← All news
Seshat 1.8.1 converts to CMYK exactly as the separations print, tested against the Ghent PDF Output Suite
Seshat 1.8.1 converts RGB, Lab and spot colours to CMYK with overprint, spot plates, rendering intents and the file's own output intent kept, and prints every patch of the Ghent PDF Output Suite 5.0 test pages as separated.
By Rekhu
2 min read
A conversion to CMYK is only right when the converted file prints as the original's separations would: where colours overprint, where a spot colour lies on CMYK, where a process colour knocks out. Seshat 1.8.1 does that, and is tested against the Ghent PDF Output Suite 5.0, the print industry's own test pages for PDF/X workflows.
What changed
- Every patch of the Ghent suite passes after conversion. Its six pages of overprint, spot colour, DeviceN, transparency, colour management, 16-bit and JPEG 2000 patches print the same converted as separated, with no cross on any patch. This is our own test with the suite's public test pages, not a certification by the Ghent Workgroup.
- Spot colours become their plate. A spot colour converted to CMYK is laid over the CMYK plates as inks mix on press: one that overprints adds to what is below, one that knocks out replaces it. Its colour comes from its own definition in the file.
- Overprint keeps its meaning. Process inks in DeviceN, overprint mode and shadings that overprint print as before; an RGB colour that overprints still knocks out, as it does on press.
- The file's own press. Without a profile in the order, Seshat converts for the file's own output intent. The plates and proof are made for it too.
- CMYK stays exact when it means the same colours. CMYK made for the same printing condition keeps its values, also through another maker's profile (FOGRA39 by Adobe or by the ECI). CMYK made for another printing condition is converted, as PDF/X asks.
- Each object's rendering intent is used: perceptual, relative colorimetric, saturation or absolute.
- Smaller files. What the conversion replaced is taken out of the file.
Also in this version
- Spelling now uses the same dictionaries as LibreOffice and Firefox, which know every form of a word, in seven languages; a word another of those languages knows is not reported.
- Faster checks of files with many small shapes.
- A compression a picture uses that Seshat cannot unpack itself is no longer reported as a damaged picture.
Nothing changes in your code: convert_cmyk and convert_profile work as before.