|
| Author | Thread |
|
|
11/23/2004 04:07:34 AM · #1 |
At some point, on some site, I saw a method of using the 300D as a monitor calibration device sort of like using a spyder. Now, today, I have searched high, low, far, and wide but have not been able to find this page again.
Has anyone see this site? Can you help me remember how to get there?
-Dan |
|
|
|
11/23/2004 04:16:42 AM · #2 |
Well THAT would be interesting to see! Hope somebody can find it :-)
|
|
|
|
11/23/2004 11:52:47 AM · #3 |
Hasn't anyone else seen this? I know that I wasn't dreaming!
(In other words, BUMP!) |
|
|
|
11/23/2004 12:26:19 PM · #4 |
I saw the same article. I think it was on DP Review. I'll look around for it and let you know if I can locate it.
Edit...
Located something that might be along the lines.
//forums.dpreview.com/forums/read.asp?forum=1031&message=10543613
Message edited by author 2004-11-23 17:32:04. |
|
|
|
12/19/2004 08:14:42 AM · #5 |
Ok, so I wasn't able to find the original document, but luckily I printed it out and recently found that print amongst the mess on my desk...
Please note that these are not my words and I am reproducing them here exactly as seen on the print out of the original document for the benefit of all man- (and, to be PC woman-) kind. If you recognize the document and you can point me to the original, please do so as I would like to properly credit this work.
__________________________________
Calibrating your monitor using a digital camera.
Theory: Expensive calibration pods work because they analyze the light coming from the monitor against an objective standard. The software knows what it should be seeing, and what it is seeing. Knowing the difference between expected and the observed allows for a correction to be made.
What it does is uses different patterns of light tests and adjustments until the standard of 6500K is reached. Another device that measures light is a digital camera. While it does quite a bit more than that as well, it can certainly pull a lowly duty such as the one described here. I used a Canon Digital Rebel (EOS 300D), but the theory works for any camera with little modification of the techique.
What you will need:
-A digital camera with manual white balance
-A test chart available from //www.construnet.hu/nokia/Monitors/TEST/monitor_test.html
-Remote capture software for your camera is a MAJOR plus
-Simple editing software like Photoshop, Paint Shop Pro, etc...
-A monitor with independently adjustable RED, BLUE, and GREEN level controls (see note1)
-A desktop color curves adjuster. nVIDIA cards come with this, but other brands can use utilities available on the web
-About a half hour of time
Please read the entire procedure before stsarting, so you are sure to understand what is going on...
The procedure:
Basically, you display the test pattern on the screen, and take a picture of your monitor in a darkened room.
The resulting file is then opened in Photoshop (or whatever) and the eyedropper tool is used to look at the color of the various grey bars of the test pattern. Using canon remote capture to dump the files directly on the pc saves a tremendous amount of time, as well as compact flash connector life. Use your best lens, and focus it too close to the monitor so as to blur the pattern, and prevents monitor grime from appearing, as well as to avoid any moiré effects from the monitor aperture grille. Also make sure that your parameter is set correctly. Contrast, saturation, and color tone should all be the 0 or default position. Also remember to set the camera white balance to Cloudy, as that gives you a 6500K white balance. 6500K is what we want the monitor to look like at the end, so if you show white on the monitor, it is 6500K white. When I used the remote capture, I could have them all go into a specified folder, and PSP remembers the last open folder, so the actual image acquisition -> analysis is quite fast.
[Missing screen shot of the remote capture utility]
Open the capture in PSP, and then look at the various grey bars.
[Missing screen shot of the eyedropper tool in use]
These values should match as closely as possible, which brings us to the adjusting phase of our endeavor.
THis part requires a lot of trial and error. I cannot figure out any faster way to do it, but using remote capture, one can at least go through the cycle in a minute, so it actually does not take too long.
Looking at the various levels of white, you should notice that some are higher than others.
If your monitor does not have independent RGB level adjustments, skip ahead to the next section after you place the color as close to 6500K as possible. This might be labeled as 6500, imaging, standard, warm, daylight, etc. Consult your monitor manual for full details, as I don't know the exact procedure for every monitor out there.
Now then, once you figured that say RED is too bright by looking at the file in PSP (e.g. R155, G140, B142), lower that color on your monitor setup. YOu have to guess and check. Get all values as close as possible to each other. After adjusting, take another capture, and re-compare. Often you will find that some colors may be too bright in the bright whites, and too dark in the shadows. This required further adjustment using desktop curves correction. This utility is standard with nVIDIA based video cards, and access by going to Control Panel > Display > Settings > Advanced > geForce something > Color Correction. Place the mode to Advanced, enabling the curves to be adjusted.
[Missing screen shot of the nVIDIA Color Correction panel]
If you are not sure how curves work, I would suggest you read up on them somewhere online, or in the Photoshop manual. Basically, you will be using these curves as if you were adjusting a photo. Let's wlak through the process.
1. I notice in PSP that my dark whites have lower red values than the other colors.
2. I go to curves, select the red channel, and move the left side of the curve up slightly.
3. I re-shoot, and compare the results.
4. Repeat for all of the white bars, one at a time.
5. I realize that due to the lack of precision in the monitor curves adjustment, you will probably not be able to zero in on exact matches. However, I can get it a HECK of a lot closer to matching.
So, now that you have seen the workflow, you know the principles involved. While this is specific to windows, the rebel, and nVIDIA cares, the methods can be applied to other systems quite easily.
Some tips to make the process smoother:
-Use jpeg on the camera so you can open the file right up. You don't need a lot of detail in these pics.
-Use the keyboard shortcut "ALT+TAB" to switch between remote capture, the pattern and PSP.
-Drag the remote capture window to the lower part of the screen, and use spacebar to release the cutter.
-Only frame the test pattern in your camera. You don't need a monitor frame, remote capture windows, or anything like that getting in the picture.
-Be patient, and don't overcorrect.
-Save the color profile in the nVIDIA utility, so you don't have to repeast when finished.
Conclusion:
While I don't own a calibration pod, I do own a camera. After using the routine described here, I borrowed a pod, and ran the calibration. The result: only VERY slight bit of difference, leading me to believe this method works quite well.
I can say that my photos actually look worse of the monitor. I see a lot of noise and detail that was previously hidden. In other words, my monitor reveals the limitations of the camera, rather than hiding them. This is closer to how pics look when printed, and is NOT a bad thing.
It might take some playing around to get perfect results, but I can tell you improvement is almost always possible, especially if you have never calibrated before. When I started, some of the color values were double digits apart.
Now, a few white bars have exactly the same values for all three colors, and ALL of them are within 5 points.
Finally, my pictures on the monitor look like prints, and I didn't waste any money on a calibration suite. |
|
|
|
12/19/2004 09:00:08 AM · #6 |
all well and good - have you tried it?
|
|
|
|
12/19/2004 09:53:56 AM · #7 |
| sounds like it would work...im going to try it tonight and post the results on here. |
|
|
|
12/19/2004 10:06:33 AM · #8 |
usually if you print something from the internet, the URL prints in the lower left corner of one or all of the pages. I'm assuming you alread looked there though lol. Just trying to help.
|
|
|
|
12/19/2004 12:13:19 PM · #9 |
Originally posted by deapee:
usually if you print something from the internet, the URL prints in the lower left corner of one or all of the pages. |
Yeah, was hoping to find those fields... I apparently printed from some other document type... pdf, doc, etc...
Really funny thing about the document... at the very end is a copyright message... with the name Dan Davis.... really funny cause that's my name, but I don't think I'm the original author and I'm reasonably sure that I didn't put that message there. |
|
|
|
12/19/2004 03:26:03 PM · #10 |
Interesting..., but why the 6500K temperature? Wouldn't that be a bit on the warm side? Is it an international standard?
:)atwl
|
|
|
|
Current Server Time: 04/13/2025 06:12:20 AM  |
Home -
Challenges -
Community -
League -
Photos -
Cameras -
Lenses -
Learn -
Help -
Terms of Use -
Privacy -
Top ^
DPChallenge, and website content and design, Copyright © 2001-2025 Challenging Technologies, LLC.
All digital photo copyrights belong to the photographers and may not be used without permission.
Current Server Time: 04/13/2025 06:12:20 AM EDT.
|
