Pages: [1] 2 3   Go Down

Author Topic: Profiling  (Read 13355 times)

Leonardo Barreto

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 379
    • http://leonardobarreto.com/
Profiling
« on: December 12, 2007, 08:14:31 pm »

_What program do you use to profile your digital backs?
_How often do you feel that a new profile has to be made?
_Is it true that the Leafs are more easily profileable  than the Phase?
_Do you apply the profile before raw processing the file or as a TIFF in Photoshop?(image>mode>assign profile)
_Are C1 profiles enough, or, are them so perfect that no further profiling is necessary.

I am yet to climb this shoulder of the learning curve, and I know that you have and could give me a hand, thanks.
Logged
[font=Comic Sa

digitaldog

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 20651
  • Andrew Rodney
    • http://www.digitaldog.net/
Profiling
« Reply #1 on: December 12, 2007, 08:23:18 pm »

What makes you feel you need to profile your back (of that the process is even useful or effective)?
Logged
http://www.digitaldog.net/
Author "Color Management for Photographers".

ZOG

  • Newbie
  • *
  • Offline Offline
  • Posts: 39
    • http://
Profiling
« Reply #2 on: December 12, 2007, 09:02:35 pm »

The main reason I profile my back is for artwork reproduction. I seem to have problems with my newer PhaseOne back (P45+). I have an old scanning back and profiling it is a breeze. It makes very accurate reproduction colors for artwork and densities that are right on! The secret is to have the profile applied before opening the image in Photoshop.

I seem to have difficulties to shoot a picture (with the P45+) without having any profiles attached to the image. Dumping the profile in Photoshop does not seem to solve the problem. You cannot make a decent profile from a file that has already a profile attached to it. Therefore, my images have a lot more noise and are not as color accurate....

I was told by PhaseOne that some profiling software have a problem with the newer backs because of the high quantity of color information the provide. I have my doubts.....But, they could be right.

This is why I still use the old scanning back (by PhaseOne) to make my artwork reproductions. I wish I could use my newer back. It's a lot faster..... I guess I'll have to keep experimenting.

For other pictures, C1 does the job. There is plenty of tools to get the color you want.

Andre
Logged

Streetwise

  • Full Member
  • ***
  • Offline Offline
  • Posts: 125
    • DavidAnderson.tv
Profiling
« Reply #3 on: December 12, 2007, 09:33:52 pm »

Quote
_What program do you use to profile your digital backs?
_How often do you feel that a new profile has to be made?
_Is it true that the Leafs are more easily profileable  than the Phase?
_Do you apply the profile before raw processing the file or as a TIFF in Photoshop?(image>mode>assign profile)
_Are C1 profiles enough, or, are them so perfect that no further profiling is necessary.

I am yet to climb this shoulder of the learning curve, and I know that you have and could give me a hand, thanks.
[a href=\"index.php?act=findpost&pid=160249\"][{POST_SNAPBACK}][/a]


Profiling a camera is a lot like shooting at a fast moving target. ie. lighting is usually always different in every situation....

Generally speaking, the only time I profile is when I'm doing art repro work. I use the Leaf 65 back and shoot using the ColorChecker SG chart. I basically turn off everything in the camera relating to curves, sharpness, etc and shoot an exposure with the card in it. Then I'll export the tiff using the ProPhoto RGB profile (you have to export using *something* with Leaf, but open it in Photoshop without any profile. I'll then use the PictoColor plugin and create a custom profile from the card. I'll then open the other images with no profile enabled and "apply" the custom profile, ultimately converting to Adobe 98. It's still not perfect though, I have to do a fair amount of tweaking in CS3 to get a perfect match on the printer.

I've never used a Phase back, so I don't know about ease of profiling it. In most cases though, just doing a custom gray balance gets you close to where you need to be.
Logged

jpjespersen

  • Guest
Profiling
« Reply #4 on: December 12, 2007, 09:36:59 pm »

I think that this looks very promising.  http://www.luminous-landscape.com/reviews/...ommercial.shtml
I am planning on getting it soon.
JP
http://p45plus.typepad.com
Logged

Leonardo Barreto

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 379
    • http://leonardobarreto.com/
Profiling
« Reply #5 on: December 12, 2007, 10:14:25 pm »

Yes. That is exactly the type of question that I have to address. I have a friend that just got his 22mp Leaf (he has a Betterlight and 1dsmk2) and was going to go through the process of profiling his back so he invited me to his studio. He has a target with all the colors that you have to balance with the same lighting you are using, in this case 4 Totas, and bring the image to the interface of the program, align it perfectly well and then the software makes an analysis of the way the colors should be an the way they are being captured by the sensor.

After that a ColorSync profile is stored in the Library. This is somehow used to corrects the color variations introduced by a handful of elements like light source, lens, and the sensor itself.

All of this makes sense to me, but not completely.

The Phase approach is to say that they have spent a lot of time and neurons to make "ready to wear" profiles for every camera and light situation.

I use the C1 profiles in my work all the time, for example, there is one that blocks the purple more than the others, and the Easy Gray profile is of grate use to clean white walls in gallery installation shots with almost no change in the color of the art work itself (according to someone at Phase this was made for studio seamless users and to correct lens casting).

So your question is an interesting one since it implies that some photographers believe in using this profiles and not introducing one concocted the morning of the important shoot...

 


Quote
What makes you feel you need to profile your back (of that the process is even useful or effective)?
[a href=\"index.php?act=findpost&pid=160252\"][{POST_SNAPBACK}][/a]
Logged
[font=Comic Sa

Leonardo Barreto

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 379
    • http://leonardobarreto.com/
Profiling
« Reply #6 on: December 12, 2007, 10:39:59 pm »

Thank you jpjespersen for the link.. this is what the guys at ColoreEyes say...

"Coloreyes is based on the premise that tonal values captured by your camera should not be messed with. We also firmly believe and have proven over and over again that one profile can work extremely well under a huge range of conditions. So there is no need to make a huge amount of profiles. One good one will do the trick."

That is an interesting notion.I wonder if it applies to photography of artwork..

This is what they also say...

So if you listen to the “experts” camera profiling can’t possibly work. It’s too expensive, to time consuming and on and on and on. The problem with all of these “experts” is they have never tried Coloreyes. Cameras have always been treated like scanners, so perfect reproduction of the target was the object of the exercise. Shooting the target was treated like a simple copy job with very little in depth instruction.

Three years ago we built this product from the ground up using some very different concepts. Two major changes in the way we approached this process dramatically changed the quality and universality of custom camera profiles. First, shooting the target needed to be done in a far more controlled fashion. The quality of the data captured had a dramatic impact on the viability of the profile. Hence the manual that ships with ColorEyes explains in great detail how to shoot the target.

Second, the assumption that accurate reproduction of the target was at the heart of a good profile is a fatal error. Digital cameras, like film create an image with contrast higher than our perception. We have come to expect a certain tonal range and contrast in images we like. Building a profile to reproduce a target exactly takes away that nice S curve of tonal values we have come to expect, and instead we get a profile that changes those values. This is the root of the comments about camera profiling being too flat. Allow the profiler to reproduce the target exactly and what you get will not make you happy.

It means that this is a difficult science   or art/science. Or a slippery portion of the learning curve...
Logged
[font=Comic Sa

AndreNapier

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 422
    • Andre Napier Photography
Profiling
« Reply #7 on: December 13, 2007, 01:26:43 am »

I profile my A75S for every light set up in a studio and name the profile with the name of the specific  model or client and the take number. The whole process for setting profiles in 3 folders for color, brightness and contrast , and saturation takes me about 5-7 minutes. I am shooting people so I always aim for the best and most flattering skin colors. The profiles attached by Leaf  software are totally useless and sometimes just scary. I can see how someone not knowing better can freak out by using portrait profile on A75S and exposing it just right. Your image looks like the model skin is a combination on lemon and orange. Profiling Leaf is however so easy that my 11 year old son can truly do it on his own even though ( because of the harsh USA under age labor laws ) he only works part time in my sweat shop studio. Leaf colors are however amazing due to the chip they use and this comment comes from someone who use P1 for 6 year before switching.
If you get Aptus I can help you with basic understanding of the process. It took me under a week to feel confident and happy with Leaf.
Http://Andrenapier.com
Logged

eronald

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 6642
    • My gallery on Instagram
Profiling
« Reply #8 on: December 13, 2007, 03:01:07 am »

I do have an opinion here, but it is not necessarily that of a working photographer. As a consultant, I can supply edited profiles for any camera supported by Capture One. Some members of this forum are using my profiles. These are specifically designed for imaging human skin, and supply "pleasing" color rather than "accurate" color.  I use my own custom software in this process, in addition to what is available from vendors. I also assist members of the photo industry in tuning camera color - details can be supplied on request.

At the moment color management for capture is still evolving, while color management for output eg. ink jet print or CMYK press is now a well-established process.

The industry consensus seems to be that creating profiles on site for specific stable lighting situations and product shots works well. You buy a package, set up, and have a good chance of decent color in what you're imaging.  General-purpose camera profiles, that should work for every lighting, and can be used for shooting people, are still as much of an art form than a science, I'd say, which is why I supply deliberately tweaked profiles which have a specific "look", like a film does.

I'm very interested in hearing what members of this forum are doing with respect to camera color, and can also provide advice if solicited. Andrew Rodney, who posted above, is one of the industry's leading experts on print color management.

Edmund
« Last Edit: December 13, 2007, 03:36:46 am by eronald »
Logged
If you appreciate my blog posts help me by following on https://instagram.com/edmundronald

Dinarius

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 1218
Profiling
« Reply #9 on: December 13, 2007, 07:24:44 am »

As someone who shoots fine art catalogues for a living, I find this thread fascinating.

All of my work is shot on location, so I tweak the capture to suit the specific situation.

I combine a basICColor Grey Card (read about it here http://www.colourmanagement.net/bc.html ) for neutralizing any casts, with a Bruce Fraser type calibration using a Gretag CC.

The grey card also helps with gamma and density having a perfect RGB of 143 in Adobe 1998. (Lab 60,0,0)

The results I get are better than film in that:

1. I don't have to concern myself with any vagaries down the line (processing/scanning/colour correction of the scan, when necessary)

2. Unlike film, my digital file has no inbuilt curve. It is as flat as a pancake and I can give it as much or as little as I like. I wouldn't have it any other way!  

D.
Logged

Dustbak

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2442
    • Pepperanddust
Profiling
« Reply #10 on: December 13, 2007, 09:38:21 am »

I have been wondering whether I should profile my backs but the right tools are expensive so I am still considering. Using grey cards and color cards naturally but no camera profiling software like the X-Rite camera profiler module or something like that.

At this moment,  I already come so close to what most people perceive to be correct colors for my product & still work I don't need to.

For portrait work I wonder whether I would like to have 'correct' colors, I prefer something that is pleasant instead of 'correct'.

Sofar I have not yet had significant incentive to do the back profiling, not significant enough to drop several K on it. There are other things higher on my priority list.
« Last Edit: December 13, 2007, 09:47:01 am by Dustbak »
Logged

Leonardo Barreto

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 379
    • http://leonardobarreto.com/
Profiling
« Reply #11 on: December 13, 2007, 09:42:11 am »

So, correct me if I am wrong..

* Photographers can live w/out profiling but if you want predictable colors as in skin and painting it is the way to go...

* Some do a shoot specific profile -- or set specific -- as part of their pre production procedure, others make a general profile (or use a PRET-A-PORTER profiles).

*One way to profile is by using only one light source so not to be married to multiple variable lamp set up and make distinction between color and density corrections.

*Applying the profile has two schools, one is on the program that develops the raw and makes a one generation correction before the fact, and the other, -- less intrinsically perfect -- that converts the file -- using no profile, sharpening, curves etc-- and applies: image>mode>assign profile in PS to the TIFF.

* Targets "have to be matte" say some maestros, and for other experts "the only way to get highly saturated colors is by using a GLOSSY color target" !  

*"Profiling Leaf is however so easy that my 11 year old son can truly do it on his own" or "
... if you listen to the “experts” camera profiling can’t possibly work. It’s too expensive, to time consuming and on and on and on."
« Last Edit: December 13, 2007, 09:43:42 am by Leonardo Barreto »
Logged
[font=Comic Sa

digitaldog

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 20651
  • Andrew Rodney
    • http://www.digitaldog.net/
Profiling
« Reply #12 on: December 13, 2007, 09:57:55 am »

Quote
At this moment,  I already come so close to what most people perceive to be correct colors for my product & still work I don't need to.

For portrait work I wonder whether I would like to have 'correct' colors, I prefer something that is pleasant instead of 'correct'.

Sofar I have not yet had significant incentive to do the back profiling, not significant enough to drop several K on it. There are other things higher on my priority list.
[{POST_SNAPBACK}][/a]

Sounds like you've answered your question. Profiling a camera, in some conditions (copy work comes to mind) is quite useful because the illuminate is fixed, the dynamic range is fixed etc. You CAN treat the capture device as a scanner on a stick. Change conditions, all bets are off.

Then there comes the idea, is a profile a solution in search of a problem. In your case, you seem happy with the color rendering you get. Unless you're spending a lot of time either in the Raw converter (better) or Photoshop trying to adjust global color and tone, you're probably fine.

Digital cameras produce data that's scene referred. You have to decide how to handle the conversions to output referred. The ICC ( a group that by and large, promotes the use of ICC profiles for devices) has this to say about the process:

[a href=\"http://www.color.org/icc_white_paper_20_digital_photography_color_management_basics.pdf]http://www.color.org/icc_white_paper_20_di...ment_basics.pdf[/url]

It would be great if we could profile the scene (the illuminant, the dynamic range of the scene using known spectral properties of the chip) and that may come in the foreseeable future. Until that time, treating a digital camera like a scanner to build a profile works best when you treat the digital camera like a scanner when you shoot! Scanners are simple devices when you consider the illuminant and intensity of the illuminant provided (its fixed) and that the thing you're trying to capture has a fixed gamut (film).

When the time comes we can measure the scene itself to build scene specific profiles (and that will come), shooting a target of color patches to build a profile will seem like using a kitchen knife as a screwdriver on the space shuttle.
« Last Edit: December 13, 2007, 09:58:14 am by digitaldog »
Logged
http://www.digitaldog.net/
Author "Color Management for Photographers".

jpjespersen

  • Guest
Profiling
« Reply #13 on: December 13, 2007, 01:26:45 pm »

Remember there is Profiling Scenes and Profiling a Camera.  The product ColorEyes is for profiling your camera.  Check out some examples.  http://www.integrated-color.com/cecamera/examples.html
Logged

digitaldog

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 20651
  • Andrew Rodney
    • http://www.digitaldog.net/
Profiling
« Reply #14 on: December 13, 2007, 01:37:07 pm »

Quote
Remember there is Profiling Scenes and Profiling a Camera.

What? Makes no sense.
Logged
http://www.digitaldog.net/
Author "Color Management for Photographers".

jpjespersen

  • Guest
Profiling
« Reply #15 on: December 13, 2007, 03:25:49 pm »

Maybe I am wrong, I am basing my information off of this video.  http://www.luminous-landscape.com/reviews/...ommercial.shtml
Quote
What? Makes no sense.
[a href=\"index.php?act=findpost&pid=160408\"][{POST_SNAPBACK}][/a]
Logged

digitaldog

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 20651
  • Andrew Rodney
    • http://www.digitaldog.net/
Profiling
« Reply #16 on: December 13, 2007, 03:44:45 pm »

Quote
Maybe I am wrong, I am basing my information off of this video.  http://www.luminous-landscape.com/reviews/...ommercial.shtml
[a href=\"index.php?act=findpost&pid=160439\"][{POST_SNAPBACK}][/a]

As a marketing video, it seems OK. I however wasn't at all convinced, as Michael seemed to be and found many of the various points about the process by the ColorEye's boys to be conflicting.

You can attempt to profile a camera. How do you profile a scene? A scene as you see it, someone else, some other device?

Yes, the title is appropriate, its a Commercial!
Logged
http://www.digitaldog.net/
Author "Color Management for Photographers".

Leonardo Barreto

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 379
    • http://leonardobarreto.com/
Profiling
« Reply #17 on: December 13, 2007, 04:15:37 pm »

I remember a workshop I had when student at ICP and the big thing was the Zone System. We used to make a test to get the exact ISO using one particular film/lens/film/development method so that density and contrast would remain the same whenever we used the same components and procedures.

That was some type of profiling, --the color variable was not in--.

In the present day the components are: not the "scene", but the light sources illuminating the scene -- I thinkt -- the camera, lens, digita sensor and light meter etc.

If you have a "scene" that was, for example, a wall in the MoMa, and you could set the lights exactly the same, or leave the as they where when the profile was done, then, in theory and practice that "scene" was profiled and any painting that you put on the wall will have consistent color.

On the other side, I have also noticed that the methodology of profiling (cameras or camera plus light source(ses) is as disputed territory as El Putumayo (guerrilla controlled, Colombia).

I have not reached the conclusions part of this book, on the contrary, we could go point by point, for example..

POINT ONE (not in any particular order)

Do we want o use one light source? The proponents of this notion, the friend's of jpjespersen, say, with some logic to it, that the only way to be sure that there the target is going to be illuminated by uniform light is by using a single lamp (they also prefer daylight flash).

OK, who is against POINT ONE and why//.

Quote
As a marketing video, it seems OK. I however wasn't at all convinced, as Michael seemed to be and found many of the various points about the process by the ColorEye's boys to be conflicting.

You can attempt to profile a camera. How do you profile a scene? A scene as you see it, someone else, some other device?

Yes, the title is appropriate, its a Commercial!
[a href=\"index.php?act=findpost&pid=160445\"][{POST_SNAPBACK}][/a]
Logged
[font=Comic Sa

digitaldog

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 20651
  • Andrew Rodney
    • http://www.digitaldog.net/
Profiling
« Reply #18 on: December 13, 2007, 04:46:35 pm »

Quote
Do we want o use one light source? The proponents of this notion, the friend's of jpjespersen, say, with some logic to it, that the only way to be sure that there the target is going to be illuminated by uniform light is by using a single lamp (they also prefer daylight flash).

OK, who is against POINT ONE and why//.
[a href=\"index.php?act=findpost&pid=160454\"][{POST_SNAPBACK}][/a]

You picked up on just one of the many discrepancies of this commercial (I only wish Michael had picked it up). They tell you that you should use a single light because no two are identical in terms of the illuminant (it would be interesting to actually measure the differences of two good stobes using an EyeOne Pro Spectrophotometer). OK, so now, how when I use a world of other lighting (just the differences in daylight throughout the day), does this magic profile now somehow not 'care'?
Logged
http://www.digitaldog.net/
Author "Color Management for Photographers".

jpjespersen

  • Guest
Profiling
« Reply #19 on: December 13, 2007, 06:14:05 pm »

Judging by the samples provided ( http://www.integrated-color.com/cecamera/examples.html) ,
this profiling of the cam seems like a good thing.  To me it makes sense.  It is comparative to using a canned printing profile as compared to making your own printing profile.
JP
Logged
Pages: [1] 2 3   Go Up