Luminous Landscape Forum
Raw & Post Processing, Printing => Printing: Printers, Papers and Inks => Topic started by: HumptyDumpty on November 23, 2015, 05:27:40 pm
-
As I venture further into my printing adventure I am starting to more and more desire better print quality and get more and more lured by syren song of better printer profiles.
I first looked at getting something like i1Pro 2 Photo. Pro is I would be able to create my own profiles. Cons are:
1. It's $1,200
2. It would take time to build them and I have no expertise in building them nor time to become an expert so they would probably not be as good as ones made by somebody like Andrew Rodney
3. I don't make profiles regularly so once I built several I need i1Pro would be sitting in the corner gathering dust and if I ever upgrade to printer model that has built in profiler i1Pro would become redundant
Then I considered paying somebody to make them for me. Pros of that are obvious (I would get better profile than I can ever create in less time). Cons are:
1. Cost wouldn't be as high as cost of i1Pro but it still would be good part of it
2. If I get different printer or want to try some new papers I would have to go through it all over again rising my cost, quickly exceeding cost of i1Pro, and introducing delays
Then I read somewhere that even a best profile can't completely improve print quality, that part of an issue is due to, IIRC, manufacturer's printer drivers are "not linearized" and thus only way to get best quality possible is to go around manufacturer's drivers and use RIP.
So I started looking at ImagePrint. Pros are:
1. For 17" printer (what I have) it's cheaper than i1Pro
2. I don't have to build profiles nor worry about cost of new profiles
3. Allegedly it results in best print quality one can get
Cons are:
1. It supports only Epsons. I don't see myself printing wider than what 3880 can in the near future but if I was to go wider, or if my 3880 died, my next printer would be most likely Canon and at that point in time my investment in ImagePrinmt would be lost and I would be back at square one.
So what are your thoughts / what you would do / recommend?
-
If your aim is ICC output profiles, build your own with a caveat comming. If you need the functionality of IP, that's a different story. It's a good product (well I haven't used nor needed it since version 6). But I don't see the two being on parity. You could spend less and get a ColorMunki which is limited but builds good profiles. Maybe you can live with canned profiles from the paper manufacturers and spend nothing more! Not all profiles are created equally (more about that soon). But it is possible you don't need to spend anything more....
-
If your aim is ICC output profiles, build your own with a caveat comming. If you need the functionality of IP, that's a different story. It's a good product (well I haven't used nor needed it since version 6). But I don't see the two being on parity. You could spend less and get a ColorMunki which is limited but builds good profiles. Maybe you can live with canned profiles from the paper manufacturers and spend nothing more! Not all profiles are created equally (more about that soon). But it is possible you don't need to spend anything more....
My goal is to get best print quality possible. That is why I was considering going beyond paper manufacturer profiles and I am hearing ImagePrint takes that up even further.
-
Count me as a very satisfied user of ImagePrint with my 3880. When I printed out of Photoshop with custom profiles, I was very satisfied with the results, but when I made the leap to IP, I could see the difference. Moreover, IP makes the printing process very fast and easy (e.g. resizing and sharpening files with just a few clicks of the mouse). I have found final output soft proofing inside the program to be especially useful, even though I also soft proof in PS using ImagePrint's profiles. What you see is mighty close to what you get. Now are the results better than what you could get creating custom profiles with iLPro? That is a question I cannot answer.
Rob
-
So I started looking at ImagePrint. Pros are:
1. For 17" printer (what I have) it's cheaper than i1Pro
2. I don't have to build profiles nor worry about cost of new profiles
3. Allegedly it results in best print quality one can get
Cons are:
1. It supports only Epsons. I don't see myself printing wider than what 3880 can in the near future but if I was to go wider, or if my 3880 died, my next printer would be most likely Canon and at that point in time my investment in ImagePrinmt would be lost and I would be back at square one.
So what are your thoughts / what you would do / recommend?
I do my own profiles and I use ImagePrint. The profiles that I create is for my proofing RIP, not for ImagePrint. I also do profiles for some of my clients. They make smaller prints, using the Epson driver. Then they come to me and we do the large format prints with ImagePrint. When I use the x900 printers I can see a small but significant difference between the prints. It seems like that the Epson profiles don't use that much of the orange and green inks that the Imageprint profiles do. You can see an improvement in some portraits and landscape images. This will also give the prints better light fastness, as those inks don't fade as fast as the yellow ink does.
For B&W there is a bigger difference. For me, one of the big advantages of ImagePrint is the result that I get when I use the narrow gamut tint picker. I can adjust the tint of the image, so I can emulate the effects that I got in the darkroom by choosing different papers, developers and toners. I've tried several other methods for printing B&W images, but I haven't seen anything better.
-
Stephan - My understanding is that ImagePrint allows you to print create your own profiles should you wish. You have to print the charts using ImagePrint, but thereafter you scan and create the profiles the normal way. Have you ever created your own profiles for ImagePrint, and then compared them with those supplied with ImagePrint?
-
Thank you guys, this was helpful!
-
Stephan - My understanding is that ImagePrint allows you to print create your own profiles should you wish. You have to print the charts using ImagePrint, but thereafter you scan and create the profiles the normal way. Have you ever created your own profiles for ImagePrint, and then compared them with those supplied with ImagePrint?
That's correct (at least back in V6, the last copy I owned). You can build your own and I found them better than the supplied profiles. At least at the time, the IP profiles had the annoying habit of blues shifting magenta.
-
My goal is to get best print quality possible. That is why I was considering going beyond paper manufacturer profiles and I am hearing ImagePrint takes that up even further.
I have never seen anything to come even close to convincing me that custom/homemade profiles offer any advantage. They may well give different results from the paper manufacturer profiles, but different does not mean better - it just means different. Of course the photographer who puts a lot of time and money into custom profiles will be psychologically inclined to prefer them! That's just human nature. And given the huge vested interest that Epson, Canson, etc. have in photographers getting great prints from their papers, and their ability to afford the best profile-making equipment, is it likely that you or "Joe Shmoe's Custom Profiles" can do better?
You want to make better prints? Work at it. See as many original prints as you can, at galleries, museums, and so on (web images don't count). Take a workshop. Sit at your computer and make print after print after print, compare them, figure out what works and what doesn't.
-
I have never seen anything to come even close to convincing me that custom/homemade profiles offer any advantage. They may well give different results from the paper manufacturer profiles, but different does not mean better - it just means different.
Sorry man, that's simply not the case. Here's a prefect example: Epson canned profile vs. custom. These are photo's of actual prints made on a P600. The only difference is the profile, both using RelCol. Night and day:
Look at the mapping of blues, examine the dynamic range and lack thereof from the Epson profile.
-
Huge difference Andrew - and I must say kind of surprising. What paper?
-
Huge difference Andrew - and I must say kind of surprising. What paper?
Premium Luster.
-
OK that rules out the narrow paper gamut hypothesis.
-
OK that rules out the narrow paper gamut hypothesis.
Oh it's the profile, no question. And it shows up in other profiles generated with the same software used to build it. The color engine isn't so good. And that's one reason why not all profiles are created equally! The name of my next video. Just the difference sending the same spectral data to say ProfileMaker Pro vs. MonacoPROFILER vs. Copra vs. i1Profiler do result in differences on the print, often quite visible, due to their color engines.
-
Agree - I've seen this in profiling application comparisons before. But I would have expected Epson would find it in their interest to use and make only the best for their papers in their printers, and as I said, I found their HPN profile for the P800 to be quite OK (in real world photos). So what's going on?
-
Agree - I've seen this in profiling application comparisons before. But I would have expected Epson would find it in their interest to use and make only the best for their papers in their printers, and as I said, I found their HPN profile for the P800 to be quite OK (in real world photos). So what's going on?
Not all Epson profiles are created the same. That's the first issue. You're on a Mac? Double click on Luster, then Exhibition Fiber or as in this example, Metallic:
-
I see that - can't argue with the facts, but the motivation seems counter-intuitive.
-
Stephan - My understanding is that ImagePrint allows you to print create your own profiles should you wish. You have to print the charts using ImagePrint, but thereafter you scan and create the profiles the normal way. Have you ever created your own profiles for ImagePrint, and then compared them with those supplied with ImagePrint?
Several times and 9 times out of 10 it's a draw. But I have found some papers where my custom profile will give a better result than the one provided from Colorbyte.
-
That's correct (at least back in V6, the last copy I owned). You can build your own and I found them better than the supplied profiles. At least at the time, the IP profiles had the annoying habit of blues shifting magenta.
You still can do your own, at least for color. And I remember the same magenta skies with my Epson 9600, but I don't see that today.
-
Stephan - My understanding is that ImagePrint allows you to print create your own profiles should you wish. You have to print the charts using ImagePrint, but thereafter you scan and create the profiles the normal way. Have you ever created your own profiles for ImagePrint, and then compared them with those supplied with ImagePrint?
Several times and 9 times out of 10 it's a draw. But I have found some papers where my custom profile will give a better result than the one provided from Colorbyte.
Thank you. Much appreciated. May I ask a follow-up question please? In your opinion & experience, does it make a difference whether you're using an i1 V1 or V2, and/or UV-cut or non-cut for making your own profiles?
In some sense, your response is a little surprising. People often praise the ImagePrint profiles, but unless you're doing something unique, your reply suggests that the big difference is in the RIP itself rather than the profiles that they create and supply.
-
In your opinion & experience, does it make a difference whether you're using an i1 V1 or V2, and/or UV-cut or non-cut for making your own profiles?
In a word, yes, it can. Depends on a number of factors like the papers, degree of OBA's etc. In terms of UV Cut vs. Not, no hard and fast rules as far as I'm concerned and I'll often measure both and build a profile each way then decided.
-
Imageprint simplified the printing process for me. Before purchasing the rip, I built profiles with varied success on quite a few different papers, but none of my profiles yielded the precise duplication between my calibrated Eizio monitor and the print as well as Imageprint. For me the investment in the software has more than paid for itself by avoiding wasted paper and ink
Stanley
-
Thank you. Much appreciated. May I ask a follow-up question please? In your opinion & experience, does it make a difference whether you're using an i1 V1 or V2, and/or UV-cut or non-cut for making your own profiles?
I think that the i1 v2 is much improved compared to the first version. You can get similar results if you compare them, but it easier to make mistakes with the original version. Mistakes that an unexperienced profile creator doesn't notice. When I build a profile I measure my chart three or four times and then I compare the result before I create an average of the measurements. I notice that it is easier to get very similar results when I use my i1Pro2.
In some sense, your response is a little surprising. People often praise the ImagePrint profiles, but unless you're doing something unique, your reply suggests that the big difference is in the RIP itself rather than the profiles that they create and supply.
The results that I get when I build my own profiles or use the ImagePrint profiles are very similar. I don't think that the big difference is the profile, it's more the media setting (ImagePrint calls it the recipe). If you don't have a good media setting, no profile can correct the result.
-
Sorry man, that's simply not the case. Here's a prefect example: Epson canned profile vs. custom. These are photo's of actual prints made on a P600. The only difference is the profile, both using RelCol. Night and day:
Look at the mapping of blues, examine the dynamic range and lack thereof from the Epson profile.
Andrew, what would the comparison look like with one of the better (X-Rite created) Epson profiles? e.g. for Exhibition Fiber?
As you note in your follow-on post, all Epson canned profiles are not created equal. And since most of the "higher quality" (no offense to fans of Premium Luster) Epson papers appear to be made by X-Rite, making the comparison with one of those better papers would seem to make sense.
-
Seems like the problem is sort of Forrest Gump-y. You never know what quality of profile you're gonna get unless you're a colorgeek and know/understand these sort of esoteric things. That precludes about 99.7% of the users out there.
I think Epson is doing themselves a great dis-service by not riding shotgun on these profiles to give the customer a consistent quality. But doesn't that sound just like Epson?
-
Andrew, what would the comparison look like with one of the better (X-Rite created) Epson profiles? e.g. for Exhibition Fiber?
Depends on how it was measured and built (custom settings for Perceptual table). But the differences should be much, much smaller than we see between the Seiko vs. X-rite profiles. That's night and day.
Let's say you use an iSis and produce various M-series measurements, then maybe OBA correction and/or post optimization as I do with all my profiles. You could see a difference on a good suite of test reference images (mine, Bill Atkinson, Roman 16). Even when I build custom profiles in i1P using 1700 odd patches, depending on a few factors, I see improved gray balance and saturated color rendering after running a post optimization using my custom target for that task. It can be subtle but it's visible! You're just not going to see that with any 'canned profile' for obvious reasons.
-
You never know what quality of profile you're gonna get unless you're a colorgeek and know/understand these sort of esoteric things.
Actually you'll know after making prints using color reference images. One can conduct colorimetric testing too, but that requires a Spectrophotometer and something like ColorThink Pro. That data is useful and interesting but not prior to making prints!
-
What I meant is exactly that. How many people have come here (well, maybe less so) or DPR and when you tell them 'print a standard test image' they say 'I never thought of that!' or 'where can I get one'?
I think many would just suffer in silence and say 'wow, this paper sorta sucks' when the profile itself isn't really very good.
-
What I meant is exactly that.
Well I was referring to the suggestion that:
You never know what quality of profile you're gonna get unless you're a colorgeek and know/understand these sort of esoteric things.
You don't have to be a geek and I don't believe this is esoteric. One does need to now what to look for in printed output, if not, it's kind of moot: ignorance is bliss. FWIW, next video in the works goes over this exactly; what to look for, what test images to use etc.
How many people have come here (well, maybe less so) or DPR and when you tell them 'print a standard test image' they say 'I never thought of that!' or 'where can I get one'?
I don't know. But the answer is easy to provide, education is key.
I think many would just suffer in silence and say 'wow, this paper sorta sucks' when the profile itself isn't really very good.
Possible. But we can't help those who are unable or unwilling to learn. None of us were born with any knowledge of printing, color management or image quality, we learned one way or another. As for DPR, well the site is FILLED with people who don't have a clue. Not so much here. We either try to assist and teach or just ignore them. I prefer the former.
Bottom line is, the tools (images) and techniques to evaluate profiles are around and conducting such tests is pretty easy. The colorimetric tests are far more difficult.
-
I think that the i1 v2 is much improved compared to the first version. You can get similar results if you compare them, but it easier to make mistakes with the original version. Mistakes that an unexperienced profile creator doesn't notice. When I build a profile I measure my chart three or four times and then I compare the result before I create an average of the measurements. I notice that it is easier to get very similar results when I use my i1Pro2.
What sorts of mistakes are you referring to? Surely you don't mean mis-scanning the target. Do you mean too much sample variation, leading to a need to do as you suggest - averaging multiple scans?
I don't think that the big difference is the profile, it's more the media setting (ImagePrint calls it the recipe). If you don't have a good media setting, no profile can correct the result.
I've seen this with some other RIPs. I don't see the point of buying a RIP less it's using its own media profile. Otherwise I may as well save my money and use the Epson driver.
-
Depends on how it was measured and built (custom settings for Perceptual table). But the differences should be much, much smaller than we see between the Seiko vs. X-rite profiles. That's night and day.
Let's say you use an iSis and produce various M-series measurements, then maybe OBA correction and/or post optimization as I do with all my profiles. You could see a difference on a good suite of test reference images (mine, Bill Atkinson, Roman 16). Even when I build custom profiles in i1P using 1700 odd patches, depending on a few factors, I see improved gray balance and saturated color rendering after running a post optimization using my custom target for that task. It can be subtle but it's visible! You're just not going to see that with any 'canned profile' for obvious reasons.
Thanks, Andrew. I was hoping you had a comparison at-hand of Exhibition Fiber, or one of Epson's other "better" papers.
The OP came onboard with the notion that he needed either custom profiles or a RIP in order to get "better print quality." It's been my experience that the canned profiles generally work very well - but then I use Epson's "better" papers almost exclusively (the profiles for which are all, I believe, created by X-Rite; and, again, no intended insult to those who favor Premium Luster). Your example contrasting the custom-versus-canned profiles for Premium Luster - clearly quite dramatic - is, perhaps, somewhat less than representative of Epson's Signature paper lineup? Lending what I'd suggest is false credence to the OP's belief that he's not going to get good prints unless he goes down the road of custom profiles or a RIP.
The reality is that buying an i1Pro 2 or other spectrophotometer is not going to suddenly turn one into a color and profile expert. I agree with Peter that most folks out there today making profiles, either for themselves or for profit, are probably little better - and potentially far worse - in their efforts than are Epson, Canon, HP, Canson, Hahnemuhle, Ilford, etc. My guess is that of those making profiles, very few are at your level (1700 patches!).
Which is not at all to suggest that if the OP wants to learn that somewhat esoteric part of the printmaking business, by all means proceed! But to believe that the canned manufacturer profiles are all rubbish and the only way to get excellent results is to go there, seems patently false.
-
Your example contrasting the custom-versus-canned profiles for Premium Luster - clearly quite dramatic - is, perhaps, somewhat less than representative of Epson's Signature paper lineup?
All I can do is provide a list of the profiles made using X-rite vs. whatever the Seiko people use:
As I think I mentioned, if you see a "V" in the name, that's an X-rite profile.
-
While I haven't printed using each profile in the supplied list, one can soft proof the Gamut Test File (or the Roman 16 blue girl who's print I illustrated) and this problem with mapping of blues to nearly black is clearly visible on-screen. Not with any of the profiles with a "V" in the name using the X-rite color engine. The one slight outlier is the profile named Epson Stylus Pro 3880_3885_3890 Standard. Not as bad. That said, I have no idea what that's for. It's still a profile tagged with Seiko in the copyright but it's not anywhere as bad as the others.
-
Sorry man, that's simply not the case. Here's a prefect example: Epson canned profile vs. custom. These are photo's of actual prints made on a P600. The only difference is the profile, both using RelCol. Night and day:
Look at the mapping of blues, examine the dynamic range and lack thereof from the Epson profile.
Well, your photos certainly provide food for thought! I can't help but think that the Epson profile image of the woman could be a lot better processed, but even so I will have to reconsider my ideas regarding profiles. Thanks.
-
What sorts of mistakes are you referring to? Surely you don't mean mis-scanning the target. Do you mean too much sample variation, leading to a need to do as you suggest - averaging multiple scans?
No, I don't mean mis-scanning a target. But when I compare several measurements done with the first version of i1 I get a higher variation than I get with version 2. I can think of several reasons for that, but the only thing I'm sure of is that when I started using i1Pro2 my Delta E went down when I compared 3-4 measurements of the same target.