for i in 0..4 {
transform!(luma_dc; blk[i], blk[i + 4], blk[i + 8], blk[i + 12]);
}
- for row in blk.chunks_mut(4) {
+ for row in blk.chunks_exact_mut(4) {
transform!(luma_dc; row[0], row[1], row[2], row[3]);
}
}
-pub fn idct(blk: &mut [i16; 16], qp: u8, quant_dc: bool) {
+pub fn idct_skip_dc(blk: &mut [i16; 16], qp: u8) {
const BLK_INDEX: [usize; 16] = [
0, 2, 0, 2,
2, 1, 2, 1,
];
let qidx = (qp % 6) as usize;
let shift = qp / 6;
- let start = if quant_dc { 0 } else { 1 };
- for (el, &idx) in blk.iter_mut().zip(BLK_INDEX.iter()).skip(start) {
+ for (el, &idx) in blk.iter_mut().zip(BLK_INDEX.iter()).skip(1) {
*el = (*el * LEVEL_SCALE[idx][qidx]) << shift;
}
- for row in blk.chunks_mut(4) {
+ for row in blk.chunks_exact_mut(4) {
+ transform!(row[0], row[1], row[2], row[3], 0);
+ }
+ for i in 0..4 {
+ transform!(blk[i], blk[i + 4], blk[i + 8], blk[i + 12], 6);
+ }
+}
+
+pub fn idct(blk: &mut [i16; 16], qp: u8) {
+ const BLK_INDEX: [usize; 16] = [
+ 0, 2, 0, 2,
+ 2, 1, 2, 1,
+ 0, 2, 0, 2,
+ 2, 1, 2, 1
+ ];
+ let qidx = (qp % 6) as usize;
+ let shift = qp / 6;
+ for (el, &idx) in blk.iter_mut().zip(BLK_INDEX.iter()) {
+ *el = (*el * LEVEL_SCALE[idx][qidx]) << shift;
+ }
+ for row in blk.chunks_exact_mut(4) {
transform!(row[0], row[1], row[2], row[3], 0);
}
for i in 0..4 {
*dst = i32::from(src).wrapping_mul(i32::from(qmat[idx])).wrapping_add(bias) >> shift;
}
}
- for row in tmp.chunks_mut(8) {
+ for row in tmp.chunks_exact_mut(8) {
transform!(row[0], row[1], row[2], row[3], row[4], row[5], row[6], row[7]);
}
for col in 0..8 {
pub fn add_coeffs(dst: &mut [u8], offset: usize, stride: usize, coeffs: &[i16]) {
let out = &mut dst[offset..][..stride * 3 + 4];
- for (line, src) in out.chunks_mut(stride).take(4).zip(coeffs.chunks(4)) {
+ for (line, src) in out.chunks_mut(stride).take(4).zip(coeffs.chunks_exact(4)) {
for (dst, src) in line.iter_mut().take(4).zip(src.iter()) {
*dst = (i32::from(*dst) + i32::from(*src)).max(0).min(255) as u8;
}
pub fn add_coeffs8(dst: &mut [u8], offset: usize, stride: usize, coeffs: &[i16; 64]) {
let out = &mut dst[offset..];
- for (line, src) in out.chunks_mut(stride).take(8).zip(coeffs.chunks(8)) {
+ for (line, src) in out.chunks_mut(stride).take(8).zip(coeffs.chunks_exact(8)) {
for (dst, src) in line.iter_mut().take(8).zip(src.iter()) {
*dst = (i32::from(*dst) + i32::from(*src)).max(0).min(255) as u8;
}