//#ifdef HAVE_CONFIG_H

#include "config.h"

//#endif



#include "audio.h"

#include "ssc_generic.h"



#ifndef VC_PROJ

#include <xtensa/tie/xt_hifi2.h>

extern ae_int32 AE_MUL16S(ae_int16 b, ae_int16 c);

extern ae_int64 AE_MUL32_HH(ae_int32x2 d0, ae_int32x2 d1);

#endif

short win_square[108] = 

{

	0, 0, 0, 0, 1, 3, 6, 11, 18, 29, 43, 62, 86, 116, 155, 201, 257, 323, 402, 493, 598, 718, 855, 

	1010, 1182, 1375, 1589, 1823, 2081, 2362, 2667, 2997, 3353, 3733, 4140, 4572, 5031, 5514, 6025, 

	6559, 7117, 7699, 8304, 8929, 9574, 10239, 10920, 11617, 12326, 13048, 13779, 14517, 15261, 16009, 

	16758, 17506, 18249, 18988, 19719, 20441, 21150, 21847, 22528, 23193, 23838, 24463, 25068, 25649, 

	26207, 26743, 27252, 27737, 28195, 28627, 29034, 29414, 29769, 30100, 30406, 30686, 30944, 31179, 

	31392, 31584, 31757, 31911, 32049, 32168, 32273, 32365, 32444, 32510, 32566, 32612, 32650, 32682, 

	32706, 32724, 32738, 32750, 32756, 32762, 32764, 32766, 32766, 32766, 32766, 32766

};





#if 1



#ifdef FOR_VISUAL_C_NO_ASSM

static void comb_filter_const(int *y, short T, int N,

      short g10, short g11, short g12)

{

	int i;

//#if 0

#ifdef VC_PROJ

	long long Ltmp;

#else

	ae_int64 Ltmp;

#endif

	for (i=0; i<N; i++)

	{

//#if 0

#ifdef VC_PROJ

		Ltmp = SSC_MULT16x32(g10,y[i-T]) + SSC_MULT16x32(g11, SSC_ADD32(y[i-T+1],y[i-T-1])) + SSC_MULT16x32(g12, SSC_ADD32(y[i-T+2],y[i-T-2]));

#else

		Ltmp = SSC_MULT16x32(g10,y[i-T]);

		AE_MULA32_HH(Ltmp, g11,SSC_ADD32(y[i-T+1],y[i-T-1]));

		AE_MULA32_HH(Ltmp, g12,SSC_ADD32(y[i-T+2],y[i-T-2]));

#endif

		y[i] += (int)((long long)Ltmp>>15);

	}

}



#endif







void comb_filter(int *y,short T0, short T1, int N,

	short g0, short g1,int overlap)



{

//	int i;

	short g00, g01, g02, g10, g11, g12;

	short w2;

//#if 0

#ifdef VC_PROJ

	long long Ltmp;

#else

	ae_int64 Ltmp;

#endif



	short *win = win_square;

	int i;

	

	if (g0==0 && g1==0) return;

	if (g0==g1 && T0==T1) overlap=0;

	

	g00 = SSC_MULT16x16_Q15(g0, 10048);

	g01 = SSC_MULT16x16_Q15(g0, 7112);

	g02 = SSC_MULT16x16_Q15(g0, 4248);

	g10 = SSC_MULT16x16_Q15(g1, 10048);

	g11 = SSC_MULT16x16_Q15(g1, 7112);

	g12 = SSC_MULT16x16_Q15(g1, 4248);



	if(g0==0)

	{

		i = overlap;

		do

		{

			w2 = *win++;

//#if 0

#ifdef VC_PROJ

			Ltmp = SSC_MULT16x32(SSC_MULT16x16_Q15(w2,g10),y[-T1])

				+ SSC_MULT16x32(SSC_MULT16x16_Q15(w2,g11),SSC_ADD32(y[-T1+1],y[-T1-1]))

				+ SSC_MULT16x32(SSC_MULT16x16_Q15(w2,g12),SSC_ADD32(y[-T1+2],y[-T1-2]));

#else

			Ltmp = SSC_MULT16x32(SSC_MULT16x16_Q15(w2,g10),y[-T1]);

			AE_MULA32_HH(Ltmp, SSC_MULT16x16_Q15(w2,g11),SSC_ADD32(y[-T1+1],y[-T1-1]));

			AE_MULA32_HH(Ltmp, SSC_MULT16x16_Q15(w2,g12),SSC_ADD32(y[-T1+2],y[-T1-2]));

#endif

			*y += (int)((long long)Ltmp>>15);

			y++;

		}

		while(--i);



	}

	else if(g1==0)

	{

		i = overlap;

		do				

		{

			w2 = *win++;

//#if 0

#ifdef VC_PROJ

			Ltmp = SSC_MULT16x32(SSC_MULT16x16_Q15((Q15ONE-w2),g00),y[-T0])

				+ SSC_MULT16x32(SSC_MULT16x16_Q15((Q15ONE-w2),g01),SSC_ADD32(y[-T0+1],y[-T0-1]))

				+ SSC_MULT16x32(SSC_MULT16x16_Q15((Q15ONE-w2),g02),SSC_ADD32(y[-T0+2],y[-T0-2]));

#else

			Ltmp = SSC_MULT16x32(SSC_MULT16x16_Q15((Q15ONE-w2),g00),y[-T0]);

			AE_MULA32_HH(Ltmp, SSC_MULT16x16_Q15((Q15ONE-w2),g01),SSC_ADD32(y[-T0+1],y[-T0-1]));

			AE_MULA32_HH(Ltmp, SSC_MULT16x16_Q15((Q15ONE-w2),g02),SSC_ADD32(y[-T0+2],y[-T0-2]));

#endif

			*y += (int)((long long)Ltmp>>15);

			y++;

		}

		while(--i);



		return;

	}

	else

	{

		if(T0==T1)

		{

		

		for(i=0;i<overlap;i++)			

		{



				w2 = *win++;

//#if 0

#ifdef VC_PROJ

				Ltmp = SSC_MULT16x32((SSC_MULT16x16_Q15((Q15ONE-w2),g00)+SSC_MULT16x16_Q15(w2,g10)),y[-T0])

					+ SSC_MULT16x32((SSC_MULT16x16_Q15((Q15ONE-w2),g01)+SSC_MULT16x16_Q15(w2,g11)),SSC_ADD32(y[-T0+1],y[-T0-1]))

					+ SSC_MULT16x32((SSC_MULT16x16_Q15((Q15ONE-w2),g02)+SSC_MULT16x16_Q15(w2,g12)),SSC_ADD32(y[-T0+2],y[-T0-2]));

#else

				Ltmp = SSC_MULT16x32((SSC_MULT16x16_Q15((Q15ONE-w2),g00)+SSC_MULT16x16_Q15(w2,g10)),y[-T0]);

				AE_MULA32_HH(Ltmp, (SSC_MULT16x16_Q15((Q15ONE-w2),g01)+SSC_MULT16x16_Q15(w2,g11)),SSC_ADD32(y[-T0+1],y[-T0-1]));

				AE_MULA32_HH(Ltmp, (SSC_MULT16x16_Q15((Q15ONE-w2),g02)+SSC_MULT16x16_Q15(w2,g12)),SSC_ADD32(y[-T0+2],y[-T0-2]));

#endif

				*y += (int)((long long)Ltmp>>15);

				y++;				

			}



		}

		else

		{



	

		for(i=0;i<overlap;i++)			

			{



				w2 = *win++;

//#if 0

#ifdef VC_PROJ

						Ltmp = SSC_MULT16x32(SSC_MULT16x16_Q15((Q15ONE-w2),g00),y[-T0])

							+ SSC_MULT16x32(SSC_MULT16x16_Q15((Q15ONE-w2),g01),SSC_ADD32(y[-T0+1],y[-T0-1]))

							+ SSC_MULT16x32(SSC_MULT16x16_Q15((Q15ONE-w2),g02),SSC_ADD32(y[-T0+2],y[-T0-2]))

							+ SSC_MULT16x32(SSC_MULT16x16_Q15(w2,g10),y[-T1])

							+ SSC_MULT16x32(SSC_MULT16x16_Q15(w2,g11),SSC_ADD32(y[-T1+1],y[-T1-1]))

							+ SSC_MULT16x32(SSC_MULT16x16_Q15(w2,g12),SSC_ADD32(y[-T1+2],y[-T1-2]));

#else

						Ltmp = SSC_MULT16x32(SSC_MULT16x16_Q15((Q15ONE-w2),g00),y[-T0]);

						AE_MULA32_HH(Ltmp, SSC_MULT16x16_Q15((Q15ONE-w2),g01),SSC_ADD32(y[-T0+1],y[-T0-1]));

						AE_MULA32_HH(Ltmp, SSC_MULT16x16_Q15((Q15ONE-w2),g02),SSC_ADD32(y[-T0+2],y[-T0-2]));

						AE_MULA32_HH(Ltmp, SSC_MULT16x16_Q15(w2,g10),y[-T1]);

						AE_MULA32_HH(Ltmp, SSC_MULT16x16_Q15(w2,g11),SSC_ADD32(y[-T1+1],y[-T1-1]));

						AE_MULA32_HH(Ltmp, SSC_MULT16x16_Q15(w2,g12),SSC_ADD32(y[-T1+2],y[-T1-2]));

#endif

						*y += (int)((long long)Ltmp>>15);

						y++;

			}

		}

	}



#ifndef FOR_VISUAL_C_NO_ASSM

	if(N-overlap != 0)

		comb_filter_const_asm(y, T1, N-overlap, g10, g11, g12);

#else

	comb_filter_const(y, T1, N-overlap, g10, g11, g12);

#endif



}



const signed char tf_select_table[8] = {

	0, -2, 0, -3, 3, 0, 1,-1

};

#endif