summaryrefslogtreecommitdiffstats
path: root/media/ffvpx/libavcodec/simple_idct.c
diff options
context:
space:
mode:
authorDaniel Baumann <daniel.baumann@progress-linux.org>2024-04-07 19:33:14 +0000
committerDaniel Baumann <daniel.baumann@progress-linux.org>2024-04-07 19:33:14 +0000
commit36d22d82aa202bb199967e9512281e9a53db42c9 (patch)
tree105e8c98ddea1c1e4784a60a5a6410fa416be2de /media/ffvpx/libavcodec/simple_idct.c
parentInitial commit. (diff)
downloadfirefox-esr-36d22d82aa202bb199967e9512281e9a53db42c9.tar.xz
firefox-esr-36d22d82aa202bb199967e9512281e9a53db42c9.zip
Adding upstream version 115.7.0esr.upstream/115.7.0esrupstream
Signed-off-by: Daniel Baumann <daniel.baumann@progress-linux.org>
Diffstat (limited to '')
-rw-r--r--media/ffvpx/libavcodec/simple_idct.c269
1 files changed, 269 insertions, 0 deletions
diff --git a/media/ffvpx/libavcodec/simple_idct.c b/media/ffvpx/libavcodec/simple_idct.c
new file mode 100644
index 0000000000..eb13cff146
--- /dev/null
+++ b/media/ffvpx/libavcodec/simple_idct.c
@@ -0,0 +1,269 @@
+/*
+ * Simple IDCT
+ *
+ * Copyright (c) 2001 Michael Niedermayer <michaelni@gmx.at>
+ *
+ * This file is part of FFmpeg.
+ *
+ * FFmpeg is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU Lesser General Public
+ * License as published by the Free Software Foundation; either
+ * version 2.1 of the License, or (at your option) any later version.
+ *
+ * FFmpeg is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ * Lesser General Public License for more details.
+ *
+ * You should have received a copy of the GNU Lesser General Public
+ * License along with FFmpeg; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+ */
+
+/**
+ * @file
+ * simpleidct in C.
+ */
+
+#include "libavutil/intreadwrite.h"
+#include "mathops.h"
+#include "simple_idct.h"
+
+#define IN_IDCT_DEPTH 16
+
+#define BIT_DEPTH 8
+#include "simple_idct_template.c"
+#undef BIT_DEPTH
+
+#define BIT_DEPTH 10
+#include "simple_idct_template.c"
+
+#define EXTRA_SHIFT 2
+#include "simple_idct_template.c"
+
+#undef EXTRA_SHIFT
+#undef BIT_DEPTH
+
+#define BIT_DEPTH 12
+#include "simple_idct_template.c"
+#undef BIT_DEPTH
+#undef IN_IDCT_DEPTH
+
+#define IN_IDCT_DEPTH 32
+#define BIT_DEPTH 10
+#include "simple_idct_template.c"
+#undef BIT_DEPTH
+#undef IN_IDCT_DEPTH
+
+/* 2x4x8 idct */
+
+#define CN_SHIFT 12
+#define C_FIX(x) ((int)((x) * (1 << CN_SHIFT) + 0.5))
+#define C1 C_FIX(0.6532814824)
+#define C2 C_FIX(0.2705980501)
+
+/* row idct is multiple by 16 * sqrt(2.0), col idct4 is normalized,
+ and the butterfly must be multiplied by 0.5 * sqrt(2.0) */
+#define C_SHIFT (4+1+12)
+
+static inline void idct4col_put(uint8_t *dest, ptrdiff_t line_size, const int16_t *col)
+{
+ int c0, c1, c2, c3, a0, a1, a2, a3;
+
+ a0 = col[8*0];
+ a1 = col[8*2];
+ a2 = col[8*4];
+ a3 = col[8*6];
+ c0 = ((a0 + a2) * (1 << CN_SHIFT - 1)) + (1 << (C_SHIFT - 1));
+ c2 = ((a0 - a2) * (1 << CN_SHIFT - 1)) + (1 << (C_SHIFT - 1));
+ c1 = a1 * C1 + a3 * C2;
+ c3 = a1 * C2 - a3 * C1;
+ dest[0] = av_clip_uint8((c0 + c1) >> C_SHIFT);
+ dest += line_size;
+ dest[0] = av_clip_uint8((c2 + c3) >> C_SHIFT);
+ dest += line_size;
+ dest[0] = av_clip_uint8((c2 - c3) >> C_SHIFT);
+ dest += line_size;
+ dest[0] = av_clip_uint8((c0 - c1) >> C_SHIFT);
+}
+
+#define BF(k) \
+{\
+ int a0, a1;\
+ a0 = ptr[k];\
+ a1 = ptr[8 + k];\
+ ptr[k] = a0 + a1;\
+ ptr[8 + k] = a0 - a1;\
+}
+
+/* only used by DV codec. The input must be interlaced. 128 is added
+ to the pixels before clamping to avoid systematic error
+ (1024*sqrt(2)) offset would be needed otherwise. */
+/* XXX: I think a 1.0/sqrt(2) normalization should be needed to
+ compensate the extra butterfly stage - I don't have the full DV
+ specification */
+void ff_simple_idct248_put(uint8_t *dest, ptrdiff_t line_size, int16_t *block)
+{
+ int i;
+ int16_t *ptr;
+
+ /* butterfly */
+ ptr = block;
+ for(i=0;i<4;i++) {
+ BF(0);
+ BF(1);
+ BF(2);
+ BF(3);
+ BF(4);
+ BF(5);
+ BF(6);
+ BF(7);
+ ptr += 2 * 8;
+ }
+
+ /* IDCT8 on each line */
+ for(i=0; i<8; i++) {
+ idctRowCondDC_int16_8bit(block + i*8, 0);
+ }
+
+ /* IDCT4 and store */
+ for(i=0;i<8;i++) {
+ idct4col_put(dest + i, 2 * line_size, block + i);
+ idct4col_put(dest + line_size + i, 2 * line_size, block + 8 + i);
+ }
+}
+
+/* 8x4 & 4x8 WMV2 IDCT */
+#undef CN_SHIFT
+#undef C_SHIFT
+#undef C_FIX
+#undef C1
+#undef C2
+#define CN_SHIFT 12
+#define C_FIX(x) ((int)((x) * M_SQRT2 * (1 << CN_SHIFT) + 0.5))
+#define C1 C_FIX(0.6532814824)
+#define C2 C_FIX(0.2705980501)
+#define C3 C_FIX(0.5)
+#define C_SHIFT (4+1+12)
+static inline void idct4col_add(uint8_t *dest, ptrdiff_t line_size, const int16_t *col)
+{
+ int c0, c1, c2, c3, a0, a1, a2, a3;
+
+ a0 = col[8*0];
+ a1 = col[8*1];
+ a2 = col[8*2];
+ a3 = col[8*3];
+ c0 = (a0 + a2)*C3 + (1 << (C_SHIFT - 1));
+ c2 = (a0 - a2)*C3 + (1 << (C_SHIFT - 1));
+ c1 = a1 * C1 + a3 * C2;
+ c3 = a1 * C2 - a3 * C1;
+ dest[0] = av_clip_uint8(dest[0] + ((c0 + c1) >> C_SHIFT));
+ dest += line_size;
+ dest[0] = av_clip_uint8(dest[0] + ((c2 + c3) >> C_SHIFT));
+ dest += line_size;
+ dest[0] = av_clip_uint8(dest[0] + ((c2 - c3) >> C_SHIFT));
+ dest += line_size;
+ dest[0] = av_clip_uint8(dest[0] + ((c0 - c1) >> C_SHIFT));
+}
+
+#define RN_SHIFT 15
+#define R_FIX(x) ((int)((x) * M_SQRT2 * (1 << RN_SHIFT) + 0.5))
+#define R1 R_FIX(0.6532814824)
+#define R2 R_FIX(0.2705980501)
+#define R3 R_FIX(0.5)
+#define R_SHIFT 11
+static inline void idct4row(int16_t *row)
+{
+ unsigned c0, c1, c2, c3;
+ int a0, a1, a2, a3;
+
+ a0 = row[0];
+ a1 = row[1];
+ a2 = row[2];
+ a3 = row[3];
+ c0 = (a0 + a2)*R3 + (1 << (R_SHIFT - 1));
+ c2 = (a0 - a2)*R3 + (1 << (R_SHIFT - 1));
+ c1 = a1 * R1 + a3 * R2;
+ c3 = a1 * R2 - a3 * R1;
+ row[0]= (c0 + c1) >> R_SHIFT;
+ row[1]= (c2 + c3) >> R_SHIFT;
+ row[2]= (c2 - c3) >> R_SHIFT;
+ row[3]= (c0 - c1) >> R_SHIFT;
+}
+
+void ff_simple_idct84_add(uint8_t *dest, ptrdiff_t line_size, int16_t *block)
+{
+ int i;
+
+ /* IDCT8 on each line */
+ for(i=0; i<4; i++) {
+ idctRowCondDC_int16_8bit(block + i*8, 0);
+ }
+
+ /* IDCT4 and store */
+ for(i=0;i<8;i++) {
+ idct4col_add(dest + i, line_size, block + i);
+ }
+}
+
+void ff_simple_idct48_add(uint8_t *dest, ptrdiff_t line_size, int16_t *block)
+{
+ int i;
+
+ /* IDCT4 on each line */
+ for(i=0; i<8; i++) {
+ idct4row(block + i*8);
+ }
+
+ /* IDCT8 and store */
+ for(i=0; i<4; i++){
+ idctSparseColAdd_int16_8bit(dest + i, line_size, block + i);
+ }
+}
+
+void ff_simple_idct44_add(uint8_t *dest, ptrdiff_t line_size, int16_t *block)
+{
+ int i;
+
+ /* IDCT4 on each line */
+ for(i=0; i<4; i++) {
+ idct4row(block + i*8);
+ }
+
+ /* IDCT4 and store */
+ for(i=0; i<4; i++){
+ idct4col_add(dest + i, line_size, block + i);
+ }
+}
+
+void ff_prores_idct_10(int16_t *block, const int16_t *qmat)
+{
+ int i;
+
+ for (i = 0; i < 64; i++)
+ block[i] *= qmat[i];
+
+ for (i = 0; i < 8; i++)
+ idctRowCondDC_extrashift_10(block + i*8, 2);
+
+ for (i = 0; i < 8; i++) {
+ block[i] += 8192;
+ idctSparseCol_extrashift_10(block + i);
+ }
+}
+
+void ff_prores_idct_12(int16_t *block, const int16_t *qmat)
+{
+ int i;
+
+ for (i = 0; i < 64; i++)
+ block[i] *= qmat[i];
+
+ for (i = 0; i < 8; i++)
+ idctRowCondDC_int16_12bit(block + i*8, 0);
+
+ for (i = 0; i < 8; i++) {
+ block[i] += 8192;
+ idctSparseCol_int16_12bit(block + i);
+ }
+}