Some JPEGs [1] have incorrect DHT entries that map 2 codes to
the same value.
The second (last) mapping does not ever actually appear in the
code stream, therefore ignoring any mappings after the first one
fixes this.
Without this, an "mjpeg_decode_dc: bad vlc: 0:0" error is thrown.
In all known files, the 2 codes are mapped to symbol 0 so only
that case is checked.
---
[1]: Embedded JPEGs in "X7 RAW" and "X7 CinemaDNG" samples here:
https://www.dji.com/gr/zenmuse-x7/info#downloads
Signed-off-by: Nick Renieris <velocityra at gmail.com>
---
libavcodec/jpegtables.c | 17 ++++++++++++++---
1 file changed, 14 insertions(+), 3 deletions(-)
diff --git a/libavcodec/jpegtables.c b/libavcodec/jpegtables.c
index cbe5523cb4..fa5c6f9fc5 100644
--- a/libavcodec/jpegtables.c
+++ b/libavcodec/jpegtables.c
@@ -130,14 +130,25 @@ void ff_mjpeg_build_huffman_codes(uint8_t *huff_size, uint16_t *huff_code,
{
int i, j, k,nb, code, sym;
- code = 0;
+ /* Some badly encoded files [1] map 2 different codes to symbol 0.
+ Only the first one is valid, so we zero-initialize this here and
+ make sure we only set it once (the first time) in the loop below.
+
+ [1]: Embedded JPEGs in "X7 RAW" and "X7 CinemaDNG" samples here:
+ https://www.dji.com/gr/zenmuse-x7/info#downloads
+ */
+ huff_size[0] = 0;
+
k = 0;
+ code = 0;
for(i=1;i<=16;i++) {
nb = bits_table[i];
for(j=0;j<nb;j++) {
sym = val_table[k++];
- huff_size[sym] = i;
- huff_code[sym] = code;
+ if (sym != 0 || huff_size[sym] == 0) { /* see comment above */
+ huff_size[sym] = i;
+ huff_code[sym] = code;
+ }
code++;
}
code <<= 1;
--
2.17.1