Pages: [1] 2   Go Down

Author Topic: Interpreting Bill Atkinson's 14 balls  (Read 5684 times)

NAwlins_Contrarian

  • Full Member
  • ***
  • Offline Offline
  • Posts: 225
Interpreting Bill Atkinson's 14 balls
« on: May 03, 2017, 12:53:04 am »

I'm a newbie when it comes to inkjet photo printing and color management, so this may be a basic question, but please help me interpret my results with Bill Atkinson's 14 balls graphic, as set up for / printed on my Epson R280 (which I got (1) because I can learn from it and (2) because it's small enough not to take up valuable space).

The original is in ProPhoto RGB. I've tried various permutations: printing the original, converting to Abode RGB and printing that, and converting to the Epson profile for the R280 with premium glossy photo paper. The good news is that I can get my test prints to look very similar to what I see on my screen after I convert to the Epson printer+paper profile. The bad news is that I seem to have very little in the way of saturated blue available. I'm attaching two scaled-down versions of 14 balls. One I converted to the Epson printer+paper profile; the other I did that plus converted that result to sRGB so that hopefully it will display properly in browsers other than Firefox. (In both cases I used relative colorimetric rendering intent and black point compensation.) The first looks very much like the print I made.

Now the main question: the bottom-right ball is almost totally black. Am I correct in thinking that this means that this printer+paper combination cannot print a highly-saturated darker blue? Similarly but not as severely, the upper-right ball has a nearly-black outer ring before transitioning to blue. Does this mean that the printer+paper combination can only print highly-saturated blue toward the lighter end of the range? The black / white, red, and yellow balls look pretty good. The green balls have shifted from a lime / neon green to a darker green. The cyan balls are much less saturated. The magenta balls are somewhat less saturated and maybe shifted a bit. Am I correctly interpreting what this little test appears to have told me about what the R280 can and cannot print?

Thanks!

P.S.
I've mostly done the printing in DxO Optics Pro 11 Elite. I'd love a good light and free printing program--although I think I can make Windows Photo Viewer (which appears to be color-managed on Windows 7-64 bit Home Premium) do much of what I need. Lightroom is too bureaucratic and GIMP's printing is too buggy.

EDIT: The thumbnail of first attachment's color is far off; if you click on it and open the actual file in color-managed software, it is correct. Presumably this website is generating a thumbnail and subsequent intermediate-size version by taking the color information and stripping off the profile, similar to an assign-to-profile.
« Last Edit: May 03, 2017, 12:57:58 am by NAwlins_Contrarian »
Logged

schertz

  • Full Member
  • ***
  • Offline Offline
  • Posts: 171
    • Adobe Portfolio
Re: Interpreting Bill Atkinson's 14 balls
« Reply #1 on: May 03, 2017, 01:11:38 am »

The colours you're seeing here have to do with both the capabilities of the printer and the icc profile used to convert the colours for printing. Some of the colours, ESPECIALLY in the dark blue ball are both out of gamut and are not even "real" colours. I suggest you look at the following two resources:

Andrew Rodney's video in icc profiles relating to the blue balls and the quality of some of Epson's canned profiles (premium lustre):

http://digitaldog.net/files/Not_All_Profiles_are_created_equally.mp4  (jump to about 20 mins where he talks about the dark blues)

and also a long thread on DPreview discussing what the Bill's Balls actually are:

https://www.dpreview.com/forums/post/58341246 

warning: a lot of the discussion in the DPR thread breaks down into the all-too-frequent Andrew _vs_Charles2 bickering back and forth, but there's some useful nuggets of data in there if you can get through it)

Mike
Logged

Doug Gray

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2197
Re: Interpreting Bill Atkinson's 14 balls
« Reply #2 on: May 03, 2017, 02:38:56 am »

The colours you're seeing here have to do with both the capabilities of the printer and the icc profile used to convert the colours for printing. Some of the colours, ESPECIALLY in the dark blue ball are both out of gamut and are not even "real" colours. I suggest you look at the following two resources:

Andrew Rodney's video in icc profiles relating to the blue balls and the quality of some of Epson's canned profiles (premium lustre):

http://digitaldog.net/files/Not_All_Profiles_are_created_equally.mp4  (jump to about 20 mins where he talks about the dark blues)

and also a long thread on DPreview discussing what the Bill's Balls actually are:

https://www.dpreview.com/forums/post/58341246 

warning: a lot of the discussion in the DPR thread breaks down into the all-too-frequent Andrew _vs_Charles2 bickering back and forth, but there's some useful nuggets of data in there if you can get through it)

Mike

Actually not much bickering between Charles and Andrew in the thread. Pretty useful discussion.

What's interesting in that thread is that a major misconception about Bill Atkinson's Balls was put to rest. That the Balls were originally in Lab and that they are intended to be rendered from ProPhoto RGB. According to Bill Atkinson himself, the Balls were created in untagged RGB and they were designed to be printed directly in printer space. That is to say they should be printed in exactly the same way a patch chart used to create an ICC profile would be printed.

This actually makes sense. For one thing every single RGB color in in the printer's gamut. It's also a way to visually check the smoothness of the printer's native driver. Unevenness in device space indicates areas that would be difficult for a profile to correct.

This is quite a departure from the long thought belief that the Balls somehow came from Lab space and that they should be rendered as ProPhoto RGB images though ICC profiles and the result is somehow meaningful. The large majority of the balls have significant "colors" that are imaginary and thus have to somehow be turned into something physically realizable. For instance the blue ball has a luminance value (the L in Lab) of less than 1 over virtually the entire surface. Similarly, it's colors are imaginary. So how should it be actually rendered? On a monitor those ProPhoto colors are rendered as blues because they are converted, by a well defined math process in PCS XYZ space and the result is clipped at 0 and 255. This move the colors from imaginary ones into actual ones but with a huge boost to luminance. However, the ICC path from ProPhoto to printer is different and it goes through PCS Lab. The 3D LUTs look up the values to print using the cubes at L=0 with b clipped at -128.

Apparently some printers interpret this to print these near black while others print closer to what the monitor shows. It's quite arguable that the black interpretation is more correct (L=0 and all) though it certainly is a surprise just looking at how a monitor renders the colors.

Anyway, I give great credit to Andrew for taking the time to contact Bill Atkinson and publishing his response.  It went against much of what has been propagated about the origin and use of Atkinson's Balls even including some of Andrew's earlier works.  It says a great deal about Andrew's integrity.
« Last Edit: May 03, 2017, 02:46:11 am by Doug Gray »
Logged

NAwlins_Contrarian

  • Full Member
  • ***
  • Offline Offline
  • Posts: 225
Re: Interpreting Bill Atkinson's 14 balls
« Reply #3 on: May 03, 2017, 10:00:24 am »

Thank you both for the helpful responses. I will have to watch those videos; I've watched a few of the Rodney videos and read several of his articles, and they have helped me along a good bit.

Quote
What's interesting in that thread is that a major misconception about Bill Atkinson's Balls was put to rest. That the Balls were originally in Lab and that they are intended to be rendered from ProPhoto RGB. According to Bill Atkinson himself, the Balls were created in untagged RGB and they were designed to be printed directly in printer space. That is to say they should be printed in exactly the same way a patch chart used to create an ICC profile would be printed.

I don't know the history, but I did read this thread (forum.luminous-landscape.com/index.php?topic=117131) about his printer test page, wherein someone claiming to he offered anyone to download his files from his Dropbox (https://www.dropbox.com/sh/ka35g3gtyd10823/DfQqZRYNsW). I downloaded the 14 ball file directly from there. Even opeing it in GIMP--which unlike Lightroom does not automatically use ProPhoto RGB as a working space--clearly indicates that the file I downloaded is in ProPhoto RGB. That being the case, I don't know how I "print[ them] directly in printer space". I can convert to the profile for the printer and paper, or I can assign to a profile for the printer and paper--or I guess I can both (1) turn off the printer driver's color management and (2) print them from a program that is not color-managed? But if the file I downloaded from his Dropbox is in fact in ProPhoto RGB, does that really accomplish what was intended? Do you have a recommendation on how to run this test?

Quote
The large majority of the balls have significant "colors" that are imaginary and thus have to somehow be turned into something physically realizable. For instance the blue ball has a luminance value (the L in Lab) of less than 1 over virtually the entire surface. Similarly, it's colors are imaginary. So how should it be actually rendered?

Ah, well that might (?) explain why a monitor (or video card plus software) would crush them into a displayable blue but a printer would render them black or nearly so--I guess. Rendering intents remain an area where (1) I clearly don't know / understand everything, (2) some software lacks potentially-useful controls, and (3) I have found misinformation even from people who are (and are widely regarded as) experts.

Thanks!
Logged

Doug Gray

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2197
Re: Interpreting Bill Atkinson's 14 balls
« Reply #4 on: May 03, 2017, 03:15:27 pm »

I've made a more detailed breakdown of the "color" distribution on Andrew Rodney's test image "Gamut_Test_File_Flat.tiff"

Basically, I broke down the image colors into 4 groups.

  • Imaginary colors that don't exist in the human gamut. These are shown as black
  • Colors that exceed the Mac Adam limits. These are real colors that can exist, for instance, with laser monochrome sources but they don't exist as colors on surfaces illuminated with D50 hence can never be printed even if a printer had millions of distinct, and "perfect" is some sense inks. These are shown in Cyan.
  • Next are colors that are outside of my printer's gamut by Delta E of 5 or more. These will, of course vary with printers but generally are similar for printers with CYMK inks. These colors are shown in red
  • Finally, there are printable colors. All the "white" areas are in gamut, printable colors

As for printing Bill's Balls, the Adobe utility program "Adobe Color Print Utility" is what people generally use. It goes directly to device space. By definition, RGB values in device space are within the printer's gamut.
« Last Edit: May 03, 2017, 03:20:54 pm by Doug Gray »
Logged

Mark D Segal

  • Contributor
  • Sr. Member
  • *
  • Offline Offline
  • Posts: 12512
    • http://www.markdsegal.com
Re: Interpreting Bill Atkinson's 14 balls
« Reply #5 on: May 03, 2017, 03:23:31 pm »

Doug,

Thanks for your, as usual, interesting and insightful technical explanations, but on this one, at a more fundamental level for quite some time I've been having a bit of problem trying to fathom the practical, operational significance of a test like this. What is a bunch of unprintable, or imaginary colours going to really tell me about how my printer will perform in printing real world photographs. As you would know, I'm all for validation of performance using objective, repeatable methodology, but it must be somehow grounded in or well-related to reality.
Logged
Mark D Segal (formerly MarkDS)
Author: "Scanning Workflows with SilverFast 8....."

Doug Gray

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2197
Re: Interpreting Bill Atkinson's 14 balls
« Reply #6 on: May 03, 2017, 04:06:58 pm »

Doug,

Thanks for your, as usual, interesting and insightful technical explanations, but on this one, at a more fundamental level for quite some time I've been having a bit of problem trying to fathom the practical, operational significance of a test like this. What is a bunch of unprintable, or imaginary colours going to really tell me about how my printer will perform in printing real world photographs. As you would know, I'm all for validation of performance using objective, repeatable methodology, but it must be somehow grounded in or well-related to reality.

Actually, I made a program to do this some time back. Mostly I was just interested in whether an image I had was within the printer's gamut or not. Adobe soft proofing is quite limited and is stuck on using Delta E 1976 ( >6 )for OOG mask generation. Another interest is the quality of printing when colors are within the printer's gamut. For instance, smoothness of gradients, lack of banding, and such. Color accuracy can be determined reasonably well by printing in gamut patches distributed differently than those the profile was made from. But smoothness is another kettle of fish. Anyway, this is the area that I have most interest.

On a secondary level, people often suggest using reference images to test their color management workflow and I think that images which contain the more extreme colors are not as useful for checking profile quality as images that are within gamut.

This is not to say they aren't useful. People that use ProPhoto RGB in order to print colors that are available may, by various techniques, unknowingly wind up with image colors a long distance from what a printer can print. In particular, printer profiles that map OOG from PCS LAB space in a similar way to the way a ProPhoto image is mapped to  a monitor (almost always in XYZ space) will contain fewer surprises for photographers using a ProPhoto workflow. So I do value that in a profile and concur with Andrew's observations about that.

Tossing in the Mac Adam limits and imaginary colors was just a quick way for me to visualize those extremes in an image. For instance the Roman 16 image set has a few colors slightly outside my printer gamut but no imaginary or Mac Adam limited colors. I consider those the best reference images out there.
« Last Edit: May 03, 2017, 04:10:23 pm by Doug Gray »
Logged

NAwlins_Contrarian

  • Full Member
  • ***
  • Offline Offline
  • Posts: 225
Re: Interpreting Bill Atkinson's 14 balls
« Reply #7 on: May 04, 2017, 12:12:38 am »

Doug, thanks for the extended and technical explanation. Wow, not a lot of 'white' in your transformation of those 14 balls! I am definitely observing some banding (especially cyan, but significantly in the darker / lower magenta, red, and to some extent green balls. As for a printer test: I will have to try the Adobe utility (assuming I have it or can download it for free: I have Lightroom 6 and PSE 14, but not full Photoshop). Presumably (please correct me if I'm wrong) printing balls that go 128 in to 255 and 127 in to 0 (or whatever for the number of bits) without any color profile will just make the printer show you exactly how far its gamut can go in KCMYRGB.

I'm still learning a lot here. On the one hand, I guess I should be happy that the printed 14 balls look very much like what I see on-screen when I convert to the printer+paper profile. On the other hand, I guess there's no cheap and easy answer to print those saturated bright greens. Tonight I made some little prints for my younger son's school project on Montana. One was a picture (I want to say from Wikimedia?) of mountains in the distance (what he wanted to show) with bright green grass in the foreground and a lake in between. He commented, totally unsolicited, on how the lake water looked a bit gray (on screen it looks more a deeper blue). He is nine years old, and knows nothing about color except what is in a box of crayons. I noticed the somewhat muted greens in the grass. I guess that's just what an R280 can give you.
Logged

Mark D Segal

  • Contributor
  • Sr. Member
  • *
  • Offline Offline
  • Posts: 12512
    • http://www.markdsegal.com
Re: Interpreting Bill Atkinson's 14 balls
« Reply #8 on: May 04, 2017, 07:51:38 am »

Saturated bright greens is an area of gamut that challenges even the Epson P800 and Canon Pro-1000 printers. One needs the extended gamut of the Epson x900 or new SC-Pxxxx printers using wide gamut papers such as Legacy Baryta/Platine, IGFS, PhotoRag Baryta etc. to not clip those colours. One see this readily in green leaves segment of Bill Atkinson's excellent printer test image - more info in my Canon Pro-1000 review on this website.
Logged
Mark D Segal (formerly MarkDS)
Author: "Scanning Workflows with SilverFast 8....."

NAwlins_Contrarian

  • Full Member
  • ***
  • Offline Offline
  • Posts: 225
Re: Interpreting Bill Atkinson's 14 balls
« Reply #9 on: May 05, 2017, 12:38:32 am »

Okay, I downloaded the Adobe Color Printer Utility and used it to print the Atkinson fourteen balls file, basically as I downloaded it from his Dropbox,* with color management turned off at the printer driver. The CMYK parts look about as I expected. But the RGB parts are very different. The red balls print almost more orange and the blue balls are totally purple. Is this normal / as expected?

I tend to suspect that it is normal; that it's just that the untagged 'red' tells the printer to drop maximum amounts of magenta and yellow ink, and the untagged 'blue' tells the printer to drop maximum amounts of magenta and cyan ink; and that those ink colors are not perfectly balanced out to give the 'right' blended colors with this sort of crude approach. But you please tell me whether this is so.

Also:
Quote
Saturated bright greens is an area of gamut that challenges even the Epson P800 and Canon Pro-1000 printers. One needs the extended gamut of the Epson x900 or new SC-Pxxxx printers using wide gamut papers such as Legacy Baryta/Platine, IGFS, PhotoRag Baryta etc. to not clip those colours. One see this readily in green leaves segment of Bill Atkinson's excellent printer test image

I was pleasantly surprised by how well my local Costco's Fuji Frontier DL 650 Pro printed the bright green foliage on the left side of the Atkinson printer test image. Trying that (Costco printing with Dry Creek profiles) has been part of my learning and experimentation. Supposedly 'under the hood' (or 'under the bonnet' to those of you concerned with colour?) the Frontier DL 650 Pro is an Epson inkjet.

*It seems worth mentioning that this printing approach is totally 'dumb', and if you want the entire image printed, you need to scale it ahead of time to the exact size you need for the paper size and printer resolution. Also, you have to save it as a TIFF.
Logged

Doug Gray

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2197
Re: Interpreting Bill Atkinson's 14 balls
« Reply #10 on: May 05, 2017, 12:40:50 pm »

Okay, I downloaded the Adobe Color Printer Utility and used it to print the Atkinson fourteen balls file, basically as I downloaded it from his Dropbox,* with color management turned off at the printer driver. The CMYK parts look about as I expected. But the RGB parts are very different. The red balls print almost more orange and the blue balls are totally purple. Is this normal / as expected?

Yes, normal. In fact different printers models, and especially brands, will print very different tones and sometimes hues w/o color management. My Epson prints much darker than my Canon.

None of that matters when printing w/o color management because the ICC profile's job is to map requested colors to what can be printed. They don't have to be close, just smooth.  What is important to look for is banding. To the extend banding exists the interpolation algorithms have a hard time tracking the changes. Printers with the least banding tend to work better with smaller target patch sets because they don't need as many adjacent colors to track abrupt changes.

Perhaps the best use would be when using non-OEM inks. Printing non-OEM inks and comparing them to OEM ones can be used as a quick test for suitability. Any shift in color means you need  a custom profile but banding may mean the non-OEM ink just isn't suitable even with a custom profile.
« Last Edit: May 05, 2017, 01:51:28 pm by Doug Gray »
Logged

deanwork

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2400
Re: Interpreting Bill Atkinson's 14 balls
« Reply #11 on: May 05, 2017, 06:54:17 pm »

When printing the other Atkinson print test chart, the one with the strawberries, monument valley, babies, color scales, etc,  should this target be sent to the printer to be mapped by the custom profile as adobe rgb or pro photo , if pro photo is your standard working space? I use this chart to print newly created ic profiles and to test the gamut of various printers I use. Is this the best overall target out there for evaluating the practical considerations one will encounter with a variety of hues and subject matter? It is the best one I've come across and I've used it for about 12 years now. But this target was created originally long before pro photo was in use with pigment inkjet printers.
Logged

Mark D Segal

  • Contributor
  • Sr. Member
  • *
  • Offline Offline
  • Posts: 12512
    • http://www.markdsegal.com
Re: Interpreting Bill Atkinson's 14 balls
« Reply #12 on: May 05, 2017, 07:36:26 pm »

When printing the other Atkinson print test chart, the one with the strawberries, monument valley, babies, color scales, etc,  should this target be sent to the printer to be mapped by the custom profile as adobe rgb or pro photo , if pro photo is your standard working space? I use this chart to print newly created ic profiles and to test the gamut of various printers I use. Is this the best overall target out there for evaluating the practical considerations one will encounter with a variety of hues and subject matter? It is the best one I've come across and I've used it for about 12 years now. But this target was created originally long before pro photo was in use with pigment inkjet printers.

It's a very good evaluation image to print in ProPhoto space because it contains a limited spectrum of colour beyond ARGB(98), and indeed somewhat beyond the gamut of wide gamut printers such as the Epson x900 and SC-Px000 series. I agree with Doug Gray that one wants to test printers primarily with in gamut colours; to my mind the idea being that we don't want to confuse observation of printer performance for the colours it can reproduce naturally with with the characteristics of gamut compression provided by Rendering Intents. However, the limited inclusion of what one may call "stretch targets" in an evaluation image that challenges gamut boundaries can be useful to see differences of gamut rendition between different printers, papers, Rendering Intents and profiles. I think the Atkinson target accommodates all this very nicely and it is one of my primary "go to" evaluation tools in the printer and paper reviews I prepare for this website. What I really like most about this test page is that one quickly memorizes its "memory colours" to an extent that if a print one makes with it is out of whack, it is immediately and obviously noticeable.
Logged
Mark D Segal (formerly MarkDS)
Author: "Scanning Workflows with SilverFast 8....."

Wayne Fox

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 4237
    • waynefox.com
Re: Interpreting Bill Atkinson's 14 balls
« Reply #13 on: May 06, 2017, 12:05:36 am »


What's interesting in that thread is that a major misconception about Bill Atkinson's Balls was put to rest. That the Balls were originally in Lab and that they are intended to be rendered from ProPhoto RGB. According to Bill Atkinson himself, the Balls were created in untagged RGB and they were designed to be printed directly in printer space. That is to say they should be printed in exactly the same way a patch chart used to create an ICC profile would be printed.

Seems a little odd, but if Bill himself said so then I can't question it.(I assume that's what you are saying?)

  I downloaded all of bills test images over a decade ago when Bill offered them in the now defunct Apple .mac public folders. I have the two original folders in my drop box. The original folders were offered as "print with profile" and "print without profile". While the 14ball file hadn't been created at that time, the larger 28ball test print had, and is definitely in the "print with profile" folder as created by bill himself.
Logged

NAwlins_Contrarian

  • Full Member
  • ***
  • Offline Offline
  • Posts: 225
Re: Interpreting Bill Atkinson's 14 balls
« Reply #14 on: May 06, 2017, 12:45:04 am »

Quote
When printing the other Atkinson print test chart, the one with the strawberries, monument valley, babies, color scales, etc,  should this target be sent to the printer to be mapped by the custom profile as adobe rgb or pro photo , if pro photo is your standard working space?

and

Quote
It's a very good evaluation image to print in ProPhoto space ....

Indeed the Atkinson printer test page seems great--but I cannot reconcile the versions I have / make them work as expected. Presumably to test a printer, you get a version, preferably a 16-bit TIFF, in a wide working color space like ProPhoto RGB; then convert it to a profile for the specific printer and desired paper, maybe experimenting with different rendering intents; and then print that. From somewhere I'd gotten an sRGB JPEG version, which looks normal. With the recent link to the Atkinson Dropbox, I downloaded directly much larger TIFF versions, 39 MB and 100 MB. These versions look very different from the older-to-me JPEG, with the big, directly-downloaded files' grays looking very orangeish-brown. The files either have no color space info, or else actually are in sRGB. I tried assigning (not converting to) various profiles (sRGB, Adobe RBG, ProPhoto RGB, ColorMatch RGB). None of the usual suspects produced an even semi-normal looking image, and assigning ProPhoto made the grays look positively pink.
Logged

Mark D Segal

  • Contributor
  • Sr. Member
  • *
  • Offline Offline
  • Posts: 12512
    • http://www.markdsegal.com
Re: Interpreting Bill Atkinson's 14 balls
« Reply #15 on: May 06, 2017, 07:58:29 am »

According to the version of the Atkinson Printer Test Image I have and from my experience using it, Doug Gray is correct: it was originally provided in Lab mode and it prints very well in ProPhoto space if the printer and the profile are both satisfactory. Of course its appearance varies considerably between PK and MK paper types, but that's completely normal and expected.
Logged
Mark D Segal (formerly MarkDS)
Author: "Scanning Workflows with SilverFast 8....."

NAwlins_Contrarian

  • Full Member
  • ***
  • Offline Offline
  • Posts: 225
Re: Interpreting Bill Atkinson's 14 balls
« Reply #16 on: May 07, 2017, 01:58:55 am »

Quote
According to the version of the Atkinson Printer Test Image I have and from my experience using it, Doug Gray is correct: it was originally provided in Lab mode ....

Doh! Thanks. Now that you say it, I do seem to recall having heard that. Lightroom opens the big file I downloaded (at least apparently) correctly, and I exported versions of it in the profile for the printer+paper I wanted. Those exported versions also seem right. The LAB part also explains why Windows Photo Viewer was utterly clueless on how to open the file, and GIMP was off. Somewhat disappointingly, DxO Optics Pro apparently will not open it at all. As much as I often find Lightroom a bureaucratic PITA for my purposes, it does have its capabilities.
Logged

deanwork

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2400
Re: Interpreting Bill Atkinson's 14 ball
« Reply #17 on: May 07, 2017, 12:18:00 pm »

Yes I find that target an extremely useful tool for quickly evaluating the practical considerations of my custom profiles, comparing the different printers and inksets usable  gamut, gray neutrality in an rgb environment, shadow and highlight separation, and overall mid tone contrast of a particular paper. After 12 years of using it I haven't found anything more useful. It is a great thing to compare various papers with.  But as been stated here, there are a lot of versions of it out there and it would be important to use the lab verson for converting to a wide gamut space like pro photo.


quote author=Mark D Segal link=topic=117752.msg975761#msg975761 date=1494027386]
It's a very good evaluation image to print in ProPhoto space because it contains a limited spectrum of colour beyond ARGB(98), and indeed somewhat beyond the gamut of wide gamut printers such as the Epson x900 and SC-Px000 series. I agree with Doug Gray that one wants to test printers primarily with in gamut colours; to my mind the idea being that we don't want to confuse observation of printer performance for the colours it can reproduce naturally with with the characteristics of gamut compression provided by Rendering Intents. However, the limited inclusion of what one may call "stretch targets" in an evaluation image that challenges gamut boundaries can be useful to see differences of gamut rendition between different printers, papers, Rendering Intents and profiles. I think the Atkinson target accommodates all this very nicely and it is one of my primary "go to" evaluation tools in the printer and paper reviews I prepare for this website. What I really like most about this test page is that one quickly memorizes its "memory colours" to an extent that if a print one makes with it is out of whack, it is immediately and obviously noticeable.
[/quote]
Logged

rdonson

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 3263
Re: Interpreting Bill Atkinson's 14 balls
« Reply #18 on: May 07, 2017, 01:48:21 pm »

Folks, I've learned a lot from this thread but.....  I can't help but smile every time I read the subject line.... I wonder if Bill would approve  ;D
Logged
Regards,
Ron

NAwlins_Contrarian

  • Full Member
  • ***
  • Offline Offline
  • Posts: 225
Re: Interpreting Bill Atkinson's 14 balls
« Reply #19 on: May 07, 2017, 05:03:10 pm »

Quote
But as been stated here, there are a lot of versions of it out there and it would be important to use the lab verson for converting to a wide gamut space like pro photo.

I've tried to go back and delete the two or three versions I had as sRGB JPEGs. They would seem artificially limited. For the moment I've added the big LAB file to my Lightroom catalog, then I export versions in the ICC profile specific to the printer+paper combinations I want to test. By using export you lose control over the rendering intent from LAB to the color space associated with the profile. I think I can get around that by printing to a file, but I need to experiment. In the meantime, I think that Lightroom export uses relative colorimetric, which is probably what I'd want for a test anyway.

Quote
I can't help but smile every time I read the subject line.... I wonder if Bill would approve

That thought crossed my mind a couple of times, especially in the discussion of the blue ones being the most problematic. It does not help that I have a lowbrow sense of humor. But as a wise man once put it, "If we weren't all crazy, we would go insane."
« Last Edit: May 07, 2017, 05:08:11 pm by NAwlins_Contrarian »
Logged
Pages: [1] 2   Go Up