Previous news story    Next news story

Just posted: Printer Primer Part 2: Print workflow

By dpreview staff on Jul 28, 2011 at 19:00 GMT

Just Posted: Printer Primer Part 2: Print Workflow. As part of our expansion into providing printing content, we're taking the opportunity to update some of our existing content. Part 2 of our printer primer - print workflow - explains how to get the most out of your printer, whether you're an experienced print maker or you're dipping your toes into home printing for the first time.

Click here to read Part 2 of our Printer Primer: Print Workflow

Comments

Total comments: 36
Peter iNova
By Peter iNova (Aug 3, 2011)

The chestnut of specifications for highest quality printing is this:

"300 dpi". But how far can you bend that?

That's a rule of thumb carried over from dot screen printing in which 300 actual printing dots per inch is about the maximum any mechanical printing process can reliably achieve under perfect conditions.

At 12" viewing, a 4 x 6 print crosses 18° of retina. 50 line pairs per degree is approximately the human limit. Meaning at this distance, an 1800 pixel-wide image will look as sharp as can be. And it will use the full 300 ppi.

You can resample an image to any size and print out a series of the same image at different pixels per inch. Somewhere around 200 ppi, at standard 16-18 inch viewing distance for a letter page, more pixels won't add more appreciated detail.

The National Geographic's images are about 180 dpi. You'll get shots that sharp at 200 camera pixels per running inch of print. Meaning your 18 MP cam's images will look max sharp at about 25 inches big!

0 upvotes
Steve Bingham
By Steve Bingham (Aug 3, 2011)

In Color Management or Color Handling do you have "Printer Manages Colors" checked? I use PS CS5 and not LR3 but this might be the issue. Give it a try for starters. Otherwise LR3 might be tweaking your color - like printing in a color space not intended.

0 upvotes
aaaja
By aaaja (Aug 2, 2011)

I am so glad somebody with great know-how is sharing information for free.

Q: i could not find a descent answer for a problem i have with LR3 and printing. Pictures printed by any other tool are more or less ok.

LR3 did work for me under Wimndows XP - now under W7 64b, the pictures are dark and miss colourness. as mentioned same picture exported to jpg lokks ok.

It is a Epso 2880

It would be juts great to do some printing without exporting first.

thanx for any help
yes my screen is calibrated ;-)

0 upvotes
igorschutz
By igorschutz (Jul 29, 2011)

I think that DPReview should hire TerryW to write some articles...

1 upvote
TerryW
By TerryW (Jul 29, 2011)

Only if I get paid by the word.....and I can write A LOT of words! :-)

-tw

0 upvotes
PLShutterbug
By PLShutterbug (Jul 29, 2011)

The article mentions use of a colorimeter to calibrate the monitor. Any recommendations?

0 upvotes
TerryW
By TerryW (Jul 29, 2011)

You've got a choice between a *colorimeter* and a *spectrophotometer*. Key difference is that a colorimeter uses filters (RGB but sometimes more) while a spectrophotometer samples in spectral bands, typically @ 10 nanometer spacing but sometimes more (i1Pro samples at 3.3nm but reports at 10nm).

Colorimeter filters should ideally be "tuned" to the specific RGB primaries of your display, especially if it's one of the newer "wide gamut" (AdobeRGB) displays...this is VERY important...not just any colorimeter will do for all displays.

A spectrophotometer on the other hand, since it doesn't rely on filters, can be used with any display without any special "tuning" requirements.

While colorimeters are less expensive (couple hundred bucks typically), a spectro can be anywhere from about $600-1,000. I would say that a spectro is a better long-term investment but colorimeters are fine if you intend to stick with your current display for a few more years.

continued.....

1 upvote
TerryW
By TerryW (Jul 29, 2011)

....continued from above.

One thing that should be considered is that often a spectro such as the X-Rite ColorMunki and EyeOne Pro can be used for BOTH display AND print profiling. A colorimeter is a display-only device. SO, if you have any visions of doing both display and print profiling down-the-road, the spectro is the better investment. And, at least in theory, you'll have a better print-to-screen match since you can use the same instrument for both print and display profiling.

The only downside of a spectro as a display profiling device compared to a colorimeter is that they TEND to be less accurate at very low light levels ("noisier") compared to a colorimeter. The visual result is slightly better shadow rendering from a colorimeter than a spectro. This can be countered somewhat by simply raising the black luminance level when calibrating.

continued...

1 upvote
TerryW
By TerryW (Jul 29, 2011)

...continued from above.

Once you decide on an instrument (spectro vs. colorimeter), you need to decide on software....and not all display calibration and profiling software packages are created equal.

If you have a "high-end" display like an EIZO ColorEdge display or some of the NEC displays that supports *hardware* calibration (DDC support), then you'll want to make sure you get software that supports this. Hardware "LUT" calibration (10-14bit LUTs) is a big deal and produces a much smoother result than calibrating via 8bit video cards.

Two 3rd party packages that I like are ColorEyes Display Pro and basICColor Display 4....both support a wide range of colorimeters and spectrophotometers and they both support a wide range of DDC-compliant displays. ColorEyes has good support for EIZO displays while basICColor supports the EIZOs and a fair number of other displays with hardware LUTs.

If you don't have a high-end display, then X-Rite i1Profiler is good choice.

-Terry

1 upvote
TerryW
By TerryW (Jul 29, 2011)

Guess I didn't actually get around to recommending an instrument!

For spectros, the X-Rite EyeOne Pro and ColorMunki are good and relatively inexpensive.

For colorimeters, if you can find an X-Rite DTP94, these are still very well regarded.
Other choices are the Spyder3 by DataColor, the DISCUS from basICColor (expensive but very good) and the EyeOne Display colorimeters from X-Rite.

If you have a standard "sRGB" LCD display, pretty much any colorimeter will do....but, like I said earlier, if you have any of the newer wide-gamut/AdobeRGB displays, then you should first see if the manufacturer offers a special colorimeter (and software) that is "tuned" correctly for the RGB primaries of their display.

-Terry

1 upvote
PLShutterbug
By PLShutterbug (Jul 30, 2011)

Terry, thanks for the very informative reply. I'll take some time to digest what you've said then decide what to do. I don't think I have a high-gamut display so probably one of the less-expensive devices will do me.

0 upvotes
Eric Napa
By Eric Napa (Aug 1, 2011)

I'm not a color management professional. However, I have an xRite ColorMunki that I use with my Mac Cinema display and Epson 9900. It is easy to keep my monitor calibrated and very straightforward to set up an icc profile for my papers. The included software also allows you to fine tune an existing profile for a given image. I can't compare the ColorMunki to any other solutions, but I can say that I get better color with the profiles I generate than with those I download. I can definitely recommend the ColorMunki. If nothing else, its a very cost effective way to get started.

0 upvotes
TerryW
By TerryW (Jul 29, 2011)

Another inconsistency in the article...

You state (and I "snip")...
"The larger the print you want to make, the larger your digital image will need to be. ......... requires a lot more output resolution (i.e. more pixels)."

But then you go on to say in the next paragraph...
"Resolution in digital photographs is expressed in pixels per inch (ppi)."

Using that definition, the SAME resolution is required NO MATTER what size the final print is going to be. 300ppi is 300ppi no matter what size the print is. For sure it requires a higher *density* of pixels at the larger size (or more "megapixels" if you prefer) but not necessarily more resolution, at least using your definition of resolution.

2 upvotes
Amadou Diallo
By Amadou Diallo (Jul 29, 2011)

Article has been updated.

0 upvotes
TerryW
By TerryW (Jul 29, 2011)

Thanks for making those corrections...the monitor profiling reads much better.

I still feel that parts of the paragraph below about printer profiling is a bit clumsy and incorrect:
"...Each color patch has a reference RGB value which the printer profiling software uses to compare against measurements of the actual printer output...."

The use of the word "compare" in this context is incorrect....there's nothing to compare since the metrics are different...you can't compare RGB device values to measured XYZ or L*a*b* colorimetric values...two different things. It seems to still be implying that there's some *known* behavior of the printer that the measurements are being compared against. There effectively is no known or assumed behavior of the printer UNTIL the profile itself is created. With the profile, you now have a predication of what colorimetric value should result from "x" RGB device values.

-Terry

1 upvote
TerryW
By TerryW (Jul 29, 2011)

To continue my thought about errors in this article....

In reference to the error describing monitor profiling, unfortunately you made the same error a couple of paragraphs later when discussing printer profiling:

"Each color patch has an objective known value which the printer profiling software uses to compare against measurements of the actual printer output."

Again, incorrect. There is absolutely no "expectation" of what the printed patches SHOULD look like (or measure). The testchart consists of more-or-less arbitrary RGB or, in the case of software RIPs,CMYK values, which the spectrophotometer simply records as the response or "characterization" of that printer.

The only slight deviation to this is software like X-Rite i1Profiler than can perform a profile optimization. In this case, once the initial profile is created and the device characteristics become "known", a 2nd optimized chart is generated that is actually "tuned" to the behavior of the printer.

-Terry

1 upvote
TerryW
By TerryW (Jul 29, 2011)

As a color management "professional", I feel compelled to point out two similar errors in your understanding of profiling. They're not grevious errors but they do cause confusion in the lay person's understanding of how display and printer profiling actually work.

In the section about "Monitor Calibration and Profiling", you state and I quote:

"The colorimeter attaches to the monitor screen and measures a series of onscreen color patches generated by its included software. Each patch has an absolute, device-independent value which the profiling software compares to the colorimeter-measured output from the monitor. "

Each patch in fact does NOT have an "absolute, device-independent value" or known value. "Device Independent" in a color management world implies either L*a*b* or XYZ values which is not the case here. These are strictly RGB device values that when measured by an instrument are used to characterize the display (and build a ICC profile of course).

Terry

1 upvote
JEPH
By JEPH (Jul 29, 2011)

This was a useful refresher for me. I hope you continue this series on printing.

I hope the series continues.

0 upvotes
steveh0607
By steveh0607 (Jul 29, 2011)

Overall this was a very good article. However, in reality the difference between Adobe RGB and sRGB is pretty close to nothing. If you always print everything at home then you can use Adobe RGB if you want, but if you ever send anything out to a commercial printer the file needs to be sRGB in order to look right.

Also one can make a fairly large print, about 16x20 or so, from a 8-10 mp file. You'd have to stick your face up to the print to see a difference.

0 upvotes
bull-deano
By bull-deano (Jul 29, 2011)

a. sRGB and Adobe RGB are close but render different results, especially when printing large format. We use Adobe RGB here at our stores but like you say it is common practise to use sRGB in the industry. This is a good point as it could lead to unexpected results.

Your second point is most welcome. You can of course make acceptable prints from files with less than 10mp's. We still use a Nikon D2h in various photographic projects and print up to and sometimes beyond 16x20. You can print below 180 and 200ppi with ease. I wish people would stop suggesting that you get unacceptable results if you do. How on earth do you think building wraps, billboards and grand format are printed.

0 upvotes
Amadou Diallo
By Amadou Diallo (Jul 29, 2011)

As mentioned in the article, one of the factors in determining acceptable ppi specs for print is viewing distance. If you looked at a billboard graphic from 3 feet away as opposed to 3 blocks away, its image quality would obviously look different.

0 upvotes
TerryW
By TerryW (Jul 29, 2011)

Exactly right amadou diallo....

On one hand, I get really tired of hearing folks claim that with their 2mp P&S they can print an image as big as the side of a barn and it "looks great"....at the same time, resolution isn't all it's cracked up to be either.

I'd rather hear people say..."my 2mp P&S printed as big as the side of a barn....*when viewed at the proper distance* (1/2 mile away)...looks just peachy to me". THAT would be correct. :-)

0 upvotes
bull-deano
By bull-deano (Jul 29, 2011)

You guys missed the point.

0 upvotes
TerryW
By TerryW (Jul 29, 2011)

Point being?......

...it's the QUALITY of the pixels that matters, not the NUMBER of pixels that determine final print quality? I agree...if that was your point.

As far as sRGB vs. AdobeRGB (vs. other color spaces such as ProPhotoRGB)....THAT would be a whole 'nuther article. There is no one-size-fits-all color space....the quality of your display, the type of printer and papers you intend to use, etc., all come into play. Sometimes sRGB can be too "big" a color space and other times AdobeRGB can actually be too small of a color space....many of the newer printers (and not-so new) can easily exceed AdobeRGB in certain color areas and depending on the media you're using.

-Terry

1 upvote
Rickard Hansson
By Rickard Hansson (Jul 29, 2011)

As you are talking "print workflow", then i would recommend you to do an third article where you talk about RIP:s.

For professional or semiprofessional users a RIP (EFI, GMG or whatever) is a more suitable setup than printing from the drivers.

With a RIP you can setup a workflow once and be sure that you print the same way each time.
Colour management is often easier and better as well and so is the software for making printer profiles (which is often an built in option in the RIP, to make printer profiles).

2 upvotes
bull-deano
By bull-deano (Jul 29, 2011)

Your point is excellent but in terms of mass market I don't think RIPs are an affordable workflow option. Unless there are cheaper ones out there that I don't know about.

0 upvotes
TerryW
By TerryW (Jul 29, 2011)

Rickard...a man after my own heart! (I install EFI, GMG, ColorBurst and a few other RIPs). And to "bull-deano", ColorBurst is affordable..probably costs less than the last lens you bought. If you don't trust your images to a cheap lens, why would you trust the final result (your prints) to anything less than the best...or at least "better"? :-)

Software RIPs arguably produce no better color (or wider color gamut) than the RGB printer drivers...but they do bring to the table a lot of other features such as nesting/ganging, print verification, etc. that you don't get with vendor's printer drivers.

I think one of the MAJOR advantages of software RIPs is that it totally gets around a lot of the color management problems caused by operating system "upgrades". It's been pretty well documented some of the color problems introduced on either Mac or Windows platforms with each successive OS update. This almost never happens with software RIPs.

-Terry

0 upvotes
Rickard Hansson
By Rickard Hansson (Jul 31, 2011)

Sure there are RIP:s that are affordable.

We install and maintain EFI and GMG rips for our customers, the GMG rips are not cheap, but very good ones, the EFI rips are options based and therefor the price varies. One option to go for is the EFI Xpress, it is available in two versions, one for Photo and one for Proofing.
The colorbyte offers is something i have not used before, so i cannot say anything of the prices for that.

It is just a thought.

Lots of people in here are photographers, they can spend money after money on cameras, lenses, tripods, lights, flashes and on top of that maybe an expensive printer such as 17+ inches printers from Epson, canon or HP.

The startng point for a RIP such as the EFI Express is approx half the price as buying the nikon 85 f1.4 AF-S here in sweden.

0 upvotes
hbx2004
By hbx2004 (Jul 29, 2011)

Well written article. Maybe difference between dpi and ppi should be explained just a bit more deeply.
Btw. last image (Sofv-1xi-with-artwork.jpg) doesn't show up enlarged.

0 upvotes
Amadou Diallo
By Amadou Diallo (Jul 29, 2011)

Image link is fixed (make sure you refresh and/or clear browser cache). Thanks for pointing this out.

0 upvotes
kff
By kff (Jul 29, 2011)

Is it possible to stop spam like ghjktty ?

0 upvotes
CdS
By CdS (Jul 28, 2011)

It would be very helpful to have a print or pdf version of these articles for future reference. Any chance of adding this feature?

4 upvotes
bettyboo
By bettyboo (Jul 29, 2011)

highlight the yext and then print. works on my lexmark z735 that I use rather htan my epsom R265 (its cheaper)

0 upvotes
Steve Bingham
By Steve Bingham (Jul 29, 2011)

Great article and I am sure it will help many.

Just a small addition. Many will see the 1440 (1440 x 720) on your Epson sample and think they need that as a PRINT resolution. Actually, the native resolution of the Epson 3880 (and most large format Epsons) is 360 (or 720 depending on what you have checked). Whatever resolution you put in for print resolution will be converted by the Epson to 360 - whether 200, 300, 360, or even 720. As information is then extrapolated from your file to 360 it is better to use an even multiple or fraction. Careful examination will reveal that the use of 240 (2/3), 360, or even 1440 (4x) will produce almost identical results. It takes a long time and knowledge to see the difference with a lupe. Without a lupe it is close to impossible. In addition, using 300, as opposed to 240 or 360 will also produce little visible difference - but it is there.

0 upvotes
TerryW
By TerryW (Jul 29, 2011)

Steve makes a good point....there is a huge difference between the resolution used for *rendering* an images vs. *printing* resolution. The rendering resolution...or the actual image resolution that gets sent to the printer...is typically half of the output resolution.

Even the printing resolution can be misleading....in the world of prepress and imagesetters/platesetters, the term used for output resolution is the actual *addressability* of the output device. These inkjet printers do NOT have 1440, 2880, etc. addressability....it's likely closer to 720dpi or maybe only 360dpi. Typically what you'll get out of the higher resolutions is less banding, higher ink load (bigger "color gamut" in some cases) and an apparent increase in detail....I say "apparent" because a 240ppi image may look sharper at 2880 than 720...but since the image is only 240ppi, it's not an actual increase in real detail (maybe better acutance?).

-Terry

0 upvotes
yichang
By yichang (Aug 5, 2011)

I got a cannon exus. Mostly, it works well. However, when I want to focus on different distance, that is nearly impossible. Where can I find the access to adjust the focal distance?
I want to take great pictures with some outstandings.

0 upvotes
Total comments: 36