Forgot Password
Pentax Camera Forums Home
 

Reply
Show Printable Version Search this Thread
02-16-2019, 01:50 PM - 2 Likes   #31
Pentaxian
Dartmoor Dave's Avatar

Join Date: Aug 2012
Location: Dartmoor, UK
Photos: Albums
Posts: 1,894
The pinkish cast that you both mention might well be real. It could be that the inks have shifted slightly in the test card (it's five years old), or it could be a characteristic of the lens used (the choice of lens has a significant impact on custom profiles), or it could be an ambient light intrusion because I shot the flash and LED cards in my living room this morning without setting up my light tent (I was a bit hungover after a great night with good friends). The bottom line is that there are so many variables when it comes to creating custom profiles that you can drive yourself nuts trying to iron them all out. My aim was simply to show how small the differences actually are between different illuminants, and I don't feel that a slight pink cast undermines that point even if it does exist.

The daylight LED is a SAD light box that I use when too many days of Dartmoor winter gloom start getting me down. It claims to be full spectrum, but I think the test card example shows that the spectrum is actually far from flat.

As for the averaged profile: when you use the "chart" tool in Adobe Profile Editor it shows you the values that it has used for the 16 colour patches out of the total 24 on a standard Macbeth/ColorCheck card (the monochrome values are implicit in the 16 colour values). I just averaged out the values for the three different light sources used in this very basic example. If 16 values doesn't sound like enough to create a profile, remember that the software then twists the entire colour space around those coordinates.

Folks, you are never going to be able to cancel out all the variables and create an absolutely objectively correct profile on your own. You just can't. But even with the most basic work you'll end up with something much more accurate than the manufacturer's embedded profile, because accuracy isn't the embedded profile's aim. The embedded profile just wants to make you say, "Wow, look at those amazingly bright and punchy colours!" Custom profiles are for when you're after something a bit more subtle than that.

02-16-2019, 02:11 PM   #32
Senior Moderator
Loyal Site Supporter
Loyal Site Supporter
BigMackCam's Avatar

Join Date: Mar 2010
Location: North-East of England
Posts: 12,753
Original Poster
QuoteOriginally posted by Dartmoor Dave Quote
The pinkish cast that you both mention might well be real. It could be that the inks have shifted slightly in the test card (it's five years old), or it could be a characteristic of the lens used (the choice of lens has a significant impact on custom profiles), or it could be an ambient light intrusion because I shot the flash and LED cards in my living room this morning without setting up my light tent (I was a bit hungover after a great night with good friends). The bottom line is that there are so many variables when it comes to creating custom profiles that you can drive yourself nuts trying to iron them all out. My aim was simply to show how small the differences actually are between different illuminants, and I don't feel that a slight pink cast undermines that point even if it does exist.

The daylight LED is a SAD light box that I use when too many days of Dartmoor winter gloom start getting me down. It claims to be full spectrum, but I think the test card example shows that the spectrum is actually far from flat.

As for the averaged profile: when you use the "chart" tool in Adobe Profile Editor it shows you the values that it has used for the 16 colour patches out of the total 24 on a standard Macbeth/ColorCheck card (the monochrome values are implicit in the 16 colour values). I just averaged out the values for the three different light sources used in this very basic example. If 16 values doesn't sound like enough to create a profile, remember that the software then twists the entire colour space around those coordinates.

Folks, you are never going to be able to cancel out all the variables and create an absolutely objectively correct profile on your own. You just can't. But even with the most basic work you'll end up with something much more accurate than the manufacturer's embedded profile, because accuracy isn't the embedded profile's aim. The embedded profile just wants to make you say, "Wow, look at those amazingly bright and punchy colours!" Custom profiles are for when you're after something a bit more subtle than that.
Thanks again, Dave - very useful info. I tend to agree with you that trying to create an "absolutely correct" profile might be folly given all of the variables. Indeed, you mention your colour chart... my own ColorChecker Passport is a few years old by now, so the colours probably aren't 100% accurate on that alone, before lighting issues are taken into account.

And now I'm thinking... Even though there may (will?) be some spikes in the spectrum, that variable temperature light panel I originally mentioned might be "good enough", since I'm not looking for lab-quality colour accuracy but rather a decent starting point in each of my profiles. I can live with minor inaccuracies. Or I could go with the flash + gel approach helpfully suggested by others, which would seem to offer an equally useful, if - once more - not 100% colour accurate option.

Last edited by BigMackCam; 02-16-2019 at 05:33 PM.
02-16-2019, 03:19 PM - 2 Likes   #33
Senior Moderator
Loyal Site Supporter
Loyal Site Supporter
BigMackCam's Avatar

Join Date: Mar 2010
Location: North-East of England
Posts: 12,753
Original Poster
Well, I just bought the Aputure MX LED light panel, for delivery some time in the next week. I figure I have little to lose... I can try it for dual-illuminant colour profile creation, but if those profiles should prove problematic for general use, it'll serve as an alternative for close-range indoor single-flash photography. Given the cost and my non-lab-level requirements, I think it might be a useful piece of kit. Either way, I'll let you good folks know
02-19-2019, 08:49 PM   #34
Loyal Site Supporter
Loyal Site Supporter




Join Date: Dec 2017
Photos: Gallery
Posts: 527
See other thread #20 for an image of the Color Checker based on x-rite's reported D50 RGB values.

FWIW, no pinkish cast is visible in the Gimp file, the derived png, jpeg, or rendering of the link above on my newly acquired BenQ SW2700PT (factory cal but nothing tweaked yet by me).


Last edited by kaseki; 02-19-2019 at 08:59 PM.
02-20-2019, 02:15 AM - 2 Likes   #35
Pentaxian
Dartmoor Dave's Avatar

Join Date: Aug 2012
Location: Dartmoor, UK
Photos: Albums
Posts: 1,894
QuoteOriginally posted by kaseki Quote
See other thread #20 for an image of the Color Checker based on x-rite's reported D50 RGB values. FWIW, no pinkish cast is visible in the Gimp file, the derived png, jpeg, or rendering of the link above on my newly acquired BenQ SW2700PT (factory cal but nothing tweaked yet by me).

Thanks for the link to the other thread. Could I perhaps clarify a couple of things about the two different threads, just to avoid confusion for anyone who happens along here and doesn't read through the whole discussion?

My understanding is that your other thread is about your wish to be absolutely sure that your monitor is perfectly calibrated, beyond the level of calibration that most of us do with things like Spyder (in my case). So, to be sure of that, you've wanted a completely accurate image of a ColorCheck card that you can look at on your screen. And the ColorCheck image that you're referring to above is one that you've downloaded from x-rite to use as a neutral reference on your monitor, rather than one that you've shot with your own camera and lens.

That's a completely legitimate thing to do if you want to be absolutely sure about your monitor's calibration, but it's not really what this particular thread is about. And I'm sure you're already aware that the downloaded test card image from x-rite would be of no use in creating a custom camera profile.

This thread (at least as I've personally interpreted the OP) is about how to shoot test cards with our own cameras and lenses to create custom camera profiles, which has nothing at all to with monitor calibration. We've been talking about things like which illuminants to use, whether we can be sure if the colours in our test cards are still stable after a few years, how the colour characteristics of different lenses affect things -- all the many variables that we have to consider when creating our own custom camera profiles.

This is absolutely not a criticism of your other thread or your valuable contributions to this one. I've personally found the other thread fascinating and informative, but I do want to keep things clear about the different aims underlying the two discussions. Camera profiles and monitor profiles are completely different things, and I think it's important that we avoid any confusion between them, just to help out any newcomers to the concept of profiling who might come along.


Edit: I should also mention that I've discovered that I get different white balance values from the white and mid-grey patches on my own test card, when of course they should be the same. I've concluded that the inks are no longer accurate on my own card, so I'll be getting another one. Any pink cast that anyone sees in the examples I posted above is down to the card used no longer being accurate, so apologies for causing confusion with that.

Last edited by Dartmoor Dave; 02-20-2019 at 02:27 AM.
02-20-2019, 04:04 AM - 1 Like   #36
Senior Moderator
Loyal Site Supporter
Loyal Site Supporter
BigMackCam's Avatar

Join Date: Mar 2010
Location: North-East of England
Posts: 12,753
Original Poster
QuoteOriginally posted by Dartmoor Dave Quote
This thread (at least as I've personally interpreted the OP) is about how to shoot test cards with our own cameras and lenses to create custom camera profiles, which has nothing at all to with monitor calibration. We've been talking about things like which illuminants to use, whether we can be sure if the colours in our test cards are still stable after a few years, how the colour characteristics of different lenses affect things -- all the many variables that we have to consider when creating our own custom camera profiles.

This is absolutely not a criticism of your other thread or your valuable contributions to this one. I've personally found the other thread fascinating and informative, but I do want to keep things clear about the different aims underlying the two discussions. Camera profiles and monitor profiles are completely different things, and I think it's important that we avoid any confusion between them, just to help out any newcomers to the concept of profiling who might come along.
Your interpretation is correct, Dave. My primary interest here (although the thread has evolved to include other helpul and related elements) is in achieving suitable man-made illumination of a test target so that I can take shots for use in developing custom camera profiles. Relying on natural lighting conditions is patchy at best in this country and requires patience (as well you know ), especially at this time of the year.

A lighting panel with adjustable colour temperature that can illuminate a target at both 2850k and 6500k (or reasonably close to both) seems like a useful tool, even given the subsequent concerns around spectral spikes from the LEDs. Absolute accuracy isn't something I'm desperately concerned about (though that would be nice, of course). If I can produce dual-illuminant profiles that work nicely with no significant colour shifts in different lighting situations, I'll be more than happy.

I'm awaiting delivery of the panel, and looking forward to trying it out. I'll post a new thread on my findings in the near future, and provide a link here

QuoteOriginally posted by Dartmoor Dave Quote
Edit: I should also mention that I've discovered that I get different white balance values from the white and mid-grey patches on my own test card, when of course they should be the same. I've concluded that the inks are no longer accurate on my own card, so I'll be getting another one. Any pink cast that anyone sees in the examples I posted above is down to the card used no longer being accurate, so apologies for causing confusion with that.
Even on my ColorChecker Passport, I get *tiny* differences in white balance values from white versus mid grey, but in my case they're so tiny as to have no visible effect. Whether this was always the case, or it's as a result of the card being a few years old now, I can't say...
02-20-2019, 04:22 AM   #37
Pentaxian
Dartmoor Dave's Avatar

Join Date: Aug 2012
Location: Dartmoor, UK
Photos: Albums
Posts: 1,894
QuoteOriginally posted by BigMackCam Quote
Absolute accuracy isn't something I'm desperately concerned about (though that would be nice, of course). If I can produce dual-illuminant profiles that work nicely with no significant colour shifts in different lighting situations, I'll be more than happy.

Absolute accuracy has never been my own aim with camera profiles either, but of course I completely accept that for many people as much objective accuracy as possible will be the goal. I'm just interested in getting results that I personally like, although thanks to this thread I've now discovered that I've been creating profiles that I like with a test card that isn't actually accurate.
02-20-2019, 05:02 AM - 3 Likes   #38
Pentaxian
Digitalis's Avatar

Join Date: Mar 2009
Location: Adelaide.
Posts: 9,537
QuoteOriginally posted by photoptimist Quote
flash and a tungsten source would be the logic pairing.
Can confirm.I have discussed this at length with Adobe many years ago - they use dedicated light sources for developing their profiles. They also use a massive 250 swatch test target to build the profiles, using 16 swatches and expecting a flawless profile is asking a bit much.

02-20-2019, 10:15 AM - 1 Like   #39
Loyal Site Supporter
Loyal Site Supporter




Join Date: Dec 2017
Photos: Gallery
Posts: 527
QuoteOriginally posted by Dartmoor Dave Quote
Thanks for the link to the other thread. Could I perhaps clarify a couple of things about the two different threads, just to avoid confusion for anyone who happens along here and doesn't read through the whole discussion?

My understanding is that your other thread is about your wish to be absolutely sure that your monitor is perfectly calibrated, beyond the level of calibration that most of us do with things like Spyder (in my case). So, to be sure of that, you've wanted a completely accurate image of a ColorCheck card that you can look at on your screen. And the ColorCheck image that you're referring to above is one that you've downloaded from x-rite to use as a neutral reference on your monitor, rather than one that you've shot with your own camera and lens....
I would only make this adjustment -- remove the words "absolutely" and "perfectly" from the first sentence of the second paragraph. I may have naively believed something similar to that when I started the other thread, but with further research now believe it to be impossible. I would say that I want to be sure that the monitor calibration process did not result in a worse condition than before I started. It is a test that doesn't make use of the calibration equipment claims/calibration to prove the calibration.

And this thread is tied to the entire process, as the physical card needs to be illuminated in some way that allows valid comparison with the display image.
02-20-2019, 02:00 PM - 2 Likes   #40
Senior Moderator
Loyal Site Supporter
Loyal Site Supporter
BigMackCam's Avatar

Join Date: Mar 2010
Location: North-East of England
Posts: 12,753
Original Poster
My Aputure AL-MX light arrived today Quite a piece of equipment, very solid, well built and beautifully finished. I'm amazed, given the price.

I hadn't paid much attention to the advertised specifications other than lux and colour temperature ranges, but according to the manufacturer's specifications, it achieves a CRI (Color Rendering Index) score of 95+, and a TLCI (Television Lighting Consistency Index) score of, again, 95+. For a modestly-priced lighting device, I guess those measurements ought to be treated with a at least a little skepticism, Still, given the highly professional build quality and presentation, I'm hopeful that the manufacturer may not have exaggerated too much.

I hope to run some tests over the next couple of days, and intend to report back

Last edited by BigMackCam; 02-21-2019 at 05:42 AM.
02-21-2019, 12:08 PM - 1 Like   #41
Senior Member




Join Date: Apr 2016
Location: Los Angeles
Posts: 157
I built some dual-illuminant profiles for my 645Z, and what I did was to use two strobes with snoots to get the daylight, and then removed the -500K colour-correction domes and used the modelling lamps on full power still with the snoots to get the tungsten. It was a fair bit of effort to set up, and required a lot of flagging to avoid coloured reflection interference and glare, but ended up producing some good profiles.

I got the basic concept from Anders Torger's write-up of how to build dcamprof profiles, and used his Lumariver Profile Designer software for creating mine.
Reply

Bookmarks
  • Submit Thread to Facebook Facebook
  • Submit Thread to Twitter Twitter
  • Submit Thread to Digg Digg
Tags - Make this thread easier to find by adding keywords to it!
camera, chart, color, colour, colour chart, conditions, flash, light, lighting, passport color checker, photo studio, profiles, strobist, time, x-rite
Thread Tools Search this Thread
Search this Thread:

Advanced Search


Similar Threads
Thread Thread Starter Forum Replies Last Post
Colour management - browser issue, X-rite vs DisplayCAL profiles BigMackCam Digital Processing, Software, and Printing 15 04-17-2018 09:07 PM
Fuji Colour profiles on a K1 gnugent Pentax K-1 35 08-30-2016 07:31 AM
Custom colour profiles and pixel shift processing. robobob Pentax DSLR Discussion 7 08-11-2016 11:15 PM
Highly variable lens colour profiles rob8888 Pentax Medium Format 5 09-02-2014 09:20 AM
eye chart? what eye chart?? dcmsox2004 Post Your Photos! 5 08-20-2009 06:41 AM



All times are GMT -7. The time now is 10:22 PM. | See also: NikonForums.com, CanonForums.com part of our network of photo forums!
  • Red (Default)
  • Green
  • Gray
  • Dark
  • Dark Yellow
  • Dark Blue
  • Old Red
  • Old Green
  • Old Gray
  • Dial-Up Style
Hello! It's great to see you back on the forum! Have you considered joining the community?
register
Creating a FREE ACCOUNT takes under a minute, removes ads, and lets you post! [Dismiss]
Top