Profiling a printer using a scanner?

[bcachot]

Is there any way of profiling a printer using a profiled scanner? A product like eye-one photo does it all (monitor/camera/printer) using a spectrophotometer and is no doubt more accurate, but I don’t want to spend that much cash at this stage. I would prefer to stick with a product like eye-one display to ensure my monitor is calibrated. I have no idea if it’s possible to profile a printer using a scanner or if the results are halfway decent. Needless to say, the scanner would have to be profiled first.

[Schewe]

Uh, no...the profile will ONLY be good as the combination of the quality of the scanner (not a real measuring device), it's profile and the software to arrive at a printer profile from the above device. It's not a tight loop. You really would be better off having a 3rd party make real custom profiles for your printer/paper than going down that road. As long as you have a real accurate display profile and your custom perinter profile, you'll end up with superior results.

[Jonathan Wienke]

I recommend a custom profile from a service as well. It's less money than a scanner-based kludge, and much better quality profile.

[bcachot]

I recommend a custom profile from a service as well. It's less money than a scanner-based kludge, and much better quality profile.
[a href=\"index.php?act=findpost&pid=89354\"][{POST_SNAPBACK}][/a]

When you say 'from a service' do you mean taking printouts to a bureau of some kind and have them create a profile for me?

[jjlphoto]

When you say 'from a service' do you mean taking printouts to a bureau of some kind and have them create a profile for me?
[{POST_SNAPBACK}][/a]

Yes- try ink jet art- $25, and I hear good things about the quality of their profiles. They instruct you to download their targets from their website, print them out, and mail them in.

[a href=\"http://www.inkjetart.com/profiles/profiles.html]http://www.inkjetart.com/profiles/profiles.html[/url]

Heck, for that price, I'm considering selling my $1400 GMB i1 Pro spectro, and taking those proceeds and getting the $230 GMB i1 Display, and with the difference, I will be able to buy a lifetime of higher quality profiles than I am capable of making myself! (and without wasting my time either.)

BTW, I have tried Monoco EZ color years ago, (one of those highly touted scanner based profining systems) and ...well... lets just say you are better off using the color controls side of the Epson driver and making trial and error prints with the color sliders just like in a wet darkroom using an enlarger with a colorhead.

[Greg_E]

Heck, for that price, I'm considering selling my $1400 GMB i1 Pro spectro, and taking those proceeds and getting the $230 GMB i1 Display, and with the difference, I will be able to buy a lifetime of higher quality profiles than I am capable of making myself! (and without wasting my time either.)

For the thicker media he is using an i1 and the same Match software that you have. Only the thinner media that will fit through the DTP70 gets run through Profiler. And the profiles that he posts for download are also built in Profiler because they have a license to post "generic" profiles from Xrite. That might change now that it's all one company. And yes it does require a distribution license to be able to post "generic" profiles, but most places don't seem to care about the legal stuff.

I have Profiler, and it makes nice profiles, even from the 729 patch target though the 1728 might be better for most of the Epson printers.

[b2martin]

Profile Prism does what you suggest.  You print a test target with your printer (over 9000 color squares).  Scan the test target print and a standard color reference (IT-8 type standard) at the same time.  Process this using Profile Prism software and it generates a printer profile.  Cost for the software and targets is less than $100.

[jjlphoto]

As Jeff and others have mentioned, a scanner is not a measuring device, and any scanner generated profiles will introduce the limittions and characteristics of that individual scanner. A spectrophotometer is a true measuring device.

[Greg_E]

Profile Prism does what you suggest.  You print a test target with your printer (over 9000 color squares).  Scan the test target print and a standard color reference (IT-8 type standard) at the same time.  Process this using Profile Prism software and it generates a printer profile.  Cost for the software and targets is less than $100.
[a href=\"index.php?act=findpost&pid=89553\"][{POST_SNAPBACK}][/a]

I still have to wonder what Profile Prism could do if you measured the target with a spectro and let it build the profile?

[Jonathan Wienke]

The difference between a scanner and a spectrophotometer is that a scanner measures color in 3 channels, while a spectrophotometer measures in 30 or more channels. If you've ever tried to EQ a set of speakers to a room, you know that a 32-channel EQ will do a much better job than a 3-channel EQ. A scanner is like a 3-channel EQ, and will never do as good a job as the spectrophotometer.

[Greg_E]

The difference between a scanner and a spectrophotometer is that a scanner measures color in 3 channels, while a spectrophotometer measures in 30 or more channels. If you've ever tried to EQ a set of speakers to a room, you know that a 32-channel EQ will do a much better job than a 3-channel EQ. A scanner is like a 3-channel EQ, and will never do as good a job as the spectrophotometer.
[a href=\"index.php?act=findpost&pid=89741\"][{POST_SNAPBACK}][/a]

It really has more to do than the number of channels used to measure the target, those channels just make the math more precise. The measuring hardware is also much more precise than the scanner ccd which is another reason that the spectro does a better job. And the spectro has a better dynamic range than the scanner, less flare in the optics, and several other advantages. When all the debate is done it comes down to something much more simple. You start with an image that is probably 3 or 4 channels (RGB or CMYK), then it gets converted into the connection space (in this case 3 channel L*a*b*), and then to RGB (because we are dealing with the RGB printer driver). So in reality having extra channels in the measurment end may not  really be worth a lot. You can make good profiles with L*a*b* or XYZ data.

[bcachot]

The difference between a scanner and a spectrophotometer is that a scanner measures color in 3 channels, while a spectrophotometer measures in 30 or more channels. If you've ever tried to EQ a set of speakers to a room, you know that a 32-channel EQ will do a much better job than a 3-channel EQ. A scanner is like a 3-channel EQ, and will never do as good a job as the spectrophotometer.
[a href=\"index.php?act=findpost&pid=89741\"][{POST_SNAPBACK}][/a]

A spectro would be better, no doubt. I'm kind of curious though if the results from a quality scanner would be halfway decent. Good enough to satisfy a serious hobbyist. Would love to know if someone tried this.

[digitaldog]

A spectro would be better, no doubt. I'm kind of curious though if the results from a quality scanner would be halfway decent. Good enough to satisfy a serious hobbyist. Would love to know if someone tried this.
[a href=\"index.php?act=findpost&pid=89863\"][{POST_SNAPBACK}][/a]

Depends on the printer. I've seen scanner based solutions produce profiles that were on par with Spectrophotometer build profiles on well behaved devices like a Fuji Pictrography. I've seen butt ugly profiles from printers like ink jets. YMMV. Usually it's not worth the hassles in time and money using scanner based products but occasionally they do a good job.