Luminous Landscape Forum
Raw & Post Processing, Printing => Colour Management => Topic started by: Brookie on February 10, 2020, 12:19:26 pm
-
Spyder is offering an upgrade for Spyder5 owners to the newer SpyderX. So I am considering upgrading to the XPro or XElite.
Questions - do people prefer the Xrite tool or the Spyder tool, which model, and why?
Spyder’s comparison of the two new models indicates that with the Elite soft proofing of prints is a feature (but not on the Pro version) - what does this allow you to do in soft proofing that you can’t already do in Lightroom?
I am using this on a MacBookPro, a BenQ SW2700PT and I print with a Canon Pixma Pro-10. I don’t anticipate making my own paper profiles, FWIW.
Thanks for your input!
-
This kind of says it all:
https://forum.luminous-landscape.com/index.php?topic=103094.msg1004707#msg1004707
(http://digitaldog.net/files/BadSpyderImage.jpg)
-
This kind of says it all:
https://forum.luminous-landscape.com/index.php?topic=103094.msg1004707#msg1004707
(http://digitaldog.net/files/BadSpyderImage.jpg)
No mention of the SpyderX though.
The SpyderX uses a completely different sensor to the previous Spyders, so you can't really extrapolate the accuracy from them. The SpyderX seems to have a poor low light sensitivity though.
-
True but their history over the years is piss-poor. Their marketing speak has been often misleading. They have ripped off the product concepts of others. Not a lot of history that bode well for this company.
-
Andrew,
Am I to assume that ideal data points would plot on the yellow curve? Is the yellow curve a theoretical ideal or is it a best-fit curve from other data?
Thanks,
Brookie
-
Sorry Brookie. I don't understand the question.
-
There was no explanatory text in your post or at the original thread you copied from so I’m just trying to be sure I understand what the graphed data mean. I gather that departure from the yellow curve on the graph is undesirable- rather a perfect match between an instrument reading and the correct or perfect value would plot right on top of the yellow line. A point plotting off the yellow line thus indicates a mismatch. Is that the idea and therefor why you are saying that the Spyder5 is not an accurate calibration tool? Or am I missing the point?
Thanks,
Brookie
-
The plot shows a high-end reference device and how close or how far in dE each devices correlation is from this reference. The Spyder devices are in the weeds.
-
👍
-
Andrew
So the plot of the old I1 from 20 years ago was accurate and the new I2-3 or plus have like a 15-20% better accuracy?
Ps I have asked for your opinion before about upgrading my old i1pro to a newer model? You replied if mine is still working keep it, just wondering if this new edition negates that opinion?
Thanks
Bab
-
BAB--
The i1(1) and i1(2) are two different samples of the i1Display Pro (or i1D3) the i1D2 is pretty much crap like the spyders are. If you're using an i1pro to calibrate your display, you'd likely be much better off with an i1d3 because of it's accuracy down to low luminance values. However, you'd be even better off using your i1pro to create a colorimeter matrix correction for an i1d3 using your specific display and DisplayCal, THEN using the i1d3 to calibrate.
The above graph shows really admirable performance for the i1d3 on that specific display, but I'd be willing to bet that the i1pro would stay pretty consistently in that position across different displays with regard to the jeti, but the i1d3s would shuffle around a bit.
Long story short, you really need both a spectro and a colorimeter for good accuracy unless your colorimeter has been specifically mated to your display.
sorry for the kind of convoluted explanation but i think that made sense...
-
The above graph shows really admirable performance for the i1d3 on that specific display, but I'd be willing to bet that the i1pro would stay pretty consistently in that position across different displays with regard to the jeti, but the i1d3s would shuffle around a bit.
Lessbones and Andrew,
OK, now I am a bit confused again, Lessbones is suggesting the i1D3 may not be as good, right? What is the basis for that, have you seen that the i1D3 is not as good from experience or something else? Also, if I Google “i1D3” what I find looks an awful lot like the i1 Display Pro - what is the difference and which one does a guy really need to get (if I go ahead and replace my Spyder5)?
Thanks,
Brookie
-
i1D3 = i1Display Pro
It's the best colorimeter you can buy unless you're willing to drop thousands. My point was that if you really want to have a super accurate calibration, you should ALSO have a spectrophotometer, like the i1Pro/i1pro2/i1pro3 to create a correlation matrix for your specific display technology. A colorimeter reads information through 3 color filters, one each for RGB. A spectrophotometer actually breaks the light spectrum down into very small pieces (6/10nm for the i1pro series). Therefore it is capable of detecting any kind of spikes in specific parts of the spectrum which may throw off a device that is measuring with only 3 filters. You can use this to correct the measurement of the colorimeter.
This whole business gets very technical, but basically if you want a quick answer and don't want to have to dig deep into the whole physics of light, just get the i1Display Pro. It's the best device available for calibrating monitors under $1000-- my point was more that you can improve it even further by adding a spectro if you have the resources, and more importantly, the time to understand the process--
-
I agree that using a Spectrophotometer and Colorimeter can in some cases produce superior results but a lot has to do with the design (filter matrices) in the Colorimeter. Anyway, in the very, very old days (circa 2007), I’d do this with my SpectraView using two devices as follows.
Steps used in those old days:
1. Connect the iOne Pro/Monitor
2. Select the desired Target in SpectraView and calibrate the display.
3. Confirm that the calibrated white point is acceptable (judge 100% white only).
4. Disconnect the iOne Pro/Monitor and connect the iOne Display / DTP94.
5. Open the SpectraView Preferences dialog.
6. Re-detect the color sensor.
7. Change the "Primary Colors Chromaticity Source" to "Factory Measurement"
8. Click OK.
Next you will create a new Target with a custom white point that has been adjusted with the offset between the two devices:
1. In SpectraView, click the "Edit Target settings" icon.
2. Click the "Edit..." button in the "White Point" group.
3. Click the "Measure" button and measure the white patch on the screen using the iOne Display / DTP94.
4. The CIE xy values should be filled in automatically.
5. Click OK.
6. Click OK.
7. Enter a new name for the Target file.
8. Make sure the new Target is selected, and calibrate using the iOne Display / DTP94.
The white point should match what you got with the iOne Pro/Monitor, and the greyscale should be much better.
I havent seen the need to do this for years of use with my i1Display Pro. It’s a super anal step and if you have both kinds of devices and want to go that route, go that route.
-
I used to do that quite a bit too with SpectraView-- these days as I've gotten more into doing calibrations for video I've become much more baffled by the lack of colorimeter correction in some calibration programs. When making 3DLuts for Davinci Resolve for example, I've gotten MUCH better results using a correction matrix than without (with DisplayCal and with Calman) so I feel like it must be at least somewhat important to measure the full primaries as well to really properly tune the colorimeter response. It does help to shift the white measurement, but I'd imagine that measuring the RGB subpixels alone would give a clearer picture of how the devices behave at the outer limits of the display gamut.
I could be wrong about how much that matters when you're not trying to limit the display gamut to a certain standard, but I just yesterday sent an email to the folks at BasicColor asking if they'd be interested in implementing such a thing in the updated Display software.
Andrew, I know you've been using a PA271Q-- I've always felt like the one I'm working on exhibits a slight magenta cast at D65 regardless of which measurement device I've used to profile the white point (i1d3, i1pro, i1pro2) through SpectraView-- I haven't tried with any other software quite yet, but I was thinking of experimenting with creating a 3DLut in DisplayCal and uploading it via multiprofiler.... but i'm getting OT here--
-
I used to do that quite a bit too with SpectraView-- these days as I've gotten more into doing calibrations for video I've become much more baffled by the lack of colorimeter correction in some calibration programs. When making 3DLuts for Davinci Resolve for example, I've gotten MUCH better results using a correction matrix than without (with DisplayCal and with Calman) so I feel like it must be at least somewhat important to measure the full primaries as well to really properly tune the colorimeter response. It does help to shift the white measurement, but I'd imagine that measuring the RGB subpixels alone would give a clearer picture of how the devices behave at the outer limits of the display gamut.
I could be wrong about how much that matters when you're not trying to limit the display gamut to a certain standard, but I just yesterday sent an email to the folks at BasicColor asking if they'd be interested in implementing such a thing in the updated Display software.
Andrew, I know you've been using a PA271Q-- I've always felt like the one I'm working on exhibits a slight magenta cast at D65 regardless of which measurement device I've used to profile the white point (i1d3, i1pro, i1pro2) through SpectraView-- I haven't tried with any other software quite yet, but I was thinking of experimenting with creating a 3DLut in DisplayCal and uploading it via multiprofiler.... but i'm getting OT here--
It's not necessarily a good idea to make correction matrices for i1D3 and PA271Q using low res spectrophotometer. And 3DLUT calibration is definitely a bad idea in case of PA271Q.
-
It's not necessarily a good idea to make correction matrices for i1D3 and PA271Q using low res spectrophotometer. And 3DLUT calibration is definitely a bad idea in case of PA271Q.
Care to expand on that? Specifically the second part. I get that the 15nm optical vs. 4.5nm of the Jeti could be an issue if the spikes in spectral response are THAT thin...
-
Care to expand on that? Specifically the second part. I get that the 15nm optical vs. 4.5nm of the Jeti could be an issue if the spikes in spectral response are THAT thin...
i1Pros usually have ~2-3 ∆E2k errors on spiky spectra backlights, while i1D3 with proper spectral correction is close to ~1 or even less ∆E2k to reference. That means correction matrix can make things worse, it's better to create .ccss correction using 3,3nm mode or use generic X-Rite PFS correction done with high resolution CS-1000 spectroradiometer.
SpectraView II uses internal factory calibrations done with lab grade Konica-Minoltas spectroradiometers and colorimeters (CS-2000, CA-310 and CA-2500 for uniformity) worth thousands of $ each, making it perfectly linear. While factory calibration they're most probably programming the first internal 14(16)bit 3DLUT.
While creating a 17^3 3DLUT with DisplayCAL you're using much inferior sensor and you're limited to 8 bit display path, and you only put it to second "external softproofing 3DLUT" so it's impossible to achieve the factory level of accuracy that way.
-
Guys,
Thanks for all your input, but you’re way past my pay grade! Here is my situation:
I have recently started printing and am having a bit of a problem with soft proofing - in particular with matte paper and the ONSCREEN soft proof looking WORSE than the actual print does.
So I recalibrated my monitor - BenQ SW2700PT using my Spyder5. When I first got this monitor about 3 years ago it would display ~99% of sRGB gamut and high 90s% for AdobeRGB. But now it is only giving me mid-70% of AdobeRGB. It was puzzling to me that the AdobeRGB had apparently deteriorated so much. By the way, I edit in ProPhoto color space using LR and PS subscription with all current updates on a MacBookPro (current Catalina). I print from the LR print module using appropriate LCC paper profiles.
Then I coincidentally received an email offer to upgrade to a SpyderX which got me to thinking about this whole thing and lead me to make the initial post.
I called BenQ and they told me that the screen really should not degrade like that and speculated that my colorimeter might be going bad. That of course made me think even more about replacing the Spyder5.
Based on the graph that Andrew posted it looks like maybe I ought to switch to an i1 Display Pro instead of going to the newer Spyder. Anymore thoughts on that are appreciated.
Also, is it true that the gamut range of a screen should not change much over time, or at least the 2 or 3 years I’ve been using my BenQ? Not knocking BenQ or their comment to me, just interested in seeing if people think that a monitor’s gamut is likely to degrade based on their experiences.
And, is it true that a colorimeter would more likely go bad over the 2 year time frame I’ve had the Spyder5?
I doubt I will rise to the level of technical proficiency or demand as the above respondents, but I sure would like my printing to be more predictable!
Thanks for your comments as always,
Brookie
-
Guys,
Thanks for all your input, but you’re way past my pay grade! Here is my situation:
I have recently started printing and am having a bit of a problem with soft proofing - in particular with matte paper and the ONSCREEN soft proof looking WORSE than the actual print does.
So I recalibrated my monitor - BenQ SW2700PT using my Spyder5. When I first got this monitor about 3 years ago it would display ~99% of sRGB gamut and high 90s% for AdobeRGB. But now it is only giving me mid-70% of AdobeRGB. It was puzzling to me that the AdobeRGB had apparently deteriorated so much. By the way, I edit in ProPhoto color space using LR and PS subscription with all current updates on a MacBookPro (current Catalina). I print from the LR print module using appropriate LCC paper profiles.
Then I coincidentally received an email offer to upgrade to a SpyderX which got me to thinking about this whole thing and lead me to make the initial post.
I called BenQ and they told me that the screen really should not degrade like that and speculated that my colorimeter might be going bad. That of course made me think even more about replacing the Spyder5.
Based on the graph that Andrew posted it looks like maybe I ought to switch to an i1 Display Pro instead of going to the newer Spyder. Anymore thoughts on that are appreciated.
Also, is it true that the gamut range of a screen should not change much over time, or at least the 2 or 3 years I’ve been using my BenQ? Not knocking BenQ or their comment to me, just interested in seeing if people think that a monitor’s gamut is likely to degrade based on their experiences.
And, is it true that a colorimeter would more likely go bad over the 2 year time frame I’ve had the Spyder5?
I doubt I will rise to the level of technical proficiency or demand as the above respondents, but I sure would like my printing to be more predictable!
Thanks for your comments as always,
Brookie
You can change the SW2700PT gamut by changing modes or calibrating it to smaller space, so most probably you did it unintentionally. There's no way the display could deteriorate in this way, and Spyder 5 shouldn't go bad as well. Are you calibrating it with Color Palette software? Check what color mode is selected in OSD menu.
-
i1Pros usually have ~2-3 ∆E2k errors on spiky spectra backlights, while i1D3 with proper spectral correction is close to ~1 or even less ∆E2k to reference. That means correction matrix can make things worse, it's better to create .ccss correction using 3,3nm mode or use generic X-Rite PFS correction done with high resolution CS-1000 spectroradiometer.
SpectraView II uses internal factory calibrations done with lab grade Konica-Minoltas spectroradiometers and colorimeters (CS-2000, CA-310 and CA-2500 for uniformity) worth thousands of $ each, making it perfectly linear. While factory calibration they're most probably programming the first internal 14(16)bit 3DLUT.
While creating a 17^3 3DLUT with DisplayCAL you're using much inferior sensor and you're limited to 8 bit display path, and you only put it to second "external softproofing 3DLUT" so it's impossible to achieve the factory level of accuracy that way.
This is all very interesting, and while I see your point, how is one meant to compensate for drift over time in that scenario? Do you have a ccss made from when you measured the PA271q with your reference spectroradiometer? (not sure if you own it or if it was borrowed), because I'd love to at least attempt some verification reports using that with an i1d3.
-
You can change the SW2700PT gamut by changing modes or calibrating it to smaller space, so most probably you did it unintentionally. There's no way the display could deteriorate in this way, and Spyder 5 shouldn't go bad as well. Are you calibrating it with Color Palette software? Check what color mode is selected in OSD menu.
Czornyj,
Thanks for the reply. Out of curiosity I just ran another full calibration using Pallette Master Element (PME) with the Spyder5 plugged into a USB port on the side of the monitor. I don’t see an opportunity to change the color space in the calibration set up process, unless you mean changing color space using the hockey puck toggle - which I have set to Adobe RGB.
So when I do a full recalibration as noted above, then follow it up with a validation in PME, the results indicate a deltaE avg = 0.47 with a range of 0.09 to 0.78 which seems pretty good. But if I just check the calibration in the Spyder launch application it indicates 97% of sRGB but only 72% of Adobe RGB. That just doesn’t seem right??
Can you tell if I am simply doing something wrong?
Thanks,
Brookie
-
Brookie,
I have a SW2700PT - using PME for calibration and profiling I obtain a larger gamut volume than Adobe 1998, although there is some poke thru in the greens, per the attached image. The wireframe represents the monitor gamut, and the solid is Adobe 1998. Post a link to your profile and I'll check its gamut with ColorThink Pro, not sure what the Spyder launch application is or does.
Richard Southworth
Added by edit - I use a Xrite Display Pro.
-
Brookie,
I have a SW2700PT - using PME for calibration and profiling I obtain a larger gamut volume than Adobe 1998, although there is some poke thru in the greens, per the attached image. The wireframe represents the monitor gamut, and the solid is Adobe 1998. Post a link to your profile and I'll check its gamut with ColorThink Pro, not sure what the Spyder launch application is or does.
Richard Southworth
Added by edit - I use a Xrite Display Pro.
Richard,
Thanks for the reply. I did not save the previous validation report that I pulled the numbers above from. So I reran a full calibration followed by a validation in PME. Not quite as good a result this go ‘round, but it still “passes”. Here is a copy of the new validation report.
-
I thought your concern was the monitor profile gamut volume as compared to Adobe 1998 - again, post a link to the profile so I can download, and I'll check it. Not worried about validation, it looks ok, similar to mine.
Richard Southworth
-
Richard,
I am a rookie at this and am definitely confused, as you’ve probably already guessed. I’m not sure where to find what you are asking for on my computer. Do you mean the profile that results from Display Calibrator Assistant (accessible from system pref/display/color)? I’m guessing that is not what you mean since you are asking for a “link” and what I just mentioned is on my computer and not accessible via a link. Sorry, I’m in the weeds here and not sure what to find for you. I appreciate your offer to help, if you could also help me find what it is you want from my laptop that would be great. The more I try to figure this out, the more I realize how little I really understand the process.........
-
Brookie,
sorry, my fault, should have been clearer. I would like to examine the actual profile, SW2700.....icc. I think the best way is for you to zip it, i.e. compress it into a xxx.zip file, and attach it to your reply. Don't remember what os you're running, if Windows 10 you can right click on the profile file, left click on Send to, and choose compressed (zipped) folder. I'm sure there is a similar process for Macs.
Richard Southworth
-
Richard I'm using Mac OS current with Catalina. For some reason the forum isn't letting me upload the zip file even though it is only 5kb. And it won't let me upload the original 16kb file because of the .icc extension.
FWIW the file name is SW2700 2_D65_L80_G22_Rel_2020-02-15T01.55.16Z.icc.zip
Here is a screen shot of what the file looks like when I open it. Hope that does the trick.
Thanks,
Brookie
-
Brookie,
The only way I can check the profile file is to have a copy. If you want to proceed, I suggest you try:
1. Create a copy of the file, so that the original is safe
2. Rename to something simple, like test.icc - I suspect the multiple periods may be causing the problem.
3. Zip up the file, resulting in test.zip
4. Now try to attach to a reply
Richard Southworth
Added by edit - or rename the zip file you've already created to test.zip, and try to attach
-
Let's see if this works!
Thanks Richard!
-
Your gamut is ok, very similar to mine. In the attached image the wireframe represents your display profile gamut, and solid is Adobe 1998. There is slight poke-thru, but for all practical purposes your display gamut encompasses the Adobe 1998 gamut.
Richard Southworth
-
Brookie,
Another test you should perform - decompress the attached file, load the resulting tif into Photoshop, rotate to horizontal and blow it up to fit the screen. You should see little (preferably no) tint in any of the neutral slices. Also you should be able to distinguish each slice from its neighbor.
Richard Southworth
-
Richard,
That’s really cool, how did you do that? It is interesting comparing the image of your monitor gamut to mine - in particular the upper red-orange-blue-violet area behind the green peak - on the image for my monitor it appears to have lower and more extended slope, whereas, your monitor looks to have a steeper more abrupt taper. Is that just a difference in the angle which you are viewing the two 3D projections from, or does it tell us something different about our profiles?
Thanks again for taking time to help an old rookie out, it is really appreciated!!!
Brookie (aka Dave Cornue)
-
Brookie,
Another test you should perform - decompress the attached file, load the resulting tif into Photoshop, rotate to horizontal and blow it up to fit the screen. You should see little (preferably no) tint in any of the neutral slices. Also you should be able to distinguish each slice from its neighbor.
Richard Southworth
Richard,
The two blackest strips are indistinguishable from each other, the third is very barely distinguishable from 1 and 2, from there on out to white all strips are easily distinguishable from each other. I do have my profile set for a lower screen brightness to make printing more predictable, so that may be why strips are harder to see on the black end than on the white end. There is a slight olive tint to some of the midrange strips and a couple mid to high range strips have a magenta tint. Oddly enough the magenta tint does not seem to be equally distributed, two strips really show it, others do not. Also, it is interesting that the tints are more visible when viewed in LR than when viewed in PS. Wonder why that is?
-
The image has no embedded profile, so it takes on your default workspace. Try assigning various standard workspaces, see if that affects the black separation. Start in Photoshop, assign ProPhoto, save it out and load into Lightroom, should look identical. As far as color tint, you may get better results creating a LUT based profile instead of matrix. I can see very slight tint in a few of mine, it may be that as good as this display is for the money it isn't in the same league as the more expensive NEC, etc.
Also the neutral performance is affected by the accuracy of the measuring device and number of patches in the target. Make sure your room is dim, hopefully the puck seals well against the screen but better to calibrate/profile in lower light. I'm assuming you're using the advanced mode in Palette Master Element, and that you've upgraded to the latest version. I use absolute black point, have had some problems in the past using relative.
Richard Southworth
-
Brookie,
The attached image is normally used for printer profile checking, but it is also useful to assess display profiles. You can google on the name and download the original tif.
Richard Southworth
-
Richard,
Thanks for the suggestions. I probably won’t be able to get back to them for a couple days though.
I’ve been using a similar Datacolor image to test different papers. I’ll take a look at this one too.
Thanks for all your help!
Brookie
-
Andrew: Where does the Spectraview II fit in the graph? (with NEC PA242W monitor)
-
Andrew: Where does the Spectraview II fit in the graph? (with NEC PA242W monitor)
I don't understand the question.
-
I don't understand the question.
I forget who makes the sensor for the NEC Spectraview II PA242W monitor. So which of the devices shown in the chart is that puck or isn't it shown? If not, where would it fall if you know? Tks.
-
i1 Display.
-
i1 Display.
So if I understand the chart, because the i1's which are the same as the units furnished with the NEC Spectraview II are closest to the Jeti reference, they're the most accurate?
-
So if I understand the chart, because the i1's which are the same as the units furnished with the NEC Spectraview II are closest to the Jeti reference, they're the most accurate?
The Jeti is the reference device. It’s very expensive, it’s very accurate. The owner very, very likely had it recalibrated and certified yearly. Those devices that correlate to it closely are ideal because they correlate with a high end reference. Those that don’t are not accurate devices.
You buy a 25 foot tape measure at Home Depot and you assume, expect it is accurate. How accurate is that, how accurate do you need? You compare it to some $10K super duper NASA reference measuring system and see it’s accurate to within +/- 1000th of an inch, but another tape measure you bought at Harbor Freight is accurate within a 10th of an inch. You now know which of the two tape measures is more accurate. And of course it’s useful to determine how accurate you need a tape measure (maybe 10th of an inch accuracy if more than enough for the task at hand). But if the two tape measures are about the same price, you’d be foolish to not recommend or buy the tape measure from Home Depot. And this is what the graph is illustrating.
-
Well, I have the NEC monitor and Spectraview II software and puck. I bought them together figuring why mess around with differing manufacturers and making them work together especially when the Spectraview software is part of it. It's nice to know NEC used an accurate puck.
-
The image has no embedded profile, so it takes on your default workspace. Try assigning various standard workspaces, see if that affects the black separation. Start in Photoshop, assign ProPhoto, save it out and load into Lightroom, should look identical. As far as color tint, you may get better results creating a LUT based profile instead of matrix. I can see very slight tint in a few of mine, it may be that as good as this display is for the money it isn't in the same league as the more expensive NEC, etc.
Also the neutral performance is affected by the accuracy of the measuring device and number of patches in the target. Make sure your room is dim, hopefully the puck seals well against the screen but better to calibrate/profile in lower light. I'm assuming you're using the advanced mode in Palette Master Element, and that you've upgraded to the latest version. I use absolute black point, have had some problems in the past using relative.
Richard Southworth
Richard,
Sorry I haven’t gotten back on this sooner. My PS profile is set to ProPhoto - I double checked that. My room is medium bright, but I will try recallibrating with lower light. I tilt the screen slightly to be sure I have good contact between the Spyder5 and the screen. I am using Advanced mode in PME. I am using Relative black point - in the past when I’ve used Absolute the resulting calibration seems to block the blackest blacks worse than in Relative mode.
The slight tint I do see is visible on both the BenQ monitor and on my MacBookPro so I’m inclined to think it is not the monitor - though I agree, as good as the SW2700PT is, you can’t expect it to match a monitor that costs 3 to 4 times as much. We have to be reasonable after all.
Thanks again for your help,
Brookie
-
Richard,
I decided to see if the tint I see on screen holds up in the print. So I printed your grey scale step wedge on a couple of matte papers. I do see the tint, but it is less visible in print than it is on the screen. Both prints were made with appropriate paper profiles and with the B&W check box left unchecked. No soft proofing adjustments were made other than selecting Perceptual intent - blacks looked a bit better on screen with perceptual than relative. The warmer toned paper (Canon Premium Matte) revealed the tint a bit more prominently than the cooler toned paper (Red River Polar Matte) - which I think makes sense.