refactor(av1): split inverse transform into stages and passes (S3776)

Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>

crates/av1/src/transform.rs

@@ -165,6 +165,15 @@ pub(crate) fn inverse_dct<T: Lane>(t: &mut [T], n: u32, r: u32) {
     for i in 0..len {
         t[i] = copy[brev(n, i)];
     }
+    inverse_dct_stage1(t, n, r);
+    inverse_dct_stage2(t, n, r);
+    inverse_dct_stage3(t, n, r);
+    inverse_dct_stage4(t, n, r);
+    inverse_dct_stage5(t, n, r);
+    inverse_dct_stage6(t, n, r);
+}
+
+fn inverse_dct_stage1<T: Lane>(t: &mut [T], n: u32, r: u32) {
     if n == 6 {
         for i in 0..16 {
             bf(t, 32 + i, 63 - i, 63 - 4 * brev(4, i) as i32, false);
@@ -190,6 +199,9 @@ pub(crate) fn inverse_dct<T: Lane>(t: &mut [T], n: u32, r: u32) {
             hd(t, 16 + 2 * i, 17 + 2 * i, i & 1 == 1, r);
         }
     }
+}
+
+fn inverse_dct_stage2<T: Lane>(t: &mut [T], n: u32, r: u32) {
     if n == 6 {
         for i in 0..4 {
             for j in 0..2 {
@@ -221,6 +233,9 @@ pub(crate) fn inverse_dct<T: Lane>(t: &mut [T], n: u32, r: u32) {
             }
         }
     }
+}
+
+fn inverse_dct_stage3<T: Lane>(t: &mut [T], n: u32, r: u32) {
     for i in 0..2 {
         bf(t, 2 * i, 2 * i + 1, 32 + 16 * i as i32, i == 0);
     }
@@ -248,6 +263,9 @@ pub(crate) fn inverse_dct<T: Lane>(t: &mut [T], n: u32, r: u32) {
             }
         }
     }
+}
+
+fn inverse_dct_stage4<T: Lane>(t: &mut [T], n: u32, r: u32) {
     for i in 0..2 {
         hd(t, i, 3 - i, false, r);
     }
@@ -273,6 +291,9 @@ pub(crate) fn inverse_dct<T: Lane>(t: &mut [T], n: u32, r: u32) {
             }
         }
     }
+}
+
+fn inverse_dct_stage5<T: Lane>(t: &mut [T], n: u32, r: u32) {
     if n >= 3 {
         for i in 0..4 {
             hd(t, i, 7 - i, false, r);
@@ -300,6 +321,9 @@ pub(crate) fn inverse_dct<T: Lane>(t: &mut [T], n: u32, r: u32) {
             hd(t, i, 15 - i, false, r);
         }
     }
+}
+
+fn inverse_dct_stage6<T: Lane>(t: &mut [T], n: u32, r: u32) {
     if n >= 5 {
         for i in 0..4 {
             bf(t, 27 - i, 20 + i, 32, true);
@@ -531,10 +555,9 @@ pub(crate) fn inverse_transform_2d(dequant: &[i32], residual: &mut [i32], tx_sz:
     let col_clamp = (bit_depth + 6).max(16);
     let rk = row_kind(tx_type);
     let ck = col_kind(tx_type);
-    let mut t = [0i32; 64];
     let rect2 = log2w.abs_diff(log2h) == 1;
+    let p = Tx2d { w, h, log2w, log2h, row_shift, col_shift, row_clamp, col_clamp, rk, ck, rect2 };
     if !lossless && bit_depth <= 10 {
-        let p = Tx2d { w, h, log2w, log2h, row_shift, col_shift, row_clamp, col_clamp, rk, ck, rect2 };
         if w >= 8 && h >= 8 {
             transform_lanes::<8>(dequant, residual, &p);
         } else {
@@ -542,6 +565,13 @@ pub(crate) fn inverse_transform_2d(dequant: &[i32], residual: &mut [i32], tx_sz:
         }
         return;
     }
+    scalar_row_pass(dequant, residual, &p, lossless);
+    scalar_col_pass(residual, &p, lossless);
+}
+
+fn scalar_row_pass(dequant: &[i32], residual: &mut [i32], p: &Tx2d, lossless: bool) {
+    let (w, h, log2w, row_shift, row_clamp, col_clamp, rk, rect2) = (p.w, p.h, p.log2w, p.row_shift, p.row_clamp, p.col_clamp, p.rk, p.rect2);
+    let mut t = [0i32; 64];
     let in_w = w.min(32);
     for i in 0..h {
         // Every 1D transform maps an all-zero row to zeros (each step is linear with
@@ -572,6 +602,11 @@ pub(crate) fn inverse_transform_2d(dequant: &[i32], residual: &mut [i32], tx_sz:
             residual[i * w + j] = round2(t[j], row_shift).clamp(lo, hi);
         }
     }
+}
+
+fn scalar_col_pass(residual: &mut [i32], p: &Tx2d, lossless: bool) {
+    let (w, h, log2h, col_shift, col_clamp, ck) = (p.w, p.h, p.log2h, p.col_shift, p.col_clamp, p.ck);
+    let mut t = [0i32; 64];
     for j in 0..w {
         for i in 0..h {
             t[i] = residual[i * w + j];
@@ -617,11 +652,16 @@ fn transform_1d<T: Lane>(t: &mut [T], kind: Kind1d, n: u32, r: u32) {
 /// The 2D inverse transform with `L` rows (row pass) or columns (column pass) at a time;
 /// bit-identical to the one-at-a-time process for bit depths up to 10.
 fn transform_lanes<const L: usize>(dequant: &[i32], residual: &mut [i32], p: &Tx2d) {
+    let mut t = [[0i32; L]; 64];
+    lanes_row_pass(dequant, residual, p, &mut t);
+    lanes_col_pass(residual, p, &mut t);
+}
+
+fn lanes_row_pass<const L: usize>(dequant: &[i32], residual: &mut [i32], p: &Tx2d, t: &mut [[i32; L]; 64]) {
     let (w, h) = (p.w, p.h);
     let in_w = w.min(32);
     let lo = -(1i32 << (p.col_clamp - 1));
     let hi = (1i32 << (p.col_clamp - 1)) - 1;
-    let mut t = [[0i32; L]; 64];
     let mut i0 = 0;
     while i0 < h {
         // Rows without coefficients (and rows 32..63) transform to zeros.
@@ -649,6 +689,10 @@ fn transform_lanes<const L: usize>(dequant: &[i32], residual: &mut [i32], p: &Tx
         }
         i0 += L;
     }
+}
+
+fn lanes_col_pass<const L: usize>(residual: &mut [i32], p: &Tx2d, t: &mut [[i32; L]; 64]) {
+    let (w, h) = (p.w, p.h);
     let mut j0 = 0;
     while j0 < w {
         for i in 0..h {

crates/h264/src/picture.rs

@@ -250,7 +250,7 @@ impl Frame {
 
     /// Copy the w x h window of chroma component `c` at (x0, y0) (clamped) into `out`.
     #[inline]
-    pub fn chroma_window(&self, c: usize, (x0, y0): (i32, i32), (w, h): (usize, usize), out: &mut [u8], os: usize) {
+    pub fn chroma_window(&self, c: usize, x0: i32, y0: i32, w: usize, h: usize, out: &mut [u8], os: usize) {
         let (cw, ch) = (self.cwidth, self.cheight);
         for r in 0..h {
             let yy = (y0 + r as i32).clamp(0, ch as i32 - 1) as usize;

crates/h264/src/slicedec.rs

@@ -232,7 +232,7 @@ fn predict_block(f: &Frame, mv: [i16; 2], px: usize, py: usize, bw: usize, bh: u
     let (fx, fy) = ((mv[0] & 7) as u32, (mv[1] & 7) as u32);
     for (comp, out) in c.into_iter().enumerate() {
         let mut w = [0u8; 16 * 9];
-        f.chroma_window(comp, (cx, cy), (cw + 1, ch + 1), &mut w, 16);
+        f.chroma_window(comp, cx, cy, cw + 1, ch + 1, &mut w, 16);
         inter::mc_chroma_win(&w, 16, fx, fy, cw, ch, out, cs);
     }
 }

crates/hevc/src/inter.rs

@@ -128,7 +128,7 @@ pub fn mc_chroma(f: &Frame, c: usize, x: i32, y: i32, mv: [i16; 2], w: usize, h:
     let yi = y + (mv[1] as i32 >> 3);
     let shift1 = bd.min(12) - 8;
     let shift3 = 14 - bd;
-    f.chroma_window(c, (xi - 1, yi - 1), (w + 3, h + 3), &mut sc.win, WIN);
+    f.chroma_window(c, xi - 1, yi - 1, w + 3, h + 3, &mut sc.win, WIN);
     interpolate::<4>(&sc.win, &mut sc.tmp, fx, fy, &CHROMA_FILTER[fx], &CHROMA_FILTER[fy], w, h, shift1, shift3, out);
 }
 
@@ -158,7 +158,7 @@ pub fn put_bi(p0: &[i16], p1: &[i16], w: usize, h: usize, bd: u32, dst: &mut [u1
 }
 
 /// Explicit weighted prediction, one list (8-265). `o` is the offset already scaled to the bit depth.
-pub fn put_weighted_uni(p: &[i16], (w, h): (usize, usize), bd: u32, (log2wd, wt, o): (u32, i32, i32), dst: &mut [u16], ds: usize) {
+pub fn put_weighted_uni(p: &[i16], w: usize, h: usize, bd: u32, log2wd: u32, wt: i32, o: i32, dst: &mut [u16], ds: usize) {
     let max = (1i32 << bd) - 1;
     for r in 0..h {
         for c in 0..w {
@@ -173,9 +173,14 @@ pub fn put_weighted_uni(p: &[i16], (w, h): (usize, usize), bd: u32, (log2wd, wt,
 pub fn put_weighted_bi(
     p0: &[i16],
     p1: &[i16],
-    (w, h): (usize, usize),
+    w: usize,
+    h: usize,
     bd: u32,
-    (log2wd, w0, w1, o0, o1): (u32, i32, i32, i32, i32),
+    log2wd: u32,
+    w0: i32,
+    w1: i32,
+    o0: i32,
+    o1: i32,
     dst: &mut [u16],
     ds: usize,
 ) {

crates/hevc/src/mvpred.rs

@@ -36,23 +36,9 @@ fn same_motion(a: &MvField, b: &MvField) -> bool {
     true
 }
 
-/// Coding block and prediction block geometry (6.4.2): coding block (xc, yc, ncb), prediction block (xp, yp, w, h, part_idx).
-#[derive(Clone, Copy)]
-pub struct PbGeom {
-    pub xc: i32,
-    pub yc: i32,
-    pub ncb: i32,
-    pub xp: i32,
-    pub yp: i32,
-    pub w: i32,
-    pub h: i32,
-    pub part_idx: usize,
-}
-
 impl<'a> SliceDecoder<'a> {
     /// Prediction block availability (6.4.2).
-    fn pb_avail(&self, g: &PbGeom, xn: i32, yn: i32) -> bool {
-        let PbGeom { xc, yc, ncb, xp, yp, w, h, part_idx } = *g;
+    fn pb_avail(&self, xc: i32, yc: i32, ncb: i32, xp: i32, yp: i32, w: i32, h: i32, part_idx: usize, xn: i32, yn: i32) -> bool {
         let same_cb = xc <= xn && yc <= yn && xc + ncb > xn && yc + ncb > yn;
         let avail = if !same_cb {
             self.pic.zavail(xp, yp, xn, yn, self.slice_addr)
@@ -73,13 +59,23 @@ impl<'a> SliceDecoder<'a> {
     }
 
     /// Merge mode motion (8.5.3.2.2).
-    pub fn derive_merge(&self, g: PbGeom, part: PartMode, merge_idx: usize) -> MvField {
-        let PbGeom { xc, yc, ncb, xp: xp0, yp: yp0, w: w0, h: h0, part_idx: part_idx0 } = g;
+    pub fn derive_merge(
+        &self,
+        xc: i32,
+        yc: i32,
+        ncb: i32,
+        xp0: i32,
+        yp0: i32,
+        w0: i32,
+        h0: i32,
+        part_idx0: usize,
+        part: PartMode,
+        merge_idx: usize,
+    ) -> MvField {
         let pml = self.pps.log2_parallel_merge_level;
         let (xp, yp, w, h, part_idx) = if pml > 2 && ncb == 8 { (xc, yc, ncb, ncb, 0) } else { (xp0, yp0, w0, h0, part_idx0) };
-        let g = PbGeom { xc, yc, ncb, xp, yp, w, h, part_idx };
         let same_mer = |xn: i32, yn: i32| (xp >> pml) == (xn >> pml) && (yp >> pml) == (yn >> pml);
-        let avail = |xn: i32, yn: i32| self.pb_avail(&g, xn, yn) && !same_mer(xn, yn);
+        let avail = |xn: i32, yn: i32| self.pb_avail(xc, yc, ncb, xp, yp, w, h, part_idx, xn, yn) && !same_mer(xn, yn);
         let mut list: [MvField; 5] = [MvField::default(); 5];
         let mut n = 0;
         let max = self.sh.max_num_merge_cand as usize;
@@ -241,8 +237,20 @@ impl<'a> SliceDecoder<'a> {
     }
 
     /// AMVP predictor (8.5.3.2.6, 8.5.3.2.7).
-    pub fn derive_amvp(&self, g: PbGeom, l: usize, ref_idx: usize, mvp_flag: usize) -> [i16; 2] {
-        let PbGeom { xp, yp, w, h, .. } = g;
+    pub fn derive_amvp(
+        &self,
+        xc: i32,
+        yc: i32,
+        ncb: i32,
+        xp: i32,
+        yp: i32,
+        w: i32,
+        h: i32,
+        part_idx: usize,
+        l: usize,
+        ref_idx: usize,
+        mvp_flag: usize,
+    ) -> [i16; 2] {
         let target = &self.refs[l][ref_idx];
         let (tpoc, tlt) = (target.poc, target.long_term);
         let cur_poc = self.pic.poc;
@@ -269,7 +277,10 @@ impl<'a> SliceDecoder<'a> {
             pick.map(|(mv, r)| if r.poc != tpoc && !r.long_term && !tlt { scale_mv(mv, cur_poc - r.poc, cur_poc - tpoc) } else { mv })
         };
         let pa = [(xp - 1, yp + h), (xp - 1, yp + h - 1)];
-        let av_a = [self.pb_avail(&g, pa[0].0, pa[0].1), self.pb_avail(&g, pa[1].0, pa[1].1)];
+        let av_a = [
+            self.pb_avail(xc, yc, ncb, xp, yp, w, h, part_idx, pa[0].0, pa[0].1),
+            self.pb_avail(xc, yc, ncb, xp, yp, w, h, part_idx, pa[1].0, pa[1].1),
+        ];
         let is_scaled = av_a[0] || av_a[1];
         let mut a = None;
         for k in 0..2 {
@@ -285,7 +296,7 @@ impl<'a> SliceDecoder<'a> {
             }
         }
         let pb = [(xp + w, yp - 1), (xp + w - 1, yp - 1), (xp - 1, yp - 1)];
-        let av_b = [0, 1, 2].map(|k| self.pb_avail(&g, pb[k].0, pb[k].1));
+        let av_b = [0, 1, 2].map(|k| self.pb_avail(xc, yc, ncb, xp, yp, w, h, part_idx, pb[k].0, pb[k].1));
         let mut b = None;
         for k in 0..3 {
             if av_b[k] && b.is_none() {

crates/hevc/src/picture.rs

@@ -125,7 +125,7 @@ impl Frame {
     }
 
     /// Copy a w x h window of chroma component `c` (0 = Cb, 1 = Cr) with edge replication.
-    pub fn chroma_window(&self, c: usize, (x0, y0): (i32, i32), (w, h): (usize, usize), out: &mut [i16], os: usize) {
+    pub fn chroma_window(&self, c: usize, x0: i32, y0: i32, w: usize, h: usize, out: &mut [i16], os: usize) {
         let (cw, ch) = (self.cwidth as i32, self.cheight as i32);
         let sh = self.log2_ctb - 1;
         let mask = (1i32 << sh) - 1;

crates/hevc/src/slicedec.rs

@@ -7,7 +7,6 @@ use crate::cabac::{Cabac, Contexts, init_contexts};
 use crate::error::{Result, ensure, invalid};
 use crate::inter;
 use crate::intra::{self, Refs};
-use crate::mvpred::PbGeom;
 use crate::params::{Layout, Pps, Sps};
 use crate::picture::{FrameRef, RefPic};
 use crate::slice::{SliceHeader, SliceType};
@@ -669,7 +668,7 @@ impl<'a> SliceDecoder<'a> {
         self.mark_edges(x0, y0, n, n, E_TU_V | E_PU_V, E_TU_H | E_PU_H);
         if skip {
             self.cu_intra = false;
-            self.prediction_unit(PbGeom { xc: x0, yc: y0, ncb: n, xp: x0, yp: y0, w: n, h: n, part_idx: 0 }, PartMode::P2Nx2N, true)?;
+            self.prediction_unit(x0, y0, n, x0, y0, n, n, 0, PartMode::P2Nx2N, true)?;
             self.finish_cu(x0, y0, n);
             return Ok(());
         }
@@ -698,7 +697,7 @@ impl<'a> SliceDecoder<'a> {
             }
             self.parse_intra_modes(x0, y0, n, part)?;
             self.max_trafo_depth = self.sps.max_th_depth_intra + self.intra_split as u32;
-            self.transform_tree((x0, y0), (x0, y0), log2, 0, 0, [true, true])?;
+            self.transform_tree(x0, y0, x0, y0, log2, 0, 0, [true, true])?;
         } else {
             self.intra_split = false;
             let mut merge2nx2n = false;
@@ -715,8 +714,7 @@ impl<'a> SliceDecoder<'a> {
                 PartMode::PNxN => &[(0, 0, h, h), (h, 0, h, h), (0, h, h, h), (h, h, h, h)],
             };
             for (i, &(px, py, pw, ph)) in parts.iter().enumerate() {
-                let merge =
-                    self.prediction_unit(PbGeom { xc: x0, yc: y0, ncb: n, xp: x0 + px, yp: y0 + py, w: pw, h: ph, part_idx: i }, part, false)?;
+                let merge = self.prediction_unit(x0, y0, n, x0 + px, y0 + py, pw, ph, i, part, false)?;
                 if part == PartMode::P2Nx2N {
                     merge2nx2n = merge;
                 }
@@ -730,7 +728,7 @@ impl<'a> SliceDecoder<'a> {
             let root_cbf = if !merge2nx2n { self.dd(RQT_ROOT_CBF) == 1 } else { true };
             if root_cbf {
                 self.max_trafo_depth = self.sps.max_th_depth_inter;
-                self.transform_tree((x0, y0), (x0, y0), log2, 0, 0, [true, true])?;
+                self.transform_tree(x0, y0, x0, y0, log2, 0, 0, [true, true])?;
             }
         }
         self.finish_cu(x0, y0, n);
@@ -900,7 +898,7 @@ impl<'a> SliceDecoder<'a> {
         Ok(())
     }
 
-    fn transform_tree(&mut self, (x0, y0): (i32, i32), base: (i32, i32), log2: u32, depth: u32, blk_idx: u32, parent_cbf_c: [bool; 2]) -> Result<()> {
+    fn transform_tree(&mut self, x0: i32, y0: i32, xb: i32, yb: i32, log2: u32, depth: u32, blk_idx: u32, parent_cbf_c: [bool; 2]) -> Result<()> {
         let split =
             if log2 <= self.sps.log2_max_tb && log2 > self.sps.log2_min_tb && depth < self.max_trafo_depth && !(self.intra_split && depth == 0) {
                 self.dd(SPLIT_TRANSFORM + (5 - log2 as usize)) == 1
@@ -925,10 +923,10 @@ impl<'a> SliceDecoder<'a> {
         }
         if split {
             let h = 1 << (log2 - 1);
-            self.transform_tree((x0, y0), (x0, y0), log2 - 1, depth + 1, 0, cbf_c)?;
-            self.transform_tree((x0 + h, y0), (x0, y0), log2 - 1, depth + 1, 1, cbf_c)?;
-            self.transform_tree((x0, y0 + h), (x0, y0), log2 - 1, depth + 1, 2, cbf_c)?;
-            self.transform_tree((x0 + h, y0 + h), (x0, y0), log2 - 1, depth + 1, 3, cbf_c)?;
+            self.transform_tree(x0, y0, x0, y0, log2 - 1, depth + 1, 0, cbf_c)?;
+            self.transform_tree(x0 + h, y0, x0, y0, log2 - 1, depth + 1, 1, cbf_c)?;
+            self.transform_tree(x0, y0 + h, x0, y0, log2 - 1, depth + 1, 2, cbf_c)?;
+            self.transform_tree(x0 + h, y0 + h, x0, y0, log2 - 1, depth + 1, 3, cbf_c)?;
             return Ok(());
         }
         let cbf_luma =
@@ -937,18 +935,10 @@ impl<'a> SliceDecoder<'a> {
         if depth > 0 {
             self.mark_edges(x0, y0, n, n, E_TU_V, E_TU_H);
         }
-        self.transform_unit((x0, y0), base, log2, blk_idx, cbf_luma, cbf_c)
+        self.transform_unit(x0, y0, xb, yb, log2, blk_idx, cbf_luma, cbf_c)
     }
 
-    fn transform_unit(
-        &mut self,
-        (x0, y0): (i32, i32),
-        (xb, yb): (i32, i32),
-        log2: u32,
-        blk_idx: u32,
-        cbf_luma: bool,
-        cbf_c: [bool; 2],
-    ) -> Result<()> {
+    fn transform_unit(&mut self, x0: i32, y0: i32, xb: i32, yb: i32, log2: u32, blk_idx: u32, cbf_luma: bool, cbf_c: [bool; 2]) -> Result<()> {
         let n = 1i32 << log2;
         let chroma_here = log2 > 2 || blk_idx == 3;
         let cbf_chroma = cbf_c[0] || cbf_c[1];
@@ -1274,7 +1264,7 @@ impl<'a> SliceDecoder<'a> {
                 max_y = max_y.max(yc);
             }
         }
-        self.reconstruct((x0, y0), log2, c, qp, ts, (max_x, max_y))
+        self.reconstruct(x0, y0, log2, c, qp, ts, max_x, max_y)
     }
 
     fn coeff_abs_level_remaining(&mut self, rice: u32) -> Result<i32> {
@@ -1293,7 +1283,7 @@ impl<'a> SliceDecoder<'a> {
     }
 
     /// Scaling, transform and addition to the prediction.
-    fn reconstruct(&mut self, (x0, y0): (i32, i32), log2: u32, c: usize, qp: i32, ts: bool, (max_x, max_y): (usize, usize)) -> Result<()> {
+    fn reconstruct(&mut self, x0: i32, y0: i32, log2: u32, c: usize, qp: i32, ts: bool, max_x: usize, max_y: usize) -> Result<()> {
         let n = 1usize << log2;
         let bd = if c == 0 { self.pic.bd_y } else { self.pic.bd_c };
         let coeffs = &mut self.coeffs[..n * n];
@@ -1345,8 +1335,19 @@ impl<'a> SliceDecoder<'a> {
     }
 
     /// prediction_unit(): parse, derive motion, store it and predict. Returns merge_flag.
-    fn prediction_unit(&mut self, g: PbGeom, part: PartMode, skip: bool) -> Result<bool> {
-        let PbGeom { xc, yc, ncb, xp, yp, w, h, part_idx } = g;
+    fn prediction_unit(
+        &mut self,
+        xc: i32,
+        yc: i32,
+        ncb: i32,
+        xp: i32,
+        yp: i32,
+        w: i32,
+        h: i32,
+        part_idx: usize,
+        part: PartMode,
+        skip: bool,
+    ) -> Result<bool> {
         let merge = if skip { true } else { self.dd(MERGE_FLAG) == 1 };
         let mvf = if merge {
             let idx = if self.sh.max_num_merge_cand > 1 {
@@ -1361,7 +1362,7 @@ impl<'a> SliceDecoder<'a> {
             } else {
                 0
             };
-            self.derive_merge(PbGeom { xc, yc, ncb, xp, yp, w, h, part_idx }, part, idx as usize)
+            self.derive_merge(xc, yc, ncb, xp, yp, w, h, part_idx, part, idx as usize)
         } else {
             let pred_idc = if self.sh.is_b() {
                 if w + h != 12 {
@@ -1405,7 +1406,7 @@ impl<'a> SliceDecoder<'a> {
                 if ref_idx[l] < 0 {
                     continue;
                 }
-                let mvp = self.derive_amvp(PbGeom { xc, yc, ncb, xp, yp, w, h, part_idx }, l, ref_idx[l] as usize, mvp_flag[l] as usize);
+                let mvp = self.derive_amvp(xc, yc, ncb, xp, yp, w, h, part_idx, l, ref_idx[l] as usize, mvp_flag[l] as usize);
                 // mvLX = (mvp + mvd) wrapped to 16 bits (8-94)
                 f.mv[l] = [(mvp[0] as i32 + mvd[l][0]) as i16, (mvp[1] as i32 + mvd[l][1]) as i16];
             }
@@ -1503,11 +1504,11 @@ impl<'a> SliceDecoder<'a> {
             if bi {
                 let (w0, o0) = get(0);
                 let (w1, o1) = get(1);
-                inter::put_weighted_bi(&self.pred0, &self.pred1, (bw, bh), bd, (log2wd, w0, w1, o0, o1), dst, stride);
+                inter::put_weighted_bi(&self.pred0, &self.pred1, bw, bh, bd, log2wd, w0, w1, o0, o1, dst, stride);
             } else {
                 let l = if lists[0] { 0 } else { 1 };
                 let (wt, o) = get(l);
-                inter::put_weighted_uni(&self.pred0, (bw, bh), bd, (log2wd, wt, o), dst, stride);
+                inter::put_weighted_uni(&self.pred0, bw, bh, bd, log2wd, wt, o, dst, stride);
             }
         }
     }