Audio: MFCC: Fix 32 bit mode HiFi3/4 window multiply

In 32-bit FFT mode the input data is 16-bit stored in the lower half
of a 32-bit icomplex32 container. The AE_MULFP32X16X2RS_L intrinsic
performs a Q1.31 x Q1.15 fractional multiply, so the 16-bit sample
must first be shifted left by 16 to Q1.31 format. Without this shift
the multiply treats the value as having 16 zero fractional bits,
producing near-zero windowed output and a corrupt FFT result.

Add the missing AE_SLAI32S(sample, 16) before the multiply in both
HiFi3 and HiFi4 mfcc_apply_window() 32-bit paths, matching the
generic C implementation.

Signed-off-by: Seppo Ingalsuo <seppo.ingalsuo@linux.intel.com>

Author: singalsu

src/audio/mfcc/mfcc_hifi3.c

@@ -144,6 +144,8 @@ void mfcc_apply_window(struct mfcc_state *state, int input_shift)
 	for (j = 0; j < fft->fft_size; j++) {
 		AE_L32_IP(sample, fft_in, 0);
 		AE_L16_XP(win, win_in, win_inc);
+		/* Data is 16-bit in 32-bit container, shift to Q1.31 for fractional multiply */
+		sample = AE_SLAI32S(sample, 16);
 		temp = AE_MULFP32X16X2RS_L(sample, win);
 		temp = AE_SLAA32S(temp, input_shift);
 		AE_S32_L_XP(temp, fft_in, fft_inc);

src/audio/mfcc/mfcc_hifi4.c

@@ -140,6 +140,8 @@ void mfcc_apply_window(struct mfcc_state *state, int input_shift)
 	for (j = 0; j < fft->fft_size; j++) {
 		AE_L32_IP(sample, fft_in, 0);
 		AE_L16_XP(win, win_in, win_inc);
+		/* Data is 16-bit in 32-bit container, shift to Q1.31 for fractional multiply */
+		sample = AE_SLAI32S(sample, 16);
 		temp = AE_MULFP32X16X2RS_L(sample, win);
 		temp = AE_SLAA32S(temp, input_shift);
 		AE_S32_L_XP(temp, fft_in, fft_inc);