Rx Cx Bx Ax Dt Ct Bt At fmul dword ptr [edx+ 4*4] ;Dx Cx Bx Ax Dt Ct Bt At fxch st(3) ;Ax Cx Bx Dx Dt Ct Bt At faddp st(7),st ;Cx Bx Dx Dt Ct Bt At fxch st(1) ;Bx Cx Dx Dt Ct Bt At faddp st(5),st ;Cx Dx Dt Ct Bt At faddp st(3),st ;Dx Dt Ct Bt At faddp st(1),st ;Dt Ct Bt At fld dword ptr [esi+10*4+0*2048] ;Lx Dt Ct Bt At fmul dword ptr [ecx+10*4] ;Ax Dt Ct Bt At fld dword ptr [esi+10*4+0*2048] ;Lx Ax Dt Ct Bt At fmul dword ptr [edx+ 5*4] ;Bx Ax Dt Ct Bt At fld dword ptr [esi+10*4+1*2048] ;Rx Bx Ax Dt Ct Bt At fmul dword ptr [ecx+10*4] ;Cx Bx Ax Dt Ct Bt At fld dword ptr [esi+10*4+1*2048] ;Rx Cx Bx Ax Dt Ct Bt At fmul dword ptr [edx+ 5*4] ;Dx Cx Bx Ax Dt Ct Bt At fxch st(3) ;Ax Cx Bx Dx Dt Ct Bt At fsubp st(7),st ;Cx Bx Dx Dt Ct Bt At fxch st(1) ;Bx Cx Dx Dt Ct Bt At faddp st(5),st ;Cx Dx Dt Ct Bt At fsubp st(3),st ;Dx Dt Ct Bt At faddp st(1),st ;Dt Ct Bt At fld dword ptr [esi+ 9*4+0*2048] ;Lx Dt Ct Bt At fmul dword ptr [ecx+ 9*4] ;Ax Dt Ct Bt At fld dword ptr [esi+ 9*4+0*2048] ;Lx Ax Dt Ct Bt At fmul dword ptr [edx+ 6*4] ;Bx Ax Dt Ct Bt At fld dword ptr [esi+ 9*4+1*2048] ;Rx Bx Ax Dt Ct Bt At fmul dword ptr [ecx+ 9*4] ;Cx Bx Ax Dt Ct Bt At fld dword ptr [esi+ 9*4+1*2048] ;Rx Cx Bx Ax Dt Ct Bt At fmul dword ptr [edx+ 6*4] ;Dx Cx Bx Ax Dt Ct Bt At fxch st(3) ;Ax Cx Bx Dx Dt Ct Bt At faddp st(7),st ;Cx Bx Dx Dt Ct Bt At fxch st(1) ;Bx Cx Dx Dt Ct Bt At faddp st(5),st ;Cx Dx Dt Ct Bt At faddp st(3),st ;Dx Dt Ct Bt At faddp st(1),st ;Dt Ct Bt At fld dword ptr [esi+ 8*4+0*2048] ;Lx Dt Ct Bt At fmul dword ptr [ecx+ 8*4] ;Ax Dt Ct Bt At fld dword ptr [esi+ 8*4+0*2048] ;Lx Ax Dt Ct Bt At fmul dword ptr [edx+ 7*4] ;Bx Ax Dt Ct Bt At fld dword ptr [esi+ 8*4+1*2048] ;Rx Bx Ax Dt Ct Bt At fmul dword ptr [ecx+ 8*4] ;Cx Bx Ax Dt Ct Bt At fld dword ptr [esi+ 8*4+1*2048] ;Rx Cx Bx Ax Dt Ct Bt At fmul dword ptr [edx+ 7*4] ;Dx Cx Bx Ax Dt Ct Bt At fxch st(3) ;Ax Cx Bx Dx Dt Ct Bt At fsubp st(7),st ;Cx Bx Dx Dt Ct Bt At fxch st(1) ;Bx Cx Dx Dt Ct Bt At faddp st(5),st ;Cx Dx Dt Ct Bt At fsubp st(3),st ;Dx Dt Ct Bt At faddp st(1),st ;Dt Ct Bt At fld dword ptr [esi+ 7*4+0*2048] ;Lx Dt Ct Bt At fmul dword ptr [ecx+ 7*4] ;Ax Dt Ct Bt At fld dword ptr [esi+ 7*4+0*2048] ;Lx Ax Dt Ct Bt At fmul dword ptr [edx+ 8*4] ;Bx Ax Dt Ct Bt At fld dword ptr [esi+ 7*4+1*2048] ;Rx Bx Ax Dt Ct Bt At fmul dword ptr [ecx+ 7*4] ;Cx Bx Ax Dt Ct Bt At fld dword ptr [esi+ 7*4+1*2048] ;Rx Cx Bx Ax Dt Ct Bt At fmul dword ptr [edx+ 8*4] ;Dx Cx Bx Ax Dt Ct Bt At fxch st(3) ;Ax Cx Bx Dx Dt Ct Bt At faddp st(7),st ;Cx Bx Dx Dt Ct Bt At fxch st(1) ;Bx Cx Dx Dt Ct Bt At faddp st(5),st ;Cx Dx Dt Ct Bt At faddp st(3),st ;Dx Dt Ct Bt At faddp st(1),st ;Dt Ct Bt At fld dword ptr [esi+ 6*4+0*2048] ;Lx Dt Ct Bt At fmul dword ptr [ecx+ 6*4] ;Ax Dt Ct Bt At fld dword ptr [esi+ 6*4+0*2048] ;Lx Ax Dt Ct Bt At fmul dword ptr [edx+ 9*4] ;Bx Ax Dt Ct Bt At fld dword ptr [esi+ 6*4+1*2048] ;Rx Bx Ax Dt Ct Bt At fmul dword ptr [ecx+ 6*4] ;Cx Bx Ax Dt Ct Bt At fld dword ptr [esi+ 6*4+1*2048] ;Rx Cx Bx Ax Dt Ct Bt At fmul dword ptr [edx+ 9*4] ;Dx Cx Bx Ax Dt Ct Bt At fxch st(3) ;Ax Cx Bx Dx Dt Ct Bt At fsubp st(7),st ;Cx Bx Dx Dt Ct Bt At fxch st(1) ;Bx Cx Dx Dt Ct Bt At faddp st(5),st ;Cx Dx Dt Ct Bt At fsubp st(3),st ;Dx Dt Ct Bt At faddp st(1),st ;Dt Ct Bt At fld dword ptr [esi+ 5*4+0*2048] ;Lx Dt Ct Bt At fmul dword ptr [ecx+ 5*4] ;Ax Dt Ct Bt At fld dword ptr [esi+ 5*4+0*2048] ;Lx Ax Dt Ct Bt At fmul dword ptr [edx+10*4] ;Bx Ax Dt Ct Bt At fld dword ptr [esi+ 5*4+1*2048] ;Rx Bx Ax Dt Ct Bt At fmul dword ptr [ecx+ 5*4] ;Cx Bx Ax Dt Ct Bt At fld dword ptr [esi+ 5*4+1*2048] ;Rx Cx Bx Ax Dt Ct Bt At fmul dword ptr [edx+10*4] ;Dx Cx Bx Ax Dt Ct Bt At fxch st(3) ;Ax Cx Bx Dx Dt Ct Bt At faddp st(7),st ;Cx Bx Dx Dt Ct Bt At fxch st(1) ;Bx Cx Dx Dt Ct Bt At faddp st(5),st ;Cx Dx Dt Ct Bt At faddp st(3),st ;Dx Dt Ct Bt At faddp st(1),st ;Dt Ct Bt At fld dword ptr [esi+ 4*4+0*2048] ;Lx Dt Ct Bt At fmul dword ptr [ecx+ 4*4] ;Ax Dt Ct Bt At fld dword ptr [esi+ 4*4+0*2048] ;Lx Ax Dt Ct Bt At fmul dword ptr [edx+11*4] ;Bx Ax Dt Ct Bt At fld dword ptr [esi+ 4*4+1*2048] ;Rx Bx Ax Dt Ct Bt At fmul dword ptr [ecx+ 4*4] ;Cx Bx Ax Dt Ct Bt At fld dword ptr [esi+ 4*4+1*2048] ;Rx Cx Bx Ax Dt Ct Bt At fmul dword ptr [edx+11*4] ;Dx Cx Bx Ax Dt Ct Bt At fxch st(3) ;Ax Cx Bx Dx Dt Ct Bt At fsubp st(7),st ;Cx Bx Dx Dt Ct Bt At fxch st(1) ;Bx Cx Dx Dt Ct Bt At faddp st(5),st ;Cx Dx Dt Ct Bt At fsubp st(3),st ;Dx Dt Ct Bt At faddp st(1),st ;Dt Ct Bt At fld dword ptr [esi+ 3*4+0*2048] ;Lx Dt Ct Bt At fmul dword ptr [ecx+ 3*4] ;Ax Dt Ct Bt At fld dword ptr [esi+ 3*4+0*2048] ;Lx Ax Dt Ct Bt At fmul dword ptr [edx+12*4] ;Bx Ax Dt Ct Bt At fld dword ptr [esi+ 3*4+1*2048] ;Rx Bx Ax Dt Ct Bt At fmul dword ptr [ecx+ 3*4] ;Cx Bx Ax Dt Ct Bt At fld dword ptr [esi+ 3*4+1*2048] ;Rx Cx Bx Ax Dt Ct Bt At fmul dword ptr [edx+12*4] ;Dx Cx Bx Ax Dt Ct Bt At fxch st(3) ;Ax Cx Bx Dx Dt Ct Bt At faddp st(7),st ;Cx Bx Dx Dt Ct Bt At fxch st(1) ;Bx Cx Dx Dt Ct Bt At faddp st(5),st ;Cx Dx Dt Ct Bt At faddp st(3),st ;Dx Dt Ct Bt At faddp st(1),st ;Dt Ct Bt At fld dword ptr [esi+ 2*4+0*2048] ;Lx Dt Ct Bt At fmul dword ptr [ecx+ 2*4] ;Ax Dt Ct Bt At fld dword ptr [esi+ 2*4+0*2048] ;Lx Ax Dt Ct Bt At fmul dword ptr [edx+13*4] ;Bx Ax Dt Ct Bt At fld dword ptr [esi+ 2*4+1*2048] ;Rx Bx Ax Dt Ct Bt At fmul dword ptr [ecx+ 2*4] ;Cx Bx Ax Dt Ct Bt At fld dword ptr [esi+ 2*4+1*2048] ;Rx Cx Bx Ax Dt Ct Bt At fmul dword ptr [edx+13*4] ;Dx Cx Bx Ax Dt Ct Bt At fxch st(3) ;Ax Cx Bx Dx Dt Ct Bt At fsubp st(7),st ;Cx Bx Dx Dt Ct Bt At fxch st(1) ;Bx Cx Dx Dt Ct Bt At faddp st(5),st ;Cx Dx Dt Ct Bt At fsubp st(3),st ;Dx Dt Ct Bt At faddp st(1),st ;Dt Ct Bt At fld dword ptr [esi+ 1*4+0*2048] ;Lx Dt Ct Bt At fmul dword ptr [ecx+ 1*4] ;Ax Dt Ct Bt At fld dword ptr [esi+ 1*4+0*2048] ;Lx Ax Dt Ct Bt At fmul dword ptr [edx+14*4] ;Bx Ax Dt Ct Bt At fld dword ptr [esi+ 1*4+1*2048] ;Rx Bx Ax Dt Ct Bt At fmul dword ptr [ecx+ 1*4] ;Cx Bx Ax Dt Ct Bt At fld dword ptr [esi+ 1*4+1*2048] ;Rx Cx Bx Ax Dt Ct Bt At fmul dword ptr [edx+14*4] ;Dx Cx Bx Ax Dt Ct Bt At fxch st(3) ;Ax Cx Bx Dx Dt Ct Bt At faddp st(7),st ;Cx Bx Dx Dt Ct Bt At fxch st(1) ;Bx Cx Dx Dt Ct Bt At faddp st(5),st ;Cx Dx Dt Ct Bt At faddp st(3),st ;Dx Dt Ct Bt At faddp st(1),st ;Dt Ct Bt At fld dword ptr [esi+ 0*4+0*2048] ;Lx Dt Ct Bt At fmul dword ptr [ecx+ 0*4] ;Ax Dt Ct Bt At fld dword ptr [esi+ 0*4+0*2048] ;Lx Ax Dt Ct Bt At fmul dword ptr [edx+15*4] ;Bx Ax Dt Ct Bt At fld dword ptr [esi+ 0*4+1*2048] ;Rx Bx Ax Dt Ct Bt At fmul dword ptr [ecx+ 0*4] ;Cx Bx Ax Dt Ct Bt At fld dword ptr [esi+ 0*4+1*2048] ;Rx Cx Bx Ax Dt Ct Bt At fmul dword ptr [edx+15*4] ;Dx Cx Bx Ax Dt Ct Bt At fxch st(3) ;Ax Cx Bx Dx Dt Ct Bt At fsubp st(7),st ;Cx Bx Dx Dt Ct Bt At fxch st(1) ;Bx Cx Dx Dt Ct Bt At faddp st(5),st ;Cx Dx Dt Ct Bt At fsubp st(3),st ;Dx Dt Ct Bt At faddp st(1),st ;Dt Ct Bt At add ecx,32*4 add edx,32*4 sub esi,16*4 fistp dword ptr [esp+12] ;[esp+8] = Dt fistp dword ptr [esp+8] ;[esp+12] = Ct fistp dword ptr [esp+4] ;[esp+0] = Bt fistp dword ptr [esp+0] ;[esp+4] = At mov eax,[esp+4] cmp eax,-32768 jge div0_noclip1low mov eax,-32768 div0_noclip1low: cmp eax,32767 jle div0_noclip1high mov eax,32767 div0_noclip1high: mov [edi],ax mov eax,[esp+12] cmp eax,-32768 jge div0_noclip3low mov eax,-32768 div0_noclip3low: cmp eax,32767 jle div0_noclip3high mov eax,32767 div0_noclip3high: mov [edi+2],ax mov eax,[esp+0] cmp eax,-32768 jge div0_noclip2low mov eax,-32768 div0_noclip2low: cmp eax,32767 jle div0_noclip2high mov eax,32767 div0_noclip2high: mov [ebx],ax nop mov eax,[esp+8] cmp eax,-32768 jge div0_noclip4low mov eax,-32768 div0_noclip4low: cmp eax,32767 jle div0_noclip4high mov eax,32767 div0_noclip4high: mov [ebx+2],ax nop add ebx,4 sub edi,4 dec ebp jne div0_loop pop eax pop eax pop eax pop eax pop ebp pop edi mov ebx,esi pop esi ; mov window,ecx ; mov window2,edx ; mov u_ptr,ebx }; #else for (i = 1; i < 16; i++) { sum1 = 0.0F; sum2 = 0.0F; sum3 = 0.0F; sum4 = 0.0F; window = &wincoef[i][16-vb_ptr]; window2 = &wincoef[i][vb_ptr]; si = 256-i*16; sum1 -= window [ 0] * vbuf[0][si+ 0]; sum2 += window2[15] * vbuf[0][si+ 0]; sum3 -= window [ 0] * vbuf[1][si+ 0]; sum4 += window2[15] * vbuf[1][si+ 0]; sum1 += window [ 1] * vbuf[0][si+ 1]; sum2 += window2[14] * vbuf[0][si+ 1]; sum3 += window [ 1] * vbuf[1][si+ 1]; sum4 += window2[14] * vbuf[1][si+ 1]; sum1 -= window [ 2] * vbuf[0][si+ 2]; sum2 += window2[13] * vbuf[0][si+ 2]; sum3 -= window [ 2] * vbuf[1][si+ 2]; sum4 += window2[13] * vbuf[1][si+ 2]; sum1 += window [ 3] * vbuf[0][si+ 3]; sum2 += window2[12] * vbuf[0][si+ 3]; sum3 += window [ 3] * vbuf[1][si+ 3]; sum4 += window2[12] * vbuf[1][si+ 3]; sum1 -= window [ 4] * vbuf[0][si+ 4]; sum2 += window2[11] * vbuf[0][si+ 4]; sum3 -= window [ 4] * vbuf[1][si+ 4]; sum4 += window2[11] * vbuf[1][si+ 4]; sum1 += window [ 5] * vbuf[0][si+ 5]; sum2 += window2[10] * vbuf[0][si+ 5]; sum3 += window [ 5] * vbuf[1][si+ 5]; sum4 += window2[10] * vbuf[1][si+ 5]; sum1 -= window [ 6] * vbuf[0][si+ 6]; sum2 += window2[ 9] * vbuf[0][si+ 6]; sum3 -= window [ 6] * vbuf[1][si+ 6]; sum4 += window2[ 9] * vbuf[1][si+ 6]; sum1 += window [ 7] * vbuf[0][si+ 7]; sum2 += window2[ 8] * vbuf[0][si+ 7]; sum3 += window [ 7] * vbuf[1][si+ 7]; sum4 += window2[ 8] * vbuf[1][si+ 7]; sum1 -= window [ 8] * vbuf[0][si+ 8]; sum2 += window2[ 7] * vbuf[0][si+ 8]; sum3 -= window [ 8] * vbuf[1][si+ 8]; sum4 += window2[ 7] * vbuf[1][si+ 8]; sum1 += window [ 9] * vbuf[0][si+ 9]; sum2 += window2[ 6] * vbuf[0][si+ 9]; sum3 += window [ 9] * vbuf[1][si+ 9]; sum4 += window2[ 6] * vbuf[1][si+ 9]; sum1 -= window [10] * vbuf[0][si+10]; sum2 += window2[ 5] * vbuf[0][si+10]; sum3 -= window [10] * vbuf[1][si+10]; sum4 += window2[ 5] * vbuf[1][si+10]; sum1 += window [11] * vbuf[0][si+11]; sum2 += window2[ 4] * vbuf[0][si+11]; sum3 += window [11] * vbuf[1][si+11]; sum4 += window2[ 4] * vbuf[1][si+11]; sum1 -= window [12] * vbuf[0][si+12]; sum2 += window2[ 3] * vbuf[0][si+12]; sum3 -= window [12] * vbuf[1][si+12]; sum4 += window2[ 3] * vbuf[1][si+12]; sum1 += window [13] * vbuf[0][si+13]; sum2 += window2[ 2] * vbuf[0][si+13]; sum3 += window [13] * vbuf[1][si+13]; sum4 += window2[ 2] * vbuf[1][si+13]; sum1 -= window [14] * vbuf[0][si+14]; sum2 += window2[ 1] * vbuf[0][si+14]; sum3 -= window [14] * vbuf[1][si+14]; sum4 += window2[ 1] * vbuf[1][si+14]; sum1 += window [15] * vbuf[0][si+15]; sum2 += window2[ 0] * vbuf[0][si+15]; sum3 += window [15] * vbuf[1][si+15]; sum4 += window2[ 0] * vbuf[1][si+15]; tmp = float_to_long(sum1); if (tmp > 32767) tmp = 32767; else if (tmp < -32768) tmp = -32768; pcm[i*2] = tmp; tmp = float_to_long(sum3); if (tmp > 32767) tmp = 32767; else if (tmp < -32768) tmp = -32768; pcm[i*2+1] = tmp; tmp = float_to_long(sum2); if (tmp > 32767) tmp = 32767; else if (tmp < -32768) tmp = -32768; pcm[64-i*2] = tmp; tmp = float_to_long(sum4); if (tmp > 32767) tmp = 32767; else if (tmp < -32768) tmp = -32768; pcm[64-i*2+1] = tmp; } #endif } else { #ifdef USE_ASM window = &wincoef[1][16-vb_ptr]; window2 = &wincoef[1][vb_ptr]; const void *vbuf2 = (void *)&vbuf[0][272]; __asm { push esi push vbuf2 mov ecx,window mov edx,window2 mov ebx,pcm pop esi push edi mov edi,ebx add ebx,2*2 add edi,62*2 push ebp push 0 push 0 push 0 push 0 mov ebp,15 ;ESI = FP input ;EBX = PCM output (low) ;ECX = window ;EDX = window2 ;EDI = PCM output (high) ;EBP = counter ; ;A sum = window[0...15] * vbuf[0] (alternating) ;B sum = window2[15...0] * vbuf[0] ;C sum = window[0...15] * vbuf[1] (alternating) ;D sum = window2[15...0] * vbuf[1] div1_loop: fld dword ptr [esi+ 0*4+0*2048] ;Lx fmul dword ptr [ecx+ 0*4] ;Ax fld dword ptr [esi+ 0*4+0*2048] ;Lx Ax fmul dword ptr [edx+15*4] ;Bx Ax fld dword ptr [esi+ 0*4+1*2048] ;Rx Bx Ax fmul dword ptr [ecx+ 0*4] ;Cx Bx Ax fld dword ptr [esi+ 0*4+1*2048] ;Rx Cx Bx Ax fmul dword ptr [edx+15*4] ;Dx Cx Bx Ax fld dword ptr [esi+ 1*4+0*2048] ;Lx Dt Ct Bt At fmul dword ptr [ecx+ 1*4] ;Ax Dt Ct Bt At fld dword ptr [esi+ 1*4+0*2048] ;Lx Ax Dt Ct Bt At fmul dword ptr [edx+14*4] ;Bx Ax Dt Ct Bt At fld dword ptr [esi+ 1*4+1*2048] ;Rx Bx Ax Dt Ct Bt At fmul dword ptr [ecx+ 1*4] ;Cx Bx Ax Dt Ct Bt At fld dword ptr [esi+ 1*4+1*2048] ;Rx Cx Bx Ax Dt Ct Bt At fmul dword ptr [edx+14*4] ;Dx Cx Bx Ax Dt Ct Bt At fxch st(3) ;Ax Cx Bx Dx Dt Ct Bt At fsubp st(7),st ;Cx Bx Dx Dt Ct Bt At fxch st(1) ;Bx Cx Dx Dt Ct Bt At faddp st(5),st ;Cx Dx Dt Ct Bt At fsubp st(3),st ;Dx Dt Ct Bt At faddp st(1),st ;Dt Ct Bt At fld dword ptr [esi+ 2*4+0*2048] ;Lx Dt Ct Bt At fmul dword ptr [ecx+ 2*4] ;Ax Dt Ct Bt At fld dword ptr [esi+ 2*4+0*2048] ;Lx Ax Dt Ct Bt At fmul dword ptr [edx+13*4] ;Bx Ax Dt Ct Bt At fld dword ptr [esi+ 2*4+1*2048] ;Rx Bx Ax Dt Ct Bt At fmul dword ptr [ecx+ 2*4] ;Cx Bx Ax Dt Ct Bt At fld dword ptr [esi+ 2*4+1*2048] ;Rx Cx Bx Ax Dt Ct Bt At fmul dword ptr [edx+13*4] ;Dx Cx Bx Ax Dt Ct Bt At fxch st(3) ;Ax Cx Bx Dx Dt Ct Bt At faddp st(7),st ;Cx Bx Dx Dt Ct Bt At fxch st(1) ;Bx Cx Dx Dt Ct Bt At faddp st(5),st ;Cx Dx Dt Ct Bt At faddp st(3),st ;Dx Dt Ct Bt At faddp st(1),st ;Dt Ct Bt At fld dword ptr [esi+ 3*4+0*2048] ;Lx Dt Ct Bt At fmul dword ptr [ecx+ 3*4] ;Ax Dt Ct Bt At fld dword ptr [esi+ 3*4+0*2048] ;Lx Ax Dt Ct Bt At fmul dword ptr [edx+12*4] ;Bx Ax Dt Ct Bt At fld dword ptr [esi+ 3*4+1*2048] ;Rx Bx Ax Dt Ct Bt At fmul dword ptr [ecx+ 3*4] ;Cx Bx Ax Dt Ct Bt At fld dword ptr [esi+ 3*4+1*2048] ;Rx Cx Bx Ax Dt Ct Bt At fmul dword ptr [edx+12*4] ;Dx Cx Bx Ax Dt Ct Bt At fxch st(3) ;Ax Cx Bx Dx Dt Ct Bt At fsubp st(7),st ;Cx Bx Dx Dt Ct Bt At fxch st(1) ;Bx Cx Dx Dt Ct Bt At faddp st(5),st ;Cx Dx Dt Ct Bt At fsubp st(3),st ;Dx Dt Ct Bt At faddp st(1),st ;Dt Ct Bt At fld dword ptr [esi+ 4*4+0*2048] ;Lx Dt Ct Bt At fmul dword ptr [ecx+ 4*4] ;Ax Dt Ct Bt At fld dword ptr [esi+ 4*4+0*2048] ;Lx Ax Dt Ct Bt At fmul dword ptr [edx+11*4] ;Bx Ax Dt Ct Bt At fld dword ptr [esi+ 4*4+1*2048] ;Rx Bx Ax Dt Ct Bt At fmul dword ptr [ecx+ 4*4] ;Cx Bx Ax Dt Ct Bt At fld dword ptr [esi+ 4*4+1*2048] ;Rx Cx Bx Ax Dt Ct Bt At fmul dword ptr [edx+11*4] ;Dx Cx Bx Ax Dt Ct Bt At fxch st(3) ;Ax Cx Bx Dx Dt Ct Bt At faddp st(7),st ;Cx Bx Dx Dt Ct Bt At fxch st(1) ;Bx Cx Dx Dt Ct Bt At faddp st(5),st ;Cx Dx Dt Ct Bt At faddp st(3),st ;Dx Dt Ct Bt At faddp st(1),st ;Dt Ct Bt At fld dword ptr [esi+ 5*4+0*2048] ;Lx Dt Ct Bt At fmul dword ptr [ecx+ 5*4] ;Ax Dt Ct Bt At fld dword ptr [esi+ 5*4+0*2048] ;Lx Ax Dt Ct Bt At fmul dword ptr [edx+10*4] ;Bx Ax Dt Ct Bt At fld dword ptr [esi+ 5*4+1*2048] ;Rx Bx Ax Dt Ct Bt At fmul dword ptr [ecx+ 5*4] ;Cx Bx Ax Dt Ct Bt At fld dword ptr [esi+ 5*4+1*2048] ;Rx Cx Bx Ax Dt Ct Bt At fmul dword ptr [edx+10*4] ;Dx Cx Bx Ax Dt Ct Bt At fxch st(3) ;Ax Cx Bx Dx Dt Ct Bt At fsubp st(7),st ;Cx Bx Dx Dt Ct Bt At fxch st(1) ;Bx Cx Dx Dt Ct Bt At faddp st(5),st ;Cx Dx Dt Ct Bt At fsubp st(3),st ;Dx Dt Ct Bt At faddp st(1),st ;Dt Ct Bt At fld dword ptr [esi+ 6*4+0*2048] ;Lx Dt Ct Bt At fmul dword ptr [ecx+ 6*4] ;Ax Dt Ct Bt At fld dword ptr [esi+ 6*4+0*2048] ;Lx Ax Dt Ct Bt At fmul dword ptr [edx+ 9*4] ;Bx Ax Dt Ct Bt At fld dword ptr [esi+ 6*4+1*2048] ;Rx Bx Ax Dt Ct Bt At fmul dword ptr [ecx+ 6*4] ;Cx Bx Ax Dt Ct Bt At fld dword ptr [esi+ 6*4+1*2048] ;Rx Cx Bx Ax Dt Ct Bt At fmul dword ptr [edx+ 9*4] ;Dx Cx Bx Ax Dt Ct Bt At fxch st(3) ;Ax Cx Bx Dx Dt Ct Bt At faddp st(7),st ;Cx Bx Dx Dt Ct Bt At fxch st(1) ;Bx Cx Dx Dt Ct Bt At faddp st(5),st ;Cx Dx Dt Ct Bt At faddp st(3),st ;Dx Dt Ct Bt At faddp st(1),st ;Dt Ct Bt At fld dword ptr [esi+ 7*4+0*2048] ;Lx Dt Ct Bt At fmul dword ptr [ecx+ 7*4] ;Ax Dt Ct Bt At fld dword ptr [esi+ 7*4+0*2048] ;Lx Ax Dt Ct Bt At fmul dword ptr [edx+ 8*4] ;Bx Ax Dt Ct Bt At fld dword ptr [esi+ 7*4+1*2048] ;Rx Bx Ax Dt Ct Bt At fmul dword ptr [ecx+ 7*4] ;Cx Bx Ax Dt Ct Bt At fld dword ptr [esi+ 7*4+1*2048] ;Rx Cx Bx Ax Dt Ct Bt At fmul dword ptr [edx+ 8*4] ;Dx Cx Bx Ax Dt Ct Bt At fxch st(3) ;Ax Cx Bx Dx Dt Ct Bt At fsubp st(7),st ;Cx Bx Dx Dt Ct Bt At fxch st(1) ;Bx Cx Dx Dt Ct Bt At faddp st(5),st ;Cx Dx Dt Ct Bt At fsubp st(3),st ;Dx Dt Ct Bt At faddp st(1),st ;Dt Ct Bt At fld dword ptr [esi+ 8*4+0*2048] ;Lx Dt Ct Bt At fmul dword ptr [ecx+ 8*4] ;Ax Dt Ct Bt At fld dword ptr [esi+ 8*4+0*2048] ;Lx Ax Dt Ct Bt At fmul dword ptr [edx+ 7*4] ;Bx Ax Dt Ct Bt At fld dword ptr [esi+ 8*4+1*2048] ;Rx Bx Ax Dt Ct Bt At fmul dword ptr [ecx+ 8*4] ;Cx Bx Ax Dt Ct Bt At fld dword ptr [esi+ 8*4+1*2048] ;Rx Cx Bx Ax Dt Ct Bt At fmul dword ptr [edx+ 7*4] ;Dx Cx Bx Ax Dt Ct Bt At fxch st(3) ;Ax Cx Bx Dx Dt Ct Bt At faddp st(7),st ;Cx Bx Dx Dt Ct Bt At fxch st(1) ;Bx Cx Dx Dt Ct Bt At faddp st(5),st ;Cx Dx Dt Ct Bt At faddp st(3),st ;Dx Dt Ct Bt At faddp st(1),st ;Dt Ct Bt At fld dword ptr [esi+ 9*4+0*2048] ;Lx Dt Ct Bt At fmul dword ptr [ecx+ 9*4] ;Ax Dt Ct Bt At fld dword ptr [esi+ 9*4+0*2048] ;Lx Ax Dt Ct Bt At fmul dword ptr [edx+ 6*4] ;Bx Ax Dt Ct Bt At fld dword ptr [esi+ 9*4+1*2048] ;Rx Bx Ax Dt Ct Bt At fmul dword ptr [ecx+ 9*4] ;Cx Bx Ax Dt Ct Bt At fld dword ptr [esi+ 9*4+1*2048] ;Rx Cx Bx Ax Dt Ct Bt At fmul dword ptr [edx+ 6*4] ;Dx Cx Bx Ax Dt Ct Bt At fxch st(3) ;Ax Cx Bx Dx Dt Ct Bt At fsubp st(7),st ;Cx Bx Dx Dt Ct Bt At fxch st(1) ;Bx Cx Dx Dt Ct Bt At faddp st(5),st ;Cx Dx Dt Ct Bt At fsubp st(3),st ;Dx Dt Ct Bt At faddp st(1),st ;Dt Ct Bt At fld dword ptr [esi+10*4+0*2048] ;Lx Dt Ct Bt At fmul dword ptr [ecx+10*4] ;Ax Dt Ct Bt At fld dword ptr [esi+10*4+0*2048] ;Lx Ax Dt Ct Bt At fmul dword ptr [edx+ 5*4] ;Bx Ax Dt Ct Bt At fld dword ptr [esi+10*4+1*2048] ;Rx Bx Ax Dt Ct Bt At fmul dword ptr [ecx+10*4] ;Cx Bx Ax Dt Ct Bt At fld dword ptr [esi+10*4+1*2048] ;Rx Cx Bx Ax Dt Ct Bt At fmul dword ptr [edx+ 5*4] ;Dx Cx Bx Ax Dt Ct Bt At fxch st(3) ;Ax Cx Bx Dx Dt Ct Bt At faddp st(7),st ;Cx Bx Dx Dt Ct Bt At fxch st(1) ;Bx Cx Dx Dt Ct Bt At faddp st(5),st ;Cx Dx Dt Ct Bt At faddp st(3),st ;Dx Dt Ct Bt At faddp st(1),st ;Dt Ct Bt At fld dword ptr [esi+11*4+0*2048] ;Lx Dt Ct Bt At fmul dword ptr [ecx+11*4] ;Ax Dt Ct Bt At fld dword ptr [esi+11*4+0*2048] ;Lx Ax Dt Ct Bt At fmul dword ptr [edx+ 4*4] ;Bx Ax Dt Ct Bt At fld dword ptr [esi+11*4+1*2048] ;Rx Bx Ax Dt Ct Bt At fmul dword ptr [ecx+11*4] ;Cx Bx Ax Dt Ct Bt At fld dword ptr [esi+11*4+1*2048] ;Rx Cx Bx Ax Dt Ct Bt At fmul dword ptr [edx+ 4*4] ;Dx Cx Bx Ax Dt Ct Bt At fxch st(3) ;Ax Cx Bx Dx Dt Ct Bt At fsubp st(7),st ;Cx Bx Dx Dt Ct Bt At fxch st(1) ;Bx Cx Dx Dt Ct Bt At faddp st(5),st ;Cx Dx Dt Ct Bt At fsubp st(3),st ;Dx Dt Ct Bt At faddp st(1),st ;Dt Ct Bt At fld dword ptr [esi+12*4+0*2048] ;Lx Dt Ct Bt At fmul dword ptr [ecx+12*4] ;Ax Dt Ct Bt At fld dword ptr [esi+12*4+0*2048] ;Lx Ax Dt Ct Bt At fmul dword ptr [edx+ 3*4] ;Bx Ax Dt Ct Bt At fld dword ptr [esi+12*4+1*2048] ;Rx Bx Ax Dt Ct Bt At fmul dword ptr [ecx+12*4] ;Cx Bx Ax Dt Ct Bt At fld dword ptr [esi+12*4+1*2048] ;Rx Cx Bx Ax Dt Ct Bt At fmul dword ptr [edx+ 3*4] ;Dx Cx Bx Ax Dt Ct Bt At fxch st(3) ;Ax Cx Bx Dx Dt Ct Bt At faddp st(7),st ;Cx Bx Dx Dt Ct Bt At fxch st(1) ;Bx Cx Dx Dt Ct Bt At faddp st(5),st ;Cx Dx Dt Ct Bt At faddp st(3),st ;Dx Dt Ct Bt At faddp st(1),st ;Dt Ct Bt At fld dword ptr [esi+13*4+0*2048] ;Lx Dt Ct Bt At fmul dword ptr [ecx+13*4] ;Ax Dt Ct Bt At fld dword ptr [esi+13*4+0*2048] ;Lx Ax Dt Ct Bt At fmul dword ptr [edx+ 2*4] ;Bx Ax Dt Ct Bt At fld dword ptr [esi+13*4+1*2048] ;Rx Bx Ax Dt Ct Bt At fmul dword ptr [ecx+13*4] ;Cx Bx Ax Dt Ct Bt At fld dword ptr [esi+13*4+1*2048] ;Rx Cx Bx Ax Dt Ct Bt At fmul dword ptr [edx+ 2*4] ;Dx Cx Bx Ax Dt Ct Bt At fxch st(3) ;Ax Cx Bx Dx Dt Ct Bt At fsubp st(7),st ;Cx Bx Dx Dt Ct Bt At fxch st(1) ;Bx Cx Dx Dt Ct Bt At faddp st(5),st ;Cx Dx Dt Ct Bt At fsubp st(3),st ;Dx Dt Ct Bt At faddp st(1),st ;Dt Ct Bt At fld dword ptr [esi+14*4+0*2048] ;Lx Dt Ct Bt At fmul dword ptr [ecx+14*4] ;Ax Dt Ct Bt At fld dword ptr [esi+14*4+0*2048] ;Lx Ax Dt Ct Bt At fmul dword ptr [edx+ 1*4] ;Bx Ax Dt Ct Bt At fld dword ptr [esi+14*4+1*2048] ;Rx Bx Ax Dt Ct Bt At fmul dword ptr [ecx+14*4] ;Cx Bx Ax Dt Ct Bt At fld dword ptr [esi+14*4+1*2048] ;Rx Cx Bx Ax Dt Ct Bt At fmul dword ptr [edx+ 1*4] ;Dx Cx Bx Ax Dt Ct Bt At fxch st(3) ;Ax Cx Bx Dx Dt Ct Bt At faddp st(7),st ;Cx Bx Dx Dt Ct Bt At fxch st(1) ;Bx Cx Dx Dt Ct Bt At faddp st(5),st ;Cx Dx Dt Ct Bt At faddp st(3),st ;Dx Dt Ct Bt At faddp st(1),st ;Dt Ct Bt At fld dword ptr [esi+15*4+0*2048] ;Lx Dt Ct Bt At fmul dword ptr [ecx+15*4] ;Ax Dt Ct Bt At fld dword ptr [esi+15*4+0*2048] ;Lx Ax Dt Ct Bt At fmul dword ptr [edx+ 0*4] ;Bx Ax Dt Ct Bt At fld dword ptr [esi+15*4+1*2048] ;Rx Bx Ax Dt Ct Bt At fmul dword ptr [ecx+15*4] ;Cx Bx Ax Dt Ct Bt At fld dword ptr [esi+15*4+1*2048] ;Rx Cx Bx Ax Dt Ct Bt At fmul dword ptr [edx+ 0*4] ;Dx Cx Bx Ax Dt Ct Bt At fxch st(3) ;Ax Cx Bx Dx Dt Ct Bt At fsubp st(7),st ;Cx Bx Dx Dt Ct Bt At fxch st(1) ;Bx Cx Dx Dt Ct Bt At faddp st(5),st ;Cx Dx Dt Ct Bt At fsubp st(3),st ;Dx Dt Ct Bt At faddp st(1),st ;Dt Ct Bt At add ecx,32*4 add edx,32*4 add esi,16*4 fistp dword ptr [esp+12] ;[esp+8] = Dt fistp dword ptr [esp+8] ;[esp+12] = Ct fistp dword ptr [esp+4] ;[esp+0] = Bt fistp dword ptr [esp+0] ;[esp+4] = At mov eax,[esp+4] cmp eax,-32768 jge div1_noclip1low mov eax,-32768 div1_noclip1low: cmp eax,32767 jle div1_noclip1high mov eax,32767 div1_noclip1high: mov [edi],ax mov eax,[esp+12] cmp eax,-32768 jge div1_noclip3low mov eax,-32768 div1_noclip3low: cmp eax,32767 jle div1_noclip3high mov eax,32767 div1_noclip3high: mov [edi+2],ax mov eax,[esp+0] cmp eax,-32768 jge div1_noclip2low mov eax,-32768 div1_noclip2low: cmp eax,32767 jle div1_noclip2high mov eax,32767 div1_noclip2high: mov [ebx],ax nop mov eax,[esp+8] cmp eax,-32768 jge div1_noclip4low mov eax,-32768 div1_noclip4low: cmp eax,32767 jle div1_noclip4high mov eax,32767 div1_noclip4high: mov [ebx+2],ax nop add ebx,4 sub edi,4 dec ebp jne div1_loop pop eax pop eax pop eax pop eax pop ebp pop edi mov ebx,esi pop esi ; mov window,ecx ; mov window2,edx ; mov u_ptr,ebx }; #else for (i = 1; i < 16; i++) { sum1 = 0.0F; sum2 = 0.0F; sum3 = 0.0F; sum4 = 0.0F; window = &wincoef[i][16-vb_ptr]; window2 = &wincoef[i][vb_ptr & 15]; si = 256+i*16; sum1 += window [ 0] * vbuf[0][si+ 0]; sum2 += window2[15] * vbuf[0][si+ 0]; sum3 += window [ 0] * vbuf[1][si+ 0]; sum4 += window2[15] * vbuf[1][si+ 0]; sum1 -= window [ 1] * vbuf[0][si+ 1]; sum2 += window2[14] * vbuf[0][si+ 1]; sum3 -= window [ 1] * vbuf[1][si+ 1]; sum4 += window2[14] * vbuf[1][si+ 1]; sum1 += window [ 2] * vbuf[0][si+ 2]; sum2 += window2[13] * vbuf[0][si+ 2]; sum3 += window [ 2] * vbuf[1][si+ 2]; sum4 += window2[13] * vbuf[1][si+ 2]; sum1 -= window [ 3] * vbuf[0][si+ 3]; sum2 += window2[12] * vbuf[0][si+ 3]; sum3 -= window [ 3] * vbuf[1][si+ 3]; sum4 += window2[12] * vbuf[1][si+ 3]; sum1 += window [ 4] * vbuf[0][si+ 4]; sum2 += window2[11] * vbuf[0][si+ 4]; sum3 += window [ 4] * vbuf[1][si+ 4]; sum4 += window2[11] * vbuf[1][si+ 4]; sum1 -= window [ 5] * vbuf[0][si+ 5]; sum2 += window2[10] * vbuf[0][si+ 5]; sum3 -= window [ 5] * vbuf[1][si+ 5]; sum4 += window2[10] * vbuf[1][si+ 5]; sum1 += window [ 6] * vbuf[0][si+ 6]; sum2 += window2[ 9] * vbuf[0][si+ 6]; sum3 += window [ 6] * vbuf[1][si+ 6]; sum4 += window2[ 9] * vbuf[1][si+ 6]; sum1 -= window [ 7] * vbuf[0][si+ 7]; sum2 += window2[ 8] * vbuf[0][si+ 7]; sum3 -= window [ 7] * vbuf[1][si+ 7]; sum4 += window2[ 8] * vbuf[1][si+ 7]; sum1 += window [ 8] * vbuf[0][si+ 8]; sum2 += window2[ 7] * vbuf[0][si+ 8]; sum3 += window [ 8] * vbuf[1][si+ 8]; sum4 += window2[ 7] * vbuf[1][si+ 8]; sum1 -= window [ 9] * vbuf[0][si+ 9]; sum2 += window2[ 6] * vbuf[0][si+ 9]; sum3 -= window [ 9] * vbuf[1][si+ 9]; sum4 += window2[ 6] * vbuf[1][si+ 9]; sum1 += window [10] * vbuf[0][si+10]; sum2 += window2[ 5] * vbuf[0][si+10]; sum3 += window [10] * vbuf[1][si+10]; sum4 += window2[ 5] * vbuf[1][si+10]; sum1 -= window [11] * vbuf[0][si+11]; sum2 += window2[ 4] * vbuf[0][si+11]; sum3 -= window [11] * vbuf[1][si+11]; sum4 += window2[ 4] * vbuf[1][si+11]; sum1 += window [12] * vbuf[0][si+12]; sum2 += window2[ 3] * vbuf[0][si+12]; sum3 += window [12] * vbuf[1][si+12]; sum4 += window2[ 3] * vbuf[1][si+12]; sum1 -= window [13] * vbuf[0][si+13]; sum2 += window2[ 2] * vbuf[0][si+13]; sum3 -= window [13] * vbuf[1][si+13]; sum4 += window2[ 2] * vbuf[1][si+13]; sum1 += window [14] * vbuf[0][si+14]; sum2 += window2[ 1] * vbuf[0][si+14]; sum3 += window [14] * vbuf[1][si+14]; sum4 += window2[ 1] * vbuf[1][si+14]; sum1 -= window [15] * vbuf[0][si+15]; sum2 += window2[ 0] * vbuf[0][si+15]; sum3 -= window [15] * vbuf[1][si+15]; sum4 += window2[ 0] * vbuf[1][si+15]; tmp = float_to_long(sum1); if (tmp > 32767) tmp = 32767; else if (tmp < -32768) tmp = -32768; pcm[i*2] = tmp; tmp = float_to_long(sum3); if (tmp > 32767) tmp = 32767; else if (tmp < -32768) tmp = -32768; pcm[i*2+1] = tmp; tmp = float_to_long(sum2); if (tmp > 32767) tmp = 32767; else if (tmp < -32768) tmp = -32768; pcm[64-i*2] = tmp; tmp = float_to_long(sum4); if (tmp > 32767) tmp = 32767; else if (tmp < -32768) tmp = -32768; pcm[64-i*2+1] = tmp; } #endif } /*-- special case --*/ coef = 0; bx = (vb_ptr+1)&15; sum1 = sum2 = 0.0F; for (j = 0; j < 8; j++) { sum1 += wincoef2[coef] * vbuf[0][bx]; sum2 += wincoef2[coef] * vbuf[1][bx]; bx = (bx+2)&15; ++coef; } tmp = (long)sum1; if (tmp > 32767) tmp = 32767; else if (tmp < -32768) tmp = -32768; pcm[32] = tmp; tmp = (long)sum2; if (tmp > 32767) tmp = 32767; else if (tmp < -32768) tmp = -32768; pcm[33] = tmp; } void window_mono(const float *vbuf, int vb_ptr, short *pcm) { double converter; int i, j; int si, bx; int coef; float *window, *window2; float sum1, sum2; long tmp; coef = 0; /* first */ i = 0; if (vb_ptr & 1) { sum1 = 0.0F; window = &wincoef[i][(-vb_ptr) & 15]; si = 256-i*16; sum1 -= window[ 0] * vbuf[si+ 0]; sum1 += window[ 1] * vbuf[si+ 1]; sum1 -= window[ 2] * vbuf[si+ 2]; sum1 += window[ 3] * vbuf[si+ 3]; sum1 -= window[ 4] * vbuf[si+ 4]; sum1 += window[ 5] * vbuf[si+ 5]; sum1 -= window[ 6] * vbuf[si+ 6]; sum1 += window[ 7] * vbuf[si+ 7]; sum1 -= window[ 8] * vbuf[si+ 8]; sum1 += window[ 9] * vbuf[si+ 9]; sum1 -= window[10] * vbuf[si+10]; sum1 += window[11] * vbuf[si+11]; sum1 -= window[12] * vbuf[si+12]; sum1 += window[13] * vbuf[si+13]; sum1 -= window[14] * vbuf[si+14]; sum1 += window[15] * vbuf[si+15]; tmp = float_to_long(sum1); if (tmp > 32767) tmp = 32767; else if (tmp < -32768) tmp = -32768; pcm[i] = tmp; } else { sum1 = 0.0F; window = &wincoef[i][(-vb_ptr) & 15]; si = 256+i*16; sum1 += window[ 0] * vbuf[si+ 0]; sum1 -= window[ 1] * vbuf[si+ 1]; sum1 += window[ 2] * vbuf[si+ 2]; sum1 -= window[ 3] * vbuf[si+ 3]; sum1 += window[ 4] * vbuf[si+ 4]; sum1 -= window[ 5] * vbuf[si+ 5]; sum1 += window[ 6] * vbuf[si+ 6]; sum1 -= window[ 7] * vbuf[si+ 7]; sum1 += window[ 8] * vbuf[si+ 8]; sum1 -= window[ 9] * vbuf[si+ 9]; sum1 += window[10] * vbuf[si+10]; sum1 -= window[11] * vbuf[si+11]; sum1 += window[12] * vbuf[si+12]; sum1 -= window[13] * vbuf[si+13]; sum1 += window[14] * vbuf[si+14]; sum1 -= window[15] * vbuf[si+15]; tmp = float_to_long(sum1); if (tmp > 32767) tmp = 32767; else if (tmp < -32768) tmp = -32768; pcm[i] = tmp; } /*-- 1-16, last 15 --*/ if (vb_ptr & 1) { #ifdef USE_ASM window = &wincoef[1][16-vb_ptr]; window2 = &wincoef[1][vb_ptr]; const float *vbuf2 = vbuf+240; __asm { push esi push vbuf2 mov edx,window mov ecx,window2 mov ebx,pcm pop esi push edi mov edi,ebx add ebx,2 add edi,62 cmp eax,2 push ebp push 0 push 0 mov ebp,15 div0_loop: fld dword ptr [esi+ 0*4] ;x0 fmul dword ptr [ecx+15*4] ;A0 fld dword ptr [esi+ 0*4] ;x0 A0 fmul dword ptr [edx+ 0*4] ;B0 A0 fld dword ptr [esi+ 1*4] ;x1 B0 A0 fmul dword ptr [ecx+14*4] ;A1 B0 A0 fld dword ptr [esi+ 1*4] ;x1 A1 B0 A0 fmul dword ptr [edx+ 1*4] ;B1 A1 B0 A0 fld dword ptr [esi+ 2*4] ;x2 B1 A1 B0 A0 fxch st(2) ;A1 B1 x2 B0 A0 faddp st(4),st ;B1 x2 B0 At fsubp st(2),st ;x2 Bt At fmul dword ptr [ecx+13*4] ;A2 Bt At fld dword ptr [esi+ 2*4] ;x2 A2 Bt At fmul dword ptr [edx+ 2*4] ;B2 A2 Bt At fld dword ptr [esi+ 3*4] ;x3 B2 A2 Bt At fxch st(2) ;A2 B2 x3 Bt At faddp st(4),st ;B2 x3 Bt At faddp st(2),st ;x3 Bt At fmul dword ptr [ecx+12*4] ;A3 Bt At fld dword ptr [esi+ 3*4] ;x3 A3 Bt At fmul dword ptr [edx+ 3*4] ;B3 A3 Bt At fld dword ptr [esi+ 4*4] ;x3 B2 A2 Bt At fxch st(2) ;A2 B2 x3 Bt At faddp st(4),st ;B2 x3 Bt At fsubp st(2),st ;x3 Bt At fmul dword ptr [ecx+11*4] ;A3 Bt At fld dword ptr [esi+ 4*4] ;x3 A3 Bt At fmul dword ptr [edx+ 4*4] ;B3 A3 Bt At fld dword ptr [esi+ 5*4] ;x3 B2 A2 Bt At fxch st(2) ;A2 B2 x3 Bt At faddp st(4),st ;B2 x3 Bt At faddp st(2),st ;x3 Bt At fmul dword ptr [ecx+10*4] ;A3 Bt At fld dword ptr [esi+ 5*4] ;x3 A3 Bt At fmul dword ptr [edx+ 5*4] ;B3 A3 Bt At fld dword ptr [esi+ 6*4] ;x3 B2 A2 Bt At fxch st(2) ;A2 B2 x3 Bt At faddp st(4),st ;B2 x3 Bt At fsubp st(2),st ;x3 Bt At fmul dword ptr [ecx+ 9*4] ;A3 Bt At fld dword ptr [esi+ 6*4] ;x3 A3 Bt At fmul dword ptr [edx+ 6*4] ;B3 A3 Bt At fld dword ptr [esi+ 7*4] ;x3 B2 A2 Bt At fxch st(2) ;A2 B2 x3 Bt At faddp st(4),st ;B2 x3 Bt At faddp st(2),st ;x3 Bt At fmul dword ptr [ecx+ 8*4] ;A3 Bt At fld dword ptr [esi+ 7*4] ;x3 A3 Bt At fmul dword ptr [edx+ 7*4] ;B3 A3 Bt At fld dword ptr [esi+ 8*4] ;x3 B2 A2 Bt At fxch st(2) ;A2 B2 x3 Bt At faddp st(4),st ;B2 x3 Bt At fsubp st(2),st ;x3 Bt At fmul dword ptr [ecx+ 7*4] ;A3 Bt At fld dword ptr [esi+ 8*4] ;x3 A3 Bt At fmul dword ptr [edx+ 8*4] ;B3 A3 Bt At fld dword ptr [esi+ 9*4] ;x3 B2 A2 Bt At fxch st(2) ;A2 B2 x3 Bt At faddp st(4),st ;B2 x3 Bt At faddp st(2),st ;x3 Bt At fmul dword ptr [ecx+ 6*4] ;A3 Bt At fld dword ptr [esi+ 9*4] ;x3 A3 Bt At fmul dword ptr [edx+ 9*4] ;B3 A3 Bt At fld dword ptr [esi+10*4] ;x3 B2 A2 Bt At fxch st(2) ;A2 B2 x3 Bt At faddp st(4),st ;B2 x3 Bt At fsubp st(2),st ;x3 Bt At fmul dword ptr [ecx+ 5*4] ;A3 Bt At fld dword ptr [esi+10*4] ;x3 A3 Bt At fmul dword ptr [edx+10*4] ;B3 A3 Bt At fld dword ptr [esi+11*4] ;x3 B2 A2 Bt At fxch st(2) ;A2 B2 x3 Bt At faddp st(4),st ;B2 x3 Bt At faddp st(2),st ;x3 Bt At fmul dword ptr [ecx+ 4*4] ;A3 Bt At fld dword ptr [esi+11*4] ;x3 A3 Bt At fmul dword ptr [edx+11*4] ;B3 A3 Bt At fld dword ptr [esi+12*4] ;x3 B2 A2 Bt At fxch st(2) ;A2 B2 x3 Bt At faddp st(4),st ;B2 x3 Bt At fsubp st(2),st ;x3 Bt At fmul dword ptr [ecx+ 3*4] ;A3 Bt At fld dword ptr [esi+12*4] ;x3 A3 Bt At fmul dword ptr [edx+12*4] ;B3 A3 Bt At fld dword ptr [esi+13*4] ;x3 B2 A2 Bt At fxch st(2) ;A2 B2 x3 Bt At faddp st(4),st ;B2 x3 Bt At faddp st(2),st ;x3 Bt At fmul dword ptr [ecx+ 2*4] ;A3 Bt At fld dword ptr [esi+13*4] ;x3 A3 Bt At fmul dword ptr [edx+13*4] ;B3 A3 Bt At fld dword ptr [esi+14*4] ;x3 B2 A2 Bt At fxch st(2) ;A2 B2 x3 Bt At faddp st(4),st ;B2 x3 Bt At fsubp st(2),st ;x3 Bt At fmul dword ptr [ecx+ 1*4] ;A3 Bt At fld dword ptr [esi+14*4] ;x3 A3 Bt At fmul dword ptr [edx+14*4] ;B3 A3 Bt At fld dword ptr [esi+15*4] ;xF BE AE Bt At fxch st(2) ;AE BE xF Bt At faddp st(4),st ;BE xF Bt At faddp st(2),st ;xF Bt At fmul dword ptr [ecx+ 0*4] ;AF Bt At fld dword ptr [esi+15*4] ;xF AF Bt At fmul dword ptr [edx+15*4] ;BF AF Bt At add ecx,32*4 add edx,32*4 sub esi,16*4 fsubp st(2),st ;AF Bt At faddp st(2),st ;Bt At fchs fistp dword ptr [esp+0] ;[esp+0] = outf3 fistp dword ptr [esp+4] ;[esp+4] = outf1 mov eax,[esp+4] cmp eax,-32768 jge div0_noclip1low mov eax,-32768 div0_noclip1low: cmp eax,32767 jle div0_noclip1high mov eax,32767 div0_noclip1high: mov [edi],ax mov eax,[esp+0] cmp eax,-32768 jge div0_noclip2low mov eax,-32768 div0_noclip2low: cmp eax,32767 jle div0_noclip2high mov eax,32767 div0_noclip2high: mov [ebx],ax nop add ebx,2 sub edi,2 dec ebp jne div0_loop pop eax pop eax pop ebp pop edi mov ebx,esi pop esi ; mov window,ecx ; mov window2,edx ; mov u_ptr,ebx }; #else for (i = 1; i < 16; i++) { sum1 = 0.0F; sum2 = 0.0F; window = &wincoef[i][16-vb_ptr]; window2 = &wincoef[i][vb_ptr]; si = 256-i*16; sum1 -= window [ 0] * vbuf[si+ 0]; sum2 += window2[15] * vbuf[si+ 0]; sum1 += window [ 1] * vbuf[si+ 1]; sum2 += window2[14] * vbuf[si+ 1]; sum1 -= window [ 2] * vbuf[si+ 2]; sum2 += window2[13] * vbuf[si+ 2]; sum1 += window [ 3] * vbuf[si+ 3]; sum2 += window2[12] * vbuf[si+ 3]; sum1 -= window [ 4] * vbuf[si+ 4]; sum2 += window2[11] * vbuf[si+ 4]; sum1 += window [ 5] * vbuf[si+ 5]; sum2 += window2[10] * vbuf[si+ 5]; sum1 -= window [ 6] * vbuf[si+ 6]; sum2 += window2[ 9] * vbuf[si+ 6]; sum1 += window [ 7] * vbuf[si+ 7]; sum2 += window2[ 8] * vbuf[si+ 7]; sum1 -= window [ 8] * vbuf[si+ 8]; sum2 += window2[ 7] * vbuf[si+ 8]; sum1 += window [ 9] * vbuf[si+ 9]; sum2 += window2[ 6] * vbuf[si+ 9]; sum1 -= window [10] * vbuf[si+10]; sum2 += window2[ 5] * vbuf[si+10]; sum1 += window [11] * vbuf[si+11]; sum2 += window2[ 4] * vbuf[si+11]; sum1 -= window [12] * vbuf[si+12]; sum2 += window2[ 3] * vbuf[si+12]; sum1 += window [13] * vbuf[si+13]; sum2 += window2[ 2] * vbuf[si+13]; sum1 -= window [14] * vbuf[si+14]; sum2 += window2[ 1] * vbuf[si+14]; sum1 += window [15] * vbuf[si+15]; sum2 += window2[ 0] * vbuf[si+15]; tmp = float_to_long(sum1); if (tmp > 32767) tmp = 32767; else if (tmp < -32768) tmp = -32768; pcm[i] = tmp; tmp = float_to_long(sum2); if (tmp > 32767) tmp = 32767; else if (tmp < -32768) tmp = -32768; pcm[32-i] = tmp; } #endif } else { #ifdef USE_ASM window = &wincoef[1][16-vb_ptr]; window2 = &wincoef[1][vb_ptr]; const float *vbuf2 = vbuf+272; __asm { push esi push vbuf2 mov edx,window mov ecx,window2 mov ebx,pcm pop esi push edi mov edi,ebx add ebx,2 add edi,62 cmp eax,2 push ebp push 0 push 0 mov ebp,15 div1_loop: fld dword ptr [esi+ 0*4] ;x0 fmul dword ptr [ecx+15*4] ;A0 fld dword ptr [esi+ 0*4] ;x0 A0 fmul dword ptr [edx+ 0*4] ;B0 A0 fld dword ptr [esi+ 1*4] ;x1 B0 A0 fmul dword ptr [ecx+14*4] ;A1 B0 A0 fld dword ptr [esi+ 1*4] ;x1 A1 B0 A0 fmul dword ptr [edx+ 1*4] ;B1 A1 B0 A0 fld dword ptr [esi+ 2*4] ;x2 B1 A1 B0 A0 fxch st(2) ;A1 B1 x2 B0 A0 faddp st(4),st ;B1 x2 B0 At fsubp st(2),st ;x2 Bt At fmul dword ptr [ecx+13*4] ;A2 Bt At fld dword ptr [esi+ 2*4] ;x2 A2 Bt At fmul dword ptr [edx+ 2*4] ;B2 A2 Bt At fld dword ptr [esi+ 3*4] ;x3 B2 A2 Bt At fxch st(2) ;A2 B2 x3 Bt At faddp st(4),st ;B2 x3 Bt At faddp st(2),st ;x3 Bt At fmul dword ptr [ecx+12*4] ;A3 Bt At fld dword ptr [esi+ 3*4] ;x3 A3 Bt At fmul dword ptr [edx+ 3*4] ;B3 A3 Bt At fld dword ptr [esi+ 4*4] ;x3 B2 A2 Bt At fxch st(2) ;A2 B2 x3 Bt At faddp st(4),st ;B2 x3 Bt At fsubp st(2),st ;x3 Bt At fmul dword ptr [ecx+11*4] ;A3 Bt At fld dword ptr [esi+ 4*4] ;x3 A3 Bt At fmul dword ptr [edx+ 4*4] ;B3 A3 Bt At fld dword ptr [esi+ 5*4] ;x3 B2 A2 Bt At fxch st(2) ;A2 B2 x3 Bt At faddp st(4),st ;B2 x3 Bt At faddp st(2),st ;x3 Bt At fmul dword ptr [ecx+10*4] ;A3 Bt At fld dword ptr [esi+ 5*4] ;x3 A3 Bt At fmul dword ptr [edx+ 5*4] ;B3 A3 Bt At fld dword ptr [esi+ 6*4] ;x3 B2 A2 Bt At fxch st(2) ;A2 B2 x3 Bt At faddp st(4),st ;B2 x3 Bt At fsubp st(2),st ;x3 Bt At fmul dword ptr [ecx+ 9*4] ;A3 Bt At fld dword ptr [esi+ 6*4] ;x3 A3 Bt At fmul dword ptr [edx+ 6*4] ;B3 A3 Bt At fld dword ptr [esi+ 7*4] ;x3 B2 A2 Bt At fxch st(2) ;A2 B2 x3 Bt At faddp st(4),st ;B2 x3 Bt At faddp st(2),st ;x3 Bt At fmul dword ptr [ecx+ 8*4] ;A3 Bt At fld dword ptr [esi+ 7*4] ;x3 A3 Bt At fmul dword ptr [edx+ 7*4] ;B3 A3 Bt At fld dword ptr [esi+ 8*4] ;x3 B2 A2 Bt At fxch st(2) ;A2 B2 x3 Bt At faddp st(4),st ;B2 x3 Bt At fsubp st(2),st ;x3 Bt At fmul dword ptr [ecx+ 7*4] ;A3 Bt At fld dword ptr [esi+ 8*4] ;x3 A3 Bt At fmul dword ptr [edx+ 8*4] ;B3 A3 Bt At fld dword ptr [esi+ 9*4] ;x3 B2 A2 Bt At fxch st(2) ;A2 B2 x3 Bt At faddp st(4),st ;B2 x3 Bt At faddp st(2),st ;x3 Bt At fmul dword ptr [ecx+ 6*4] ;A3 Bt At fld dword ptr [esi+ 9*4] ;x3 A3 Bt At fmul dword ptr [edx+ 9*4] ;B3 A3 Bt At fld dword ptr [esi+10*4] ;x3 B2 A2 Bt At fxch st(2) ;A2 B2 x3 Bt At faddp st(4),st ;B2 x3 Bt At fsubp st(2),st ;x3 Bt At fmul dword ptr [ecx+ 5*4] ;A3 Bt At fld dword ptr [esi+10*4] ;x3 A3 Bt At fmul dword ptr [edx+10*4] ;B3 A3 Bt At fld dword ptr [esi+11*4] ;x3 B2 A2 Bt At fxch st(2) ;A2 B2 x3 Bt At faddp st(4),st ;B2 x3 Bt At faddp st(2),st ;x3 Bt At fmul dword ptr [ecx+ 4*4] ;A3 Bt At fld dword ptr [esi+11*4] ;x3 A3 Bt At fmul dword ptr [edx+11*4] ;B3 A3 Bt At fld dword ptr [esi+12*4] ;x3 B2 A2 Bt At fxch st(2) ;A2 B2 x3 Bt At faddp st(4),st ;B2 x3 Bt At fsubp st(2),st ;x3 Bt At fmul dword ptr [ecx+ 3*4] ;A3 Bt At fld dword ptr [esi+12*4] ;x3 A3 Bt At fmul dword ptr [edx+12*4] ;B3 A3 Bt At fld dword ptr [esi+13*4] ;x3 B2 A2 Bt At fxch st(2) ;A2 B2 x3 Bt At faddp st(4),st ;B2 x3 Bt At faddp st(2),st ;x3 Bt At fmul dword ptr [ecx+ 2*4] ;A3 Bt At fld dword ptr [esi+13*4] ;x3 A3 Bt At fmul dword ptr [edx+13*4] ;B3 A3 Bt At fld dword ptr [esi+14*4] ;x3 B2 A2 Bt At fxch st(2) ;A2 B2 x3 Bt At faddp st(4),st ;B2 x3 Bt At fsubp st(2),st ;x3 Bt At fmul dword ptr [ecx+ 1*4] ;A3 Bt At fld dword ptr [esi+14*4] ;x3 A3 Bt At fmul dword ptr [edx+14*4] ;B3 A3 Bt At fld dword ptr [esi+15*4] ;xF BE AE Bt At fxch st(2) ;AE BE xF Bt At faddp st(4),st ;BE xF Bt At faddp st(2),st ;xF Bt At fmul dword ptr [ecx+ 0*4] ;AF Bt At fld dword ptr [esi+15*4] ;xF AF Bt At fmul dword ptr [edx+15*4] ;BF AF Bt At add ecx,32*4 add edx,32*4 add esi,16*4 fsubp st(2),st ;AF Bt At faddp st(2),st ;Bt At fistp dword ptr [esp+0] ;[esp+0] = outf3 fistp dword ptr [esp+4] ;[esp+4] = outf1 mov eax,[esp+0] cmp eax,-32768 jge div1_noclip1low mov eax,-32768 div1_noclip1low: cmp eax,32767 jle div1_noclip1high mov eax,32767 div1_noclip1high: mov [ebx],ax mov eax,[esp+4] cmp eax,-32768 jge div1_noclip2low mov eax,-32768 div1_noclip2low: cmp eax,32767 jle div1_noclip2high mov eax,32767 div1_noclip2high: mov [edi],ax nop add ebx,2 sub edi,2 dec ebp jne div1_loop pop eax pop eax pop ebp pop edi mov ebx,esi pop esi ; mov window,ecx ; mov window2,edx ; mov u_ptr,ebx }; #else for (i = 1; i < 16; i++) { sum1 = 0.0F; sum2 = 0.0F; window = &wincoef[i][16-vb_ptr]; window2 = &wincoef[i][vb_ptr & 15]; si = 256+i*16; sum1 += window [ 0] * vbuf[si+ 0]; sum2 += window2[15] * vbuf[si+ 0]; sum1 -= window [ 1] * vbuf[si+ 1]; sum2 += window2[14] * vbuf[si+ 1]; sum1 += window [ 2] * vbuf[si+ 2]; sum2 += window2[13] * vbuf[si+ 2]; sum1 -= window [ 3] * vbuf[si+ 3]; sum2 += window2[12] * vbuf[si+ 3]; sum1 += window [ 4] * vbuf[si+ 4]; sum2 += window2[11] * vbuf[si+ 4]; sum1 -= window [ 5] * vbuf[si+ 5]; sum2 += window2[10] * vbuf[si+ 5]; sum1 += window [ 6] * vbuf[si+ 6]; sum2 += window2[ 9] * vbuf[si+ 6]; sum1 -= window [ 7] * vbuf[si+ 7]; sum2 += window2[ 8] * vbuf[si+ 7]; sum1 += window [ 8] * vbuf[si+ 8]; sum2 += window2[ 7] * vbuf[si+ 8]; sum1 -= window [ 9] * vbuf[si+ 9]; sum2 += window2[ 6] * vbuf[si+ 9]; sum1 += window [10] * vbuf[si+10]; sum2 += window2[ 5] * vbuf[si+10]; sum1 -= window [11] * vbuf[si+11]; sum2 += window2[ 4] * vbuf[si+11]; sum1 += window [12] * vbuf[si+12]; sum2 += window2[ 3] * vbuf[si+12]; sum1 -= window [13] * vbuf[si+13]; sum2 += window2[ 2] * vbuf[si+13]; sum1 += window [14] * vbuf[si+14]; sum2 += window2[ 1] * vbuf[si+14]; sum1 -= window [15] * vbuf[si+15]; sum2 += window2[ 0] * vbuf[si+15]; tmp = float_to_long(sum1); if (tmp > 32767) tmp = 32767; else if (tmp < -32768) tmp = -32768; pcm[i] = tmp; tmp = float_to_long(sum2); if (tmp > 32767) tmp = 32767; else if (tmp < -32768) tmp = -32768; pcm[32-i] = tmp; } #endif } /*-- special case --*/ coef = 0; bx = (vb_ptr+1)&15; sum1 = 0.0F; for (j = 0; j < 8; j++) { sum1 += wincoef2[coef++] * vbuf[bx]; bx = (bx+2)&15; } tmp = (long)sum1; if (tmp > 32767) tmp = 32767; else if (tmp < -32768) tmp = -32768; pcm[16] = tmp; } static bool pp_init_flag = true; static void pp_init() { int i,j; pp_init_flag = false; for(i=0; i<16; i++) for(j=0; j<16; j++) wincoef[i][j+16] = wincoef[i][j]; } void AMPDecoder::polyphase(float *band_l, float *band_r, short *samples, bool invert_odd) { if (pp_init_flag) pp_init(); fdct32(band_l, window[0] + winptr, !!(winptr & 1), invert_odd); if (band_r) { fdct32(band_r, window[1] + winptr, !!(winptr & 1), invert_odd); // window_dual(window[0], winptr, samples); // window_dual(window[1], winptr, samples + 1); window_dual(window, winptr, samples); } else window_mono(window[0], winptr, samples); winptr = (winptr - 1) & 15; } #if 0 void AMPDecoder::polyphase_single(float *band, int ch, short *samples) { if (pp_init_flag) pp_init(); fdct32(band, window[ch] + winptr, !!(winptr&1), false); window_dual(window[ch], winptr, samples); winptr = (winptr - 1) & 15; } #endif void AMPDecoder::polyphase_retract(int amount) { winptr = (winptr + amount) & 15; }