void vp9_fht8x8_c(const int16_t *input, tran_low_t *output, int stride,
int tx_type) {
if (tx_type == DCT_DCT) {
vpx_fdct8x8_c(input, output, stride);
} else {
tran_low_t out[64];
int i, j;
tran_low_t temp_in[8], temp_out[8];
const transform_2d ht = FHT_8[tx_type];
// Columns
for (i = 0; i < 8; ++i) {
for (j = 0; j < 8; ++j) temp_in[j] = input[j * stride + i] * 4;
ht.cols(temp_in, temp_out);
for (j = 0; j < 8; ++j) out[j * 8 + i] = temp_out[j];
}
// Rows
for (i = 0; i < 8; ++i) {
for (j = 0; j < 8; ++j) temp_in[j] = out[j + i * 8];
ht.rows(temp_in, temp_out);
for (j = 0; j < 8; ++j)
output[j + i * 8] = (temp_out[j] + (temp_out[j] < 0)) >> 1;
}
}
}
void vp9_fht16x16_c(const int16_t *input, tran_low_t *output, int stride,
int tx_type) {
if (tx_type == DCT_DCT) {
vpx_fdct16x16_c(input, output, stride);
} else {
tran_low_t out[256];
int i, j;
tran_low_t temp_in[16], temp_out[16];
const transform_2d ht = FHT_16[tx_type];
// Columns
for (i = 0; i < 16; ++i) {
for (j = 0; j < 16; ++j) temp_in[j] = input[j * stride + i] * 4;
ht.cols(temp_in, temp_out);
for (j = 0; j < 16; ++j)
out[j * 16 + i] = (temp_out[j] + 1 + (temp_out[j] < 0)) >> 2;
}
// Rows
for (i = 0; i < 16; ++i) {
for (j = 0; j < 16; ++j) temp_in[j] = out[j + i * 16];
ht.rows(temp_in, temp_out);
for (j = 0; j < 16; ++j) output[j + i * 16] = temp_out[j];
}
}
}
void vp9_fht4x4_c(const int16_t *input, tran_low_t *output, int stride,
int tx_type) {
if (tx_type == DCT_DCT) {
vpx_fdct4x4_c(input, output, stride);
} else {
tran_low_t out[4 * 4];
int i, j;
tran_low_t temp_in[4], temp_out[4];
const transform_2d ht = FHT_4[tx_type];
// Columns
for (i = 0; i < 4; ++i) {
for (j = 0; j < 4; ++j) temp_in[j] = input[j * stride + i] * 16;
if (i == 0 && temp_in[0]) temp_in[0] += 1;
ht.cols(temp_in, temp_out);
for (j = 0; j < 4; ++j) out[j * 4 + i] = temp_out[j];
}
// Rows
for (i = 0; i < 4; ++i) {
for (j = 0; j < 4; ++j) temp_in[j] = out[j + i * 4];
ht.rows(temp_in, temp_out);
for (j = 0; j < 4; ++j) output[j + i * 4] = (temp_out[j] + 1) >> 2;
}
}
}
void vp10_iht16x16_256_add_c(const tran_low_t *input, uint8_t *dest, int stride,
int tx_type) {
int i, j;
tran_low_t out[16 * 16];
tran_low_t *outptr = out;
tran_low_t temp_in[16], temp_out[16];
const transform_2d ht = IHT_16[tx_type];
// Rows
for (i = 0; i < 16; ++i) {
ht.rows(input, outptr);
input += 16;
outptr += 16;
}
// Columns
for (i = 0; i < 16; ++i) {
for (j = 0; j < 16; ++j) temp_in[j] = out[j * 16 + i];
ht.cols(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 vp10_iht8x8_64_add_c(const tran_low_t *input, uint8_t *dest, int stride,
int tx_type) {
int i, j;
tran_low_t out[8 * 8];
tran_low_t *outptr = out;
tran_low_t temp_in[8], temp_out[8];
const transform_2d ht = IHT_8[tx_type];
// inverse transform row vectors
for (i = 0; i < 8; ++i) {
ht.rows(input, outptr);
input += 8;
outptr += 8;
}
// inverse transform column vectors
for (i = 0; i < 8; ++i) {
for (j = 0; j < 8; ++j) temp_in[j] = out[j * 8 + i];
ht.cols(temp_in, temp_out);
for (j = 0; j < 8; ++j) {
dest[j * stride + i] = clip_pixel_add(dest[j * stride + i],
ROUND_POWER_OF_TWO(temp_out[j], 5));
}
}
}
void vp9_short_iht16x16_add_c(int16_t *input, uint8_t *dest, int dest_stride,
int tx_type) {
int i, j;
int16_t out[16 * 16];
int16_t *outptr = out;
int16_t temp_in[16], temp_out[16];
const transform_2d ht = IHT_16[tx_type];
// Rows
for (i = 0; i < 16; ++i) {
ht.rows(input, outptr);
input += 16;
outptr += 16;
}
// Columns
for (i = 0; i < 16; ++i) {
for (j = 0; j < 16; ++j)
temp_in[j] = out[j * 16 + i];
ht.cols(temp_in, temp_out);
for (j = 0; j < 16; ++j)
dest[j * dest_stride + i] = clip_pixel(ROUND_POWER_OF_TWO(temp_out[j], 6)
+ dest[j * dest_stride + i]); }
}
void vp9_short_iht8x8_add_c(int16_t *input, uint8_t *dest, int dest_stride,
int tx_type) {
int i, j;
int16_t out[8 * 8];
int16_t *outptr = out;
int16_t temp_in[8], temp_out[8];
const transform_2d ht = IHT_8[tx_type];
// inverse transform row vectors
for (i = 0; i < 8; ++i) {
ht.rows(input, outptr);
input += 8;
outptr += 8;
}
// inverse transform column vectors
for (i = 0; i < 8; ++i) {
for (j = 0; j < 8; ++j)
temp_in[j] = out[j * 8 + i];
ht.cols(temp_in, temp_out);
for (j = 0; j < 8; ++j)
dest[j * dest_stride + i] = clip_pixel(ROUND_POWER_OF_TWO(temp_out[j], 5)
+ dest[j * dest_stride + i]); }
}