liz/rawTherapee
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Fixed a bug where RT would crash sometimes after writing output file in batch queue. Plus small improvements to JPG thumbnails. Some little cleanups.

Browse Source
This commit is contained in:
Oliver Duis
2010-10-31 20:10:21 +01:00
parent c2a0a13f4e
commit 33a0c8174c
5 changed files with 18 additions and 21 deletions
Download Patch File Download Diff File Expand all files Collapse all files

11
rtengine/rtthumbnail.cc
View File

@@ -336,7 +336,7 @@ IImage8* Thumbnail::processImage (const procparams::ProcParams& params, int rhei
baseImg->b[i][j] = CLIP(val);
}
// appy highlight recovery, if needed
// apply highlight recovery, if needed
if (isRaw && params.hlrecovery.enabled) {
int maxval = 65535 / defGain;
if (params.hlrecovery.method=="Luminance" || params.hlrecovery.method=="Color")
@@ -623,6 +623,7 @@ void Thumbnail::transformPixel (int x, int y, int tran, int& tx, int& ty) {
ty/=scale;
}
// format: 1=8bit direct, 2=16bit direct, 3=JPG
bool Thumbnail::writeImage (const Glib::ustring& fname, int format) {
if (!thumbImg)
@@ -689,6 +690,14 @@ bool Thumbnail::writeImage (const Glib::ustring& fname, int format) {
cinfo.input_components = 3;
jpeg_set_defaults (&cinfo);
cinfo.write_JFIF_header = FALSE;
// compute optimal Huffman coding tables for the image. Bit slower to generate, but size of result image is a bit less (default was FALSE)
cinfo.optimize_coding = TRUE;
// Since math coprocessors are common these days, FLOAT should be a bit more accurate AND fast (default is ISLOW)
// (machine dependency is not really an issue, since we all run on x86 and having exactly the same file is not a requirement)
cinfo.dct_method = JDCT_FLOAT;
jpeg_set_quality (&cinfo, 85, true);
jpeg_start_compress(&cinfo, TRUE);
int rowlen = thumbImg->width*3;