Update for 4.2.9

Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

CREDITS

@@ -1,6 +1,6 @@
-See the Git history of the project (git://source.ffmpeg.org/ffmpeg) to
+See the Git history of the project (https://git.ffmpeg.org/ffmpeg) to
 get the names of people who have contributed to FFmpeg.
 
 To check the log, you can type the command "git log" in the FFmpeg
 source directory, or browse the online repository at
-http://source.ffmpeg.org.
+https://git.ffmpeg.org/ffmpeg

Changelog

@@ -1,6 +1,56 @@
 Entries are sorted chronologically from oldest to youngest within each release,
 releases are sorted from youngest to oldest.
 
+version 4.2.9
+ avcodec/escape124: Check that blocks are allocated before use
+ avcodec/huffyuvdec: Fix undefined behavior with shift
+ avcodec/j2kenc: Replace RGB24 special case by generic test
+ avcodec/j2kenc: remove misleading pred value
+ avcodec/j2kenc: fix 5/3 DWT identifer
+ avcodec/vp3: Check width to avoid assertion failure
+ avcodec/g729postfilter: Limit shift in long term filter
+ configure: update copyright year
+ avcodec/tests/snowenc: Fix 2nd test
+ avcodec/tests/snowenc: return a failure if DWT/IDWT mismatches
+ avcodec/snowenc: Fix visual weight calculation
+ avcodec/tests/snowenc: unbreak DWT tests
+ avcodec/vp3: Add missing check for av_malloc
+ avcodec/escape124: Fix some return codes
+ avcodec/escape124: fix signdness of end of input check
+ Use https for repository links
+ avcodec/motionpixels: Mask pixels to valid values
+ avcodec/xpmdec: Check size before allocation to avoid truncation
+ avcodec/bink: Avoid undefined out of array end pointers in binkb_decode_plane()
+ avcodec/bink: Fix off by 1 error in ref end
+ avcodec/utils: Ensure linesize for SVQ3
+ avcodec/utils: allocate a line more for VC1 and WMV3
+ avcodec/videodsp_template: Adjust pointers to avoid undefined pointer things
+ avcodec/pngdec: Check deloco index more exactly
+ avcodec/ffv1dec: Check that num h/v slices is supported
+ avformat/mov: Check samplesize and offset to avoid integer overflow
+ avcodec/pictordec: Remove mid exit branch
+ avcodec/eac3dec: avoid float noise in fixed mode addition to overflow
+ avcodec/utils: use 32pixel alignment for bink
+ avcodec/scpr3: Check bx
+ avcodec/012v: Order operations for odd size handling
+ avcodec/eatgq: : Check index increments in tgq_decode_block()
+ avcodec/scpr: Test bx before use
+ avcodec/sunrast: Fix maplength check
+ avcodec/wavpack: Avoid undefined shift in get_tail()
+ avformat/id3v2: Check taglen in read_uslt()
+ avcodec/ffv1dec: restructure slice coordinate reading a bit
+ avcodec/mlpdec: Check max matrix instead of max channel in noise check
+ swscale/input: Use more unsigned intermediates
+ avcodec/alsdec: The minimal block is at least 7 bits
+ avformat/replaygain: avoid undefined / negative abs
+ swscale/output: Bias 16bps output calculations to improve non overflowing range
+ avcodec/speedhq: Check buf_size to be big enough for DC
+ avcodec/ffv1dec: Fail earlier if prior context is corrupted
+ avcodec/nvenc: fix vbv buffer size in cq mode
+ avcodec/mjpegenc: take into account component count when writing the SOF header size
+ checkasm: float_dsp: Scale FLT/DBL_EPSILON sufficiently when comparing
+ swscale: aarch64: Fix yuv2rgb with negative strides
+
 version 4.2.8
  avformat/vividas: Check packet size
  avcodec/dstdec: Check for overflow in build_filter()

RELEASE

@@ -1 +1 @@
-4.2.8
+4.2.9

configure

@@ -7397,7 +7397,7 @@ cat > $TMPH <<EOF
 #define FFMPEG_CONFIG_H
 #define FFMPEG_CONFIGURATION "$(c_escape $FFMPEG_CONFIGURATION)"
 #define FFMPEG_LICENSE "$(c_escape $license)"
-#define CONFIG_THIS_YEAR 2022
+#define CONFIG_THIS_YEAR 2023
 #define FFMPEG_DATADIR "$(eval c_escape $datadir)"
 #define AVCONV_DATADIR "$(eval c_escape $datadir)"
 #define CC_IDENT "$(c_escape ${cc_ident:-Unknown compiler})"

doc/Doxyfile

@@ -38,7 +38,7 @@ PROJECT_NAME           = FFmpeg
 # could be handy for archiving the generated documentation or if some version
 # control system is used.
 
-PROJECT_NUMBER         = 4.2.8
+PROJECT_NUMBER         = 4.2.9
 
 # Using the PROJECT_BRIEF tag one can provide an optional one line description
 # for a project that appears at the top of each page and should give viewer a

doc/authors.texi

@@ -3,9 +3,9 @@
 The FFmpeg developers.
 
 For details about the authorship, see the Git history of the project
-(git://source.ffmpeg.org/ffmpeg), e.g. by typing the command
+(https://git.ffmpeg.org/ffmpeg), e.g. by typing the command
 @command{git log} in the FFmpeg source directory, or browsing the
-online repository at @url{http://source.ffmpeg.org}.
+online repository at @url{https://git.ffmpeg.org/ffmpeg}.
 
 Maintainers for the specific components are listed in the file
 @file{MAINTAINERS} in the source code tree.

doc/git-howto.texi

@@ -53,7 +53,7 @@ Most distribution and operating system provide a package for it.
 @section Cloning the source tree
 
 @example
-git clone git://source.ffmpeg.org/ffmpeg <target>
+git clone https://git.ffmpeg.org/ffmpeg.git <target>
 @end example
 
 This will put the FFmpeg sources into the directory @var{<target>}.

libavcodec/012v.c

@@ -131,8 +131,8 @@ static int zero12v_decode_frame(AVCodecContext *avctx, void *data,
             u = x/2 + (uint16_t *)(pic->data[1] + line * pic->linesize[1]);
             v = x/2 + (uint16_t *)(pic->data[2] + line * pic->linesize[2]);
             memcpy(y, y_temp, sizeof(*y) * (width - x));
-            memcpy(u, u_temp, sizeof(*u) * (width - x + 1) / 2);
-            memcpy(v, v_temp, sizeof(*v) * (width - x + 1) / 2);
+            memcpy(u, u_temp, sizeof(*u) * ((width - x + 1) / 2));
+            memcpy(v, v_temp, sizeof(*v) * ((width - x + 1) / 2));
         }
 
         line_end += stride;

libavcodec/ac3.h

@@ -75,6 +75,7 @@
 #define AC3_DYNAMIC_RANGE1      0
 
 typedef int                     INTFLOAT;
+typedef unsigned int            UINTFLOAT;
 typedef int16_t                 SHORTFLOAT;
 
 #else /* USE_FIXED */
@@ -94,6 +95,7 @@ typedef int16_t                 SHORTFLOAT;
 #define AC3_DYNAMIC_RANGE1      1.0f
 
 typedef float                   INTFLOAT;
+typedef float                   UINTFLOAT;
 typedef float                   SHORTFLOAT;
 
 #endif /* USE_FIXED */

libavcodec/alsdec.c

@@ -1016,7 +1016,7 @@ static int read_block(ALSDecContext *ctx, ALSBlockData *bd)
 
     *bd->shift_lsbs = 0;
 
-    if (get_bits_left(gb) < 1)
+    if (get_bits_left(gb) < 7)
         return AVERROR_INVALIDDATA;
 
     // read block type flag and read the samples accordingly

libavcodec/bink.c

@@ -838,7 +838,7 @@ static int binkb_decode_plane(BinkContext *c, AVFrame *frame, GetBitContext *gb,
 
     binkb_init_bundles(c);
     ref_start = frame->data[plane_idx];
-    ref_end   = frame->data[plane_idx] + (bh * frame->linesize[plane_idx] + bw) * 8;
+    ref_end   = frame->data[plane_idx] + ((bh - 1) * frame->linesize[plane_idx] + bw - 1) * 8;
 
     for (i = 0; i < 64; i++)
         coordmap[i] = (i & 7) + (i >> 3) * stride;
@@ -894,7 +894,7 @@ static int binkb_decode_plane(BinkContext *c, AVFrame *frame, GetBitContext *gb,
                 xoff = binkb_get_value(c, BINKB_SRC_X_OFF);
                 yoff = binkb_get_value(c, BINKB_SRC_Y_OFF) + ybias;
                 ref = dst + xoff + yoff * stride;
-                if (ref < ref_start || ref + 8*stride > ref_end) {
+                if (ref < ref_start || ref > ref_end) {
                     av_log(c->avctx, AV_LOG_WARNING, "Reference block is out of bounds\n");
                 } else if (ref + 8*stride < dst || ref >= dst + 8*stride) {
                     c->hdsp.put_pixels_tab[1][0](dst, ref, stride, 8);
@@ -910,7 +910,7 @@ static int binkb_decode_plane(BinkContext *c, AVFrame *frame, GetBitContext *gb,
                 xoff = binkb_get_value(c, BINKB_SRC_X_OFF);
                 yoff = binkb_get_value(c, BINKB_SRC_Y_OFF) + ybias;
                 ref = dst + xoff + yoff * stride;
-                if (ref < ref_start || ref + 8 * stride > ref_end) {
+                if (ref < ref_start || ref > ref_end) {
                     av_log(c->avctx, AV_LOG_WARNING, "Reference block is out of bounds\n");
                 } else if (ref + 8*stride < dst || ref >= dst + 8*stride) {
                     c->hdsp.put_pixels_tab[1][0](dst, ref, stride, 8);
@@ -942,7 +942,7 @@ static int binkb_decode_plane(BinkContext *c, AVFrame *frame, GetBitContext *gb,
                 xoff = binkb_get_value(c, BINKB_SRC_X_OFF);
                 yoff = binkb_get_value(c, BINKB_SRC_Y_OFF) + ybias;
                 ref = dst + xoff + yoff * stride;
-                if (ref < ref_start || ref + 8 * stride > ref_end) {
+                if (ref < ref_start || ref > ref_end) {
                     av_log(c->avctx, AV_LOG_WARNING, "Reference block is out of bounds\n");
                 } else if (ref + 8*stride < dst || ref >= dst + 8*stride) {
                     c->hdsp.put_pixels_tab[1][0](dst, ref, stride, 8);

libavcodec/eac3dec.c

@@ -139,9 +139,11 @@ static void ff_eac3_apply_spectral_extension(AC3DecodeContext *s)
             // spx_noise_blend and spx_signal_blend are both FP.23
             nscale *= 1.0 / (1<<23);
             sscale *= 1.0 / (1<<23);
+            if (nscale < -1.0)
+                nscale = -1.0;
 #endif
             for (i = 0; i < s->spx_band_sizes[bnd]; i++) {
-                float noise  = nscale * (int32_t)av_lfg_get(&s->dith_state);
+                UINTFLOAT noise = (INTFLOAT)(nscale * (int32_t)av_lfg_get(&s->dith_state));
                 s->transform_coeffs[ch][bin]   *= sscale;
                 s->transform_coeffs[ch][bin++] += noise;
             }

libavcodec/eatgq.c

@@ -58,7 +58,7 @@ static av_cold int tgq_decode_init(AVCodecContext *avctx)
     return 0;
 }
 
-static void tgq_decode_block(TgqContext *s, int16_t block[64], GetBitContext *gb)
+static int tgq_decode_block(TgqContext *s, int16_t block[64], GetBitContext *gb)
 {
     uint8_t *perm = s->scantable.permutated;
     int i, j, value;
@@ -66,6 +66,8 @@ static void tgq_decode_block(TgqContext *s, int16_t block[64], GetBitContext *gb
     for (i = 1; i < 64;) {
         switch (show_bits(gb, 3)) {
         case 4:
+            if (i >= 63)
+                return AVERROR_INVALIDDATA;
             block[perm[i++]] = 0;
         case 0:
             block[perm[i++]] = 0;
@@ -75,6 +77,8 @@ static void tgq_decode_block(TgqContext *s, int16_t block[64], GetBitContext *gb
         case 1:
             skip_bits(gb, 2);
             value = get_bits(gb, 6);
+            if (value > 64 - i)
+                return AVERROR_INVALIDDATA;
             for (j = 0; j < value; j++)
                 block[perm[i++]] = 0;
             break;
@@ -102,6 +106,7 @@ static void tgq_decode_block(TgqContext *s, int16_t block[64], GetBitContext *gb
         }
     }
     block[0] += 128 << 4;
+    return 0;
 }
 
 static void tgq_idct_put_mb(TgqContext *s, int16_t (*block)[64], AVFrame *frame,
@@ -161,8 +166,11 @@ static int tgq_decode_mb(TgqContext *s, AVFrame *frame, int mb_y, int mb_x)
         if (ret < 0)
             return ret;
 
-        for (i = 0; i < 6; i++)
-            tgq_decode_block(s, s->block[i], &gb);
+        for (i = 0; i < 6; i++) {
+            int ret = tgq_decode_block(s, s->block[i], &gb);
+            if (ret < 0)
+                return ret;
+        }
         tgq_idct_put_mb(s, s->block, frame, mb_x, mb_y);
         bytestream2_skip(&s->gb, mode);
     } else {

libavcodec/escape124.c

@@ -88,11 +88,6 @@ static CodeBook unpack_codebook(GetBitContext* gb, unsigned depth,
     unsigned i, j;
     CodeBook cb = { 0 };
 
-    if (size >= INT_MAX / 34 || get_bits_left(gb) < size * 34)
-        return cb;
-
-    if (size >= INT_MAX / sizeof(MacroBlock))
-        return cb;
     cb.blocks = av_malloc(size ? size * sizeof(MacroBlock) : 1);
     if (!cb.blocks)
         return cb;
@@ -162,7 +157,7 @@ static MacroBlock decode_macroblock(Escape124Context* s, GetBitContext* gb,
 
     // This condition can occur with invalid bitstreams and
     // *codebook_index == 2
-    if (block_index >= s->codebooks[*codebook_index].size)
+    if (block_index >= s->codebooks[*codebook_index].size || !s->codebooks[*codebook_index].blocks)
         return (MacroBlock) { { 0 } };
 
     return s->codebooks[*codebook_index].blocks[block_index];
@@ -226,7 +221,7 @@ static int escape124_decode_frame(AVCodecContext *avctx,
     // represent a lower bound of the space needed for skipped superblocks. Non
     // skipped SBs need more space.
     if (get_bits_left(&gb) < 64 + s->num_superblocks * 23LL / 4320)
-        return -1;
+        return AVERROR_INVALIDDATA;
 
     frame_flags = get_bits_long(&gb, 32);
     frame_size  = get_bits_long(&gb, 32);
@@ -277,9 +272,14 @@ static int escape124_decode_frame(AVCodecContext *avctx,
             }
 
             av_freep(&s->codebooks[i].blocks);
+            if (cb_size >= INT_MAX / 34 || get_bits_left(&gb) < (int)cb_size * 34)
+                return AVERROR_INVALIDDATA;
+
+            if (cb_size >= INT_MAX / sizeof(MacroBlock))
+                return AVERROR_INVALIDDATA;
             s->codebooks[i] = unpack_codebook(&gb, cb_depth, cb_size);
             if (!s->codebooks[i].blocks)
-                return -1;
+                return AVERROR(ENOMEM);
         }
     }
 

libavcodec/ffv1dec.c

@@ -169,24 +169,31 @@ static int decode_slice_header(FFV1Context *f, FFV1Context *fs)
     RangeCoder *c = &fs->c;
     uint8_t state[CONTEXT_SIZE];
     unsigned ps, i, context_count;
+    int sx, sy, sw, sh;
+
     memset(state, 128, sizeof(state));
+    sx = get_symbol(c, state, 0);
+    sy = get_symbol(c, state, 0);
+    sw = get_symbol(c, state, 0) + 1U;
+    sh = get_symbol(c, state, 0) + 1U;
 
     av_assert0(f->version > 2);
 
-    fs->slice_x      =  get_symbol(c, state, 0)      * f->width ;
-    fs->slice_y      =  get_symbol(c, state, 0)      * f->height;
-    fs->slice_width  = (get_symbol(c, state, 0) + 1) * f->width  + fs->slice_x;
-    fs->slice_height = (get_symbol(c, state, 0) + 1) * f->height + fs->slice_y;
 
-    fs->slice_x /= f->num_h_slices;
-    fs->slice_y /= f->num_v_slices;
-    fs->slice_width  = fs->slice_width /f->num_h_slices - fs->slice_x;
-    fs->slice_height = fs->slice_height/f->num_v_slices - fs->slice_y;
-    if ((unsigned)fs->slice_width > f->width || (unsigned)fs->slice_height > f->height)
-        return -1;
-    if (    (unsigned)fs->slice_x + (uint64_t)fs->slice_width  > f->width
-         || (unsigned)fs->slice_y + (uint64_t)fs->slice_height > f->height)
-        return -1;
+    if (sx < 0 || sy < 0 || sw <= 0 || sh <= 0)
+        return AVERROR_INVALIDDATA;
+    if (sx > f->num_h_slices - sw || sy > f->num_v_slices - sh)
+        return AVERROR_INVALIDDATA;
+
+    fs->slice_x      =  sx       * (int64_t)f->width  / f->num_h_slices;
+    fs->slice_y      =  sy       * (int64_t)f->height / f->num_v_slices;
+    fs->slice_width  = (sx + sw) * (int64_t)f->width  / f->num_h_slices - fs->slice_x;
+    fs->slice_height = (sy + sh) * (int64_t)f->height / f->num_v_slices - fs->slice_y;
+
+    av_assert0((unsigned)fs->slice_width  <= f->width &&
+                (unsigned)fs->slice_height <= f->height);
+    av_assert0 (   (unsigned)fs->slice_x + (uint64_t)fs->slice_width  <= f->width
+                && (unsigned)fs->slice_y + (uint64_t)fs->slice_height <= f->height);
 
     if (fs->ac == AC_GOLOMB_RICE && fs->slice_width >= (1<<23))
         return AVERROR_INVALIDDATA;
@@ -304,8 +311,11 @@ static int decode_slice(AVCodecContext *c, void *arg)
     }
     if ((ret = ff_ffv1_init_slice_state(f, fs)) < 0)
         return ret;
-    if (f->cur->key_frame || fs->slice_reset_contexts)
+    if (f->cur->key_frame || fs->slice_reset_contexts) {
         ff_ffv1_clear_slice_state(f, fs);
+    } else if (fs->slice_damaged) {
+        return AVERROR_INVALIDDATA;
+    }
 
     width  = fs->slice_width;
     height = fs->slice_height;
@@ -468,6 +478,11 @@ static int read_extra_header(FFV1Context *f)
         return AVERROR_INVALIDDATA;
     }
 
+    if (f->num_h_slices > MAX_SLICES / f->num_v_slices) {
+        av_log(f->avctx, AV_LOG_ERROR, "slice count unsupported\n");
+        return AVERROR_PATCHWELCOME;
+    }
+
     f->quant_table_count = get_symbol(c, state, 0);
     if (f->quant_table_count > (unsigned)MAX_QUANT_TABLES || !f->quant_table_count) {
         av_log(f->avctx, AV_LOG_ERROR, "quant table count %d is invalid\n", f->quant_table_count);
@@ -770,21 +785,25 @@ static int read_header(FFV1Context *f)
         fs->slice_damaged = 0;
 
         if (f->version == 2) {
-            fs->slice_x      =  get_symbol(c, state, 0)      * f->width ;
-            fs->slice_y      =  get_symbol(c, state, 0)      * f->height;
-            fs->slice_width  = (get_symbol(c, state, 0) + 1) * f->width  + fs->slice_x;
-            fs->slice_height = (get_symbol(c, state, 0) + 1) * f->height + fs->slice_y;
+            int sx = get_symbol(c, state, 0);
+            int sy = get_symbol(c, state, 0);
+            int sw = get_symbol(c, state, 0) + 1U;
+            int sh = get_symbol(c, state, 0) + 1U;
 
-            fs->slice_x     /= f->num_h_slices;
-            fs->slice_y     /= f->num_v_slices;
-            fs->slice_width  = fs->slice_width  / f->num_h_slices - fs->slice_x;
-            fs->slice_height = fs->slice_height / f->num_v_slices - fs->slice_y;
-            if ((unsigned)fs->slice_width  > f->width ||
-                (unsigned)fs->slice_height > f->height)
+            if (sx < 0 || sy < 0 || sw <= 0 || sh <= 0)
                 return AVERROR_INVALIDDATA;
-            if (   (unsigned)fs->slice_x + (uint64_t)fs->slice_width  > f->width
-                || (unsigned)fs->slice_y + (uint64_t)fs->slice_height > f->height)
+            if (sx > f->num_h_slices - sw || sy > f->num_v_slices - sh)
                 return AVERROR_INVALIDDATA;
+
+            fs->slice_x      =  sx       * (int64_t)f->width  / f->num_h_slices;
+            fs->slice_y      =  sy       * (int64_t)f->height / f->num_v_slices;
+            fs->slice_width  = (sx + sw) * (int64_t)f->width  / f->num_h_slices - fs->slice_x;
+            fs->slice_height = (sy + sh) * (int64_t)f->height / f->num_v_slices - fs->slice_y;
+
+            av_assert0((unsigned)fs->slice_width  <= f->width &&
+                       (unsigned)fs->slice_height <= f->height);
+            av_assert0 (   (unsigned)fs->slice_x + (uint64_t)fs->slice_width  <= f->width
+                        && (unsigned)fs->slice_y + (uint64_t)fs->slice_height <= f->height);
         }
 
         for (i = 0; i < f->plane_count; i++) {

libavcodec/g729postfilter.c

@@ -350,7 +350,7 @@ static int16_t long_term_filter(AudioDSPContext *adsp, int pitch_delay_int,
         if (tmp > 0)
             L_temp0 >>= tmp;
         else
-            L_temp1 >>= -tmp;
+            L_temp1 >>= FFMIN(-tmp, 31);
 
         /* Check if longer filter increases the values of R'(k). */
         if (L_temp1 > L_temp0) {

libavcodec/huffyuvdec.c

@@ -691,9 +691,9 @@ static void decode_422_bitstream(HYuvContext *s, int count)
 /* TODO instead of restarting the read when the code isn't in the first level
  * of the joint table, jump into the 2nd level of the individual table. */
 #define READ_2PIX_PLANE16(dst0, dst1, plane){\
-    dst0 = get_vlc2(&s->gb, s->vlc[plane].table, VLC_BITS, 3)<<2;\
+    dst0 = get_vlc2(&s->gb, s->vlc[plane].table, VLC_BITS, 3)*4;\
     dst0 += get_bits(&s->gb, 2);\
-    dst1 = get_vlc2(&s->gb, s->vlc[plane].table, VLC_BITS, 3)<<2;\
+    dst1 = get_vlc2(&s->gb, s->vlc[plane].table, VLC_BITS, 3)*4;\
     dst1 += get_bits(&s->gb, 2);\
 }
 static void decode_plane_bitstream(HYuvContext *s, int width, int plane)

libavcodec/j2kenc.c

@@ -1003,6 +1003,7 @@ static int encode_frame(AVCodecContext *avctx, AVPacket *pkt,
     int tileno, ret;
     Jpeg2000EncoderContext *s = avctx->priv_data;
     uint8_t *chunkstart, *jp2cstart, *jp2hstart;
+    const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(avctx->pix_fmt);
 
     if ((ret = ff_alloc_packet2(avctx, pkt, avctx->width*avctx->height*9 + AV_INPUT_BUFFER_MIN_SIZE, 0)) < 0)
         return ret;
@@ -1055,7 +1056,7 @@ static int encode_frame(AVCodecContext *avctx, AVPacket *pkt,
         bytestream_put_byte(&s->buf, 1);
         bytestream_put_byte(&s->buf, 0);
         bytestream_put_byte(&s->buf, 0);
-        if (avctx->pix_fmt == AV_PIX_FMT_RGB24 || avctx->pix_fmt == AV_PIX_FMT_PAL8) {
+        if ((desc->flags & AV_PIX_FMT_FLAG_RGB) || avctx->pix_fmt == AV_PIX_FMT_PAL8) {
             bytestream_put_be32(&s->buf, 16);
         } else if (s->ncomponents == 1) {
             bytestream_put_be32(&s->buf, 17);
@@ -1151,7 +1152,7 @@ FF_ENABLE_DEPRECATION_WARNINGS
 
     if (avctx->pix_fmt == AV_PIX_FMT_PAL8 && (s->pred != FF_DWT97_INT || s->format != CODEC_JP2)) {
         av_log(s->avctx, AV_LOG_WARNING, "Forcing lossless jp2 for pal8\n");
-        s->pred = FF_DWT97_INT;
+        s->pred = 1;
         s->format = CODEC_JP2;
     }
 
@@ -1229,7 +1230,7 @@ static const AVOption options[] = {
     { "tile_height",   "Tile Height",       OFFSET(tile_height),   AV_OPT_TYPE_INT,   { .i64 = 256         }, 1,     1<<30,           VE, },
     { "pred",          "DWT Type",          OFFSET(pred),          AV_OPT_TYPE_INT,   { .i64 = 0           }, 0,         1,           VE, "pred"        },
     { "dwt97int",      NULL,                0,                     AV_OPT_TYPE_CONST, { .i64 = 0           }, INT_MIN, INT_MAX,       VE, "pred"        },
-    { "dwt53",         NULL,                0,                     AV_OPT_TYPE_CONST, { .i64 = 0           }, INT_MIN, INT_MAX,       VE, "pred"        },
+    { "dwt53",         NULL,                0,                     AV_OPT_TYPE_CONST, { .i64 = 1           }, INT_MIN, INT_MAX,       VE, "pred"        },
 
     { NULL }
 };

libavcodec/mjpegenc_common.c

@@ -275,7 +275,7 @@ void ff_mjpeg_encode_picture_header(AVCodecContext *avctx, PutBitContext *pb,
     default: av_assert0(0);
     }
 
-    put_bits(pb, 16, 17);
+    put_bits(pb, 16, 8 + 3 * components);
     if (lossless && (  avctx->pix_fmt == AV_PIX_FMT_BGR0
                     || avctx->pix_fmt == AV_PIX_FMT_BGRA
                     || avctx->pix_fmt == AV_PIX_FMT_BGR24))

libavcodec/mlpdec.c

@@ -520,7 +520,7 @@ static int read_restart_header(MLPDecodeContext *m, GetBitContext *gbp,
 
     /* This should happen for TrueHD streams with >6 channels and MLP's noise
      * type. It is not yet known if this is allowed. */
-    if (max_channel > MAX_MATRIX_CHANNEL_MLP && !noise_type) {
+    if (max_matrix_channel > MAX_MATRIX_CHANNEL_MLP && !noise_type) {
         avpriv_request_sample(m->avctx,
                               "%d channels (more than the "
                               "maximum supported by the decoder)",

libavcodec/motionpixels.c

@@ -185,7 +185,7 @@ static YuvPixel mp_get_yuv_from_rgb(MotionPixelsContext *mp, int x, int y)
     int color;
 
     color = *(uint16_t *)&mp->frame->data[0][y * mp->frame->linesize[0] + x * 2];
-    return mp_rgb_yuv_table[color];
+    return mp_rgb_yuv_table[color & 0x7FFF];
 }
 
 static void mp_set_rgb_from_yuv(MotionPixelsContext *mp, int x, int y, const YuvPixel *p)

libavcodec/nvenc.c

@@ -920,8 +920,9 @@ static av_cold void nvenc_setup_rate_control(AVCodecContext *avctx)
 
         av_log(avctx, AV_LOG_VERBOSE, "CQ(%d) mode enabled.\n", tmp_quality);
 
-        //CQ mode shall discard avg bitrate & honor max bitrate;
+        // CQ mode shall discard avg bitrate/vbv buffer size and honor only max bitrate
         ctx->encode_config.rcParams.averageBitRate = avctx->bit_rate = 0;
+        ctx->encode_config.rcParams.vbvBufferSize = avctx->rc_buffer_size = 0;
         ctx->encode_config.rcParams.maxBitRate = avctx->rc_max_rate;
     }
 }

libavcodec/pictordec.c

@@ -231,8 +231,6 @@ static int decode_frame(AVCodecContext *avctx,
                         run = bytestream2_get_le16(&s->g);
                     val = bytestream2_get_byte(&s->g);
                 }
-                if (!bytestream2_get_bytes_left(&s->g))
-                    break;
 
                 if (bits_per_plane == 8) {
                     picmemset_8bpp(s, frame, val, run, &x, &y);

libavcodec/pngdec.c

@@ -309,7 +309,7 @@ static void png_filter_row(PNGDSPContext *dsp, uint8_t *dst, int filter_type,
 static void deloco_ ## NAME(TYPE *dst, int size, int alpha) \
 { \
     int i; \
-    for (i = 0; i < size; i += 3 + alpha) { \
+    for (i = 0; i < size - 2; i += 3 + alpha) { \
         int g = dst [i + 1]; \
         dst[i + 0] += g; \
         dst[i + 2] += g; \

libavcodec/scpr.c

@@ -459,6 +459,9 @@ static int decompress_p(AVCodecContext *avctx,
                 int run, bx = x * 16 + sx1, by = y * 16 + sy1;
                 uint32_t r, g, b, clr, ptype = 0;
 
+                if (bx >= avctx->width)
+                    return AVERROR_INVALIDDATA;
+
                 for (; by < y * 16 + sy2 && by < avctx->height;) {
                     ret = decode_value(s, s->op_model[ptype], 6, 1000, &ptype);
                     if (ret < 0)

libavcodec/scpr3.c

@@ -1184,6 +1184,9 @@ static int decompress_p3(AVCodecContext *avctx,
                 int run, bx = x * 16 + sx1, by = y * 16 + sy1;
                 uint32_t clr, ptype = 0, r, g, b;
 
+                if (bx >= avctx->width)
+                    return AVERROR_INVALIDDATA;
+
                 for (; by < y * 16 + sy2 && by < avctx->height;) {
                     ret = decode_value3(s, 5, &s->op_model3[ptype].cntsum,
                                         s->op_model3[ptype].freqs[0],

libavcodec/snowenc.c

@@ -1543,10 +1543,10 @@ static void calculate_visual_weight(SnowContext *s, Plane *p){
     int level, orientation, x, y;
 
     for(level=0; level<s->spatial_decomposition_count; level++){
+        int64_t error=0;
         for(orientation=level ? 1 : 0; orientation<4; orientation++){
             SubBand *b= &p->band[level][orientation];
             IDWTELEM *ibuf= b->ibuf;
-            int64_t error=0;
 
             memset(s->spatial_idwt_buffer, 0, sizeof(*s->spatial_idwt_buffer)*width*height);
             ibuf[b->width/2 + b->height/2*b->stride]= 256*16;
@@ -1557,9 +1557,13 @@ static void calculate_visual_weight(SnowContext *s, Plane *p){
                     error += d*d;
                 }
             }
-
+            if (orientation == 2)
+                error /= 2;
             b->qlog= (int)(QROOT * log2(352256.0/sqrt(error)) + 0.5);
+            if (orientation != 1)
+                error = 0;
         }
+        p->band[level][1].qlog = p->band[level][2].qlog;
     }
 }
 

libavcodec/speedhq.c

@@ -426,6 +426,8 @@ static int speedhq_decode_frame(AVCodecContext *avctx,
 
     if (buf_size < 4 || avctx->width < 8 || avctx->width % 8 != 0)
         return AVERROR_INVALIDDATA;
+    if (buf_size < avctx->width*avctx->height / 64 / 4)
+        return AVERROR_INVALIDDATA;
 
     quality = buf[0];
     if (quality >= 100) {

libavcodec/sunrast.c

@@ -19,6 +19,7 @@
  * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
  */
 
+#include "libavutil/avassert.h"
 #include "libavutil/common.h"
 #include "libavutil/intreadwrite.h"
 #include "libavutil/imgutils.h"
@@ -75,6 +76,12 @@ static int sunrast_decode_frame(AVCodecContext *avctx, void *data,
         return AVERROR_PATCHWELCOME;
     }
 
+    if (maplength > 768) {
+        av_log(avctx, AV_LOG_WARNING, "invalid colormap length\n");
+        return AVERROR_INVALIDDATA;
+    }
+
+    // This also checks depth to be valid
     switch (depth) {
         case 1:
             avctx->pix_fmt = maplength ? AV_PIX_FMT_PAL8 : AV_PIX_FMT_MONOWHITE;
@@ -96,15 +103,23 @@ static int sunrast_decode_frame(AVCodecContext *avctx, void *data,
             return AVERROR_INVALIDDATA;
     }
 
+    // This checks w and h to be valid in the sense that bytes of a padded bitmap are addressable with 32bit int
     ret = ff_set_dimensions(avctx, w, h);
     if (ret < 0)
         return ret;
 
+    // ensured by ff_set_dimensions()
+    av_assert0(w <= (INT32_MAX - 7) / depth);
+
     /* scanlines are aligned on 16 bit boundaries */
     len  = (depth * w + 7) >> 3;
     alen = len + (len & 1);
 
-    if (buf_end - buf < maplength + (len * h) * 3 / 256)
+    // ensured by ff_set_dimensions()
+    av_assert0(h  <= INT32_MAX / (3 * len));
+
+    // maplength is limited to 768 and the right term is limited to INT32_MAX / 256 so the add needs no check
+    if (buf_end - buf < (uint64_t)maplength + (len * h) * 3 / 256)
         return AVERROR_INVALIDDATA;
 
     if ((ret = ff_get_buffer(avctx, p, 0)) < 0)
@@ -118,7 +133,7 @@ static int sunrast_decode_frame(AVCodecContext *avctx, void *data,
     } else if (maplength) {
         unsigned int len = maplength / 3;
 
-        if (maplength % 3 || maplength > 768) {
+        if (maplength % 3) {
             av_log(avctx, AV_LOG_WARNING, "invalid colormap length\n");
             return AVERROR_INVALIDDATA;
         }

libavcodec/tests/snowenc.c

@@ -31,11 +31,13 @@ int main(void){
 #define width  256
 #define height 256
     int buffer[2][width*height];
+    short obuffer[width*height];
     SnowContext s;
     int i;
     AVLFG prng;
     s.spatial_decomposition_count=6;
     s.spatial_decomposition_type=1;
+    int ret = 0;
 
     s.temp_dwt_buffer  = av_mallocz_array(width, sizeof(DWTELEM));
     s.temp_idwt_buffer = av_mallocz_array(width, sizeof(IDWTELEM));
@@ -49,24 +51,34 @@ int main(void){
 
     printf("testing 5/3 DWT\n");
     for(i=0; i<width*height; i++)
-        buffer[0][i] = buffer[1][i] = av_lfg_get(&prng) % 54321 - 12345;
+        buffer[0][i] = buffer[1][i] = av_lfg_get(&prng) % 19000 - 9000;
 
     ff_spatial_dwt(buffer[0], s.temp_dwt_buffer, width, height, width, s.spatial_decomposition_type, s.spatial_decomposition_count);
-    ff_spatial_idwt((IDWTELEM*)buffer[0], s.temp_idwt_buffer, width, height, width, s.spatial_decomposition_type, s.spatial_decomposition_count);
+    for(i=0; i<width*height; i++)
+        obuffer[i] = buffer[0][i];
+    ff_spatial_idwt(obuffer, s.temp_idwt_buffer, width, height, width, s.spatial_decomposition_type, s.spatial_decomposition_count);
 
     for(i=0; i<width*height; i++)
-        if(buffer[0][i]!= buffer[1][i]) printf("fsck: %6d %12d %7d\n",i, buffer[0][i], buffer[1][i]);
+        if(buffer[1][i]!= obuffer[i]) {
+            printf("fsck: %4dx%4dx %12d %7d\n",i%width, i/width, buffer[1][i], obuffer[i]);
+            ret = 1;
+        }
 
     printf("testing 9/7 DWT\n");
     s.spatial_decomposition_type=0;
     for(i=0; i<width*height; i++)
-        buffer[0][i] = buffer[1][i] = av_lfg_get(&prng) % 54321 - 12345;
+        buffer[0][i] = buffer[1][i] = av_lfg_get(&prng) % 11000 - 5000;
 
     ff_spatial_dwt(buffer[0], s.temp_dwt_buffer, width, height, width, s.spatial_decomposition_type, s.spatial_decomposition_count);
-    ff_spatial_idwt((IDWTELEM*)buffer[0], s.temp_idwt_buffer, width, height, width, s.spatial_decomposition_type, s.spatial_decomposition_count);
+    for(i=0; i<width*height; i++)
+        obuffer[i] = buffer[0][i];
+    ff_spatial_idwt(obuffer, s.temp_idwt_buffer, width, height, width, s.spatial_decomposition_type, s.spatial_decomposition_count);
 
     for(i=0; i<width*height; i++)
-        if(FFABS(buffer[0][i] - buffer[1][i])>20) printf("fsck: %6d %12d %7d\n",i, buffer[0][i], buffer[1][i]);
+        if(FFABS(buffer[1][i] - obuffer[i])>20) {
+            printf("fsck: %4dx%4d %12d %7d\n",i%width, i/width, buffer[1][i], obuffer[i]);
+            ret = 1;
+        }
 
     {
     int level, orientation, x, y;
@@ -81,18 +93,18 @@ int main(void){
                 int w= width  >> (s.spatial_decomposition_count-level);
                 int h= height >> (s.spatial_decomposition_count-level);
                 int stride= width  << (s.spatial_decomposition_count-level);
-                DWTELEM *buf= buffer[0];
+                IDWTELEM *buf= obuffer;
                 int64_t error=0;
 
                 if(orientation&1) buf+=w;
                 if(orientation>1) buf+=stride>>1;
 
-                memset(buffer[0], 0, sizeof(int)*width*height);
-                buf[w/2 + h/2*stride]= 256*256;
-                ff_spatial_idwt((IDWTELEM*)buffer[0], s.temp_idwt_buffer, width, height, width, s.spatial_decomposition_type, s.spatial_decomposition_count);
+                memset(obuffer, 0, sizeof(short)*width*height);
+                buf[w/2 + h/2*stride]= 8*256;
+                ff_spatial_idwt(obuffer, s.temp_idwt_buffer, width, height, width, s.spatial_decomposition_type, s.spatial_decomposition_count);
                 for(y=0; y<height; y++){
                     for(x=0; x<width; x++){
-                        int64_t d= buffer[0][x + y*width];
+                        int64_t d= obuffer[x + y*width];
                         error += d*d;
                         if(FFABS(width/2-x)<9 && FFABS(height/2-y)<9 && level==2) printf("%8"PRId64" ", d);
                     }
@@ -143,5 +155,5 @@ int main(void){
         }
 
     }
-    return 0;
+    return ret;
 }

libavcodec/utils.c

@@ -256,6 +256,8 @@ void avcodec_align_dimensions2(AVCodecContext *s, int *width, int *height,
     case AV_PIX_FMT_GBRAP16BE:
         w_align = 16; //FIXME assume 16 pixel per macroblock
         h_align = 16 * 2; // interlaced needs 2 macroblocks height
+        if (s->codec_id == AV_CODEC_ID_BINKVIDEO)
+            w_align = 16*2;
         break;
     case AV_PIX_FMT_YUV411P:
     case AV_PIX_FMT_YUVJ411P:
@@ -327,6 +329,7 @@ void avcodec_align_dimensions2(AVCodecContext *s, int *width, int *height,
     *width  = FFALIGN(*width, w_align);
     *height = FFALIGN(*height, h_align);
     if (s->codec_id == AV_CODEC_ID_H264 || s->lowres ||
+        s->codec_id == AV_CODEC_ID_VC1  || s->codec_id == AV_CODEC_ID_WMV3 ||
         s->codec_id == AV_CODEC_ID_VP5  || s->codec_id == AV_CODEC_ID_VP6 ||
         s->codec_id == AV_CODEC_ID_VP6F || s->codec_id == AV_CODEC_ID_VP6A
     ) {
@@ -340,6 +343,9 @@ void avcodec_align_dimensions2(AVCodecContext *s, int *width, int *height,
         // the next rounded up width is 32
         *width = FFMAX(*width, 32);
     }
+    if (s->codec_id == AV_CODEC_ID_SVQ3) {
+        *width = FFMAX(*width, 32);
+    }
 
     for (i = 0; i < 4; i++)
         linesize_align[i] = STRIDE_ALIGN;

libavcodec/videodsp_template.c

@@ -60,7 +60,7 @@ void FUNC(ff_emulated_edge_mc)(uint8_t *buf, const uint8_t *src,
     av_assert2(start_x < end_x && block_w);
 
     w    = end_x - start_x;
-    src += start_y * src_linesize + start_x * sizeof(pixel);
+    src += start_y * src_linesize + start_x * (ptrdiff_t)sizeof(pixel);
     buf += start_x * sizeof(pixel);
 
     // top
@@ -83,7 +83,7 @@ void FUNC(ff_emulated_edge_mc)(uint8_t *buf, const uint8_t *src,
         buf += buf_linesize;
     }
 
-    buf -= block_h * buf_linesize + start_x * sizeof(pixel);
+    buf -= block_h * buf_linesize + start_x * (ptrdiff_t)sizeof(pixel);
     while (block_h--) {
         pixel *bufp = (pixel *) buf;
 

libavcodec/vp3.c

@@ -2336,6 +2336,8 @@ static av_cold int vp3_decode_init(AVCodecContext *avctx)
     s->avctx  = avctx;
     s->width  = FFALIGN(avctx->coded_width, 16);
     s->height = FFALIGN(avctx->coded_height, 16);
+    if (s->width < 18)
+        return AVERROR_PATCHWELCOME;
     if (avctx->codec_id != AV_CODEC_ID_THEORA)
         avctx->pix_fmt = AV_PIX_FMT_YUV420P;
     avctx->chroma_sample_location = AVCHROMA_LOC_CENTER;
@@ -2740,8 +2742,13 @@ static int vp3_decode_frame(AVCodecContext *avctx,
     if (ff_thread_get_buffer(avctx, &s->current_frame, AV_GET_BUFFER_FLAG_REF) < 0)
         goto error;
 
-    if (!s->edge_emu_buffer)
+    if (!s->edge_emu_buffer) {
         s->edge_emu_buffer = av_malloc(9 * FFABS(s->current_frame.f->linesize[0]));
+        if (!s->edge_emu_buffer) {
+            ret = AVERROR(ENOMEM);
+            goto error;
+        }
+    }
 
     if (s->keyframe) {
         if (!s->theora) {
@@ -3002,7 +3009,9 @@ static int theora_decode_header(AVCodecContext *avctx, GetBitContext *gb)
     /* sanity check */
     if (av_image_check_size(visible_width, visible_height, 0, avctx) < 0 ||
         visible_width  + offset_x > s->width ||
-        visible_height + offset_y > s->height) {
+        visible_height + offset_y > s->height ||
+        visible_width < 18
+    ) {
         av_log(avctx, AV_LOG_ERROR,
                "Invalid frame dimensions - w:%d h:%d x:%d y:%d (%dx%d).\n",
                visible_width, visible_height, offset_x, offset_y,
@@ -3048,6 +3057,8 @@ static int theora_decode_header(AVCodecContext *avctx, GetBitContext *gb)
     } else
         avctx->pix_fmt = AV_PIX_FMT_YUV420P;
 
+    if (s->width < 18)
+        return AVERROR_PATCHWELCOME;
     ret = ff_set_dimensions(avctx, s->width, s->height);
     if (ret < 0)
         return ret;

libavcodec/wavpack.c

@@ -95,7 +95,7 @@ static av_always_inline unsigned get_tail(GetBitContext *gb, int k)
     e   = (1 << (p + 1)) - k - 1;
     res = get_bitsz(gb, p);
     if (res >= e)
-        res = (res << 1) - e + get_bits1(gb);
+        res = res * 2U - e + get_bits1(gb);
     return res;
 }
 

libavcodec/xpmdec.c

@@ -355,6 +355,9 @@ static int xpm_decode_frame(AVCodecContext *avctx, void *data,
         return AVERROR_INVALIDDATA;
     }
 
+    if (size > SIZE_MAX / 4)
+        return AVERROR(ENOMEM);
+
     size *= 4;
 
     ptr += mod_strcspn(ptr, ",") + 1;

libavformat/id3v2.c

@@ -377,10 +377,10 @@ static void read_uslt(AVFormatContext *s, AVIOContext *pb, int taglen,
     lang[3] = '\0';
     taglen -= 3;
 
-    if (decode_str(s, pb, encoding, &descriptor, &taglen) < 0)
+    if (decode_str(s, pb, encoding, &descriptor, &taglen) < 0 || taglen < 0)
         goto error;
 
-    if (decode_str(s, pb, encoding, &text, &taglen) < 0)
+    if (decode_str(s, pb, encoding, &text, &taglen) < 0 || taglen < 0)
         goto error;
 
     // FFmpeg does not support hierarchical metadata, so concatenate the keys.

libavformat/mov.c

@@ -3899,6 +3899,13 @@ static void mov_build_index(MOVContext *mov, AVStream *st)
                 if (keyframe)
                     distance = 0;
                 sample_size = sc->stsz_sample_size > 0 ? sc->stsz_sample_size : sc->sample_sizes[current_sample];
+                if (current_offset > INT64_MAX - sample_size) {
+                    av_log(mov->fc, AV_LOG_ERROR, "Current offset %"PRId64" or sample size %u is too large\n",
+                           current_offset,
+                           sample_size);
+                    return;
+                }
+
                 if (sc->pseudo_stream_id == -1 ||
                    sc->stsc_data[stsc_index].id - 1 == sc->pseudo_stream_id) {
                     AVIndexEntry *e;

libavformat/replaygain.c

@@ -61,7 +61,7 @@ static int32_t parse_value(const char *value, int32_t min)
         }
     }
 
-    if (abs(db) > (INT32_MAX - mb) / 100000)
+    if (llabs(db) > (INT32_MAX - mb) / 100000)
         return min;
 
     return db * 100000 + sign * mb;

libswscale/aarch64/yuv2rgb_neon.S

@@ -118,8 +118,8 @@
 .endm
 
 .macro increment_yuv422p
-    add                 x6,  x6,  w7, UXTW                              // srcU += incU
-    add                 x13, x13, w14, UXTW                             // srcV += incV
+    add                 x6,  x6,  w7, SXTW                              // srcU += incU
+    add                 x13, x13, w14, SXTW                             // srcV += incV
 .endm
 
 .macro compute_rgba r1 g1 b1 a1 r2 g2 b2 a2
@@ -188,8 +188,8 @@ function ff_\ifmt\()_to_\ofmt\()_neon, export=1
     st4                 {v16.8B,v17.8B,v18.8B,v19.8B}, [x2], #32
     subs                w8, w8, #16                                     // width -= 16
     b.gt                2b
-    add                 x2, x2, w3, UXTW                                // dst  += padding
-    add                 x4, x4, w5, UXTW                                // srcY += paddingY
+    add                 x2, x2, w3, SXTW                                // dst  += padding
+    add                 x4, x4, w5, SXTW                                // srcY += paddingY
     increment_\ifmt
     subs                w1, w1, #1                                      // height -= 1
     b.gt                1b

libswscale/input.c

@@ -84,9 +84,9 @@ rgb64ToUV_half_c_template(uint16_t *dstU, uint16_t *dstV,
     int32_t rv = rgb2yuv[RV_IDX], gv = rgb2yuv[GV_IDX], bv = rgb2yuv[BV_IDX];
     av_assert1(src1==src2);
     for (i = 0; i < width; i++) {
-        int r_b = (input_pixel(&src1[8 * i + 0]) + input_pixel(&src1[8 * i + 4]) + 1) >> 1;
-        int   g = (input_pixel(&src1[8 * i + 1]) + input_pixel(&src1[8 * i + 5]) + 1) >> 1;
-        int b_r = (input_pixel(&src1[8 * i + 2]) + input_pixel(&src1[8 * i + 6]) + 1) >> 1;
+        unsigned r_b = (input_pixel(&src1[8 * i + 0]) + input_pixel(&src1[8 * i + 4]) + 1) >> 1;
+        unsigned   g = (input_pixel(&src1[8 * i + 1]) + input_pixel(&src1[8 * i + 5]) + 1) >> 1;
+        unsigned b_r = (input_pixel(&src1[8 * i + 2]) + input_pixel(&src1[8 * i + 6]) + 1) >> 1;
 
         dstU[i]= (ru*r + gu*g + bu*b + (0x10001<<(RGB2YUV_SHIFT-1))) >> RGB2YUV_SHIFT;
         dstV[i]= (rv*r + gv*g + bv*b + (0x10001<<(RGB2YUV_SHIFT-1))) >> RGB2YUV_SHIFT;
@@ -156,9 +156,9 @@ static av_always_inline void rgb48ToUV_c_template(uint16_t *dstU,
     int32_t rv = rgb2yuv[RV_IDX], gv = rgb2yuv[GV_IDX], bv = rgb2yuv[BV_IDX];
     av_assert1(src1 == src2);
     for (i = 0; i < width; i++) {
-        int r_b = input_pixel(&src1[i * 3 + 0]);
-        int g   = input_pixel(&src1[i * 3 + 1]);
-        int b_r = input_pixel(&src1[i * 3 + 2]);
+        unsigned r_b = input_pixel(&src1[i * 3 + 0]);
+        unsigned g   = input_pixel(&src1[i * 3 + 1]);
+        unsigned b_r = input_pixel(&src1[i * 3 + 2]);
 
         dstU[i] = (ru*r + gu*g + bu*b + (0x10001 << (RGB2YUV_SHIFT - 1))) >> RGB2YUV_SHIFT;
         dstV[i] = (rv*r + gv*g + bv*b + (0x10001 << (RGB2YUV_SHIFT - 1))) >> RGB2YUV_SHIFT;
@@ -178,12 +178,12 @@ static av_always_inline void rgb48ToUV_half_c_template(uint16_t *dstU,
     int32_t rv = rgb2yuv[RV_IDX], gv = rgb2yuv[GV_IDX], bv = rgb2yuv[BV_IDX];
     av_assert1(src1 == src2);
     for (i = 0; i < width; i++) {
-        int r_b = (input_pixel(&src1[6 * i + 0]) +
-                   input_pixel(&src1[6 * i + 3]) + 1) >> 1;
-        int g   = (input_pixel(&src1[6 * i + 1]) +
-                   input_pixel(&src1[6 * i + 4]) + 1) >> 1;
-        int b_r = (input_pixel(&src1[6 * i + 2]) +
-                   input_pixel(&src1[6 * i + 5]) + 1) >> 1;
+        unsigned r_b = (input_pixel(&src1[6 * i + 0]) +
+                        input_pixel(&src1[6 * i + 3]) + 1) >> 1;
+        unsigned g   = (input_pixel(&src1[6 * i + 1]) +
+                        input_pixel(&src1[6 * i + 4]) + 1) >> 1;
+        unsigned b_r = (input_pixel(&src1[6 * i + 2]) +
+                        input_pixel(&src1[6 * i + 5]) + 1) >> 1;
 
         dstU[i] = (ru*r + gu*g + bu*b + (0x10001 << (RGB2YUV_SHIFT - 1))) >> RGB2YUV_SHIFT;
         dstV[i] = (rv*r + gv*g + bv*b + (0x10001 << (RGB2YUV_SHIFT - 1))) >> RGB2YUV_SHIFT;

libswscale/output.c

@@ -1038,8 +1038,8 @@ yuv2rgba64_X_c_template(SwsContext *c, const int16_t *lumFilter,
         Y2 -= c->yuv2rgb_y_offset;
         Y1 *= c->yuv2rgb_y_coeff;
         Y2 *= c->yuv2rgb_y_coeff;
-        Y1 += 1 << 13; // 21
-        Y2 += 1 << 13;
+        Y1 += (1 << 13) - (1 << 29); // 21
+        Y2 += (1 << 13) - (1 << 29);
         // 8 bits: 17 + 13 bits = 30 bits, 16 bits: 17 + 13 bits = 30 bits
 
         R = V * c->yuv2rgb_v2r_coeff;
@@ -1047,20 +1047,20 @@ yuv2rgba64_X_c_template(SwsContext *c, const int16_t *lumFilter,
         B =                            U * c->yuv2rgb_u2b_coeff;
 
         // 8 bits: 30 - 22 = 8 bits, 16 bits: 30 bits - 14 = 16 bits
-        output_pixel(&dest[0], av_clip_uintp2(R_B + Y1, 30) >> 14);
-        output_pixel(&dest[1], av_clip_uintp2(  G + Y1, 30) >> 14);
-        output_pixel(&dest[2], av_clip_uintp2(B_R + Y1, 30) >> 14);
+        output_pixel(&dest[0], av_clip_uintp2(((R_B + Y1) >> 14) + (1<<15), 16));
+        output_pixel(&dest[1], av_clip_uintp2(((  G + Y1) >> 14) + (1<<15), 16));
+        output_pixel(&dest[2], av_clip_uintp2(((B_R + Y1) >> 14) + (1<<15), 16));
         if (eightbytes) {
             output_pixel(&dest[3], av_clip_uintp2(A1      , 30) >> 14);
-            output_pixel(&dest[4], av_clip_uintp2(R_B + Y2, 30) >> 14);
-            output_pixel(&dest[5], av_clip_uintp2(  G + Y2, 30) >> 14);
-            output_pixel(&dest[6], av_clip_uintp2(B_R + Y2, 30) >> 14);
+            output_pixel(&dest[4], av_clip_uintp2(((R_B + Y2) >> 14) + (1<<15), 16));
+            output_pixel(&dest[5], av_clip_uintp2(((  G + Y2) >> 14) + (1<<15), 16));
+            output_pixel(&dest[6], av_clip_uintp2(((B_R + Y2) >> 14) + (1<<15), 16));
             output_pixel(&dest[7], av_clip_uintp2(A2      , 30) >> 14);
             dest += 8;
         } else {
-            output_pixel(&dest[3], av_clip_uintp2(R_B + Y2, 30) >> 14);
-            output_pixel(&dest[4], av_clip_uintp2(  G + Y2, 30) >> 14);
-            output_pixel(&dest[5], av_clip_uintp2(B_R + Y2, 30) >> 14);
+            output_pixel(&dest[3], av_clip_uintp2(((R_B + Y2) >> 14) + (1<<15), 16));
+            output_pixel(&dest[4], av_clip_uintp2(((  G + Y2) >> 14) + (1<<15), 16));
+            output_pixel(&dest[5], av_clip_uintp2(((B_R + Y2) >> 14) + (1<<15), 16));
             dest += 6;
         }
     }
@@ -1097,8 +1097,8 @@ yuv2rgba64_2_c_template(SwsContext *c, const int32_t *buf[2],
         Y2 -= c->yuv2rgb_y_offset;
         Y1 *= c->yuv2rgb_y_coeff;
         Y2 *= c->yuv2rgb_y_coeff;
-        Y1 += 1 << 13;
-        Y2 += 1 << 13;
+        Y1 += (1 << 13) - (1 << 29);
+        Y2 += (1 << 13) - (1 << 29);
 
         R = V * c->yuv2rgb_v2r_coeff;
         G = V * c->yuv2rgb_v2g_coeff + U * c->yuv2rgb_u2g_coeff;
@@ -1112,20 +1112,20 @@ yuv2rgba64_2_c_template(SwsContext *c, const int32_t *buf[2],
             A2 += 1 << 13;
         }
 
-        output_pixel(&dest[0], av_clip_uintp2(R_B + Y1, 30) >> 14);
-        output_pixel(&dest[1], av_clip_uintp2(  G + Y1, 30) >> 14);
-        output_pixel(&dest[2], av_clip_uintp2(B_R + Y1, 30) >> 14);
+        output_pixel(&dest[0], av_clip_uintp2(((R_B + Y1) >> 14) + (1<<15), 16));
+        output_pixel(&dest[1], av_clip_uintp2(((  G + Y1) >> 14) + (1<<15), 16));
+        output_pixel(&dest[2], av_clip_uintp2(((B_R + Y1) >> 14) + (1<<15), 16));
         if (eightbytes) {
             output_pixel(&dest[3], av_clip_uintp2(A1      , 30) >> 14);
-            output_pixel(&dest[4], av_clip_uintp2(R_B + Y2, 30) >> 14);
-            output_pixel(&dest[5], av_clip_uintp2(  G + Y2, 30) >> 14);
-            output_pixel(&dest[6], av_clip_uintp2(B_R + Y2, 30) >> 14);
+            output_pixel(&dest[4], av_clip_uintp2(((R_B + Y2) >> 14) + (1<<15), 16));
+            output_pixel(&dest[5], av_clip_uintp2(((  G + Y2) >> 14) + (1<<15), 16));
+            output_pixel(&dest[6], av_clip_uintp2(((B_R + Y2) >> 14) + (1<<15), 16));
             output_pixel(&dest[7], av_clip_uintp2(A2      , 30) >> 14);
             dest += 8;
         } else {
-            output_pixel(&dest[3], av_clip_uintp2(R_B + Y2, 30) >> 14);
-            output_pixel(&dest[4], av_clip_uintp2(  G + Y2, 30) >> 14);
-            output_pixel(&dest[5], av_clip_uintp2(B_R + Y2, 30) >> 14);
+            output_pixel(&dest[3], av_clip_uintp2(((R_B + Y2) >> 14) + (1<<15), 16));
+            output_pixel(&dest[4], av_clip_uintp2(((  G + Y2) >> 14) + (1<<15), 16));
+            output_pixel(&dest[5], av_clip_uintp2(((B_R + Y2) >> 14) + (1<<15), 16));
             dest += 6;
         }
     }
@@ -1153,8 +1153,8 @@ yuv2rgba64_1_c_template(SwsContext *c, const int32_t *buf0,
             Y2 -= c->yuv2rgb_y_offset;
             Y1 *= c->yuv2rgb_y_coeff;
             Y2 *= c->yuv2rgb_y_coeff;
-            Y1 += 1 << 13;
-            Y2 += 1 << 13;
+            Y1 += (1 << 13) - (1 << 29);
+            Y2 += (1 << 13) - (1 << 29);
 
             if (hasAlpha) {
                 A1 = abuf0[i * 2    ] << 11;
@@ -1168,20 +1168,20 @@ yuv2rgba64_1_c_template(SwsContext *c, const int32_t *buf0,
             G = V * c->yuv2rgb_v2g_coeff + U * c->yuv2rgb_u2g_coeff;
             B =                            U * c->yuv2rgb_u2b_coeff;
 
-            output_pixel(&dest[0], av_clip_uintp2(R_B + Y1, 30) >> 14);
-            output_pixel(&dest[1], av_clip_uintp2(  G + Y1, 30) >> 14);
-            output_pixel(&dest[2], av_clip_uintp2(B_R + Y1, 30) >> 14);
+            output_pixel(&dest[0], av_clip_uintp2(((R_B + Y1) >> 14) + (1<<15), 16));
+            output_pixel(&dest[1], av_clip_uintp2(((  G + Y1) >> 14) + (1<<15), 16));
+            output_pixel(&dest[2], av_clip_uintp2(((B_R + Y1) >> 14) + (1<<15), 16));
             if (eightbytes) {
                 output_pixel(&dest[3], av_clip_uintp2(A1      , 30) >> 14);
-                output_pixel(&dest[4], av_clip_uintp2(R_B + Y2, 30) >> 14);
-                output_pixel(&dest[5], av_clip_uintp2(  G + Y2, 30) >> 14);
-                output_pixel(&dest[6], av_clip_uintp2(B_R + Y2, 30) >> 14);
+                output_pixel(&dest[4], av_clip_uintp2(((R_B + Y2) >> 14) + (1<<15), 16));
+                output_pixel(&dest[5], av_clip_uintp2(((  G + Y2) >> 14) + (1<<15), 16));
+                output_pixel(&dest[6], av_clip_uintp2(((B_R + Y2) >> 14) + (1<<15), 16));
                 output_pixel(&dest[7], av_clip_uintp2(A2      , 30) >> 14);
                 dest += 8;
             } else {
-                output_pixel(&dest[3], av_clip_uintp2(R_B + Y2, 30) >> 14);
-                output_pixel(&dest[4], av_clip_uintp2(  G + Y2, 30) >> 14);
-                output_pixel(&dest[5], av_clip_uintp2(B_R + Y2, 30) >> 14);
+                output_pixel(&dest[3], av_clip_uintp2(((R_B + Y2) >> 14) + (1<<15), 16));
+                output_pixel(&dest[4], av_clip_uintp2(((  G + Y2) >> 14) + (1<<15), 16));
+                output_pixel(&dest[5], av_clip_uintp2(((B_R + Y2) >> 14) + (1<<15), 16));
                 dest += 6;
             }
         }
@@ -1199,8 +1199,8 @@ yuv2rgba64_1_c_template(SwsContext *c, const int32_t *buf0,
             Y2 -= c->yuv2rgb_y_offset;
             Y1 *= c->yuv2rgb_y_coeff;
             Y2 *= c->yuv2rgb_y_coeff;
-            Y1 += 1 << 13;
-            Y2 += 1 << 13;
+            Y1 += (1 << 13) - (1 << 29);
+            Y2 += (1 << 13) - (1 << 29);
 
             if (hasAlpha) {
                 A1 = abuf0[i * 2    ] << 11;
@@ -1214,20 +1214,20 @@ yuv2rgba64_1_c_template(SwsContext *c, const int32_t *buf0,
             G = V * c->yuv2rgb_v2g_coeff + U * c->yuv2rgb_u2g_coeff;
             B =                            U * c->yuv2rgb_u2b_coeff;
 
-            output_pixel(&dest[0], av_clip_uintp2(R_B + Y1, 30) >> 14);
-            output_pixel(&dest[1], av_clip_uintp2(  G + Y1, 30) >> 14);
-            output_pixel(&dest[2], av_clip_uintp2(B_R + Y1, 30) >> 14);
+            output_pixel(&dest[0], av_clip_uintp2(((R_B + Y1) >> 14) + (1<<15), 16));
+            output_pixel(&dest[1], av_clip_uintp2(((  G + Y1) >> 14) + (1<<15), 16));
+            output_pixel(&dest[2], av_clip_uintp2(((B_R + Y1) >> 14) + (1<<15), 16));
             if (eightbytes) {
                 output_pixel(&dest[3], av_clip_uintp2(A1      , 30) >> 14);
-                output_pixel(&dest[4], av_clip_uintp2(R_B + Y2, 30) >> 14);
-                output_pixel(&dest[5], av_clip_uintp2(  G + Y2, 30) >> 14);
-                output_pixel(&dest[6], av_clip_uintp2(B_R + Y2, 30) >> 14);
+                output_pixel(&dest[4], av_clip_uintp2(((R_B + Y2) >> 14) + (1<<15), 16));
+                output_pixel(&dest[5], av_clip_uintp2(((  G + Y2) >> 14) + (1<<15), 16));
+                output_pixel(&dest[6], av_clip_uintp2(((B_R + Y2) >> 14) + (1<<15), 16));
                 output_pixel(&dest[7], av_clip_uintp2(A2      , 30) >> 14);
                 dest += 8;
             } else {
-                output_pixel(&dest[3], av_clip_uintp2(R_B + Y2, 30) >> 14);
-                output_pixel(&dest[4], av_clip_uintp2(  G + Y2, 30) >> 14);
-                output_pixel(&dest[5], av_clip_uintp2(B_R + Y2, 30) >> 14);
+                output_pixel(&dest[3], av_clip_uintp2(((R_B + Y2) >> 14) + (1<<15), 16));
+                output_pixel(&dest[4], av_clip_uintp2(((  G + Y2) >> 14) + (1<<15), 16));
+                output_pixel(&dest[5], av_clip_uintp2(((B_R + Y2) >> 14) + (1<<15), 16));
                 dest += 6;
             }
         }
@@ -1278,7 +1278,7 @@ yuv2rgba64_full_X_c_template(SwsContext *c, const int16_t *lumFilter,
         // 8bit: 27 -> 17bit, 16bit: 31 - 14 = 17bit
         Y -= c->yuv2rgb_y_offset;
         Y *= c->yuv2rgb_y_coeff;
-        Y += 1 << 13; // 21
+        Y += (1 << 13) - (1<<29); // 21
         // 8bit: 17 + 13bit = 30bit, 16bit: 17 + 13bit = 30bit
 
         R = V * c->yuv2rgb_v2r_coeff;
@@ -1286,9 +1286,9 @@ yuv2rgba64_full_X_c_template(SwsContext *c, const int16_t *lumFilter,
         B =                            U * c->yuv2rgb_u2b_coeff;
 
         // 8bit: 30 - 22 = 8bit, 16bit: 30bit - 14 = 16bit
-        output_pixel(&dest[0], av_clip_uintp2(R_B + Y, 30) >> 14);
-        output_pixel(&dest[1], av_clip_uintp2(  G + Y, 30) >> 14);
-        output_pixel(&dest[2], av_clip_uintp2(B_R + Y, 30) >> 14);
+        output_pixel(&dest[0], av_clip_uintp2(((R_B + Y)>>14) + (1<<15), 16));
+        output_pixel(&dest[1], av_clip_uintp2(((  G + Y)>>14) + (1<<15), 16));
+        output_pixel(&dest[2], av_clip_uintp2(((B_R + Y)>>14) + (1<<15), 16));
         if (eightbytes) {
             output_pixel(&dest[3], av_clip_uintp2(A, 30) >> 14);
             dest += 4;
@@ -1326,7 +1326,7 @@ yuv2rgba64_full_2_c_template(SwsContext *c, const int32_t *buf[2],
 
         Y -= c->yuv2rgb_y_offset;
         Y *= c->yuv2rgb_y_coeff;
-        Y += 1 << 13;
+        Y += (1 << 13) - (1 << 29);
 
         R = V * c->yuv2rgb_v2r_coeff;
         G = V * c->yuv2rgb_v2g_coeff + U * c->yuv2rgb_u2g_coeff;
@@ -1338,9 +1338,9 @@ yuv2rgba64_full_2_c_template(SwsContext *c, const int32_t *buf[2],
             A += 1 << 13;
         }
 
-        output_pixel(&dest[0], av_clip_uintp2(R_B + Y, 30) >> 14);
-        output_pixel(&dest[1], av_clip_uintp2(  G + Y, 30) >> 14);
-        output_pixel(&dest[2], av_clip_uintp2(B_R + Y, 30) >> 14);
+        output_pixel(&dest[0], av_clip_uintp2(((R_B + Y) >> 14) + (1<<15), 16));
+        output_pixel(&dest[1], av_clip_uintp2(((  G + Y) >> 14) + (1<<15), 16));
+        output_pixel(&dest[2], av_clip_uintp2(((B_R + Y) >> 14) + (1<<15), 16));
         if (eightbytes) {
             output_pixel(&dest[3], av_clip_uintp2(A, 30) >> 14);
             dest += 4;
@@ -1369,7 +1369,7 @@ yuv2rgba64_full_1_c_template(SwsContext *c, const int32_t *buf0,
 
             Y -= c->yuv2rgb_y_offset;
             Y *= c->yuv2rgb_y_coeff;
-            Y += 1 << 13;
+            Y += (1 << 13) - (1 << 29);
 
             if (hasAlpha) {
                 A = abuf0[i] << 11;
@@ -1381,9 +1381,9 @@ yuv2rgba64_full_1_c_template(SwsContext *c, const int32_t *buf0,
             G = V * c->yuv2rgb_v2g_coeff + U * c->yuv2rgb_u2g_coeff;
             B =                            U * c->yuv2rgb_u2b_coeff;
 
-            output_pixel(&dest[0], av_clip_uintp2(R_B + Y, 30) >> 14);
-            output_pixel(&dest[1], av_clip_uintp2(  G + Y, 30) >> 14);
-            output_pixel(&dest[2], av_clip_uintp2(B_R + Y, 30) >> 14);
+            output_pixel(&dest[0], av_clip_uintp2(((R_B + Y) >> 14) + (1<<15), 16));
+            output_pixel(&dest[1], av_clip_uintp2(((  G + Y) >> 14) + (1<<15), 16));
+            output_pixel(&dest[2], av_clip_uintp2(((B_R + Y) >> 14) + (1<<15), 16));
             if (eightbytes) {
                 output_pixel(&dest[3], av_clip_uintp2(A, 30) >> 14);
                 dest += 4;
@@ -1402,7 +1402,7 @@ yuv2rgba64_full_1_c_template(SwsContext *c, const int32_t *buf0,
 
             Y -= c->yuv2rgb_y_offset;
             Y *= c->yuv2rgb_y_coeff;
-            Y += 1 << 13;
+            Y += (1 << 13) - (1 << 29);
 
             if (hasAlpha) {
                 A = abuf0[i] << 11;
@@ -1414,9 +1414,9 @@ yuv2rgba64_full_1_c_template(SwsContext *c, const int32_t *buf0,
             G = V * c->yuv2rgb_v2g_coeff + U * c->yuv2rgb_u2g_coeff;
             B =                            U * c->yuv2rgb_u2b_coeff;
 
-            output_pixel(&dest[0], av_clip_uintp2(R_B + Y, 30) >> 14);
-            output_pixel(&dest[1], av_clip_uintp2(  G + Y, 30) >> 14);
-            output_pixel(&dest[2], av_clip_uintp2(B_R + Y, 30) >> 14);
+            output_pixel(&dest[0], av_clip_uintp2(((R_B + Y) >> 14) + (1<<15), 16));
+            output_pixel(&dest[1], av_clip_uintp2(((  G + Y) >> 14) + (1<<15), 16));
+            output_pixel(&dest[2], av_clip_uintp2(((B_R + Y) >> 14) + (1<<15), 16));
             if (eightbytes) {
                 output_pixel(&dest[3], av_clip_uintp2(A, 30) >> 14);
                 dest += 4;

tests/checkasm/float_dsp.c

@@ -51,7 +51,8 @@ static void test_vector_fmul(const float *src0, const float *src1)
     call_ref(cdst, src0, src1, LEN);
     call_new(odst, src0, src1, LEN);
     for (i = 0; i < LEN; i++) {
-        if (!float_near_abs_eps(cdst[i], odst[i], FLT_EPSILON)) {
+        double t = fabs(src0[i]) + fabs(src1[i]) + fabs(src0[i] * src1[i]) + 1.0;
+        if (!float_near_abs_eps(cdst[i], odst[i], t * 2 * FLT_EPSILON)) {
             fprintf(stderr, "%d: %- .12f - %- .12f = % .12g\n",
                     i, cdst[i], odst[i], cdst[i] - odst[i]);
             fail();
@@ -73,7 +74,8 @@ static void test_vector_dmul(const double *src0, const double *src1)
     call_ref(cdst, src0, src1, LEN);
     call_new(odst, src0, src1, LEN);
     for (i = 0; i < LEN; i++) {
-        if (!double_near_abs_eps(cdst[i], odst[i], DBL_EPSILON)) {
+        double t = fabs(src0[i]) + fabs(src1[i]) + fabs(src0[i] * src1[i]) + 1.0;
+        if (!double_near_abs_eps(cdst[i], odst[i], t * 2 * DBL_EPSILON)) {
             fprintf(stderr, "%d: %- .12f - %- .12f = % .12g\n",
                     i, cdst[i], odst[i], cdst[i] - odst[i]);
             fail();
@@ -117,7 +119,8 @@ static void test_vector_fmul_scalar(const float *src0, const float *src1)
     call_ref(cdst, src0, src1[0], LEN);
     call_new(odst, src0, src1[0], LEN);
         for (i = 0; i < LEN; i++) {
-            if (!float_near_abs_eps(cdst[i], odst[i], FLT_EPSILON)) {
+            double t = fabs(src0[i]) + fabs(src1[0]) + fabs(src0[i] * src1[0]) + 1.0;
+            if (!float_near_abs_eps(cdst[i], odst[i], t * 2 * FLT_EPSILON)) {
                 fprintf(stderr, "%d: %- .12f - %- .12f = % .12g\n",
                         i, cdst[i], odst[i], cdst[i] - odst[i]);
                 fail();

tests/ref/seek/vsynth_lena-snow

@@ -2,45 +2,45 @@ ret: 0         st: 0 flags:1 dts: 0.000000 pts: 0.000000 pos:   5652 size:  3035
 ret: 0         st:-1 flags:0  ts:-1.000000
 ret: 0         st: 0 flags:1 dts: 0.000000 pts: 0.000000 pos:   5652 size:  3035
 ret: 0         st:-1 flags:1  ts: 1.894167
-ret: 0         st: 0 flags:1 dts: 1.440000 pts: 1.440000 pos:  39806 size:  3640
+ret: 0         st: 0 flags:1 dts: 1.440000 pts: 1.440000 pos:  39690 size:  3640
 ret: 0         st: 0 flags:0  ts: 0.800000
-ret: 0         st: 0 flags:1 dts: 0.960000 pts: 0.960000 pos:  27442 size:  3494
+ret: 0         st: 0 flags:1 dts: 0.960000 pts: 0.960000 pos:  27382 size:  3493
 ret:-1         st: 0 flags:1  ts:-0.320000
 ret:-1         st:-1 flags:0  ts: 2.576668
 ret: 0         st:-1 flags:1  ts: 1.470835
-ret: 0         st: 0 flags:1 dts: 1.440000 pts: 1.440000 pos:  39806 size:  3640
+ret: 0         st: 0 flags:1 dts: 1.440000 pts: 1.440000 pos:  39690 size:  3640
 ret: 0         st: 0 flags:0  ts: 0.360000
-ret: 0         st: 0 flags:1 dts: 0.480000 pts: 0.480000 pos:  16134 size:  3244
+ret: 0         st: 0 flags:1 dts: 0.480000 pts: 0.480000 pos:  16074 size:  3245
 ret:-1         st: 0 flags:1  ts:-0.760000
 ret:-1         st:-1 flags:0  ts: 2.153336
 ret: 0         st:-1 flags:1  ts: 1.047503
-ret: 0         st: 0 flags:1 dts: 0.960000 pts: 0.960000 pos:  27442 size:  3494
+ret: 0         st: 0 flags:1 dts: 0.960000 pts: 0.960000 pos:  27382 size:  3493
 ret: 0         st: 0 flags:0  ts:-0.040000
 ret: 0         st: 0 flags:1 dts: 0.000000 pts: 0.000000 pos:   5652 size:  3035
 ret: 0         st: 0 flags:1  ts: 2.840000
-ret: 0         st: 0 flags:1 dts: 1.920000 pts: 1.920000 pos:  52608 size:  3582
+ret: 0         st: 0 flags:1 dts: 1.920000 pts: 1.920000 pos:  52538 size:  3582
 ret: 0         st:-1 flags:0  ts: 1.730004
-ret: 0         st: 0 flags:1 dts: 1.920000 pts: 1.920000 pos:  52608 size:  3582
+ret: 0         st: 0 flags:1 dts: 1.920000 pts: 1.920000 pos:  52538 size:  3582
 ret: 0         st:-1 flags:1  ts: 0.624171
-ret: 0         st: 0 flags:1 dts: 0.480000 pts: 0.480000 pos:  16134 size:  3244
+ret: 0         st: 0 flags:1 dts: 0.480000 pts: 0.480000 pos:  16074 size:  3245
 ret: 0         st: 0 flags:0  ts:-0.480000
 ret: 0         st: 0 flags:1 dts: 0.000000 pts: 0.000000 pos:   5652 size:  3035
 ret: 0         st: 0 flags:1  ts: 2.400000
-ret: 0         st: 0 flags:1 dts: 1.920000 pts: 1.920000 pos:  52608 size:  3582
+ret: 0         st: 0 flags:1 dts: 1.920000 pts: 1.920000 pos:  52538 size:  3582
 ret: 0         st:-1 flags:0  ts: 1.306672
-ret: 0         st: 0 flags:1 dts: 1.440000 pts: 1.440000 pos:  39806 size:  3640
+ret: 0         st: 0 flags:1 dts: 1.440000 pts: 1.440000 pos:  39690 size:  3640
 ret: 0         st:-1 flags:1  ts: 0.200839
 ret: 0         st: 0 flags:1 dts: 0.000000 pts: 0.000000 pos:   5652 size:  3035
 ret: 0         st: 0 flags:0  ts:-0.920000
 ret: 0         st: 0 flags:1 dts: 0.000000 pts: 0.000000 pos:   5652 size:  3035
 ret: 0         st: 0 flags:1  ts: 2.000000
-ret: 0         st: 0 flags:1 dts: 1.920000 pts: 1.920000 pos:  52608 size:  3582
+ret: 0         st: 0 flags:1 dts: 1.920000 pts: 1.920000 pos:  52538 size:  3582
 ret: 0         st:-1 flags:0  ts: 0.883340
-ret: 0         st: 0 flags:1 dts: 0.960000 pts: 0.960000 pos:  27442 size:  3494
+ret: 0         st: 0 flags:1 dts: 0.960000 pts: 0.960000 pos:  27382 size:  3493
 ret:-1         st:-1 flags:1  ts:-0.222493
 ret:-1         st: 0 flags:0  ts: 2.680000
 ret: 0         st: 0 flags:1  ts: 1.560000
-ret: 0         st: 0 flags:1 dts: 1.440000 pts: 1.440000 pos:  39806 size:  3640
+ret: 0         st: 0 flags:1 dts: 1.440000 pts: 1.440000 pos:  39690 size:  3640
 ret: 0         st:-1 flags:0  ts: 0.460008
-ret: 0         st: 0 flags:1 dts: 0.480000 pts: 0.480000 pos:  16134 size:  3244
+ret: 0         st: 0 flags:1 dts: 0.480000 pts: 0.480000 pos:  16074 size:  3245
 ret:-1         st:-1 flags:1  ts:-0.645825

tests/ref/vsynth/vsynth1-snow

@@ -1,4 +1,4 @@
-67c10f8d52fcd1103caa675a1408bf6e *tests/data/fate/vsynth1-snow.avi
-136088 tests/data/fate/vsynth1-snow.avi
-bfc0bcc4bc7b956933aa58acc587018d *tests/data/fate/vsynth1-snow.out.rawvideo
-stddev:   22.77 PSNR: 20.98 MAXDIFF:  175 bytes:  7603200/  7603200
+c4c77a6fb926b89fe6591c398f5cd4db *tests/data/fate/vsynth1-snow.avi
+136160 tests/data/fate/vsynth1-snow.avi
+dcf8b3f62d9c3ae2b2d0fbbacbf83e4e *tests/data/fate/vsynth1-snow.out.rawvideo
+stddev:   22.74 PSNR: 20.99 MAXDIFF:  173 bytes:  7603200/  7603200

tests/ref/vsynth/vsynth1-snow-hpel

@@ -1,4 +1,4 @@
-e62ae25d5040d04622a965bcb27fdb1e *tests/data/fate/vsynth1-snow-hpel.avi
-138446 tests/data/fate/vsynth1-snow-hpel.avi
-57c914cd150f8fc260b5989ce3e5884c *tests/data/fate/vsynth1-snow-hpel.out.rawvideo
-stddev:   22.74 PSNR: 20.99 MAXDIFF:  172 bytes:  7603200/  7603200
+5c9eb93646eb0e5570d37e9adc9625e4 *tests/data/fate/vsynth1-snow-hpel.avi
+138580 tests/data/fate/vsynth1-snow-hpel.avi
+3382bdde624d8bb4af206a5ac6614605 *tests/data/fate/vsynth1-snow-hpel.out.rawvideo
+stddev:   22.71 PSNR: 21.00 MAXDIFF:  171 bytes:  7603200/  7603200

tests/ref/vsynth/vsynth2-snow

@@ -1,4 +1,4 @@
-0a41e73ddd2f54936490655b46dad4a3 *tests/data/fate/vsynth2-snow.avi
-72868 tests/data/fate/vsynth2-snow.avi
-34a75f5cf8a71159f1a572d9cedcfef9 *tests/data/fate/vsynth2-snow.out.rawvideo
-stddev:   13.73 PSNR: 25.37 MAXDIFF:  162 bytes:  7603200/  7603200
+5e130d6a48b69348eee7f7c76c5869a3 *tests/data/fate/vsynth2-snow.avi
+72942 tests/data/fate/vsynth2-snow.avi
+9b6cee60e3ec0d1f312a8a25a7878fcc *tests/data/fate/vsynth2-snow.out.rawvideo
+stddev:   13.39 PSNR: 25.59 MAXDIFF:  154 bytes:  7603200/  7603200

tests/ref/vsynth/vsynth2-snow-hpel

@@ -1,4 +1,4 @@
-9bc409e4794ee50691a26c9c836d31a7 *tests/data/fate/vsynth2-snow-hpel.avi
-79728 tests/data/fate/vsynth2-snow-hpel.avi
-2cc64d8171175a1532fd7d3ed3011fbf *tests/data/fate/vsynth2-snow-hpel.out.rawvideo
-stddev:   13.70 PSNR: 25.39 MAXDIFF:  162 bytes:  7603200/  7603200
+8edcf0fd7f066972ff77d5b891ed6dde *tests/data/fate/vsynth2-snow-hpel.avi
+79798 tests/data/fate/vsynth2-snow-hpel.avi
+7e0f2a24feda6fb3e54b85511a28c45f *tests/data/fate/vsynth2-snow-hpel.out.rawvideo
+stddev:   13.35 PSNR: 25.62 MAXDIFF:  157 bytes:  7603200/  7603200

tests/ref/vsynth/vsynth_lena-snow

@@ -1,4 +1,4 @@
-8e96f337e8f4ccac7d72ef517e1d2208 *tests/data/fate/vsynth_lena-snow.avi
-57680 tests/data/fate/vsynth_lena-snow.avi
-90963cfd2359d460001c94d94256dc2b *tests/data/fate/vsynth_lena-snow.out.rawvideo
-stddev:   10.48 PSNR: 27.72 MAXDIFF:  119 bytes:  7603200/  7603200
+bf2cf9cacc1d98388798be98872049ee *tests/data/fate/vsynth_lena-snow.avi
+57604 tests/data/fate/vsynth_lena-snow.avi
+707a42eb20195913be55ba8dfadf72fb *tests/data/fate/vsynth_lena-snow.out.rawvideo
+stddev:   10.37 PSNR: 27.81 MAXDIFF:  120 bytes:  7603200/  7603200

tests/ref/vsynth/vsynth_lena-snow-hpel

@@ -1,4 +1,4 @@
-56b14cb1cbb637536233982e87f7ac3e *tests/data/fate/vsynth_lena-snow-hpel.avi
-61764 tests/data/fate/vsynth_lena-snow-hpel.avi
-244b0266127fa354d8485234b2c388e4 *tests/data/fate/vsynth_lena-snow-hpel.out.rawvideo
-stddev:   10.45 PSNR: 27.74 MAXDIFF:  119 bytes:  7603200/  7603200
+c6ec87a11415a99b1a781f9f5bacb722 *tests/data/fate/vsynth_lena-snow-hpel.avi
+61814 tests/data/fate/vsynth_lena-snow-hpel.avi
+40f330397b7acf6bdbb3ec6d908be451 *tests/data/fate/vsynth_lena-snow-hpel.out.rawvideo
+stddev:   10.34 PSNR: 27.83 MAXDIFF:  118 bytes:  7603200/  7603200