Pages: [1]   Go Down

Author Topic: Camera Calibration  (Read 10615 times)

Hans Kruse

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2106
    • Hans Kruse Photography
Camera Calibration
« on: April 27, 2008, 06:15:57 pm »

I found the video clip on camera calibration quite interesting. Some questions come from that:

1) I would like to hear from others who have done a camera calibration, what parameters their calibration came up with. Both Steve's and Michaels 1Ds mk3 came up close relative to the ACR default. Does the Adobe ACR camera calibration actually fall in the middle of the camera population of a particular camera model? Probably it could fall in the middle or at the edge. Of the calibration parameters I  have seen now from 1Ds mk3 anbd 5D the all fall pretty far from the ACR calibration.

2) It was not clear to me from the video, if a camera calibration using this method and the parameters available in ACR (and Lightroom) can properly calibrate a camera. Meaning that e.g. a Nikon or Pentax camera can be calibrated to look exactly as a Canon 5D or 1Ds mk3 etc. camera in terms of colors.

3) How could jpg in camera be calibrated?

Regards,
Hans

Schewe

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 6229
    • http:www.schewephoto.com
Camera Calibration
« Reply #1 on: April 27, 2008, 08:44:56 pm »

Quote
2) It was not clear to me from the video, if a camera calibration using this method and the parameters available in ACR (and Lightroom) can properly calibrate a camera. Meaning that e.g. a Nikon or Pentax camera can be calibrated to look exactly as a Canon 5D or 1Ds mk3 etc. camera in terms of colors.
[a href=\"index.php?act=findpost&pid=192176\"][{POST_SNAPBACK}][/a]


The intent and design of th Calibrate function is to use the ColorChecker charts as a basis to achieve "accurate" colors based on the color samples. There's nothing about calibration that is designed to make a Nikon look like a Canon. That's not the point.

As for #3, no, of course not...the calibration is for the camera supplied RAW files...trying to calibrate the camera JPEGS would be pretty far afield (and I doubt the results would be worthwhile).

Oh, and my name if Jeff...

:~)
Logged

Hans Kruse

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2106
    • Hans Kruse Photography
Camera Calibration
« Reply #2 on: April 28, 2008, 05:56:41 am »

Quote
The intent and design of th Calibrate function is to use the ColorChecker charts as a basis to achieve "accurate" colors based on the color samples. There's nothing about calibration that is designed to make a Nikon look like a Canon. That's not the point.

As for #3, no, of course not...the calibration is for the camera supplied RAW files...trying to calibrate the camera JPEGS would be pretty far afield (and I doubt the results would be worthwhile).

Oh, and my name if Jeff...

:~)
[a href=\"index.php?act=findpost&pid=192213\"][{POST_SNAPBACK}][/a]

First, sorry about the name mixup    I had a collegue of that name so probably this was sticking in my memory.

The question about the ACR calibration was about how accurate it is. Since the calibration is on top of the calibration profile already in ACR the additional paramters (shadow tint, red primary hue, saturation, etc.) seems to be intended to tweak the calibration curve into a correct shape. I'm not a color management expert, but the theory says that in color management icc profiles and calibration is used to make sure that input and output icc profiles will convert colors correctly between source and destination. We use icc profiles in monitor calibration and printer calibration, so why don't we use icc profiles in ACR? Wouldn't that be the optimal assuming a reliable calibration and profiling process was at hand? Compared to an optimal calibration, how accurate is the ACR calibration?

I'm also wondering about the ACR default profile compared to the camera population. Is the calibration in the middle of that population or possibly at the edge, meaning that most camera owners have mildly incorrect color rendering from ACR?

Assuming that ACR calibration would be 100% accurate, wouldn't the consequence of that be that colors from different camera models and brands look exactly the same? There is much talk about Canon and Nikon colors, but my view is that given a proper and accurate calibration there would be no such thing as Canon or Nikon colors. I don't need 100% accurate colors, but I like to have a normalized base to work from, just like I calibrate my monitor and each type of paper I use in my printer is profiled as well.

The point about jpg calibration was not about ACR at all, but about the possibility to calibrate a camera for in camera jpg processing. Is that possible? I assume that my in-camera jpg processing is not calibrated from the factory (even for a 1Ds mk3), but if course, I see a significant difference in colors between the ~5MP jpg thumbnail in RAW files from the 1Ds mk3 compared to the results from Lightroom (or ACR in PS CS3). As I don't use jpg's from the camera, it is a theoretical question for me.

But many thanks for the videos, they were informative, inspiring and humorous  

digitaldog

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 20630
  • Andrew Rodney
    • http://www.digitaldog.net/
Camera Calibration
« Reply #3 on: April 28, 2008, 09:32:52 am »

Quote
Assuming that ACR calibration would be 100% accurate, wouldn't the consequence of that be that colors from different camera models and brands look exactly the same?

A few points. First, the process is really useful if your goal is to produce a colorimetrically correct, output referred image of a color checker under a fixed illuminant. If the term output referred is new to you, read this:
http://www.color.org/ICC_white_paper_20_Di...ment_basics.pdf

This also doesn't take into account an issue/problem called camera metamerism. It doesn't take into account altering the illuminant and popping vastly different colored objects into the scene.

And lastly, its important to remove the idea that doing anything here is going to produce 100% accurate color, the term is unobtainable until you define what 100% accurate is (I'd submit its the measured color in the scene). To some degree, that's kind of what you end up with, in terms of 24 patches (plus or minus a value of difference in the measured color checker, the illuminant used in the measuring device and the values you see in ACR). Remove the color checker and pop something totally different into the scene, shoot and measure the object and what you get in ACR. Do they match? Very often they do not.
Logged
http://www.digitaldog.net/
Author "Color Management for Photographers".
Pages: [1]   Go Up