make_tuple(&vpx_fdct16x16_c, &wrapper<vpx_idct16x16_256_add_c>,
&wrapper<vpx_idct16x16_256_add_c>, TX_16X16, 256, 8, 1),
make_tuple(&vpx_fdct16x16_c, &wrapper<vpx_idct16x16_256_add_c>,
+ &wrapper<vpx_idct16x16_38_add_c>, TX_16X16, 38, 8, 1),
+ make_tuple(&vpx_fdct16x16_c, &wrapper<vpx_idct16x16_256_add_c>,
&wrapper<vpx_idct16x16_10_add_c>, TX_16X16, 10, 8, 1),
make_tuple(&vpx_fdct16x16_c, &wrapper<vpx_idct16x16_256_add_c>,
&wrapper<vpx_idct16x16_1_add_c>, TX_16X16, 1, 8, 1),
}
}
+void vpx_idct16x16_38_add_c(const tran_low_t *input, uint8_t *dest,
+ int stride) {
+ int i, j;
+ tran_low_t out[16 * 16] = { 0 };
+ tran_low_t *outptr = out;
+ tran_low_t temp_in[16], temp_out[16];
+
+ // First transform rows. Since all non-zero dct coefficients are in
+ // upper-left 8x8 area, we only need to calculate first 8 rows here.
+ for (i = 0; i < 8; ++i) {
+ idct16_c(input, outptr);
+ input += 16;
+ outptr += 16;
+ }
+
+ // Then transform columns
+ for (i = 0; i < 16; ++i) {
+ for (j = 0; j < 16; ++j) temp_in[j] = out[j * 16 + i];
+ idct16_c(temp_in, temp_out);
+ for (j = 0; j < 16; ++j) {
+ dest[j * stride + i] = clip_pixel_add(dest[j * stride + i],
+ ROUND_POWER_OF_TWO(temp_out[j], 6));
+ }
+ }
+}
+
void vpx_idct16x16_1_add_c(const tran_low_t *input, uint8_t *dest, int stride) {
int i, j;
tran_high_t a1;
add_proto qw/void vpx_idct16x16_256_add/, "const tran_low_t *input, uint8_t *dest, int stride";
+ add_proto qw/void vpx_idct16x16_38_add/, "const tran_low_t *input, uint8_t *dest, int stride";
+
add_proto qw/void vpx_idct16x16_10_add/, "const tran_low_t *input, uint8_t *dest, int stride";
add_proto qw/void vpx_idct16x16_1_add/, "const tran_low_t *input, uint8_t *dest, int stride";
add_proto qw/void vpx_idct16x16_256_add/, "const tran_low_t *input, uint8_t *dest, int stride";
specialize qw/vpx_idct16x16_256_add neon sse2/;
+ add_proto qw/void vpx_idct16x16_38_add/, "const tran_low_t *input, uint8_t *dest, int stride";
+ specialize qw/vpx_idct16x16_38_add neon sse2/;
+ $vpx_idct16x16_38_add_neon=vpx_idct16x16_256_add_neon;
+ $vpx_idct16x16_38_add_sse2=vpx_idct16x16_256_add_sse2;
+
add_proto qw/void vpx_idct16x16_10_add/, "const tran_low_t *input, uint8_t *dest, int stride";
specialize qw/vpx_idct16x16_10_add neon sse2/;
add_proto qw/void vpx_idct16x16_256_add/, "const tran_low_t *input, uint8_t *dest, int stride";
+ add_proto qw/void vpx_idct16x16_38_add/, "const tran_low_t *input, uint8_t *dest, int stride";
+
add_proto qw/void vpx_idct16x16_10_add/, "const tran_low_t *input, uint8_t *dest, int stride";
add_proto qw/void vpx_idct32x32_1024_add/, "const tran_low_t *input, uint8_t *dest, int stride";
add_proto qw/void vpx_idct16x16_256_add/, "const tran_low_t *input, uint8_t *dest, int stride";
specialize qw/vpx_idct16x16_256_add sse2 neon dspr2 msa/;
+ add_proto qw/void vpx_idct16x16_38_add/, "const tran_low_t *input, uint8_t *dest, int stride";
+ specialize qw/vpx_idct16x16_38_add sse2 neon dspr2 msa/;
+ $vpx_idct16x16_38_add_neon=vpx_idct16x16_256_add_neon;
+ $vpx_idct16x16_38_add_sse2=vpx_idct16x16_256_add_sse2;
+ $vpx_idct16x16_38_add_dspr2=vpx_idct16x16_256_add_dspr2;
+ $vpx_idct16x16_38_add_msa=vpx_idct16x16_256_add_msa;
+
add_proto qw/void vpx_idct16x16_10_add/, "const tran_low_t *input, uint8_t *dest, int stride";
specialize qw/vpx_idct16x16_10_add sse2 neon dspr2 msa/;