[Bf-blender-cvs] SVN commit: /data/svn/bf-blender [58269] trunk/blender/source/blender/imbuf /intern/cineon/logImageCore.c: Fix #36127: 10 bit DPX render would crash when doing rendering animations from
Brecht Van Lommel
brechtvanlommel at pandora.be
Mon Jul 15 16:33:19 CEST 2013
Revision: 58269
http://projects.blender.org/scm/viewvc.php?view=rev&root=bf-blender&revision=58269
Author: blendix
Date: 2013-07-15 14:33:19 +0000 (Mon, 15 Jul 2013)
Log Message:
-----------
Fix #36127: 10 bit DPX render would crash when doing rendering animations from
the UI on OS X.
The problem is due to a large variable on the stack, and pthreads have a smaller
stack size than the main thread by default. On Linux the pthread stack size seems
to be 2MB, OS X 512KB and Windows 256KB - 512KB.
Modified Paths:
--------------
trunk/blender/source/blender/imbuf/intern/cineon/logImageCore.c
Modified: trunk/blender/source/blender/imbuf/intern/cineon/logImageCore.c
===================================================================
--- trunk/blender/source/blender/imbuf/intern/cineon/logImageCore.c 2013-07-15 14:19:02 UTC (rev 58268)
+++ trunk/blender/source/blender/imbuf/intern/cineon/logImageCore.c 2013-07-15 14:33:19 UTC (rev 58269)
@@ -918,25 +918,35 @@
}
}
-static void getLinToLogLut(float *lut, LogImageFile *logImage, LogImageElement logElement)
+static float *getLinToLogLut(LogImageFile *logImage, LogImageElement logElement)
{
+ float *lut;
float gain, negativeFilmGamma, offset, step;
+ unsigned int lutsize = (unsigned int)(logElement.maxValue + 1);
unsigned int i;
+
+ lut = MEM_mallocN(sizeof(float)*lutsize, "getLinToLogLut");
negativeFilmGamma = 0.6;
step = logElement.refHighQuantity / logElement.maxValue;
gain = logElement.maxValue / (1.0f - powf(10, (logImage->referenceBlack - logImage->referenceWhite) * step / negativeFilmGamma * logImage->gamma / 1.7f));
offset = gain - logElement.maxValue;
- for (i = 0; i < (int)(logElement.maxValue + 1); i++)
+ for (i = 0; i < lutsize; i++)
lut[i] = (logImage->referenceWhite + log10f(powf((i + offset) / gain, 1.7f / logImage->gamma)) / (step / negativeFilmGamma)) / logElement.maxValue;
+
+ return lut;
}
-static void getLogToLinLut(float *lut, LogImageFile *logImage, LogImageElement logElement)
+static float *getLogToLinLut(LogImageFile *logImage, LogImageElement logElement)
{
+ float *lut;
float breakPoint, gain, kneeGain, kneeOffset, negativeFilmGamma, offset, step, softClip;
/* float filmGamma; unused */
+ unsigned int lutsize = (unsigned int)(logElement.maxValue + 1);
unsigned int i;
+
+ lut = MEM_mallocN(sizeof(float)*lutsize, "getLogToLinLut");
/* Building the Log -> Lin LUT */
step = logElement.refHighQuantity / logElement.maxValue;
@@ -952,7 +962,7 @@
kneeOffset = powf(10, (breakPoint - logImage->referenceWhite) * step / negativeFilmGamma * logImage->gamma / 1.7f) * gain - offset;
kneeGain = (logElement.maxValue - kneeOffset) / powf(5 * softClip, softClip / 100);
- for (i = 0; i < (int)(logElement.maxValue + 1); i++) {
+ for (i = 0; i < lutsize; i++) {
if (i < logImage->referenceBlack)
lut[i] = 0.0f;
else if (i > breakPoint)
@@ -960,61 +970,76 @@
else
lut[i] = (powf(10, ((float)i - logImage->referenceWhite) * step / negativeFilmGamma * logImage->gamma / 1.7f) * gain - offset) / logElement.maxValue;
}
+
+ return lut;
}
-static void getLinToSrgbLut(float *lut, LogImageElement logElement)
+static float *getLinToSrgbLut(LogImageElement logElement)
{
+ float col, *lut;
+ unsigned int lutsize = (unsigned int)(logElement.maxValue + 1);
unsigned int i;
- float col;
- for (i = 0; i < (int)(logElement.maxValue + 1); i++) {
+ lut = MEM_mallocN(sizeof(float)*lutsize, "getLogToLinLut");
+
+ for (i = 0; i < lutsize; i++) {
col = (float)i / logElement.maxValue;
if (col < 0.0031308f)
lut[i] = (col < 0.0f) ? 0.0f : col * 12.92f;
else
lut[i] = 1.055f * powf(col, 1.0f / 2.4f) - 0.055f;
}
+
+ return lut;
}
-static void getSrgbToLinLut(float *lut, LogImageElement logElement)
+static float *getSrgbToLinLut(LogImageElement logElement)
{
+ float col, *lut;
+ unsigned int lutsize = (unsigned int)(logElement.maxValue + 1);
unsigned int i;
- float col;
- for (i = 0; i < (int)(logElement.maxValue + 1); i++) {
+ lut = MEM_mallocN(sizeof(float)*lutsize, "getLogToLinLut");
+
+ for (i = 0; i < lutsize; i++) {
col = (float)i / logElement.maxValue;
if (col < 0.04045f)
lut[i] = (col < 0.0f) ? 0.0f : col * (1.0f / 12.92f);
else
lut[i] = powf((col + 0.055f) * (1.0f / 1.055f), 2.4f);
}
+
+ return lut;
}
static int convertRGBA_RGB(float *src, float *dst, LogImageFile *logImage,
LogImageElement logElement, int elementIsSource)
{
unsigned int i;
- float lut[65536];
float *src_ptr = src;
float *dst_ptr = dst;
switch (logElement.transfer) {
case transfer_UserDefined:
case transfer_Linear:
- case transfer_Logarithmic:
+ case transfer_Logarithmic: {
for (i = 0; i < logImage->width * logImage->height; i++) {
*(dst_ptr++) = *(src_ptr++);
*(dst_ptr++) = *(src_ptr++);
*(dst_ptr++) = *(src_ptr++);
src_ptr++;
}
+
return 0;
+ }
- case transfer_PrintingDensity:
+ case transfer_PrintingDensity: {
+ float *lut;
+
if (elementIsSource == 1)
- getLogToLinLut((float *)&lut, logImage, logElement);
+ lut = getLogToLinLut(logImage, logElement);
else
- getLinToLogLut((float *)&lut, logImage, logElement);
+ lut = getLinToLogLut(logImage, logElement);
for (i = 0; i < logImage->width * logImage->height; i++) {
*(dst_ptr++) = lut[float_uint(*(src_ptr++), logElement.maxValue)];
@@ -1022,7 +1047,11 @@
*(dst_ptr++) = lut[float_uint(*(src_ptr++), logElement.maxValue)];
src_ptr++;
}
+
+ MEM_freeN(lut);
+
return 0;
+ }
default:
return 1;
@@ -1033,27 +1062,30 @@
LogImageElement logElement, int elementIsSource)
{
unsigned int i;
- float lut[65536];
float *src_ptr = src;
float *dst_ptr = dst;
switch (logElement.transfer) {
case transfer_UserDefined:
case transfer_Linear:
- case transfer_Logarithmic:
+ case transfer_Logarithmic: {
for (i = 0; i < logImage->width * logImage->height; i++) {
*(dst_ptr++) = *(src_ptr++);
*(dst_ptr++) = *(src_ptr++);
*(dst_ptr++) = *(src_ptr++);
*(dst_ptr++) = 1.0f;
}
+
return 0;
+ }
- case transfer_PrintingDensity:
+ case transfer_PrintingDensity: {
+ float *lut;
+
if (elementIsSource == 1)
- getLogToLinLut((float *)&lut, logImage, logElement);
+ lut = getLogToLinLut(logImage, logElement);
else
- getLinToLogLut((float *)&lut, logImage, logElement);
+ lut = getLinToLogLut(logImage, logElement);
for (i = 0; i < logImage->width * logImage->height; i++) {
*(dst_ptr++) = lut[float_uint(*(src_ptr++), logElement.maxValue)];
@@ -1061,7 +1093,11 @@
*(dst_ptr++) = lut[float_uint(*(src_ptr++), logElement.maxValue)];
*(dst_ptr++) = 1.0f;
}
+
+ MEM_freeN(lut);
+
return 0;
+ }
default:
return 1;
@@ -1072,22 +1108,24 @@
LogImageElement logElement, int elementIsSource)
{
unsigned int i;
- float lut[65536];
float *src_ptr = src;
float *dst_ptr = dst;
switch (logElement.transfer) {
case transfer_UserDefined:
case transfer_Linear:
- case transfer_Logarithmic:
+ case transfer_Logarithmic: {
memcpy(dst, src, 4 * logImage->width * logImage->height * sizeof(float));
return 0;
+ }
- case transfer_PrintingDensity:
+ case transfer_PrintingDensity: {
+ float *lut;
+
if (elementIsSource == 1)
- getLogToLinLut((float *)&lut, logImage, logElement);
+ lut = getLogToLinLut(logImage, logElement);
else
- getLinToLogLut((float *)&lut, logImage, logElement);
+ lut = getLinToLogLut(logImage, logElement);
for (i = 0; i < logImage->width * logImage->height; i++) {
*(dst_ptr++) = lut[float_uint(*(src_ptr++), logElement.maxValue)];
@@ -1095,7 +1133,11 @@
*(dst_ptr++) = lut[float_uint(*(src_ptr++), logElement.maxValue)];
*(dst_ptr++) = *(src_ptr++);
}
+
+ MEM_freeN(lut);
+
return 0;
+ }
default:
return 1;
@@ -1106,14 +1148,13 @@
LogImageElement logElement, int elementIsSource)
{
unsigned int i;
- float lut[65536];
float *src_ptr = src;
float *dst_ptr = dst;
switch (logElement.transfer) {
case transfer_UserDefined:
case transfer_Linear:
- case transfer_Logarithmic:
+ case transfer_Logarithmic: {
for (i = 0; i < logImage->width * logImage->height; i++) {
src_ptr += 4;
*(dst_ptr++) = *(src_ptr--);
@@ -1123,12 +1164,15 @@
src_ptr += 4;
}
return 0;
+ }
- case transfer_PrintingDensity:
+ case transfer_PrintingDensity: {
+ float *lut;
+
if (elementIsSource == 1)
- getLogToLinLut((float *)&lut, logImage, logElement);
+ lut = getLogToLinLut(logImage, logElement);
else
- getLinToLogLut((float *)&lut, logImage, logElement);
+ lut = getLinToLogLut(logImage, logElement);
for (i = 0; i < logImage->width * logImage->height; i++) {
src_ptr += 4;
@@ -1138,7 +1182,11 @@
*(dst_ptr++) = *(src_ptr--);
src_ptr += 4;
}
+
+ MEM_freeN(lut);
+
return 0;
+ }
default:
return 1;
@@ -1309,7 +1357,6 @@
unsigned int i;
float *src_ptr;
float *dst_ptr;
- float lut[65536];
/* Convert data in src to linear RGBA in dst */
switch (logElement.descriptor) {
@@ -1357,7 +1404,7 @@
return 1;
else if (dstIsLinearRGB) {
/* convert data from sRGB to Linear RGB via lut */
- getSrgbToLinLut((float *)&lut, logElement);
+ float *lut = getSrgbToLinLut(logElement);
src_ptr = dst; // no error here
dst_ptr = dst;
for (i = 0; i < logImage->width * logImage->height; i++) {
@@ -1366,6 +1413,7 @@
*(dst_ptr++) = lut[float_uint(*(src_ptr++), logElement.maxValue)];
dst_ptr++; src_ptr++;
}
+ MEM_freeN(lut);
}
return 0;
}
@@ -1378,7 +1426,7 @@
float *srgbSrc;
float *srgbSrc_ptr;
float *src_ptr = src;
- float lut[65536];
+ float *lut;
if (srcIsLinearRGB != 0) {
/* we need to convert src to sRGB */
@@ -1390,13 +1438,14 @@
srgbSrc_ptr = srgbSrc;
/* convert data from Linear RGB to sRGB via lut */
- getLinToSrgbLut((float *)&lut, logElement);
+ lut = getLinToSrgbLut(logElement);
for (i = 0; i < logImage->width * logImage->height; i++) {
*(srgbSrc_ptr++) = lut[float_uint(*(src_ptr++), logElement.maxValue)];
*(srgbSrc_ptr++) = lut[float_uint(*(src_ptr++), logElement.maxValue)];
*(srgbSrc_ptr++) = lut[float_uint(*(src_ptr++), logElement.maxValue)];
srgbSrc_ptr++; src_ptr++;
}
+ MEM_freeN(lut);
}
else
srgbSrc = src;
More information about the Bf-blender-cvs
mailing list