I first bought ImagePrint 9, years ago, by seeing the claim on Luminous Landscape web site, that Imageprint ICC profiles are better than the vendor provided ones using the printer driver. I was quite happy with it for all these years using it on my Epson large format printers (first Epson 7900 and then Epson SC p-7000).
Last week I had to replace my 16 month old Epson SC p-7000 printer for stubborn printhead clog (that's another pathetic story :-() with a Canon PRO-2000 printer. So far I have been very happy with the printer. I think the Canon Printer is way ahead in terms of precision, sophistication, ethical design and customer support!
Anyways, sorry about the digression. Next step was printing Bill Atkinson printer targets with Imageprint ICC profile on my paper of choice for printing, Canson Infinity Platine Fiber Rag paper. I also compared it with the ICC profile provided by Canson with the media settings file. I am little disappointed to report that Canson ICC profile for that paper-printer combination is better than the Imageprint ICC profile. After running many tests of printing using both ICC profiles, here are the summarry of my observations, so others can benefit from my experience:
1. The Imagprint profile-printed target is little inferior to Canson profile-printed target in terms of:
- saturation in red
- skin tones
2. Imageprint profile on Epson is still slightly better than Imageprint profile on Canon in terms of:
- satuaration in red
3. However, Imageprint profile on Canon is slightly better in gray tone seperation than Canson profile
I contacted Colorbyte technical support team and they were very prompt in their reply. I found out that Imageprint profiles were generated after calibrating the Canon printer with Canon Photo Glossy 170 gsm paper. So, all canon PRO-2000/4000 users, who want to use Imageprint profile, must re-calibrate their printer using that photo paper.
After recalibrating my printer with the above paper, the print targets look slightly improved. But, the above observations still hold true.
I challenged Colorbyte to generate an ICC profile that is better than the Canson ICC profile. If they accept the challenge and provide me a better profile, I would be happy to update this thread. Until then, I would continue using Canson proviedd ICC profile.
An Update on this issue. I had been working with ColorByte on this issue and sent them test targets printed with ImagePrint and Canson Profiles for their comparison. ColorByte technical support has been very responsive with this issue. I had couple of issues with ImagePrint:
1. Saturation in Red was higher in the Canson Profile print
2. Skin tone in Canson looked better than Imageprint.
Here is how ColorByte responded to it:
"Your print does look extremely close to what we get here. That was a bit surprising because in [our] judgement our print was the better one compared to the Canon driver print you sent.
The Canon driver's print is slightly more saturated for some landscapes and objects than the ImagePrint print--but slightly over-saturated for skin tones -- and across the board it lacks tonal separation and shadow detail of ImagePrint's output.
One area we took a very close look at was the strawberries. A good section of the colors here are out of gamut in both profiles so neither reproduction can be 100% accurate. (And when bringing out-of-gamut color into gamut there’s always going to be some subjectivity). Our color management engine tends to favor preserving the tonal separation when bringing colors into gamut which can result in a bit less saturation compared to the Canon driver's method which is clearly favoring loss of tonal separation and bumping saturation. In our experience that latter method leads to blocked shadows throughout the image and a tendency to oversaturate.
In these areas of subjectivity you may prefer some things with the driver profile -- but they do come at an expense. In that regard, if I were to evaluate both prints not side-by-side but on their own merit, there are areas in the Canon print (like the loss of detail) which would be actually problematic--not just a subjective difference. There is no such areas in the print made with ImagePrint and that’s how we engineer our color management components.
I realize this is a generalized test, but when printing images of varying keys and out-of-gamut components as a group using one overall setting, there will always be sacrifices made to the potential reproductive quality of each component. For example, I personally would never print a high key image with the same shadow point that I use with a low key image. I would instead change the shadow point slider from our default value to a lower one to boost the saturation on the images where it was needed, while leaving others alone or even increasing the shadow point slightly to bring out more shadow detail. One of ImagePrint’s key values as a printing application is that it is made to allow you to adjust things like this on an image by image basis quickly and without having to edit the image.
In this case, simply bringing our shadow slider down to 1 while using Perceptual rendering (to avoid clipping of those out of gamut colors) brings that red strawberry much closer to the Canon level of saturation while still maintaining more shadow separation than the Canon print and keeping the skin tones correct. That will give you a print that gives you the saturation you want without the compromises the Canon driver is making regarding tonality so I’d recommend giving that a try and let us know what you think."
After this response from ColorByte, I re-printed the test target with the shadow slider brought down to 1. Now, the ImagePrint test target looks very close to Canson. So, apparently the shadow point has a corelation with color reproduction of out of gamut colors. This is good to know! For now, I will go back to using my ImagePrint application.