Color Space Reconsideration for Epson 2200

GM
Posted By
Glenn_Mitchell
Dec 12, 2003
Views
453
Replies
10
Status
Closed
I’ve been reconsidering my preference for editing in AdobeRGB color space.

I was persuaded by a color engineer this week that the graphs of the gamut for the Epson 2200 around the web are erroneous. The evidence is that the Ultrachome inks do have a very wide gamut for pigmented inks and the dye inks in the 1280 have an even wider gamut, but the gamuts of both are within the Colormatch color space.

I was also told that under the hood, the Epson driver for the 2200 uses sRGB. It clips the heck out of greens, for example. That was a big surprise!

The filters for all of the current DSLRs are also designed for sRGB, not wider color spaces like AdobeRGB. The fact that RAW converts can apply a profile for AdobeRGB or some other color space does not mean the cameras have a gamut that even approaches those color spaces.

The ImagePrint RIP gives a wider gamut on the 2200 by bypassing the Epson driver. IP can print about 15,000 colors that the Epson driver cannot print. An added benefit from using IP as my RIP. The gamut is still within the Colormatch color space.

So, all of this now has me reconsidering my color space decisions. I am inclined to use AdobeRGB for archiving my images (i.e., the color space used by my RAW TIFFs) but then converting to ColorMatch for my editing work space.

The idea is this . . . If you use AdobeRGB and then print using a profile for your output, the profile is likely converting a lot of out of gamut colors (especially if you use Perceptual rendering intent). That is definitely not what output profiles are intended to accomplish. You can get posterization and unexpected color shifts as a result.

There really is no advantage to editing in a color space that’s broader than the gamut of your intended output device, and there are some important disadvantages.

I’ll be interested to hear other thoughts.

Cheers,

Mitch
Related Tags
color epson space

How to Improve Photoshop Performance

Learn how to optimize Photoshop for maximum speed, troubleshoot common issues, and keep your projects organized so that you can work faster than ever before!

PC
Philo_Calhoun
Dec 12, 2003
Yes, Epson 2200 has a driver for sRGB but not Adobe RGB, but you get a much more accurate print if you use a custom camera profile rather than try to use the sRGB driver. sRGB will clip the greens, as you suggest.

Colormatch is probably closer than Adobe RGB to the printers gamut, but if you reduce your image to Colormatch you won’t be able to get that back for any other output that might have a wider gamut.
You can easily preview the image in the printer’s gamut (closely, but not exactly due to monitor gamut issues) with view>proof and selecting the custom printer profile. You can even specify the look with different rendering intents this way. You don’t need to convert to Colormatch to see how it would look with a different gamut.

I don’t agree with DSLR’s having approximately the gamut of sRGB. This is not true for a D1x.
GM
Glenn_Mitchell
Dec 12, 2003
The engineer I spoke to was John at ColorByte Software (the manufacturer of the ImagePrint RIP). He made the comment about DSLRs have filters optimized for sRGB.

His claim was that while DSLRs might support conversion to other color working spaces, their gamuts were designed around sRGB.

Now, I do not claim the expertise to contest his expert opinion. Perhaps you do, and I’ll be very interested in your comment if you do have the expertise to challenge John’s comment. 😉

RAW files have no color space profile applied. RAW conversion software uses the tag in the RAW file as a default, but the color space conversion occurs during RAW conversion.

I’m not suggesting a conversion to Colormatch to see how my output will look. I use soft proofing.

I’m instead suggesting the use of Colormatch as my editing color space. The conversion would be done up front by the RAW converter. Using ColorMatch for editing would avoid more out of gamut color conversions when ready to print with a profile.

I was concerned that I would lose many saturated colors by editing in Colormatch rather than AdobeRGB. That would be true only if the Epson 2200 could print those additional colors. John was quite adamant it could not, since the Epson 2200 (and even the Epson 1280) do not have a gamut wider than the Colormatch color space.

Cheers,

Mitch
JS
Jeff_Schewe
Dec 13, 2003
John is simply wrong-in fact the 2200 even using the Epson drivers does indeed have gamut outside of even Adobe RGB. . .not a lot mind you, but several colors peak outside of any monitor space. And if he’s talking about the bonded tri-color separation filters on the camera chips, again, he’s wrong. The separation filters are optimised for the sensor. The built in jpg algorithyms may be designed primarily for sRGB and not Adobe RGB since up till just recently, there were only two color conditions in EXIF color spaces, sRGB and unspecified.

Not sure why John would be saying what he’s saying, or what YOU are saying he’s saying but he might want to check with Bruce Fraser or Bill Atkinson before making such statements.
LH
Lawrence_Hudetz
Dec 13, 2003
Hubboy. Here we go again!

I opened Color Settings. I changed Working Space to ColorMatch. I next opened a color image. PS asked how to proceed. I said to stay with Adobe RGB. Next, I duped it, then exited the original. I reopened the original and now allowed PS to change it to the new color space, ColorMatch. They both looked the same. (I knew they would) Next, I printed them, each from their own color space. They looked the same. (I knew they would).

The image has much out of gamut, especially in the greens. No difference anywhere.

Perhaps Assign Profile makes a difference. Voila! Difference! (I expected that). Now, as I look at the reassignment, compare it to the print from the first step, there is no comparison. The crt image is lighter, less saturated. I’m sure it would print that way.

So what! Immediately, upon opening the reassignment, I am into the controls, and can make it look like the Adobe RGB image. Reprint, and now I see that the intense reds have grayed down slightly, but no huge difference.

Tentatively, I conclude that not much difference exists between the two color spaces, and Convert is a waste of time. So is
Assign.

These are practical matters to me. I’m sure that to the absolutist, to the color scientist and engineer, there are differences. But I sure don’t see what it has to day to day operations, especially if you are mostly printing to a desktop inkjet.

I’ll keep Adobe RGB. That way, when I send the image out to say, a Chromira, there are no surprises.
PC
Philo_Calhoun
Dec 15, 2003
Glen: I can send you the gamut info for sRGB, Adobe RGB, and my Epson 2200 that Dry Creek did. You’ll have to add in ColorMatch yourself, but that is available on the WEB. email me if you are interested.
PC
Philo_Calhoun
Dec 15, 2003
Lawrence: there are several issues with your test. First, looking the same on a monitor obviously depends on the colour range of your monitor. Secondly, whether printing after converting to one or another colour space will look different depends on the particular colours in your test image. Assign makes an obvious difference: take a colour correct image that is in AdobeRGB and assign it (falsly) sRGB and it will look washed out. That happens all the time for people who use AdobeRGB as their workspace and forget to convert to sRGB and use the "save for Web" feature. They then post "why do my pictures look fine in Photoshop but look washed out in IE?"
PF
Peter_Figen
Dec 15, 2003
Remember that the gamut of the 2200 (and all Epson’s AFAIK) is MUCH larger when measuring a target printed through the NCA option in the driver than those measured through any of the automated options. In the earlier printers (1270 etc.), profiling through PhotoRealistic would give smoother gradations at the expense of gamut, which would map to be quite a bit smaller than sRGB.
PC
Philo_Calhoun
Dec 15, 2003
FWIW, the 2200 has a greater gamut than either sRGB or AdobeRGB in the green blue direction ( being close to AdobeRGB and much more than sRGB) but has much less in the blue indigo direction than either colour space. So it is neither quite more nor less than AdobeRGB/sRGB. It is just different. Whether this matters would depend on the colours in the image.
IL
Ian_Lyons
Dec 15, 2003
Peter,

profiling through PhotoRealistic would give smoother gradations at the expense of gamut, which would map to be quite a bit smaller than sRGB.

So far as the gradations are concerned I’ve always found the opposite to be true. In fact I’ve read comments from many others who’ve made the same observation. A colleague recently posted a grayscale and colour gradation image that showed just how bad this method of profiling the printer can actually be. I’ll try and find where he’s now hiding it. Personally I don’t think think that any REAL benefits are accrued from profiling the printer in PhotoRealistic mode.
PF
Peter_Figen
Dec 15, 2003
Ian,

I’m referring to the 1270/1280 series, where quite a few people report the opposite of what you’re saying. It was also a good way to help get more neutral black and white prints due to the compressed gamut. If you’re referring to the 2200, I’ve only profiled it wide open.

Master Retouching Hair

Learn how to rescue details, remove flyaways, add volume, and enhance the definition of hair in any photo. We break down every tool and technique in Photoshop to get picture-perfect hair, every time.

Related Discussion Topics

Nice and short text about related topics in discussion sections