<
  News     DIFP   Benefits Profiles Account  


forums

Recent Posts



(New threads or threads with new replies - ↑ Links to top of thread)

 Confused about layers in PK Sharpener
Apr 21, 2004 9:01 PM
Posted by Bruce Fraser

 Sony Artisan and Windows
Apr 21, 2004 8:44 PM
Posted by Thomas Fors

 Preferred printing direction (Enhanced Matte) & IP with roll paper
Apr 21, 2004 7:31 PM
Posted by John Vitollo

 Kodak 14n or Fuji S2 for stock?
Apr 21, 2004 7:20 PM
Posted by Andrew Rodney

 Sony Artisan with Monaca Optic XR
Apr 21, 2004 5:05 PM
Posted by Jeff Schewe

 Epson R 800
Apr 21, 2004 4:43 PM
Posted by Sharon Brunson

 Fuji S3 and verticals
Apr 21, 2004 3:18 PM
Posted by Peter Burian

 Applying in camera parameters and RAW files
Apr 21, 2004 3:11 PM
Posted by Peter Burian

 Is the D70 a good deal?
Apr 21, 2004 3:08 PM
Posted by Peter Burian

 Camera for raw
Apr 21, 2004 3:02 PM
Posted by Peter Burian

...see more posts

articles

Weekly Chatter: InCamera and/or ACR
Saturday August 23, 2003

Section 1: inCamera and/or ACR





Color Management/WorkFlow > Hardware and Software > inCamera and/or ACR







AuthorSubject: inCamera and/or ACR  
Eric Hiss
Posted on Wed Aug 20, 2003 2:22 AM

Hi,

Now that I've got my printer and display profiles under control, I'm thinking about camera profiles for my Canon 1D. I've read very favorable reviews about a sofware called inCamera. What do you guys think about that? Also I use ACR mostly for the RAW conversion....can that be configured to accept the custom input profile? I know that C1 can but I'm on a Mac and don't think its as fast as ACR (yet).

Thanks,
Eric Hiss


Jeff Schewe
Posted on Wed Aug 20, 2003 3:05 AM

Eric Hiss wrote:

> Also I use ACR mostly for the RAW conversion....can that be configured to
> accept the custom input profile?

Nope. . .sorry, Thomas Knoll doesn't like camera profiles. . .he prefers to
dynamically adjust between two color profiles (3200K and 6500K) to get
optimum white balancing of conversions. . .which is pretty much a
breakthrough (depending on who you talk to and who's trying to sell what).

I have no opinion of "InCamera" specifically, nor any of the current crop of
camera profiling. I do have some opinions about camera profiles in general.
The further away from the exact conditions that existed at the time you MADE
the target capture to make the profile, the less useful that profile will
be.

So, if you have a "standard lighting setup" for certain types of shooting, a
camera profile may work very well. Change the lighting substantially and all
bets are off (so will the profiles be).

If you shoot in changing light (outside) and have no control over things
such as color temp or dynamic range, I seriously think camera profiles are
useless. . .

My preference is to use the small Macbeth digital camera target and shoot a
frame at the beginning of a series of shots. I'll use the second from the
brightest patch to set the white balance tool in Camera Raw. . .then use
droplets to batch convert whole series of shots using the custom settings.


Jeff Schewe
www.schewephoto.com - www.pixelgenius.com - www.imagingrevue.com




Andrew Rodney
Posted on Wed Aug 20, 2003 7:50 AM

on 8/20/03 1:25 AM, IR - Eric Hiss wrote:

> I've read very favorable reviews about a sofware called inCamera. What do you
> guys think about that?

I've had reasonably good results from inCamera but keep in mind I've yet to
find ANY camera profiling solution that's even close to 100%. InCamera is
easy to use and it's damn inexpensive so I'd give it a shot since as I said,
I've seen it work more than not. I'd check and see if you can get say a 30
day money back guarantee (I'd want that with any camera profile package).
You might ask them if they recommend using the GretagMacbeth ColorChecker DC
since that raises the cost quite a bit more. InCamera does work with the
standard 24 patch Macbeth and you might be happy with the results using that
target. But camera profiling seems to be hit or miss...


Keep in mind that if you wish to use Adobe Camera RAW, you don't need (and
can't use) custom profiles. But that's not an InCamera issue.


Ian Lyons
Posted on Wed Aug 20, 2003 8:21 AM

>
> I've had reasonably good results from inCamera but keep in mind I've yet to
> find ANY camera profiling solution that's even close to 100%. InCamera is
> easy to use and it's damn inexpensive so I'd give it a shot since as I said,
> I've seen it work more than not. I'd check and see if you can get say a 30
> day money back guarantee (I'd want that with any camera profile package).
> You might ask them if they recommend using the GretagMacbeth ColorChecker DC
> since that raises the cost quite a bit more. InCamera does work with the
> standard 24 patch Macbeth and you might be happy with the results using that
> target. But camera profiling seems to be hit or miss...
>


inCamera is one of the better solutions and for the Mac I would suggest the Plug-in rather than the standalone version. The plug-in also allows you to build scanner profiles.

I've not found any benefits to using the more expensive DC target and so would suggest trying your luck with the basic ColorChecker.

You can see what I thought of inCamera here:

http://www.computer-darkroom.com/incamera-pi/incamera-pi.htm


John MacLean
Posted on Wed Aug 20, 2003 12:21 PM

> My preference is to use the small Macbeth digital camera target ...
>
> Jeff Schewe


Jeff,

Do you mean the Mini MacB
www.bhphotovideo.com/bnh/controller/home?O=productlist&A;=details&Q;=&sku;=222176&is;=REG

or the DC
www.bhphotovideo.com/bnh/controller/home?O=productlist&A;=details&Q;=&sku;=222175&is;=REG

I'm assuming the former.

I've got the full size regular Macbeth, but it is a pain to carry around.

Andrew, A while back you mentioned on one of the forums about a larger sized card of just the #20 - Neutral 8 patch color, or something similar. Is there such a beast?

Thanks,
John

www.nmdigital.com


Jeff Schewe
Posted on Wed Aug 20, 2003 12:55 PM

John MacLean asked:

> Do you mean the Mini MacB

Yes. . .the Macbeth MINI COLOR CHECKER is what I use. . .I just hold it in
front of the lens and do a quick shot-as long as the lighting hitting it is
the same as on the subject, it's an excellent tool for setting the Camera
Raw white balance tool. I also use the black and white patch as reference of
scene luminance range-although I do NOT try to force the black patch to be
zero RGB and I don't try to force the white patch to be 255. I like to keep
a lot of head room in CR and do final point setting in Photoshop.


Jeff Schewe
www.schewephoto.com - www.pixelgenius.com - www.imagingrevue.com




Andrew Rodney
Posted on Wed Aug 20, 2003 1:00 PM

on 8/20/03 11:25 AM, IR - John MacLean wrote:

> Andrew, A while back you mentioned on one of the forums about a larger sized
> card of just the #20 - Neutral 8 patch color, or something similar. Is there
> such a beast?

Yes but no one outside of GretagMacbeth and a few testes have one but it's
coming...


John MacLean
Posted on Thu Aug 21, 2003 10:23 AM

> Yes but no one outside of GretagMacbeth and a few testes have one but it's
> coming...

Andrew,

Can you alert us when it's available?

Thanks,
John
www.nmdigital.com


Andrew Rodney
Posted on Thu Aug 21, 2003 10:26 AM

-->Can you alert us when it's available?

But of course! I'll also ping the boys and girls in Europe and see if they have a date in mind yet.


Andrew Rodney
Posted on Thu Aug 21, 2003 4:13 PM

The targets will be released by Photo Expo or sooner according to my sources inside of Gretagmacbeth.


John MacLean
Posted on Fri Aug 22, 2003 10:23 AM

Thanks,

When is P.E.? I'm looking forward to your alert!

Best,
John


John Vitollo
Posted on Fri Aug 22, 2003 10:34 AM

http://www.photoplusexpo.com/photoplusexpo/index.jsp

New York, NY
Jacob Javits Convention Center
October 30 - November 1, 2003


Allen Pacheco
Posted on Fri Aug 22, 2003 10:49 AM

I remember reading somewhere that you could use a very wide space like ProPhoto as the output destination for Camera Raw when processing a camera profiling target. Then you'd develop a profile and and in order to use it you'd continue to use ProPhoto and simply apply the custom profile "over" it in PhotoShop. Is this far-fetched or does it carry any weight?


Also, although I have a good deal of respect for those who have given inCamera glowing reviews, they do not demonstrate several very distinct limitations of the program. Limitations which I have never been able to overcome. I have the stand-alone rather than the plug-in so I am assuming that the profile construction routines are very much alike. The ability to reproduce the target that is used to actually build the profile does not show the whole picture, so to speak. I would like to see one demonstration of a decent profile constructed from inCamera that could be used with confidence.





Bruce Fraser
Posted on Fri Aug 22, 2003 10:55 AM

-->Is this far-fetched or does it carry any weight?

I'd say it's far-fetched. To the extent that it worked at all, you're essentially tossing all the advantages of shooting raw. Better to simply use ACR as designed.


Ian Lyons
Posted on Fri Aug 22, 2003 11:18 AM

Allen,


>
> Also, although I have a good deal of respect for those who have given inCamera glowing reviews, they do not demonstrate several very distinct limitations of the program. Limitations which I have never been able to overcome. I have the stand-alone rather than the plug-in so I am assuming that the profile construction routines are very much alike. The ability to reproduce the target that is used to actually build the profile does not show the whole picture, so to speak. I would like to see one demonstration of a decent profile constructed from inCamera that could be used with confidence.
>
>

Maybe you'd be so kind as to describe what you see as limitations. What part of the process are you missing or don't understand, etc. etc. I can't speak for Bruce, Andrew or Jeff but I lack Spocks ability to mind meld. Help us help YOU!




Andrew Rodney
Posted on Fri Aug 22, 2003 11:28 AM

on 8/22/03 9:50 AM, IR - Allen Pacheco wrote:

> I remember reading somewhere that you could use a very wide space like
> ProPhoto as the output destination for Camera Raw when processing a camera
> profiling target. Then you'd develop a profile and and in order to use it
> you'd continue to use ProPhoto and simply apply the custom profile "over" it
> in PhotoShop. Is this far-fetched or does it carry any weight?

But why? I'm not sure what the advantage would be to take all the time an energy to do this when ACR doesn't really require it.

> I have the stand-alone rather than the plug-in so I am assuming that the
> profile construction routines are very much alike.

The newer versions of the plug-in are significantly better than the older stand-alone. Not that you'll get perfect results 100% of the time but I'd see what it takes to move from the stand-alone to the plug-in.


Andrew Rodney
Posted on Fri Aug 22, 2003 11:40 AM

on 8/22/03 9:50 AM, IR - Allen Pacheco wrote:

> I remember reading somewhere that you could use a very wide space like
> ProPhoto as the output destination for Camera Raw when processing a camera
> profiling target. Then you'd develop a profile and and in order to use it
> you'd continue to use ProPhoto and simply apply the custom profile "over" it
> in PhotoShop. Is this far-fetched or does it carry any weight?

But why? I'm not sure what the advantage would be to take all the time an
energy to do this when ACR doesn't really require it.

> I have the stand-alone rather than the plug-in so I am assuming that the
> profile construction routines are very much alike.

The newer versions of the plug-in are significantly better than the older
stand-alone. Not that you'll get perfect results 100% of the time but I'd
see what it takes to move from the stand-alone to the plug-in.


Allen Pacheco
Posted on Fri Aug 22, 2003 2:17 PM

Ian,

Discussing inCamera...

First, I'm going to refer to the 24-tile chart, not the ColorChecker DC.
And for additional reference I use Bruce Lindbloom's computer generated file of the ColorChecker chart based upon the Lab values of the chart....
http://www.brucelindbloom.com/index.html?ColorCheckerRGB.html

If you open the reference file that comes with inCamera you will see similar values.

So far so good.
The Pictographics profiling engine is going to use these values as a basis for computing the profile for the target image provided.

Assuming you use even lighting and well-controlled neutral balance then the target exposure now becomes the main variable, correct?

And the key values to determine exposure are the six neutral tiles on the target.

So, if you make a series of exposures, gross over and under exposures are easy to dismiss by just viewing the results. You end up with a range of exposures which could be considered quite good. So I tried to zero in on an image that satisfied the profile requirements.

After studying the results and making comparisons to the "desired" values in the reference file, it's clear to me that the best results are obtained when using an exposure that matches as close as possible as many of the neutral tiles as possible.
In my case, using Canon cameras, I get the best results when I match tile #20 (neutral 8) which is approximately L81 (Lab). However, the resultant profile wants the bright white tile to be L96. And when you apply this profile to an image the highlights are dealt a pronounced clipping. This soft shoulder of the processed data is straightened right out.

Deep shadows and three-quarter tones are also lifted and the contrast in these areas reduced but that can be addressed in an editor and may or may not be a bad thing.

But the highlights are gone. And to make things worse the clipping is not equal across all three channels...the blue channel is worse than the others and the result is a greenish off-white effect in the highlights.

I experimented using target images with more exposure to compensate and the resultant profile curve pulls and tugs at the image data. In other words, if the data in the target image deviates by a large amount from the reference values the resultant profile will end up distorting the image. In my mind I want the least amount of distortion possible. A profile without contrast bias.


As I mentioned, I have version 3.1.0.0 of inCamera. Perhaps, as Andrew mentions, the plug-in produces different results.
I spent some time talking with a rep at this year's PMA and he felt the best use of the product was for low contrast scenes with controlled lighting so as not to end up with highlight data too far beyond L90 (Lab). Is this practical for some? Perhaps. But it is still a limitation. Unless you have a way around it.

If I'm off in left field then show me the error in my method and I'll praise this product loud and clear.



Allen Pacheco
Posted on Fri Aug 22, 2003 2:52 PM

Hi Andrew,
You have been a staunch supporter of custom printer and monitor profiles. To deal with the unique output characteristics of a printer or monitor via generic profiles is almost always shunned, correct?

So why the reversal with using the generic T.Knoll profiling in Camera Raw? True, it delivers the output to a standardized color space but doesn't this fly straight into the face of precise color-management?


Bruce Fraser
Posted on Fri Aug 22, 2003 3:19 PM

Andrew I'm sure has his own opinion. I'll simply point out that profiling digital cameras is fundamentally different from profiling other devices.

Monitors and printers have a fixed color gamut. Capture devices do not, but in the case of scanners it's rendered moot because the film that we scan does have a fixed color gamut.

Digital cameras have something much more slippery—they have a color mixing function. They DO have a somewhat fixed dynamic range, modulated by the exposure controls, but they don't have a fixed color gamut.

ACR uses not one but two 'generic' profiles for each camera, one created under tungsten, the other under D65, but more importantly, it allows interpolation between, and extrapolation beyond these profiles.

In a fixed lighting situation with locked exposure controls, a conventional profile makes sense. Outside of these conditions it becomes progressively less reliable the further away the conditions are from the ones in which the target was captured.

So it makes sense to recognize cameras as a special case, and deal with them accordingly. Thomas' approach in ACR is one that makes a lot of sense, but more importantly, it works rather well when the camera's behavior bears a reasonable resemblance to the one he profiled.


Jeff Schewe
Posted on Fri Aug 22, 2003 4:10 PM

Allen Pacheco wrote:

> So why the reversal with using the generic T.Knoll profiling in Camera Raw?
> True, it delivers the output to a standardized color space but doesn't this
> fly straight into the face of precise color-management?

There is no such things as a "precise color-management" ability when dealing
with a device that always changes. . .Yes, one CAN profile a camera. . .and
that profile will be "useful" in direct proportion to the exact nature of
the state of the camera when the profile was made. Change _ANYTHING_
relating to the camera such as exposure, lighting, light balance, lens and
even exposure time, and the usefulness of the profile diminishes.

If you have a "standard" studio set up with a typical lens, a typical light
source and typical exposure, a profile can be quite useful. However, change
anything and it's a crap shoot.

This is akin to difficulties in profiling various scanners over the years
that contained a non-changeable "auto-setting" that made consistent and
repeatable scans impossible. Those scanners, one of the Polaroids was
particularly vexing, in effect were NOT "profilable" because the scanner
changed based upon the images scanned.

Another analogy is that while you CAN scan an E-6 target and create a
profile for E-6 film, there are no targets for color neg film. Why? Because
each film and each process batch can vary. . .making profiling impossible.

The geeks & gurus who have advocated color management practices have long
lamented the limitations and difficulties of camera profiling. But the
simple truth is, unless there are no variables, color management simply will
not work.

Jeff Schewe
www.schewephoto.com - www.pixelgenius.com - www.imagingrevue.com





Andrew Rodney
Posted on Fri Aug 22, 2003 4:13 PM

> Hi Andrew,
> You have been a staunch supporter of custom printer and monitor profiles. To deal with the unique output characteristics of a printer or monitor via generic profiles is almost always shunned, correct?

Not always. I've found the canned profiles in ImagePrint are quite good and while I can (and do) make custom profiles for it, I could see where a large majority of users would be quite happy with them. But that's an exception to the rule.

> So why the reversal with using the generic T.Knoll profiling in Camera Raw? True, it delivers the output to a standardized color space but doesn't this fly straight into the face of precise color-management?

You didn't read my "Point-Counter Point" on ACR in PEI mag ("Jane, you ignorant slut" ).

What ACR does, on a calibrated display is give you virtually all the controls to render an image as you wish, and then bring that into Photoshop with a Working Space you pick and honors that appearance and the numbers you saw in ACR. That's all I really need.

Without this kind of workflow, the numbers you get from a camera are undefined (without RAW, you can kind of believe that you're getting sRGB or Adobe RGB but I don't really buy into that). So with this RAW file you have numbers with no meaning and Photoshop hasn't a clue how to preview or convert that data. So you need a profile. But you don't with ACR. So in a nutshell, there's no reason with ACR to worry about building a profile. What you see is what you get. What more do you want?


Bruce Fraser
Posted on Fri Aug 22, 2003 4:52 PM

>But the
>simple truth is, unless there are no variables, color management simply will
>not work.

Make that uncontrolled variables. There are always variables....


Jeff Schewe
Posted on Fri Aug 22, 2003 5:40 PM

Bruce Fraser wrote:

> Make that uncontrolled variables. There are always variables....

Yeah. . .what Bruce said. . .

:-)


Jeff Schewe
www.schewephoto.com - www.pixelgenius.com - www.imagingrevue.com





Herbert Gibson
Posted on Fri Aug 22, 2003 6:09 PM

Unless I have lost my marbles (not out of the question) the geeks and gurus BCR (Before Camera Raw) used to advocate camera profiling. There was a minor argument, for instance, that the EOS D30 did not produce EXACTLY sRGB but the major selling point made by these folk for profiling hardware and software was variation across individual items of equipment.

As soon as Adobe launched ACR, the field seemed to split into Pals of Phase One and Mates Of Adobe. At that same time ACR advocates started to downplay the variation argument.

Sorry; but if the variation argument was valid before, it remains valid as an argument against Camera Raw. If it was not valid, then some folk were urging the purchase of expensive profiling gear where it was not necessary.

I realise that in some cases (reputedly the 2100 printer) variation may have been reduced, but that case has not been made for cameras.

There is an argument that ACR tries to balance profiles between 2 color temp end points whereas a user profile is only good for one point. But there is no reason why the user cannot make a number of profiles for different conditions and call them up as required. No great hardship.

Unfortunately, it appears that those resident in the Adobe camp see their role as cheerleaders for the Camera Raw implementation rather than presenting a balanced view. Ironically, that has only served to weaken your argument.


Andrew Rodney
Posted on Fri Aug 22, 2003 6:21 PM

There was nothing like ACR before ACR! We're dealing with RAW data here (which is a plus). Other than RAW, the camera spits out what it thinks is "right" and often, the numbers are undefined (or not defined correctly). So you had to have some profile to describe the numbers to Photoshop so it could preview and convert that data correctly.

That's just not necessary with ACR. Think iof RAW and ACR as having a color neg and you get to decide what filter pack you want.

Think of non RAW files as chromes that came out of the soup with the color and density fixed. You got what you got (but imagine if you will that the light in which you viewed the chrome was so undefined and could be virtually any color until you told the light source "I want daylight" (the profile).

If you want to use ACR, you pretty much forget using profiles for the portion of the workflow where you bring the numbers into Photohop. If you want to use another method of producing the color you'll likely need a profile or at the very least the right meaning of the numbers for Photoshop (which is all a profile does anyway).

Why do I like ACR? Well if I don't have to spend $500 on a good target and then a few hundred for software, then spend the time to make a profile (which is hit or miss half the time).

As I've said elsewhere; people keep asking for easier/quicker/simplier solutions for color management. When we get one, a group complains that we color geeks are happy there' less work to produce better results. I don't get it.

If someone came out with a self calibrating and profiling printer, I'd be super happy. You think I LIKE measuring patches?


Jeff Schewe
Posted on Fri Aug 22, 2003 6:25 PM

Herbert Gibson wrote:

> Unfortunately, it appears that those resident in the Adobe camp see their role
> as cheerleaders for the Camera Raw implementation rather than presenting a
> balanced view. Ironically, that has only served to weaken your argument.

You never saw me advocate profiling a camera. . .I tried it with several
packages. . .and to my eyes and way of working, no profile has ever worked.
. .I advocate 16 bit color correction, in large part BECAUSE camera
profiling sucks. . .

Quite honestly, I don't give a rat's ass about a "balanced view". . .I've
tried C1, and it sucks for usability and UI on the Mac. I've tried the Canon
shit and that sucks. . .I'm a photographer and user of this stuff. . .I say
what works and doesn't work for me.

Camera Raw suck a lot less than any of the other applications out there if
you have a Photoshop centric workflow. . .one that I've ALWAYS indicated
that I use. . .

I think Mr. Gibson, that your anti-Adobe slant is peaking through?

Use whatever works for you. . .or quite working.


Jeff Schewe
www.schewephoto.com - www.pixelgenius.com - www.imagingrevue.com




Bruce Fraser
Posted on Fri Aug 22, 2003 6:30 PM

Herbert,

I don't think you'll find any published piece or documented conversation where I have advocated camera profiling with any greater force than saying "if you shoot in the studio under tightly-controlled lighting, it may be worthwhile." (Which in my book falls substantially short of advocacy...)

Camera-to-camera variation isn't even slightly the point here. The point is that a single camera's response varies quite dramatically under different lighting conditions. It's a fundamentally different problem from those we face in profiling printers, displays, or even scanners.

Yes, you could make profiles for a camera under different lighting conditions. Conceivably, you could even write actions that would let you blend between renditions offered by the different profiles. But the profiles in ACR aren't endpoints—ACR can extrapolate beyond them. I can't think of a way to do that with static profiles and images in Photoshop. ACR offers a lot of functionality for $99. Its weakness, which I've pointed out on many occasions, is that a good many people out there seem to own cameras that don't behave the same way as the sample that Thomas profiled—particularly Nikons. That's being looked at by the people who may be able to do something about it.

I'm not in any particular camp, and I outgrew my cheerleader outfit decades ago. Not am I making any particular argument. I just explain how the damn things work, and let people draw their own conclusions. I personally find Camera RAW much more useful than Phase One Capture, but I only shoot under available light, and don't even pretend to be a photographer.

I think the Camera RAW approach is a conceptual breakthrough in addressing the central problem of characterizing the color produced by digital cameras, which is that they have a very variable response depending on the scene dynamic range and color temperature. Is it the be-all and end-all? Hell, no. But it seems to make more sense in more situations than the alternative, which is to try to lock down all the lighting and exposure variables, then profile the result. That's the philosophical argument, such as it is.

But the only reason I use ACR is because it works better than anything else I've tried.


Jeff Schewe
Posted on Fri Aug 22, 2003 7:55 PM

Bruce Fraser wrote:

> I'm not in any particular camp, and I outgrew my cheerleader outfit decades
> ago.

You got a picture of you in your cheerleading outfut? I would pay good money
to see that. . .I've seen you in a skirt, er, kilt, but you in a
cheerleader, that WOULD be rich. . .

Jeff


Bruce Fraser
Posted on Fri Aug 22, 2003 8:19 PM

We should do it together...


Ian Lyons
Posted on Sat Aug 23, 2003 9:55 AM

Allen,


>
> Assuming you use even lighting and well-controlled neutral balance then the target exposure now becomes the main variable, correct?

To a large extent exposure is the main variable, but not the only one.

Building a profile for the non linear images converted via Canons FVU or Phase One is relatively easy. So I assume that you're trying to build a profile based on the linear raw data. In such circumstances and without some aid to get the max white close to optimum then you'll likely struggle with every digicam profiling app currently available. If the conversion is to a gamma corrected (some prefer to call it toned) value and it results in highlight end spikes then the chances are that you will either have blown highlights or will cause the profiling app to produce less than ideal profiles. For such images you've got to find a compromise between linear and the gamma/tone corrected image from Canon FVU or Phase One. I described how I went about it in the following tutorial http://www.computer-darkroom.com/d30-profiling/d30_1.htm


With the Canon D30/60 my compromise meant that the camera profile tends not to be totally accurate. Simply applying a standard gamma curve to to the linear image appears to produce an image with its own set of problems; one of which is the green cast in the highlights that you mention. Since I work on the principle of holding the highlights as best as posible I use my own compensation curve. I build the profile based on this curve and apply it to those images what will benefit from it. The image will tend to be low in contrast but I hold everything that was capatured. I then fix the remaining in Photoshop. If the image isn't going to benefit from this curve/profile I'll now use Adobe Camera Raw and save myself a headache.

I don't have a clue about highlight characteristics of other cameras so the curve and how to determine the optium exposure to build it isn't something I can help you with.

If the image doesn't have one or two channels hitting level 255 or blown highlights then I'l just use Adobe Camera Raw.

Unfortunately there are currently no conversion apps that allow me to totally avoid building profiles. However, I do find that there is lot more flexibility in my approac



>
> And the key values to determine exposure are the six neutral tiles on the target.
>

Max white and lightest grey (patches 19 and 20 ) are the only ones I look at nowadays. The other 4 patches aren't really worth the card they're printed on for determing the optimum exposure for building a profile.


> So, if you make a series of exposures, gross over and under exposures are easy to dismiss by just viewing the results. You end up with a range of exposures which could be considered quite good. So I tried to zero in on an image that satisfied the profile requirements.
>
> After studying the results and making comparisons to the "desired" values in the reference file, it's clear to me that the best results are obtained when using an exposure that matches as close as possible as many of the neutral tiles as possible.

The gross over exposure theory doesn't work as it screws up everything else on the target. I simply choose the one that places patch 19 between level 240 and 250. I'll use patch 20 to determine/adjust the white balance. Patch 19 is used to choose the image with the optimum exposure. Once I capture the colorchecker with patch 19 lying between level 240 and 250 I'm done buggering around. If it falls a tad short or over this value I'll cheat.

> In my case, using Canon cameras, I get the best results when I match tile #20 (neutral 8) which is approximately L81 (Lab). However, the resultant profile wants the bright white tile to be L96. And when you apply this profile to an image the highlights are dealt a pronounced clipping. This soft shoulder of the processed data is straightened right out.
>

That suggest that you are profiling the raw linear image. A linear raw image doesn't have a shoulder. There are no soft landings with digital capture. Basing your exposure on patch 20 means that you will b leaving a lot up to chance. If necessary forget tonal accuracy - just get the damned highlights right.


> Deep shadows and three-quarter tones are also lifted and the contrast in these areas reduced but that can be addressed in an editor and may or may not be a bad thing.
>

You're expecting miracles from the current targets. Max white is at or around level 242 and max black at or around level 50. Everything above and below are pure invention on the part of the profiling apps.


> But the highlights are gone. And to make things worse the clipping is not equal across all three channels...the blue channel is worse than the others and the result is a greenish off-white effect in the highlights.
>

I only see this when building profiles for the for linear raw images. Other vendors have a diferent colour cast or worse - total reversal (i.e. near white turns black!) Again, I'm talking about the Canon D30, D60 and 10D!

I'm told that if you want into profiling nirvana you should speak to Jack Bingham or get his ColoreEyes 20/20. The top secret manual is supposed to be the answer to all our profiling problems. Me? I'll make do with inCamera and Adobe Camera Raw








Other Sections Available if you login:
calendar

April 2004
S M T W T F S
        1 2 3
4 5 6 7 8 9 10
11 12 13 14 15 16 17
18
19
20
21
22
23
24
25
26
27
28
29
30
 
May 2004
S M T W T F S
            1
2 3 4
5
6
7
8
9 10 11 12
13
14 15
16
17
18
19
20
21
22
23
24 25 26 27 28 29
30 31          
 

Upcoming Events


MAKE YOUR WORKflow (A D-65 event - led by IR expert Seth Resnick)
    4/18/04 at 9:00 AM to 4/24/04
The Digital Portfolio, (at John Paul Caponigro's studio, Maine)
    4/19/04 at 9:00 AM to 4/23/04
Article - Book Review
    4/24/04 at 5:00 PM
Santa Fe Workshops - Master Photoshop
    4/26/04 at 9:00 AM to 4/30/04
Article - Tutorial
    4/30/04 at 5:00 PM

...see more upcoming events

imagingrevue Members receive special discounts on most events listed in calendar

  


Copyright ©2002-2003
Imaging Revue and Virtuaserve Inc. Online Media Division
Privacy Policy
Terms of Service