Ad
Ad
Ad
Pages: [1] 2 3 »   Bottom of Page
Print
Author Topic: ColorChecker Passport Profiles - vs White Balance  (Read 10156 times)
Mike Guilbault
Sr. Member
****
Offline Offline

Posts: 817



WWW
« on: February 15, 2012, 02:10:53 PM »
ReplyReply

I.E. do you create a new profile for each 'shoot'?  If I was shooting interiors, for example, the white balance may change from room to room depending on how much daylight is available, but do you need a different profile?  I'm still trying to get my head wrapped around profiles vs wb and seem to be confusing myself even more.
Logged

Schewe
Sr. Member
****
Offline Offline

Posts: 5474


WWW
« Reply #1 on: February 15, 2012, 04:52:43 PM »
ReplyReply

I'm not sure you understand the roll of custom DNG profiles. While it's useful to have a custom profile for your camera, unless you are shooting under really weird lights (fluorescent or mercury vapor), you really only need to make two profiles; one for tungsten and one for daylight. Ideally, you can cover both by shooting two targets, tungsten and daylight and make a dual illuminate profile.

Making a custom profile for each and every lighting setup isn't needed unless there's something special about the spectral output of the lights. The rest of the load is covered by proper white balancing.
Logged
JRSmit
Sr. Member
****
Offline Offline

Posts: 369


WWW
« Reply #2 on: February 16, 2012, 06:00:26 AM »
ReplyReply

I.E. do you create a new profile for each 'shoot'?  If I was shooting interiors, for example, the white balance may change from room to room depending on how much daylight is available, but do you need a different profile?  I'm still trying to get my head wrapped around profiles vs wb and seem to be confusing myself even more.

In essence, yes. Just white balancing does not cater for spectral differences that can, therefore, will change the "look" of recorded colors. By creating a profile for each room in your case this is reduced, and a better color match will be achieved.

Same goes for lenses, even the same make, can have a quite visible different color behaviour.
Also studio strobes differ in color with their power settings, how much different depends on the device, as a rule i make a image of my ccp with every change in light setup of power setting.

Logged

Fine art photography: www.janrsmit.com
Courses and workshops: www.centrumbeeldbeleving.nl

Jan R. Smit
Schewe
Sr. Member
****
Offline Offline

Posts: 5474


WWW
« Reply #3 on: February 16, 2012, 04:48:01 PM »
ReplyReply

Just to be perfectly clear, I'm really glad X-Rite came out with the whole Passport/DNG Profile product. It is a great boon to users and really helps support ACR/LR but...I think the spawning off of multiple DNG profiles has been over sold. If you understand how ACR/LR use DNG profiles, their purpose is to correct for the rendering of colors from the sensor. DNG Profiles are NOT designed as a color correction tool not as a white balance tool. The DNG profile is designed to optimize the spectral response of the sensor under "different" illumination.

The way a sensor responds to daylight is vastly different then to tungsten. But the way a sensor responds to daylight (D55) is pretty the much the same under cloudy (D65) or other similar illumination (down to the lower 4000K's which is about when ACR/LR starts tweening between illuminates).

Same deal for tungsten...tungsten DNG profiles are set to Standard Illuminate A (which is 2856K I think). So the spectral response of a sensor at 2850K and 3200K would be very similar. So you would get no benefit from doing a separate 2856K and 3200K DNG profile.

The same deal for special spectral output like fluorescent or mercury vapor...in these cases it WOULD behoove you to make special DNG profiles.

So, doing custom profiles for multiple, similar lighting won't get you much. Same for lenses...and while lenses can and do cause color shifts, those shifts are not due to spectral response of the sensor but to casts added due to the glass. But that's not something that DNG profiles were designed to correct. In the future we will prolly see a flat field correction (LCC) for color casts caused by lenses, but that's not something DNG profiles can really handle.

Sure, if you want to make a single illuminate profile for your strobes, go right ahead...but I wouldn't bother to do a strobe DNG profile for each of your various packs. That would get out of hand very quickly...

Again, I kinda think the multi-profile making is over the top based on what DNG profiles were meant to do.

If I'm at all wrong about this, perhaps Eric Chan can address the DNG profiles questions–but I don't think I am wrong :~)
Logged
digitaldog
Sr. Member
****
Offline Offline

Posts: 9097



WWW
« Reply #4 on: February 16, 2012, 05:21:55 PM »
ReplyReply

Just to be perfectly clear, I'm really glad X-Rite came out with the whole Passport/DNG Profile product. It is a great boon to users and really helps support ACR/LR but...I think the spawning off of multiple DNG profiles has been over sold.

Amen to that! A group based on some vastly differing lighting as you point out, but one for every shot? Messy, seems quite unnecessary.
Logged

Andrew Rodney
Author “Color Management for Photographers”
http://digitaldog.net/
bill t.
Sr. Member
****
Offline Offline

Posts: 2693


WWW
« Reply #5 on: February 16, 2012, 06:36:57 PM »
ReplyReply

I was just looking at some of the Passport DNG profiles I have generated...

Sunday_drab
111201_east_sideways
embudito_moon_2
joannie_left_03_late

All of them useful profiles made under lighting conditions varying from challenging to bad.  Those dng profiles almost instantly gave me images out of LR that were substantially better than anything I could have dialed up myself in any reasonable amount of time, if ever.

The real problem...it's all so untidy and dissociated!  The LR/ACR paradigm simply does not smoothly embrace the use of zillions of little profiles, however much I love them.

Is there any way that exists now that I can wrap up dng profiles with their intended image and image sets, so that I don't need separate copies of the profiles somewhere else and to remember what goes with what?  Am I missing something?  New format, anybody?  Come on Xrite, this is all your fault!  You owe us.
Logged
Schewe
Sr. Member
****
Offline Offline

Posts: 5474


WWW
« Reply #6 on: February 16, 2012, 06:51:50 PM »
ReplyReply

All of them useful profiles made under lighting conditions varying from challenging to bad.  Those dng profiles almost instantly gave me images out of LR that were substantially better than anything I could have dialed up myself in any reasonable amount of time, if ever.

You sure? That's not been my experience...I think you are falling into the trap of trying to use DNG profiles for color correction. That's not what was intended...
Logged
WayneLarmon
Jr. Member
**
Offline Offline

Posts: 51


« Reply #7 on: February 16, 2012, 09:15:30 PM »
ReplyReply

Quote
Is there any way that exists now that I can wrap up dng profiles with their intended image and image sets, so that I don't need separate copies of the profiles somewhere else and to remember what goes with what?  Am I missing something?  New format, anybody?  Come on Xrite, this is all your fault!  You owe us.

Which is why XRite wrote a profile manager:

The X-Rite ColorChecker Passport DNG ProfileManager software provides a single list of all installed DNG profiles, including non-printing image data such as file name, camera, light-source(s), and creation date. It also filters by camera or it can sort by any column header to provide the ability to correctly rename profiles (both internal and external names).  Users can easily enable/disable profiles to prevent profile overload in Lightroom and Adobe Camera Raw (ACR) and quickly collect and export DNG profiles for backup or distribution.
DNG ProfileManager

Wayne


Logged
Mike Guilbault
Sr. Member
****
Offline Offline

Posts: 817



WWW
« Reply #8 on: February 16, 2012, 09:35:53 PM »
ReplyReply

You sure? That's not been my experience...I think you are falling into the trap of trying to use DNG profiles for color correction. That's not what was intended...

There you go Jeff... that's exactly what I thought DNG profiles (and the ColorChecker Passport) were for.... to get accurate colour.  Is there an article somewhere that explains this in detail - that and the difference between what you get with a profile and WB?
Logged

bill t.
Sr. Member
****
Offline Offline

Posts: 2693


WWW
« Reply #9 on: February 16, 2012, 09:40:59 PM »
ReplyReply

It may be a trap, but it feels so good!

All I know is, when I apply those weird-lighting-specific dng profiles and take a white balance there ain't much color correction left to do.  I'm not seeing a downside that I can identify, from an image point of view.  Color channel histograms don't look any weirder than I'd expect for the light quality, etc.

Yes, I know I'm being Very Bad!  Maybe I'll soon reach my Waterloo, who knows?   Smiley

Yeah, the Xrite Profile Manager is a life saver!  Its ability to edit profile names is extremely useful, I only wish a comment field was available.  Also allows you to select which profiles will be visible from LR so you don't see every profile you ever made when trying to find a recent one.
Logged
Schewe
Sr. Member
****
Offline Offline

Posts: 5474


WWW
« Reply #10 on: February 16, 2012, 10:02:59 PM »
ReplyReply

There you go Jeff... that's exactly what I thought DNG profiles (and the ColorChecker Passport) were for.... to get accurate colour.  Is there an article somewhere that explains this in detail - that and the difference between what you get with a profile and WB?

No, it's to correct the sensor response for more accurate color rendering, not color correction. That's why a dual illuminate profile DNG profile will cover most situations where the light source is standard with a full spectrum (note, tungsten is very blue deficient which is why its useful to make either a tungsten or dual illuminate profile). Other light sources will be a lot different.
Logged
Ellis Vener
Sr. Member
****
Offline Offline

Posts: 1801



WWW
« Reply #11 on: February 17, 2012, 12:21:54 AM »
ReplyReply

"Also studio strobes differ in color with their power settings, how much different depends on the device,"

Not all do. The Paul C Buff Einstein 640 in constant color mode does not. The Broncolor Scoro and Grafit systems are very good in this regard as well. Of course different light modification methods and tools can have an effect on color.
Logged

Ellis Vener
http://www.ellisvener.com
Creating photographs for advertising, corporate and industrial clients since 1984.
Schewe
Sr. Member
****
Offline Offline

Posts: 5474


WWW
« Reply #12 on: February 17, 2012, 12:34:35 AM »
ReplyReply

But...does this have anything to do with sensor response? No...not really. Look you either buy into making a DNG profile for each and every shooting situation or you know better...I know better.
« Last Edit: February 17, 2012, 12:36:41 AM by Schewe » Logged
ErikKaffehr
Sr. Member
****
Offline Offline

Posts: 7410


WWW
« Reply #13 on: February 17, 2012, 12:48:56 AM »
ReplyReply

Hi,

My take is that you don't need a lot of different illumination based profiles. If you are shooting in a studio it makes a lot of sense to make a profile for your illumination. But as long as the lighting is continuous spectrum a pair of profiles and correct white balance will work fine. Would you for instance have very cold light shadow illuminated by sky on a sunny day it may also be useful to create a profile for that shooting condition. It may be a good idea to make a shoot of the CC card, just in case, you can use it for white balance and you can make a profile matched to the light.

A profile will probably not help you much regarding mixed light sources in my humble view.

If you have a spiky illumination spectrum, like energy saving lamps, I'd say that the color checker may not be that helpful, unless it's color checkers you want to shoot. The subject may have very different spectral absorption/reflection characteristics than your subject.

Best regards
Erik


I.E. do you create a new profile for each 'shoot'?  If I was shooting interiors, for example, the white balance may change from room to room depending on how much daylight is available, but do you need a different profile?  I'm still trying to get my head wrapped around profiles vs wb and seem to be confusing myself even more.
Logged

hjulenissen
Sr. Member
****
Offline Offline

Posts: 1679


« Reply #14 on: February 17, 2012, 12:49:12 AM »
ReplyReply

The camera sensor have a set of 3 spectral sensitivities. When making images, we want to relate those spectral responses to something known, so that we can map the raw color data into a standardized representation.

The illumination of a scene may have narrow peaks and bumps with far greater spectral resolution than the 3 channels that humans tends to see and think in. The object that is reflecting this illumination might also have a very "bumpy" spectral response.

I think that I see the difference between WB and camera profiling. But that supposes that the photographer is able to identify which camera profile applies, before doing WB. What is so bad about using a colorchecker in the start of every session instead of a grey card? The only technical reason I can see is that poor light may lead to a noisy color checker estimate, and the low number of colors may be worse than what Canon/Adobe/... might do in the lab.

I am unable to get believable saturated reds in my monitor system (Canon, Lightroom, WG-LCD, i1d3) when using the Adobe profiles. That is the reason that I bought the cc, and so far results have been promising.

-h
Logged
JRSmit
Sr. Member
****
Offline Offline

Posts: 369


WWW
« Reply #15 on: February 17, 2012, 10:55:51 AM »
ReplyReply

Jeff, just a question: how can ccp just correct the sensor color response if there is always a lens in front of the sensor to take a picture of the ccp card?
Logged

Fine art photography: www.janrsmit.com
Courses and workshops: www.centrumbeeldbeleving.nl

Jan R. Smit
Schewe
Sr. Member
****
Offline Offline

Posts: 5474


WWW
« Reply #16 on: February 17, 2012, 12:36:04 PM »
ReplyReply

Jeff, just a question: how can ccp just correct the sensor color response if there is always a lens in front of the sensor to take a picture of the ccp card?

Obviously, the whole optical system comes into play when you shoot a chart...but it's the differing response of the sensor to different light sources that is being used to create a DNG profile and correct for the sensor. The same profile will work with any lens...yes, there may be lens casts, but that's color correction, not sensor response differences.
Logged
deejjjaaaa
Sr. Member
****
Offline Offline

Posts: 860



« Reply #17 on: February 17, 2012, 06:57:11 PM »
ReplyReply

Jeff, just a question: how can ccp just correct the sensor color response if there is always a lens in front of the sensor to take a picture of the ccp card?


you can get a monochromator ( http://en.wikipedia.org/wiki/Monochromator ), remove a lens and illuminate the sensor directly w/ series of different light (each time a very narrow spectrum) - that is how really profiles are done by those who know...

granted you can't use off the shelf Adobe DNG Profile Editor then... or you can - if you write a program to generate a synthetic DNG from that series of your raw files...
Logged
Schewe
Sr. Member
****
Offline Offline

Posts: 5474


WWW
« Reply #18 on: February 17, 2012, 07:22:22 PM »
ReplyReply

granted you can't use off the shelf Adobe DNG Profile Editor then... or you can - if you write a program to generate a synthetic DNG from that series of your raw files...

Pretty sure that's what Thomas did in making up the whole DNG profile stuff...I'm not 100% sure but I think Eric (who is doing most of the newer profiles for Adobe Standard and vender matching) is doing...although Eric often just goes out and shoots with the camera/back.

I saw Thomas sitting around and working on one of the early Canons raw file support (a 1D or 1Ds as I recall) and it was a LOT of work...just trying to decode the white balance was the major hurdle. Then Nikon screwed the pooch with their "encrypted" white balance business...

Pretty long road we've come down really. Ya gotta give Thomas a lot of credit (and more recently Eric Chan).

But I still think X-Rite's incorporation of making DNG profiles is a really good thing. Just don't take it overboard.
Logged
Ellis Vener
Sr. Member
****
Offline Offline

Posts: 1801



WWW
« Reply #19 on: February 17, 2012, 07:28:53 PM »
ReplyReply

I did some testing of how much difference different lenses can make a while back when i was just starting to work with DNG profile creation , using Adobe's DNG profile editor and Xrite's ColorChecker Passport software.

Using the same body (a Canon EOS-1Ds Mark III), the same lights and modifiers (two Paul C. Buff Einstein 640s in color constant mode fitted with matched Chimera SuperPRO Medium softboxes) and three different lenses. The three lenses were:

1) a Canon 100mm EF 100mm f/2.8L micro

2) A Zeiss 100mm f/2 Makro Planar

3) A 20 plus year old Nikon 105mm f/4 AI-S Micro-Nikkor ( with Nikon F to Canon EF body adapter)

While there were slight color differences, and the lenses clearly had different but not hugely different t-stop values for the marked f-stop, the Camera Calibration profiles were virtually identical.

My conclusion is that Jeff is right.

The way i use custom camera specific profiles is dead easy i apply the camera profile for the body in use as part of the import process and then apply a white balance, shot with the white balance target in the CCPp, to the photos as needed. Any major change in lighting gets a new test frame with that target in it.

« Last Edit: February 17, 2012, 07:31:06 PM by Ellis Vener » Logged

Ellis Vener
http://www.ellisvener.com
Creating photographs for advertising, corporate and industrial clients since 1984.
Pages: [1] 2 3 »   Top of Page
Print
Jump to:  

Ad
Ad
Ad