std/vp8: calculate the predicted values

This commit, part of a series, is inspired by pull request PR #168 by
Damian Kaczmarek (with AI assistance), with some modifications.

Updates PR #168
This commit is contained in:
Nigel Tao committed 2026-06-16 14:00:28 +10:00
1 parent 3a74d542ab
commit bb3c85e2ef
8 files changed
+2276 -11

No files matched your search

+74
View File
@@ -111,6 +111,80 @@ pri const CLAMP_NO_MORE_THAN_9 : roarray[16] base.u8[..= 9] = [
0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 9, 9, 9, 9, 9, 9,
]
pri const CLAMP : roarray[1024] base.u8 = [
0x00, 0x01, 0x02, 0x03, 0x04, 0x05, 0x06, 0x07, 0x08, 0x09, 0x0A, 0x0B, 0x0C, 0x0D, 0x0E, 0x0F,
0x10, 0x11, 0x12, 0x13, 0x14, 0x15, 0x16, 0x17, 0x18, 0x19, 0x1A, 0x1B, 0x1C, 0x1D, 0x1E, 0x1F,
0x20, 0x21, 0x22, 0x23, 0x24, 0x25, 0x26, 0x27, 0x28, 0x29, 0x2A, 0x2B, 0x2C, 0x2D, 0x2E, 0x2F,
0x30, 0x31, 0x32, 0x33, 0x34, 0x35, 0x36, 0x37, 0x38, 0x39, 0x3A, 0x3B, 0x3C, 0x3D, 0x3E, 0x3F,
0x40, 0x41, 0x42, 0x43, 0x44, 0x45, 0x46, 0x47, 0x48, 0x49, 0x4A, 0x4B, 0x4C, 0x4D, 0x4E, 0x4F,
0x50, 0x51, 0x52, 0x53, 0x54, 0x55, 0x56, 0x57, 0x58, 0x59, 0x5A, 0x5B, 0x5C, 0x5D, 0x5E, 0x5F,
0x60, 0x61, 0x62, 0x63, 0x64, 0x65, 0x66, 0x67, 0x68, 0x69, 0x6A, 0x6B, 0x6C, 0x6D, 0x6E, 0x6F,
0x70, 0x71, 0x72, 0x73, 0x74, 0x75, 0x76, 0x77, 0x78, 0x79, 0x7A, 0x7B, 0x7C, 0x7D, 0x7E, 0x7F,
0x80, 0x81, 0x82, 0x83, 0x84, 0x85, 0x86, 0x87, 0x88, 0x89, 0x8A, 0x8B, 0x8C, 0x8D, 0x8E, 0x8F,
0x90, 0x91, 0x92, 0x93, 0x94, 0x95, 0x96, 0x97, 0x98, 0x99, 0x9A, 0x9B, 0x9C, 0x9D, 0x9E, 0x9F,
0xA0, 0xA1, 0xA2, 0xA3, 0xA4, 0xA5, 0xA6, 0xA7, 0xA8, 0xA9, 0xAA, 0xAB, 0xAC, 0xAD, 0xAE, 0xAF,
0xB0, 0xB1, 0xB2, 0xB3, 0xB4, 0xB5, 0xB6, 0xB7, 0xB8, 0xB9, 0xBA, 0xBB, 0xBC, 0xBD, 0xBE, 0xBF,
0xC0, 0xC1, 0xC2, 0xC3, 0xC4, 0xC5, 0xC6, 0xC7, 0xC8, 0xC9, 0xCA, 0xCB, 0xCC, 0xCD, 0xCE, 0xCF,
0xD0, 0xD1, 0xD2, 0xD3, 0xD4, 0xD5, 0xD6, 0xD7, 0xD8, 0xD9, 0xDA, 0xDB, 0xDC, 0xDD, 0xDE, 0xDF,
0xE0, 0xE1, 0xE2, 0xE3, 0xE4, 0xE5, 0xE6, 0xE7, 0xE8, 0xE9, 0xEA, 0xEB, 0xEC, 0xED, 0xEE, 0xEF,
0xF0, 0xF1, 0xF2, 0xF3, 0xF4, 0xF5, 0xF6, 0xF7, 0xF8, 0xF9, 0xFA, 0xFB, 0xFC, 0xFD, 0xFE, 0xFF,
0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF,
0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF,
0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF,
0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF,
0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF,
0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF,
0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF,
0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF,
0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF,
0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF,
0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF,
0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF,
0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF,
0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF,
0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF,
0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF,
0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
]
// RFC 6386 Section 11.5. Subblock Mode Probability Table.
//
// Note that the specification (the RFC) differs from the implementation
+1 -2
View File
@@ -13,8 +13,7 @@ pri func decoder.decode_coefficients!(
mbx: base.u32[..= 0x3FF],
mby: base.u32[..= 0x3FF],
seg: base.u32[..= 3],
luma_mode: base.u32[..= 4],
chroma_mode: base.u32[..= 3]) {
luma_mode: base.u32[..= 4]) {
var bc : base.u32
var by : base.u32
var bx : base.u32
+5 -1
View File
@@ -82,8 +82,12 @@ pri func decoder.decode_one_macroblock!(workbuf: slice base.u8, mbx: base.u32[..
} else {
this.decode_coefficients!(
workbuf: args.workbuf, mbx: args.mbx, mby: args.mby,
seg: seg, luma_mode: luma_mode, chroma_mode: chroma_mode)
seg: seg, luma_mode: luma_mode)
}
this.reconstruct!(
workbuf: args.workbuf, mbx: args.mbx, mby: args.mby,
luma_mode: luma_mode, chroma_mode: chroma_mode)
}
// RFC 6386 Section 11.2. Luma Modes.
+675
View File
@@ -0,0 +1,675 @@
// Copyright 2026 The Wuffs Authors.
//
// Licensed under the Apache License, Version 2.0 <LICENSE-APACHE or
// https://www.apache.org/licenses/LICENSE-2.0> or the MIT license
// <LICENSE-MIT or https://opensource.org/licenses/MIT>, at your
// option. This file may not be copied, modified, or distributed
// except according to those terms.
//
// SPDX-License-Identifier: Apache-2.0 OR MIT
// --------
// Predicts a 4×4 luma block, the Xs below, based on samples top and left.
//
// a b c d e f g h i
// p X X X X
// q X X X X
// r X X X X
// s X X X X
pri func decoder.predict_y4!(b: base.u32[..= 15]) {
var cachey : base.u32[..= 13]
var cachex : base.u32[..= 20]
var mode : base.u32[..= 9]
var z : base.u32
var avg : base.u8
var tm0 : base.u32
var tm1 : base.u32
var tm2 : base.u32
var a : base.u32
var b : base.u32
var c : base.u32
var d : base.u32
var e : base.u32
var f : base.u32
var g : base.u32
var h : base.u32
var i : base.u32
var p : base.u32
var q : base.u32
var r : base.u32
var s : base.u32
var ab : base.u8
var bc : base.u8
var cd : base.u8
var de : base.u8
var ef : base.u8
var pa : base.u8
var qp : base.u8
var rq : base.u8
var sr : base.u8
var abc : base.u8
var bcd : base.u8
var cde : base.u8
var def : base.u8
var efg : base.u8
var fgh : base.u8
var ghi : base.u8
var hii : base.u8
var pab : base.u8
var qpa : base.u8
var rqp : base.u8
var srq : base.u8
var ssr : base.u8
var sss : base.u8
cachey = ((args.b & 0xC) * 1) + 1
cachex = ((args.b & 0x3) * 4) + 8
mode = this.mb_subblock_modes[args.b] as base.u32
if mode == 0 { // DC.
avg = ((4 +
(this.yuv_cache[cachey - 1][cachex + 0] as base.u32) +
(this.yuv_cache[cachey - 1][cachex + 1] as base.u32) +
(this.yuv_cache[cachey - 1][cachex + 2] as base.u32) +
(this.yuv_cache[cachey - 1][cachex + 3] as base.u32) +
(this.yuv_cache[cachey + 0][cachex - 1] as base.u32) +
(this.yuv_cache[cachey + 1][cachex - 1] as base.u32) +
(this.yuv_cache[cachey + 2][cachex - 1] as base.u32) +
(this.yuv_cache[cachey + 3][cachex - 1] as base.u32)) / 8) as base.u8
z = 0
while z < 16 {
this.yuv_cache[cachey + (z >> 2)][cachex + (z & 3)] = avg
z += 1
}
} else if mode == 1 { // TM.
tm0 = this.yuv_cache[cachey - 1][cachex - 1] as base.u32
z = 0
while z < 16,
inv cachey >= 1,
inv cachex >= 8,
{
tm1 = this.yuv_cache[cachey + (z >> 2)][cachex - 1] as base.u32
tm2 = this.yuv_cache[cachey - 1][cachex + (z & 3)] as base.u32
this.yuv_cache[cachey + (z >> 2)][cachex + (z & 3)] =
CLAMP[((tm1 + tm2) ~mod- tm0) & 1023]
z += 1
}
} else if mode == 2 { // VE.
a = this.yuv_cache[cachey - 1][cachex - 1] as base.u32
b = this.yuv_cache[cachey - 1][cachex + 0] as base.u32
c = this.yuv_cache[cachey - 1][cachex + 1] as base.u32
d = this.yuv_cache[cachey - 1][cachex + 2] as base.u32
e = this.yuv_cache[cachey - 1][cachex + 3] as base.u32
f = this.yuv_cache[cachey - 1][cachex + 4] as base.u32
def = (((d + (2 * e) + f + 2) / 4) & 0xFF) as base.u8
cde = (((c + (2 * d) + e + 2) / 4) & 0xFF) as base.u8
bcd = (((b + (2 * c) + d + 2) / 4) & 0xFF) as base.u8
abc = (((a + (2 * b) + c + 2) / 4) & 0xFF) as base.u8
z = 0
while z < 4 {
this.yuv_cache[cachey + z][cachex + 0] = abc
this.yuv_cache[cachey + z][cachex + 1] = bcd
this.yuv_cache[cachey + z][cachex + 2] = cde
this.yuv_cache[cachey + z][cachex + 3] = def
z += 1
}
} else if mode == 3 { // HE.
s = this.yuv_cache[cachey + 3][cachex - 1] as base.u32
r = this.yuv_cache[cachey + 2][cachex - 1] as base.u32
q = this.yuv_cache[cachey + 1][cachex - 1] as base.u32
p = this.yuv_cache[cachey + 0][cachex - 1] as base.u32
a = this.yuv_cache[cachey - 1][cachex - 1] as base.u32
qpa = (((q + (2 * p) + a + 2) / 4) & 0xFF) as base.u8
rqp = (((r + (2 * q) + p + 2) / 4) & 0xFF) as base.u8
srq = (((s + (2 * r) + q + 2) / 4) & 0xFF) as base.u8
ssr = (((s + (2 * s) + r + 2) / 4) & 0xFF) as base.u8
z = 0
while z < 4 {
this.yuv_cache[cachey + 0][cachex + z] = qpa
this.yuv_cache[cachey + 1][cachex + z] = rqp
this.yuv_cache[cachey + 2][cachex + z] = srq
this.yuv_cache[cachey + 3][cachex + z] = ssr
z += 1
}
} else if mode == 4 { // RD.
s = this.yuv_cache[cachey + 3][cachex - 1] as base.u32
r = this.yuv_cache[cachey + 2][cachex - 1] as base.u32
q = this.yuv_cache[cachey + 1][cachex - 1] as base.u32
p = this.yuv_cache[cachey + 0][cachex - 1] as base.u32
a = this.yuv_cache[cachey - 1][cachex - 1] as base.u32
b = this.yuv_cache[cachey - 1][cachex + 0] as base.u32
c = this.yuv_cache[cachey - 1][cachex + 1] as base.u32
d = this.yuv_cache[cachey - 1][cachex + 2] as base.u32
e = this.yuv_cache[cachey - 1][cachex + 3] as base.u32
cde = (((c + (2 * d) + e + 2) / 4) & 0xFF) as base.u8
bcd = (((b + (2 * c) + d + 2) / 4) & 0xFF) as base.u8
abc = (((a + (2 * b) + c + 2) / 4) & 0xFF) as base.u8
pab = (((p + (2 * a) + b + 2) / 4) & 0xFF) as base.u8
qpa = (((q + (2 * p) + a + 2) / 4) & 0xFF) as base.u8
rqp = (((r + (2 * q) + p + 2) / 4) & 0xFF) as base.u8
srq = (((s + (2 * r) + q + 2) / 4) & 0xFF) as base.u8
this.yuv_cache[cachey + 0][cachex + 0] = pab
this.yuv_cache[cachey + 0][cachex + 1] = abc
this.yuv_cache[cachey + 0][cachex + 2] = bcd
this.yuv_cache[cachey + 0][cachex + 3] = cde
this.yuv_cache[cachey + 1][cachex + 0] = qpa
this.yuv_cache[cachey + 1][cachex + 1] = pab
this.yuv_cache[cachey + 1][cachex + 2] = abc
this.yuv_cache[cachey + 1][cachex + 3] = bcd
this.yuv_cache[cachey + 2][cachex + 0] = rqp
this.yuv_cache[cachey + 2][cachex + 1] = qpa
this.yuv_cache[cachey + 2][cachex + 2] = pab
this.yuv_cache[cachey + 2][cachex + 3] = abc
this.yuv_cache[cachey + 3][cachex + 0] = srq
this.yuv_cache[cachey + 3][cachex + 1] = rqp
this.yuv_cache[cachey + 3][cachex + 2] = qpa
this.yuv_cache[cachey + 3][cachex + 3] = pab
} else if mode == 5 { // VR.
r = this.yuv_cache[cachey + 2][cachex - 1] as base.u32
q = this.yuv_cache[cachey + 1][cachex - 1] as base.u32
p = this.yuv_cache[cachey + 0][cachex - 1] as base.u32
a = this.yuv_cache[cachey - 1][cachex - 1] as base.u32
b = this.yuv_cache[cachey - 1][cachex + 0] as base.u32
c = this.yuv_cache[cachey - 1][cachex + 1] as base.u32
d = this.yuv_cache[cachey - 1][cachex + 2] as base.u32
e = this.yuv_cache[cachey - 1][cachex + 3] as base.u32
ab = (((a + b + 1) / 2) & 0xFF) as base.u8
bc = (((b + c + 1) / 2) & 0xFF) as base.u8
cd = (((c + d + 1) / 2) & 0xFF) as base.u8
de = (((d + e + 1) / 2) & 0xFF) as base.u8
cde = (((c + (2 * d) + e + 2) / 4) & 0xFF) as base.u8
bcd = (((b + (2 * c) + d + 2) / 4) & 0xFF) as base.u8
abc = (((a + (2 * b) + c + 2) / 4) & 0xFF) as base.u8
pab = (((p + (2 * a) + b + 2) / 4) & 0xFF) as base.u8
qpa = (((q + (2 * p) + a + 2) / 4) & 0xFF) as base.u8
rqp = (((r + (2 * q) + p + 2) / 4) & 0xFF) as base.u8
this.yuv_cache[cachey + 0][cachex + 0] = ab
this.yuv_cache[cachey + 0][cachex + 1] = bc
this.yuv_cache[cachey + 0][cachex + 2] = cd
this.yuv_cache[cachey + 0][cachex + 3] = de
this.yuv_cache[cachey + 1][cachex + 0] = pab
this.yuv_cache[cachey + 1][cachex + 1] = abc
this.yuv_cache[cachey + 1][cachex + 2] = bcd
this.yuv_cache[cachey + 1][cachex + 3] = cde
this.yuv_cache[cachey + 2][cachex + 0] = qpa
this.yuv_cache[cachey + 2][cachex + 1] = ab
this.yuv_cache[cachey + 2][cachex + 2] = bc
this.yuv_cache[cachey + 2][cachex + 3] = cd
this.yuv_cache[cachey + 3][cachex + 0] = rqp
this.yuv_cache[cachey + 3][cachex + 1] = pab
this.yuv_cache[cachey + 3][cachex + 2] = abc
this.yuv_cache[cachey + 3][cachex + 3] = bcd
} else if mode == 6 { // LD.
b = this.yuv_cache[cachey - 1][cachex + 0] as base.u32
c = this.yuv_cache[cachey - 1][cachex + 1] as base.u32
d = this.yuv_cache[cachey - 1][cachex + 2] as base.u32
e = this.yuv_cache[cachey - 1][cachex + 3] as base.u32
f = this.yuv_cache[cachey - 1][cachex + 4] as base.u32
g = this.yuv_cache[cachey - 1][cachex + 5] as base.u32
h = this.yuv_cache[cachey - 1][cachex + 6] as base.u32
i = this.yuv_cache[cachey - 1][cachex + 7] as base.u32
hii = (((h + (2 * i) + i + 2) / 4) & 0xFF) as base.u8
ghi = (((g + (2 * h) + i + 2) / 4) & 0xFF) as base.u8
fgh = (((f + (2 * g) + h + 2) / 4) & 0xFF) as base.u8
efg = (((e + (2 * f) + g + 2) / 4) & 0xFF) as base.u8
def = (((d + (2 * e) + f + 2) / 4) & 0xFF) as base.u8
cde = (((c + (2 * d) + e + 2) / 4) & 0xFF) as base.u8
bcd = (((b + (2 * c) + d + 2) / 4) & 0xFF) as base.u8
this.yuv_cache[cachey + 0][cachex + 0] = bcd
this.yuv_cache[cachey + 0][cachex + 1] = cde
this.yuv_cache[cachey + 0][cachex + 2] = def
this.yuv_cache[cachey + 0][cachex + 3] = efg
this.yuv_cache[cachey + 1][cachex + 0] = cde
this.yuv_cache[cachey + 1][cachex + 1] = def
this.yuv_cache[cachey + 1][cachex + 2] = efg
this.yuv_cache[cachey + 1][cachex + 3] = fgh
this.yuv_cache[cachey + 2][cachex + 0] = def
this.yuv_cache[cachey + 2][cachex + 1] = efg
this.yuv_cache[cachey + 2][cachex + 2] = fgh
this.yuv_cache[cachey + 2][cachex + 3] = ghi
this.yuv_cache[cachey + 3][cachex + 0] = efg
this.yuv_cache[cachey + 3][cachex + 1] = fgh
this.yuv_cache[cachey + 3][cachex + 2] = ghi
this.yuv_cache[cachey + 3][cachex + 3] = hii
} else if mode == 7 { // VL.
b = this.yuv_cache[cachey - 1][cachex + 0] as base.u32
c = this.yuv_cache[cachey - 1][cachex + 1] as base.u32
d = this.yuv_cache[cachey - 1][cachex + 2] as base.u32
e = this.yuv_cache[cachey - 1][cachex + 3] as base.u32
f = this.yuv_cache[cachey - 1][cachex + 4] as base.u32
g = this.yuv_cache[cachey - 1][cachex + 5] as base.u32
h = this.yuv_cache[cachey - 1][cachex + 6] as base.u32
i = this.yuv_cache[cachey - 1][cachex + 7] as base.u32
ef = (((e + f + 1) / 2) & 0xFF) as base.u8
de = (((d + e + 1) / 2) & 0xFF) as base.u8
cd = (((c + d + 1) / 2) & 0xFF) as base.u8
bc = (((b + c + 1) / 2) & 0xFF) as base.u8
ghi = (((g + (2 * h) + i + 2) / 4) & 0xFF) as base.u8
fgh = (((f + (2 * g) + h + 2) / 4) & 0xFF) as base.u8
efg = (((e + (2 * f) + g + 2) / 4) & 0xFF) as base.u8
def = (((d + (2 * e) + f + 2) / 4) & 0xFF) as base.u8
cde = (((c + (2 * d) + e + 2) / 4) & 0xFF) as base.u8
bcd = (((b + (2 * c) + d + 2) / 4) & 0xFF) as base.u8
this.yuv_cache[cachey + 0][cachex + 0] = bc
this.yuv_cache[cachey + 0][cachex + 1] = cd
this.yuv_cache[cachey + 0][cachex + 2] = de
this.yuv_cache[cachey + 0][cachex + 3] = ef
this.yuv_cache[cachey + 1][cachex + 0] = bcd
this.yuv_cache[cachey + 1][cachex + 1] = cde
this.yuv_cache[cachey + 1][cachex + 2] = def
this.yuv_cache[cachey + 1][cachex + 3] = efg
this.yuv_cache[cachey + 2][cachex + 0] = cd
this.yuv_cache[cachey + 2][cachex + 1] = de
this.yuv_cache[cachey + 2][cachex + 2] = ef
this.yuv_cache[cachey + 2][cachex + 3] = fgh
this.yuv_cache[cachey + 3][cachex + 0] = cde
this.yuv_cache[cachey + 3][cachex + 1] = def
this.yuv_cache[cachey + 3][cachex + 2] = efg
this.yuv_cache[cachey + 3][cachex + 3] = ghi
} else if mode == 8 { // HD.
s = this.yuv_cache[cachey + 3][cachex - 1] as base.u32
r = this.yuv_cache[cachey + 2][cachex - 1] as base.u32
q = this.yuv_cache[cachey + 1][cachex - 1] as base.u32
p = this.yuv_cache[cachey + 0][cachex - 1] as base.u32
a = this.yuv_cache[cachey - 1][cachex - 1] as base.u32
b = this.yuv_cache[cachey - 1][cachex + 0] as base.u32
c = this.yuv_cache[cachey - 1][cachex + 1] as base.u32
d = this.yuv_cache[cachey - 1][cachex + 2] as base.u32
sr = (((s + r + 1) / 2) & 0xFF) as base.u8
rq = (((r + q + 1) / 2) & 0xFF) as base.u8
qp = (((q + p + 1) / 2) & 0xFF) as base.u8
pa = (((p + a + 1) / 2) & 0xFF) as base.u8
bcd = (((b + (2 * c) + d + 2) / 4) & 0xFF) as base.u8
abc = (((a + (2 * b) + c + 2) / 4) & 0xFF) as base.u8
pab = (((p + (2 * a) + b + 2) / 4) & 0xFF) as base.u8
qpa = (((q + (2 * p) + a + 2) / 4) & 0xFF) as base.u8
rqp = (((r + (2 * q) + p + 2) / 4) & 0xFF) as base.u8
srq = (((s + (2 * r) + q + 2) / 4) & 0xFF) as base.u8
this.yuv_cache[cachey + 0][cachex + 0] = pa
this.yuv_cache[cachey + 0][cachex + 1] = pab
this.yuv_cache[cachey + 0][cachex + 2] = abc
this.yuv_cache[cachey + 0][cachex + 3] = bcd
this.yuv_cache[cachey + 1][cachex + 0] = qp
this.yuv_cache[cachey + 1][cachex + 1] = qpa
this.yuv_cache[cachey + 1][cachex + 2] = pa
this.yuv_cache[cachey + 1][cachex + 3] = pab
this.yuv_cache[cachey + 2][cachex + 0] = rq
this.yuv_cache[cachey + 2][cachex + 1] = rqp
this.yuv_cache[cachey + 2][cachex + 2] = qp
this.yuv_cache[cachey + 2][cachex + 3] = qpa
this.yuv_cache[cachey + 3][cachex + 0] = sr
this.yuv_cache[cachey + 3][cachex + 1] = srq
this.yuv_cache[cachey + 3][cachex + 2] = rq
this.yuv_cache[cachey + 3][cachex + 3] = rqp
} else { // HU.
s = this.yuv_cache[cachey + 3][cachex - 1] as base.u32
r = this.yuv_cache[cachey + 2][cachex - 1] as base.u32
q = this.yuv_cache[cachey + 1][cachex - 1] as base.u32
p = this.yuv_cache[cachey + 0][cachex - 1] as base.u32
sr = (((s + r + 1) / 2) & 0xFF) as base.u8
rq = (((r + q + 1) / 2) & 0xFF) as base.u8
qp = (((q + p + 1) / 2) & 0xFF) as base.u8
rqp = (((r + (2 * q) + p + 2) / 4) & 0xFF) as base.u8
srq = (((s + (2 * r) + q + 2) / 4) & 0xFF) as base.u8
ssr = (((s + (2 * s) + r + 2) / 4) & 0xFF) as base.u8
sss = (s & 0xFF) as base.u8
this.yuv_cache[cachey + 0][cachex + 0] = qp
this.yuv_cache[cachey + 0][cachex + 1] = rqp
this.yuv_cache[cachey + 0][cachex + 2] = rq
this.yuv_cache[cachey + 0][cachex + 3] = srq
this.yuv_cache[cachey + 1][cachex + 0] = rq
this.yuv_cache[cachey + 1][cachex + 1] = srq
this.yuv_cache[cachey + 1][cachex + 2] = sr
this.yuv_cache[cachey + 1][cachex + 3] = ssr
this.yuv_cache[cachey + 2][cachex + 0] = sr
this.yuv_cache[cachey + 2][cachex + 1] = ssr
this.yuv_cache[cachey + 2][cachex + 2] = sss
this.yuv_cache[cachey + 2][cachex + 3] = sss
this.yuv_cache[cachey + 3][cachex + 0] = sss
this.yuv_cache[cachey + 3][cachex + 1] = sss
this.yuv_cache[cachey + 3][cachex + 2] = sss
this.yuv_cache[cachey + 3][cachex + 3] = sss
}
}
pri func decoder.predict_y16!(mode: base.u32[..= 12]) {
var x : base.u32
var y : base.u32
var z : base.u32
var avg : base.u8
var val : base.u8
var tm0 : base.u32
var tm1 : base.u32
var tm2 : base.u32
if args.mode == 0 { // DC.
avg = ((16 +
(this.yuv_cache[0x00][0x08] as base.u32) +
(this.yuv_cache[0x00][0x09] as base.u32) +
(this.yuv_cache[0x00][0x0A] as base.u32) +
(this.yuv_cache[0x00][0x0B] as base.u32) +
(this.yuv_cache[0x00][0x0C] as base.u32) +
(this.yuv_cache[0x00][0x0D] as base.u32) +
(this.yuv_cache[0x00][0x0E] as base.u32) +
(this.yuv_cache[0x00][0x0F] as base.u32) +
(this.yuv_cache[0x00][0x10] as base.u32) +
(this.yuv_cache[0x00][0x11] as base.u32) +
(this.yuv_cache[0x00][0x12] as base.u32) +
(this.yuv_cache[0x00][0x13] as base.u32) +
(this.yuv_cache[0x00][0x14] as base.u32) +
(this.yuv_cache[0x00][0x15] as base.u32) +
(this.yuv_cache[0x00][0x16] as base.u32) +
(this.yuv_cache[0x00][0x17] as base.u32) +
(this.yuv_cache[0x01][0x07] as base.u32) +
(this.yuv_cache[0x02][0x07] as base.u32) +
(this.yuv_cache[0x03][0x07] as base.u32) +
(this.yuv_cache[0x04][0x07] as base.u32) +
(this.yuv_cache[0x05][0x07] as base.u32) +
(this.yuv_cache[0x06][0x07] as base.u32) +
(this.yuv_cache[0x07][0x07] as base.u32) +
(this.yuv_cache[0x08][0x07] as base.u32) +
(this.yuv_cache[0x09][0x07] as base.u32) +
(this.yuv_cache[0x0A][0x07] as base.u32) +
(this.yuv_cache[0x0B][0x07] as base.u32) +
(this.yuv_cache[0x0C][0x07] as base.u32) +
(this.yuv_cache[0x0D][0x07] as base.u32) +
(this.yuv_cache[0x0E][0x07] as base.u32) +
(this.yuv_cache[0x0F][0x07] as base.u32) +
(this.yuv_cache[0x10][0x07] as base.u32)) / 32) as base.u8
z = 0
while z < 256 {
this.yuv_cache[0x01 + (z >> 4)][0x08 + (z & 15)] = avg
z += 1
}
} else if args.mode == 1 { // TM.
tm0 = this.yuv_cache[0x00][0x07] as base.u32
z = 0
while z < 256 {
tm1 = this.yuv_cache[0x01 + (z >> 4)][0x07] as base.u32
tm2 = this.yuv_cache[0x00][0x08 + (z & 15)] as base.u32
this.yuv_cache[0x01 + (z >> 4)][0x08 + (z & 15)] =
CLAMP[((tm1 + tm2) ~mod- tm0) & 1023]
z += 1
}
} else if args.mode == 2 { // VE.
x = 0
while x < 16 {
val = this.yuv_cache[0x00][0x08 + x]
y = 0
while y < 16,
inv x < 16,
{
this.yuv_cache[0x01 + y][0x08 + x] = val
y += 1
}
x += 1
}
} else if args.mode == 3 { // HE.
y = 0
while y < 16 {
val = this.yuv_cache[0x01 + y][0x07]
x = 0
while x < 16,
inv y < 16,
{
this.yuv_cache[0x01 + y][0x08 + x] = val
x += 1
}
y += 1
}
} else if args.mode <= 10 { // DCTop.
avg = ((8 +
(this.yuv_cache[0x01][0x07] as base.u32) +
(this.yuv_cache[0x02][0x07] as base.u32) +
(this.yuv_cache[0x03][0x07] as base.u32) +
(this.yuv_cache[0x04][0x07] as base.u32) +
(this.yuv_cache[0x05][0x07] as base.u32) +
(this.yuv_cache[0x06][0x07] as base.u32) +
(this.yuv_cache[0x07][0x07] as base.u32) +
(this.yuv_cache[0x08][0x07] as base.u32) +
(this.yuv_cache[0x09][0x07] as base.u32) +
(this.yuv_cache[0x0A][0x07] as base.u32) +
(this.yuv_cache[0x0B][0x07] as base.u32) +
(this.yuv_cache[0x0C][0x07] as base.u32) +
(this.yuv_cache[0x0D][0x07] as base.u32) +
(this.yuv_cache[0x0E][0x07] as base.u32) +
(this.yuv_cache[0x0F][0x07] as base.u32) +
(this.yuv_cache[0x10][0x07] as base.u32)) / 16) as base.u8
z = 0
while z < 256 {
this.yuv_cache[0x01 + (z >> 4)][0x08 + (z & 15)] = avg
z += 1
}
} else if args.mode == 11 { // DCLeft.
avg = ((8 +
(this.yuv_cache[0x00][0x08] as base.u32) +
(this.yuv_cache[0x00][0x09] as base.u32) +
(this.yuv_cache[0x00][0x0A] as base.u32) +
(this.yuv_cache[0x00][0x0B] as base.u32) +
(this.yuv_cache[0x00][0x0C] as base.u32) +
(this.yuv_cache[0x00][0x0D] as base.u32) +
(this.yuv_cache[0x00][0x0E] as base.u32) +
(this.yuv_cache[0x00][0x0F] as base.u32) +
(this.yuv_cache[0x00][0x10] as base.u32) +
(this.yuv_cache[0x00][0x11] as base.u32) +
(this.yuv_cache[0x00][0x12] as base.u32) +
(this.yuv_cache[0x00][0x13] as base.u32) +
(this.yuv_cache[0x00][0x14] as base.u32) +
(this.yuv_cache[0x00][0x15] as base.u32) +
(this.yuv_cache[0x00][0x16] as base.u32) +
(this.yuv_cache[0x00][0x17] as base.u32)) / 16) as base.u8
z = 0
while z < 256 {
this.yuv_cache[0x01 + (z >> 4)][0x08 + (z & 15)] = avg
z += 1
}
} else { // DCTopLeft.
z = 0
while z < 256 {
this.yuv_cache[0x01 + (z >> 4)][0x08 + (z & 15)] = 0x80
z += 1
}
}
}
pri func decoder.predict_uv8!(b: base.u32[..= 1], mode: base.u32[..= 12]) {
var cachey : base.u32[..= 18]
var cachex : base.u32[..= 24]
var x : base.u32
var y : base.u32
var z : base.u32
var avg : base.u8
var val : base.u8
var tm0 : base.u32
var tm1 : base.u32
var tm2 : base.u32
cachey = 0x12
cachex = (args.b * 16) + 8
if args.mode == 0 { // DC.
avg = ((8 +
(this.yuv_cache[cachey - 1][cachex + 0] as base.u32) +
(this.yuv_cache[cachey - 1][cachex + 1] as base.u32) +
(this.yuv_cache[cachey - 1][cachex + 2] as base.u32) +
(this.yuv_cache[cachey - 1][cachex + 3] as base.u32) +
(this.yuv_cache[cachey - 1][cachex + 4] as base.u32) +
(this.yuv_cache[cachey - 1][cachex + 5] as base.u32) +
(this.yuv_cache[cachey - 1][cachex + 6] as base.u32) +
(this.yuv_cache[cachey - 1][cachex + 7] as base.u32) +
(this.yuv_cache[cachey + 0][cachex - 1] as base.u32) +
(this.yuv_cache[cachey + 1][cachex - 1] as base.u32) +
(this.yuv_cache[cachey + 2][cachex - 1] as base.u32) +
(this.yuv_cache[cachey + 3][cachex - 1] as base.u32) +
(this.yuv_cache[cachey + 4][cachex - 1] as base.u32) +
(this.yuv_cache[cachey + 5][cachex - 1] as base.u32) +
(this.yuv_cache[cachey + 6][cachex - 1] as base.u32) +
(this.yuv_cache[cachey + 7][cachex - 1] as base.u32)) / 16) as base.u8
z = 0
while z < 64 {
this.yuv_cache[cachey + (z >> 3)][cachex + (z & 7)] = avg
z += 1
}
} else if args.mode == 1 { // TM.
tm0 = this.yuv_cache[cachey - 1][cachex - 1] as base.u32
z = 0
while z < 64,
inv cachey == 0x12,
inv cachex >= 8,
{
tm1 = this.yuv_cache[cachey + (z >> 3)][cachex - 1] as base.u32
tm2 = this.yuv_cache[cachey - 1][cachex + (z & 7)] as base.u32
this.yuv_cache[cachey + (z >> 3)][cachex + (z & 7)] =
CLAMP[((tm1 + tm2) ~mod- tm0) & 1023]
z += 1
}
} else if args.mode == 2 { // VE.
x = 0
while x < 8,
inv cachey == 0x12,
inv cachex >= 8,
{
val = this.yuv_cache[cachey - 1][cachex + x]
y = 0
while y < 8,
inv cachey == 0x12,
inv cachex >= 8,
inv x < 8,
{
this.yuv_cache[cachey + y][cachex + x] = val
y += 1
}
x += 1
}
} else if args.mode == 3 { // HE.
y = 0
while y < 8,
inv cachey == 0x12,
inv cachex >= 8,
{
val = this.yuv_cache[cachey + y][cachex - 1]
x = 0
while x < 8,
inv cachey == 0x12,
inv cachex >= 8,
inv y < 8,
{
this.yuv_cache[cachey + y][cachex + x] = val
x += 1
}
y += 1
}
} else if args.mode <= 10 { // DCTop.
avg = ((4 +
(this.yuv_cache[cachey + 0][cachex - 1] as base.u32) +
(this.yuv_cache[cachey + 1][cachex - 1] as base.u32) +
(this.yuv_cache[cachey + 2][cachex - 1] as base.u32) +
(this.yuv_cache[cachey + 3][cachex - 1] as base.u32) +
(this.yuv_cache[cachey + 4][cachex - 1] as base.u32) +
(this.yuv_cache[cachey + 5][cachex - 1] as base.u32) +
(this.yuv_cache[cachey + 6][cachex - 1] as base.u32) +
(this.yuv_cache[cachey + 7][cachex - 1] as base.u32)) / 8) as base.u8
z = 0
while z < 64 {
this.yuv_cache[cachey + (z >> 3)][cachex + (z & 7)] = avg
z += 1
}
} else if args.mode == 11 { // DCLeft.
avg = ((4 +
(this.yuv_cache[cachey - 1][cachex + 0] as base.u32) +
(this.yuv_cache[cachey - 1][cachex + 1] as base.u32) +
(this.yuv_cache[cachey - 1][cachex + 2] as base.u32) +
(this.yuv_cache[cachey - 1][cachex + 3] as base.u32) +
(this.yuv_cache[cachey - 1][cachex + 4] as base.u32) +
(this.yuv_cache[cachey - 1][cachex + 5] as base.u32) +
(this.yuv_cache[cachey - 1][cachex + 6] as base.u32) +
(this.yuv_cache[cachey - 1][cachex + 7] as base.u32)) / 8) as base.u8
z = 0
while z < 64 {
this.yuv_cache[cachey + (z >> 3)][cachex + (z & 7)] = avg
z += 1
}
} else { // DCTopLeft.
z = 0
while z < 64 {
this.yuv_cache[cachey + (z >> 3)][cachex + (z & 7)] = 0x80
z += 1
}
}
}
+176
View File
@@ -0,0 +1,176 @@
// Copyright 2026 The Wuffs Authors.
//
// Licensed under the Apache License, Version 2.0 <LICENSE-APACHE or
// https://www.apache.org/licenses/LICENSE-2.0> or the MIT license
// <LICENSE-MIT or https://opensource.org/licenses/MIT>, at your
// option. This file may not be copied, modified, or distributed
// except according to those terms.
//
// SPDX-License-Identifier: Apache-2.0 OR MIT
pri func decoder.reconstruct!(
workbuf: slice base.u8,
mbx: base.u32[..= 0x3FF],
mby: base.u32[..= 0x3FF],
luma_mode: base.u32[..= 4],
chroma_mode: base.u32[..= 3]) {
var mode : base.u32[..= 12]
var b : base.u32
this.prepare_yuv_cache!(workbuf: args.workbuf, mbx: args.mbx, mby: args.mby)
if args.luma_mode < 4 {
mode = this.substitute_dc_top_left(mode: args.luma_mode, mbx: args.mbx, mby: args.mby)
this.predict_y16!(mode: mode)
} else {
b = 0
while b < 16 {
this.predict_y4!(b: b)
b += 1
}
}
mode = this.substitute_dc_top_left(mode: args.chroma_mode, mbx: args.mbx, mby: args.mby)
this.predict_uv8!(b: 0, mode: mode)
this.predict_uv8!(b: 1, mode: mode)
}
pri func decoder.substitute_dc_top_left(
mode: base.u32[..= 3],
mbx: base.u32[..= 0x3FF],
mby: base.u32[..= 0x3FF]) base.u32[..= 12] {
if args.mode == 0 { // DC.
if args.mby == 0 {
if args.mbx == 0 {
return 11 // DCLeft.
} else {
return 12 // DCTopLeft.
}
} else if args.mby == 0 {
return 10 // DCTop.
}
}
return args.mode
}
pri func decoder.prepare_yuv_cache!(
workbuf: roslice base.u8,
mbx: base.u32[..= 0x3FF],
mby: base.u32[..= 0x3FF]) {
var i : base.u32
var j : base.u32
var offset : base.u64[..= 0x17FE_1FF8]
var o : base.u64
if args.mbx <= 0 {
j = 0x00
while j < 0x11 {
this.yuv_cache[j][0x07] = 0x81
j += 1
}
while j < 0x1A {
this.yuv_cache[j][0x07] = 0x81
this.yuv_cache[j][0x17] = 0x81
j += 1
}
} else {
j = 0x00
while j < 0x11 {
this.yuv_cache[j][0x07] = this.yuv_cache[j][0x17]
j += 1
}
while j < 0x1A {
this.yuv_cache[j][0x07] = this.yuv_cache[j][0x0F]
this.yuv_cache[j][0x17] = this.yuv_cache[j][0x1F]
j += 1
}
}
if args.mby <= 0 {
i = 0x00
while i < 0x20 {
this.yuv_cache[0x00][i] = 0x7F
i += 1
}
i = 0x00
while i < 0x20 {
this.yuv_cache[0x11][i] = 0x7F
i += 1
}
} else {
offset = 0 +
(((((args.mby - 1) * this.workbuf_yuv_y_stride) + args.mbx) * 16) as base.u64)
i = 0x00
while i < 0x10,
inv args.mby > 0,
{
o = offset + (i as base.u64)
if o < args.workbuf.length() {
this.yuv_cache[0x00][0x08 + i] = args.workbuf[o]
}
i += 1
}
if (args.mbx + 1) < this.mbw {
while i < 0x14,
inv args.mby > 0,
{
o = offset + (i as base.u64)
if o < args.workbuf.length() {
this.yuv_cache[0x00][0x08 + i] = args.workbuf[o]
}
i += 1
}
} else {
while i < 0x14,
inv args.mby > 0,
{
o = offset + (0x0F as base.u64)
if o < args.workbuf.length() {
this.yuv_cache[0x00][0x08 + i] = args.workbuf[o]
}
i += 1
}
}
offset = this.workbuf_yuv_y_end +
(((((args.mby - 1) * this.workbuf_yuv_uv_stride) + args.mbx) * 8) as base.u64)
i = 0x00
while i < 0x08,
inv args.mby > 0,
{
o = offset + (i as base.u64)
if o < args.workbuf.length() {
this.yuv_cache[0x11][0x08 + i] = args.workbuf[o]
}
i += 1
}
offset = this.workbuf_yuv_u_end +
(((((args.mby - 1) * this.workbuf_yuv_uv_stride) + args.mbx) * 8) as base.u64)
i = 0x00
while i < 0x08,
inv args.mby > 0,
{
o = offset + (i as base.u64)
if o < args.workbuf.length() {
this.yuv_cache[0x11][0x18 + i] = args.workbuf[o]
}
i += 1
}
}
j = 0x04
while j < 0x10 {
this.yuv_cache[j][0x18] = this.yuv_cache[0][0x18]
this.yuv_cache[j][0x19] = this.yuv_cache[0][0x19]
this.yuv_cache[j][0x1A] = this.yuv_cache[0][0x1A]
this.yuv_cache[j][0x1B] = this.yuv_cache[0][0x1B]
j += 4
}
}
+2
View File
@@ -100,6 +100,8 @@ pub struct decoder? implements base.image_decoder(
mb_subblock_modes : array[16] base.u8[..= 9],
yuv_cache : array[26] array[32] base.u8,
dst_x : base.u32,
dst_y : base.u32,