ryan/mozjpeg
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Fix overflow issue #157

Browse Source 
DCT coefficients are clipped when preprocessing for deringing is used
as they can overflow
This commit is contained in:
Frank Bossen
2015-02-12 12:50:06 +01:00
parent 39b950076a
commit 02939f53a0
1 changed files with 31 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files

32
jcdctmgr.c
View File

@@ -22,6 +22,7 @@
#include "jpeglib.h" #include "jpeglib.h"
#include "jdct.h" /* Private declarations for DCT subsystem */ #include "jdct.h" /* Private declarations for DCT subsystem */
#include "jsimddct.h" #include "jsimddct.h"
#include "jchuff.h"
#include <assert.h> #include <assert.h>
#include <math.h> #include <math.h>
@@ -726,6 +727,17 @@ forward_DCT (j_compress_ptr cinfo, jpeg_component_info * compptr,
/* Quantize/descale the coefficients, and store into coef_blocks[] */ /* Quantize/descale the coefficients, and store into coef_blocks[] */
(*do_quantize) (coef_blocks[bi], divisors, workspace); (*do_quantize) (coef_blocks[bi], divisors, workspace);
if (do_preprocess) {
int i;
int maxval = (1 << MAX_COEF_BITS) - 1;
for (i = 0; i < 64; i++) {
if (coef_blocks[bi][i] < -maxval)
coef_blocks[bi][i] = -maxval;
if (coef_blocks[bi][i] > maxval)
coef_blocks[bi][i] = maxval;
}
}
} }
} }
@@ -845,12 +857,22 @@ forward_DCT_float (j_compress_ptr cinfo, jpeg_component_info * compptr,
/* Quantize/descale the coefficients, and store into coef_blocks[] */ /* Quantize/descale the coefficients, and store into coef_blocks[] */
(*do_quantize) (coef_blocks[bi], divisors, workspace); (*do_quantize) (coef_blocks[bi], divisors, workspace);
if (do_preprocess) {
int i;
int maxval = (1 << MAX_COEF_BITS) - 1;
for (i = 0; i < 64; i++) {
if (coef_blocks[bi][i] < -maxval)
coef_blocks[bi][i] = -maxval;
if (coef_blocks[bi][i] > maxval)
coef_blocks[bi][i] = maxval;
}
}
} }
} }
#endif /* DCT_FLOAT_SUPPORTED */ #endif /* DCT_FLOAT_SUPPORTED */
#include "jchuff.h"
#include "jpeg_nbits_table.h" #include "jpeg_nbits_table.h"
static const float jpeg_lambda_weights_flat[64] = { static const float jpeg_lambda_weights_flat[64] = {
@@ -997,6 +1019,11 @@ quantize_trellis(j_compress_ptr cinfo, c_derived_tbl *dctbl, c_derived_tbl *actb
int bits; int bits;
dc_candidate[k][bi] = qval - DC_TRELLIS_CANDIDATES/2 + k; dc_candidate[k][bi] = qval - DC_TRELLIS_CANDIDATES/2 + k;
if (dc_candidate[k][bi] >= (1<<MAX_COEF_BITS))
dc_candidate[k][bi] = (1<<MAX_COEF_BITS)-1;
if (dc_candidate[k][bi] <= -(1<<MAX_COEF_BITS))
dc_candidate[k][bi] = -(1<<MAX_COEF_BITS)+1;
delta = dc_candidate[k][bi] * q - x; delta = dc_candidate[k][bi] * q - x;
dc_candidate_dist = delta * delta * lambda_dc; dc_candidate_dist = delta * delta * lambda_dc;
dc_candidate[k][bi] *= 1 + 2*sign; dc_candidate[k][bi] *= 1 + 2*sign;
@@ -1076,6 +1103,9 @@ quantize_trellis(j_compress_ptr cinfo, c_derived_tbl *dctbl, c_derived_tbl *actb
continue; continue;
} }
if (qval >= (1<<MAX_COEF_BITS))
qval = (1<<MAX_COEF_BITS)-1;
num_candidates = jpeg_nbits_table[qval]; num_candidates = jpeg_nbits_table[qval];
for (k = 0; k < num_candidates; k++) { for (k = 0; k < num_candidates; k++) {
int delta; int delta;