Upload & Sell: Off
Peter Figen wrote:
Chuck - If, like you say, the Noritsu, can't use profiles, and is only passing whatever RGB is gets, straight to the paper, your ProPhoto files should actually look desaturated compared to sRGB - because the RGB numbers for saturated colors are so much lower for a given color in ProPhoto than in sRGB. There must be something else going on here that you left out.
You are assuming I was comparing print to monitor. I wasn't. My previous comment related to comparing two prints of the same file, one made by converting to the machine profile, the other from the unconverted file. In that situation if the converted print max saturation doesn't match the managed one the fault lies in the profile not matching actual conditions.
As you opine, the QA at Costco and skill level of the operators vary, but I factor that into my decision to print there. I agree is not the ideal lab, but then not everyone's requirements are as critical as yours are. It's nice to be able to do your Christmas cards on-line at 9 PM and get 50 of them, with envelopes, the next morning at 10:30AM for $15
When I was running offset printing plants for a living every proof and production press sheet had measurable color control targets and visible indicators for press variables such as dot gain and slur. Whenever I print at Costco do something similar by either embedding an image of my test target into the middle copy of the print I'm making as a control or printing to the next size with a 128,128, 128 expanded canvas on the image with the control targets embedded into that border. Then when I get the prints I can easily tell how well the printer and processor is calibrated that day. The test prints allow me to communicate the problems, if any, to the operator by showing him a good reference print of the target.
Mostly I print portraits and part of my shooting workflow is setting Custom WB and shooting a test shot of the subject holding the target...
By doing the technical part of capture "by the numbers" — that is to say recording a full range of detail with neutral WB — the results I get, for better or for worse, are the best my camera can produce with it's default settings. The reproduction of the MacBeth target isn't an exact match side-by-side with it appearance on the monitor and when I print that control file in the same batch as the final portraits the target in the print doesn't exactly match the original target either. But the that really doesn't matter in practical terms for that type of work. What matters is that the camera exposure control winds up revealing a full range of detail in shadows and highlights on both final JPG screen images and the prints, and that both the monitor and print images stay neutral. The MacBeth target test image makes it easy to see and measure both with the only measurement tools I have nowadays: the eye dropper measurement tool and by eyes by side-by-side comparison with a known reference.
I put a good deal of trust in the fact that the profile based color management will control the gray balance over the tonal scale, but also have a realistic expectation that the colors are somewhat of a moving target between gamuts in the middle values due to the physical differences in the gamuts. The 3D wireframes, while not a precision instrument, provide a "ballpark" understanding of how different colors will change between my monitor and the print. I have an iMac and here's how it compares to the Lustre profile at my local Costco..
As with most monitor > printer comparisons some colors are less saturated on the print, others like reds are even more saturated, which works out nicely for the Christmas cards, when the full gamut of the printer is utilized by letting it manage the color...
My goal is maxing out that gamut, relying on the color management to sort out the rest as best it can give the difference between the gamut sizes and shapes. I trust the color management technology will do that, by understand the physical limitations of the process and device the means to monitor them via my test image protocol.
My personal needs are not that critical with respect to color, but let's say I was selling clothing on line. In terms of my workflow making the prints I'd want a wide-gamut monitor because it would, with experience of actual printing results under my belt, help me during editing, eliminating the need to make as many test prints. So in that sense montitor > printer match is desirable. But in the case of clothing I'd want the color of the clothes in the photo to match the actual clothing in the image the customer sees in the web or print catalog to make the selection.
That creates a bit of a dilemma because I must assume the potential client will be using an sRGB monitor and printing an offset catalog will shift the colors. That introduces the potential for a person buying the print or pair of pants, comparing them to the on-line image and being disappointed when they don't match and returning them.
One way clothing and other retailers deal with that problem is by starting with a palette of colors that both an sRGB monitor and SWOP CYMK gamut can both reproduce with colorimetric accuracy. In other words the colors in the colored SWOP gamut which don't hang outside of the sRGB wireframe below...
There are orange and cyan colors the monitor can reproduce better, and greens the press can print better but the "safe" colors are those where the two gamuts overlap. What they do then is create a press printed swatch book of all those colors and use it to guide the selection of the fabrics used for the clothing. The color is managed from the lowest common denominator — the colors monitor and offset can match — backwards to the fabric dyes. When managed that way the person buying a pair of green pants on-line or from a catalog will not be disappointed when they get them and they don't match. Not everyone manages their color that way, but those who do probably have a lower rate of customer returns.
Color management for photography is much simpler. All you really need to do to make the color look "real" is capture and reproduce a full tonal range and keep the neutral grays R=G=B in the files on screen and neutral on the prints per the "recipe" for grays in the printer profile. In that regard Costco works for me for my photos. But I buy my clothes off the rack where I can judge the color in person, both under the store lights and over by a window in the daylight