thanks but I was unable to clarify my whole point: only CMYK conversion issue. My bad
Actually I finished the biggest job, using Gimp only for path selection & special curves, and combining 3 layers /photos
Before sending a tiff to a print service, I used 305glossyfujisupreme icc only for softproofing. The printed pics are way way
better than I expected, since the softproof was very similar but worse than the isocoatedv2eci softproof. Which is strange and suggests doubts
about softproofing as well. Btw, a bug in 2.10.38: it remains stuck with the last softproof. So you must edit/preferences etc.Close Gimp. Restart Gimp.
So all ok? yes but with a good service, and with a RGB printer.
The publisher wants a CMYK Fogra39 file. AND they use tens of print services, with different machines: digital printers.
AND Fogra39 is obsolete and meant for offset printing only.
So, no I never expected a good result from a RGB to CMYK conversion. Quite on the contrary, I was awfully worried. Now much more.
I trust https://www.textcompare.org/image/ after a huge number of tests
But, Rich2005 made me think of a further test with Gimp, layers, and ...I was getting crazy: files almost identical! Then, however Gimp converts to srgb!!
So I did many further tests with https://www.diffchecker.com/ https://www.img2go.com/compare-image https://www.imagetools.org/compare
Well the results are not 100% identical among all these tools, but maybe one says 83% another 95%, or one 98% vs 94% : difference between the files to compare.
(btw, Rich2005, you might want to dig into imagemagik: could it be a member of the company? no idea)
Let's divide RGBs in 2 groups:
Monitor RGBs: Adobe1998, NikonAdobe 4...,Nikon srgb 4..., IEC srgb..., srgb elle v2 etc etc.
Printer RGBs: e.g. 305glossyfujisupreme and a bunch of Saaldigital icc s, from basiccolor Germany
My findings, per type of conversion:
1 across monitor RGBs: zero or less than 1% difference. But ... yes icc tag changed!
2 from monitor to printer RGBs: maybe 20% difference
3 from monitor RGBs to CMYKs : Conversion done yes. But between tif and jpg output: difference up to 99%. Id est: TOTALLY different.
At 80, an obsolete logic:
1 maybe all the fuss on color spaces, conversion can mess colors etc.etc. is "much ado about nothing". Or, maybe, in nice words, the conversion
by krita/cyan/lcms is broken. In rude words, pieces of junk. for conversion
2 I have no printer to test
3 UFO thing. Quality unknown: tif ok, jpg junk? OR jpg ok, tif junk? OR jpg and tif 50% junk?
Plus, serious doubts on softproofing as well. But its another piece of cake.
I am curious yes, but right now I think that a full explanation is anyway unlikely
For me somebody who had a file converted to CMYK (possibly isocoatedv2eci) and printed successfully, would be a real help
But this is Gimp. There is another forum more inclined to colour conversion, including CMYK ?
(Photoshop: (sure this is a hard job: can it be trusted?) install and learning new stuff, which I'll never use again )
Well, maybe the whole point is a bit more clear now. Or is it a... mess
enri
Actually I finished the biggest job, using Gimp only for path selection & special curves, and combining 3 layers /photos
Before sending a tiff to a print service, I used 305glossyfujisupreme icc only for softproofing. The printed pics are way way
better than I expected, since the softproof was very similar but worse than the isocoatedv2eci softproof. Which is strange and suggests doubts
about softproofing as well. Btw, a bug in 2.10.38: it remains stuck with the last softproof. So you must edit/preferences etc.Close Gimp. Restart Gimp.
So all ok? yes but with a good service, and with a RGB printer.
The publisher wants a CMYK Fogra39 file. AND they use tens of print services, with different machines: digital printers.
AND Fogra39 is obsolete and meant for offset printing only.
So, no I never expected a good result from a RGB to CMYK conversion. Quite on the contrary, I was awfully worried. Now much more.
I trust https://www.textcompare.org/image/ after a huge number of tests
But, Rich2005 made me think of a further test with Gimp, layers, and ...I was getting crazy: files almost identical! Then, however Gimp converts to srgb!!
So I did many further tests with https://www.diffchecker.com/ https://www.img2go.com/compare-image https://www.imagetools.org/compare
Well the results are not 100% identical among all these tools, but maybe one says 83% another 95%, or one 98% vs 94% : difference between the files to compare.
(btw, Rich2005, you might want to dig into imagemagik: could it be a member of the company? no idea)
Let's divide RGBs in 2 groups:
Monitor RGBs: Adobe1998, NikonAdobe 4...,Nikon srgb 4..., IEC srgb..., srgb elle v2 etc etc.
Printer RGBs: e.g. 305glossyfujisupreme and a bunch of Saaldigital icc s, from basiccolor Germany
My findings, per type of conversion:
1 across monitor RGBs: zero or less than 1% difference. But ... yes icc tag changed!
2 from monitor to printer RGBs: maybe 20% difference
3 from monitor RGBs to CMYKs : Conversion done yes. But between tif and jpg output: difference up to 99%. Id est: TOTALLY different.
At 80, an obsolete logic:
1 maybe all the fuss on color spaces, conversion can mess colors etc.etc. is "much ado about nothing". Or, maybe, in nice words, the conversion
by krita/cyan/lcms is broken. In rude words, pieces of junk. for conversion
2 I have no printer to test
3 UFO thing. Quality unknown: tif ok, jpg junk? OR jpg ok, tif junk? OR jpg and tif 50% junk?
Plus, serious doubts on softproofing as well. But its another piece of cake.
I am curious yes, but right now I think that a full explanation is anyway unlikely
For me somebody who had a file converted to CMYK (possibly isocoatedv2eci) and printed successfully, would be a real help
But this is Gimp. There is another forum more inclined to colour conversion, including CMYK ?
(Photoshop: (sure this is a hard job: can it be trusted?) install and learning new stuff, which I'll never use again )
Well, maybe the whole point is a bit more clear now. Or is it a... mess
enri