Testing Your Pro Lab v1

Testing your Pro Lab v1.0  | Using sRGB images

For this "testing" I have chosen to use images embedded with the sRGB colorspace. If you'd like to do your own test with your Pro Lab then you may choose to use the Adobe RGB (1998) colorspace. Either way, it is a legitimate test. I have also geared my testing in a way that should be easily duplicated. This is not about the science of color, but simple ways to help you achieve better results.

Here are some of the images I used in my testing. There were several synthetic targets used for evaluation as well: those were used for color geek purposes.

 

 

For the test, I sent two sets of files to my Pro Lab to be printed on Lustre photo paper with no color correction. Set <1> was embedded (tagged) with the sRGB colorspace and Set <2> were converted using Relative Colorimetric w/Black Point Compensation to a generic RA-4 processor-based ICC profile and then converted back and embedded with the sRGB profile. This test had no untagged images. (On a side note: after comparing several online RA-4 processor-based ICC profiles I determined they all have the same basic color gamut).

In our lab, I printed Set <3> using the same sRGB embedded images above through an EFI XF RIP to an Epson Stylus Pro 9900 HDR printer using Epson Premium Photo Glossy Paper.

A                 B

Diagram A shows a 2-D overlay of the typical Fuji Frontier output profile on top of the common sRGB working colorspace. If we just looked at a 2-D diagram it would seem that the majority of Frontier colorspace is smaller than our sRGB working colorspace. Yet, to get a better understanding of how two colorspaces compare to each other you should always look at them in 3-D as seen in Diagram B above. Although this is a static image (you can't rotate this image) you can see the Frontier colorspace (solid red) is protruding out (more saturated) from the sRGB colorspace in the darker blues and reds; while being completely absorbed in the highlight areas. Just for reference, even the Adobe RGB (1988) colorspace doesn't fully encompass the Frontier ICC Profile. It too had less saturation in the dark blues and reds.

What I Received Back
Hardware/software being used: a calibrated Apple 15" MBP and CS5. My ambient photo viewing conditions included being in a gray room and using a D50 light booth.

With a calibrated monitor and proper viewing conditions the two sets of Lab prints and the Epson prints were all very close to the monitor. The Pro Lab prints had a slight red cast noticeable in the skin tones, browns and purples, but overall, very good. As expected, the yellows from the RA-4 process had a slight orange cast to them compared to the clean yellow from the Epson print. I did adjust my monitor's brightness level slightly and if I had a monitor capable of manually adjusting the RGB levels I could have easily matched my Pro Labs prints perfectly.

To each other, there were slight differences, yet these differences were noticeable. Nothing was unexceptable, it really started to come down to personal preference. 

<1>. Set <1> which were straight sRGB image had the least saturation. Most notably the girl in the red image; the reds were the most desaturated of the 3 sets. There were also notable saturation difference in the picture of the young girl painting. The saturated paint cans and flowers on the wall were visually less saturated through the RA-4 process. And as stated above, the yellows had more of an orangish-yellow hue. Yet I will go back and say, perceptually, the entire Set <1> to itself looked great, the orangish-yellows looked yellow if only looking at the Set <1> images. And the reds looked red, you wouldn't necessarily look at them and say where did my reds go.

<2>. What was more interesting is Set <2>. So these are images were converted from sRGB to a generic Frontier profile then converted back to sRGB. I converted it back to sRGB incase the Pro Labs workflow assumes all files are sRGB--leaving it tagged with a Frontier profile would have resulted in a set of terrible looking prints. You could try tagging your files with an output device colorspace, I know people who do, but you need to make sure it is a "good" profile before doing so. 

Not surprisingly the majority of the images looked the same as Set <1> (straight sRGB). I thought this would be true since the majority of the Frontier profile fit into the sRGB colorspace. Yet, where  there were differences, they were very different. The saturation level of the reds and blues were stronger in Set <2>. Especially the girl with the red background and the young painter images produced more saturated colors and better matched the Epson prints. The other pictures may have had a slight improvement, but not nearly as noticeable as in the reds and super saturated colors. If we take into account the 3-D diagram above, by first converting it to our Output Colorspace it expanded the red and blue outward towards the out boundary of the sRGB colorspace. It didn't take it out of sRGB colorspace so when converting it back to sRGB colorspace all we are really doing it putting a nice wrapper on our image we know the workflow will understand.

Overall, the saturation of the reds and blues was a little too much for me. Sure I received a much more "red" and "blue" print, but I think to the detriment of the overall image. I personally preferred Set <1> over Set <2>. But I could see a need at some point to send pre-converted files if reds and highly saturated colors where important to hold in an image.

<3>. And then there was Set <3>, the Epson prints. By far the most saturated of all three sets. The yellows were crisp and the reds saturated. I was surprised the greens and magentas looked the same. I figured the whole set of prints would have a greater saturation level, but not really. I know my images were all clipped to the sRGB colorspace, but I was still impressed to how close all three sets compared to each other.

I loved the colors of Set <3> and would pick it overall. Yet, I know you don't necessarily have that option. I was impressed with my Pro Labs prints overall and after doing some readings with my new X-Rite eXact spectrodensitometer I found my Pro Labs prints to be in perfect Gray Balance. In the end if your Pro Lab can achieve gray balance, having a calibrated monitor and working in sRGB will help keep you efficient and effective.

For further testing you may want to take a range of RAW photos and apply/tag the sRGB colorspace to one set and the Adobe RGB (1998) colorspace to the other set and see what you get back. Make sure you name then sRGB and Adobe at the beginning of the file name so you will be able to identify them from the back printing.

**The Adobe RGB (1998) colorspace is what I recommend, but only after evaluating your computer hardware and Pro Labs color workflow ability.**

The Granger Rainbow  | A visualization tool for evaluating ICC profiles

Reading through the different photography forums there seems to be one topic leading the way: HELP! I am getting unexpected results from my pro lab. Although this is a common issue we have all encountered, in many ways, it is impossible to solve with a generic answer. Your entire workflow starting with the image capturing device, to the imaging editing software, to the final output device, would need to be looked at to help specifically answer this question.

With this blog, what I'd like to do is give you helpful, generic, tips and tools that you can use to gain an understanding of your workflow, especially once you upload it to your pro lab. This can even help you if are having similar problems with your home printer.

First, I'd like to give credit to the book Real World Color Management, 2nd Edition, by Bruce Fraser, Chris Murphy and Fred Bunting. This would be my initial tip to you, BUY THIS BOOK! It's a little out dated, as it was published in 2004, but still a top notch reference book on color theory and color management. For reference, I found out about the Granger Rainbow on page 238.

To download my Granger Rainbow, developed by Dr. Ed Granger, right-click on the Granger Rainbow image in Figure 1 and select "Download Linked File". The book also tells you how to build one in Photoshop which is what I did. Depending on your browser, browser settings, monitor, etc. you may or may not see exactly what I will be describing. If you were to download all of these images and open them up in Photoshop, keeping their embedded profile, you'd see what I will be describing.

 

Figure 1 | The Granger Rainbow with the sRGB profile embedded

Figure 1 represents the Granger Rainbow embedded with the standard sRGB IEC61966-2.1.icc profile found in the Adobe Creative Suite. To the right is a 2-D Lab view of the sRGB color gamut showing the outer most points at L*50. For this Blog, my working colorspace will be sRGB.

Figure 2 | sRGB converted using a Durst Lambda ICC profile found on the internet

Figure 2 shows the Granger Rainbow converted using a Durst Lambda profile. A Durst Lambda is an RGB laser output device that uses light sensitive photographic paper through an RA-4 wet processor. Devices such as these also include printers from Fuji, Mitsubishi and Agfa and are used in almost every pro lab. To the right is a 2-D Lab view of the Durst Lambda profile (red) imposed on top of the sRGB profile at L*50. For a complete comparison between the two profiles we would need to look at them in 3-D. This is more to help show the expected results on a typical RA-4 wet process device.

First you will notice the overall dynamic range (contrast) has diminished. The photographic wet process cannot produce a very deep black. Secondly, the overall colors are muted (not as vivid) due to most of the output device's gamut being smaller than our working colorspace.

Have you ever wondered why your reds never come out as bright as you remembered? Not only does the 2-D diagram show the smaller gamut in the red area, the Granger Rainbow quickly shows you the clipping (desaturation) that occurs in the reds.

Something else you may have noticed with your pro lab prints; are your yellows as pure as you remembered? Perceptually the yellow may appear yellow in an image, but as seen in the Granger Rainbow and in the 2-D diagram your yellows are printing on the orange side. See how the red line of the lambda diagram falls short of the yellow in sRGB colorspace? Yellow through this process actually creates more of an orangish-yellow then a pure yellow. You'd know right away way your sunflowers did not print as expected if you had applied your pro lab's ICC profile to the Granger Rainbow.

Figure 3 | sRGB converted using an Epson x900 Luster ICC profile from the Epson DVD

Figure 3 shows the Granger Rainbow converted using the Epson 9900 Luster 260 ICC profile that comes on the Epson Driver disk. The first thing you will notice is how vivid this profile is over the RGB laser device, no muted red here. The Epson is not removing color like a laser device, it is adding color making for a larger gamut throughout all levels of L*. You can also see the higher d-max capability of an inkjet printer. Blacks can be made up of multiple colors to add depth/richness to the black. Where as an RGB laser device removes color to achieve d-max limiting the overall depth. And look at yellow, your sunflowers be more like you remember because this printer is capable of achieving a more pure yellow--look at the difference between Figure 2 and Figure 3 at yellow's peak.

These are just two examples of using a Granger Rainbow to help evaluate the output device you are sending your files to. Not all pro labs will have an ICC profile for you, but you can typically find just about anything on-line. It's more important you know the process in which you are sending your files to so you can at least find they same type of profile on-line. In my research, I compared a Fuji Frontier profile, multiple Durst Lambda profiles, and an Agfa profile in ColorThink (a different blog) and they were all very similar in nature. In some cases they looked like the same profile. Generically speaking, a well profiled HP, Cannon and Epson printer will look very similar when it comes to photographic images based on perception.

Click Here for more information about The Advantage Group.

Next... I am going to "Test" my pro lab's workflow and then tell you how you can do the same.

Welcome

First of all, I'd like to welcome you to The Advantage Group's new "Printing Tips" blog.

I've created this blog as a vessel to pass along answers to commonly asked questions. Questions pertaining to color management, software, printers, and medias.