I just got my Epson Stylus Pro 7900 with the SpectroProofer option. One thing the SpectroProofer enables is bypassing everything to talk to the printer directly. I made up a sample target (based on Bill Atkinson’s RGB 5×12 Hues Target) and printed/measured this with the SpectroProofer. I then took the output printed by the SpectroProofer and measured this manually with MeasureTool and my Eye-One Pro (Rev B). The average/worst correlation was 1.00/2.02 Delta E 2000, good enough for me given two different instruments and the SpectroProofer patches are a tad narrow for the Eye-One.
I then printed the same target with Photoshop CS3 and No Color Management and this compares well with the SpectroProofer output (0.69/1.90). So I have confidence that CS3 is passing the data directly.
I then printed the target again with Photoshop CS4 and No Color Management, same driver settings and get a CS3/CS4 difference of 3.22/10.25. I repeated the CS4 readings and get 3.21/10.29.
The odd man out here is CS4. The differences between the CS3 and CS4 printouts (remember both with No Color Management) are clearly visible. GIven that the Epson 7900 drivers are brand spanking new, I’m not expecting architectural changes in the short term so I plan to stick with CS3 for profiling/printing.
You may want to explore the solution given and explained by Eric Chan:
Eric Chan, "PS3 prints fine, PS4 prints too dark, same Mac" #12, 12 Dec 2008 6:07 am </webx?14/11>
Eric Chan – 6:07am Dec 12, 08 PST (#12 of 16) Edited: 12-Dec-2008 at 06:07am
Camera Raw Engineer
Essentially, with the newer Leopard APIs, the recommended approach is for a printer driver to register its printer profiles to the system (i.e., Leopard). That is, the driver needs to provide a default profile to Leopard so that Leopard can provide the appropriate color matching. The default profile provided usually is a function of the color mode (e.g., RGB color or black-and-white), media type, resolution, etc.
CS4 was updated to use the newer printing APIs (a requirement moving forward) and hence relies on the driver providing the correct default profile. If the driver doesn’t do this, then Leopard will provide its own default profile, which is obviously not the correct printing profile and will have the side effect of making prints appear too dark and possibly with a strong color cast.
(BTW, this is exactly what was going on with some Canon iPF printers recently with many users blaming Adobe for introducing printing bugs in CS4, when the real issue was the driver not correctly providing the default profile to Leopard. The issue was resolved a couple of days ago when Canon released an updated driver.)
The problem appears to be Leopard indeed, ironically because CS4 was optimized for Leopard, why early versions are not.
The solution/workaround has worked for several users already, and it’s certainly easy to implement.
1. Quit Photoshop 2. Go to System Preferences > Print & Fax 3. Set the printer you are going to use as the Default Printer 4. Close the System Preferences and open Photoshop
This is very interesting. I have a new 9900 on the way and am about to upgrade to a new MacPro and CS4. It sounds like there’s no way to send raw data through the driver. Not good for making profiles at at. This could be the beginning of another printing nightmare. Great.
It would be nice to have an official data flow chart from Apple, Adobe and Epson to diagram what goes where. I think we need to see one printer at a time before jumping to the conclusion that it affects them all. In a week or so I’ll have both a 9800 and 9900 in the studio, along with the latest hardware/software. I don’t have the SpectroProof version coming as I already have a standalone Spectrolino but I will be able to directly compare Tiger to Leopard on these two printers. I also have contacts at Epson that I’ll be talking to directly. We’ll see what they say too.
If Canon can fix this Epson can too. But while your at it I have never understood why Epson does not provide a print plugin for PS. Seems they are at the mercy of all this, Adobe, Apple, Printer/Driver silliness.
I’m also curious if anyone at Adobe even noticed this, and if they did, was anything said. To be sure, they wouldn’t have had the latest printer, but they did have the latest OS.
It may also be the case that you actually can profile it anyway the way it is, but those profiles would only work under that specific condition – OS, printer driver, etc. Oh, but why confuse something even more that already confuses the hell out of most people.
"It may also be the case that you actually can profile it anyway the way it is, but those profiles would only work under that specific condition – OS, printer driver, etc."
You definitely wouldn’t want to do this. The intention is see what gamut etc you can get from the printer/media, not some distorted path. I have a 4800 sitting here and could repeat the test but I’m not prepared to waste any more time with CS4, at least until it’s updated or Epson announce CS4 compatible drivers.
BTW, you’ll notice a huge change in the linearization curves between the x800 and x900, with the new printers having a significantly more perceptual linearization. The new printers really excel at pastels so it’s understandable that Epson would want to open up this region.
I figured it out. I had brought over the old twain driver from CS3 and this was causing it. Once I installed the twain driver from the extras in CS4 it was fine.
