Maybe I'm being dense, but I don't understand how I'm supposed to expose the target. Is proper exposure of no consequence?
E.g., I'm going calibration at ISO 400 800 1600 3200 for a 10D. I am doing exposures of 1/6, 1/15, 1/60, 1/125. Camera on Manual mode, f/4 (fwiw).
All in the same lighting conditions. Obviously, some of those are going to be terribly overexposed, and some underexposed.
Does over/underexpsure not matter?
John
Exposure Settings for Calibration Target
In-camera NR below 1/30 sec
Dear Vlad,
My camera makes its own noise reduction when using shutter speeds longer than 1/30 sec. I therefore want two versions of each profile. My original profiles were made with 1/2000 exposure time (I used flash, which also unfortunately forced me to use Default WB) and around 1/2 sec. To have NI to select the correct profile for each new image I thought it would be smart to rename the profiles to file names containing 1by25 and 1by30, respectively.
DSC-F828 (ISO64; 8bpp; SharpSoft; 3264x2448; WB Default; 1by30s).dnp
DSC-F828 (ISO64; 8bpp; SharpSoft; 3264x2448; WB Daylight; 1by25s).dnp
"Auto Match Profiles" did still not, however, pick the correct profile (it picked the 1/25 one for a 1/250 sec image).
I then edited the "Exposure time:" in the "Device Mode:" of each profile so that they also contained 1/25 and 1/30, respectively, but still it does not work.
I can now see from the "Profile Viewer" that the profile EXIF data still contians the old (1/2 sec) exposure time. Do you think that that is the problem or could it be that the images differ in white balance (Default vs. Daylight)?
In any case it looks as if I should edit the EXIF data of the profiles, or what do you think? If so, can it be done, and in such a case, what would be the easiest way to do it?
Anders
My camera makes its own noise reduction when using shutter speeds longer than 1/30 sec. I therefore want two versions of each profile. My original profiles were made with 1/2000 exposure time (I used flash, which also unfortunately forced me to use Default WB) and around 1/2 sec. To have NI to select the correct profile for each new image I thought it would be smart to rename the profiles to file names containing 1by25 and 1by30, respectively.
DSC-F828 (ISO64; 8bpp; SharpSoft; 3264x2448; WB Default; 1by30s).dnp
DSC-F828 (ISO64; 8bpp; SharpSoft; 3264x2448; WB Daylight; 1by25s).dnp
"Auto Match Profiles" did still not, however, pick the correct profile (it picked the 1/25 one for a 1/250 sec image).
I then edited the "Exposure time:" in the "Device Mode:" of each profile so that they also contained 1/25 and 1/30, respectively, but still it does not work.
I can now see from the "Profile Viewer" that the profile EXIF data still contians the old (1/2 sec) exposure time. Do you think that that is the problem or could it be that the images differ in white balance (Default vs. Daylight)?
In any case it looks as if I should edit the EXIF data of the profiles, or what do you think? If so, can it be done, and in such a case, what would be the easiest way to do it?
Anders
Anders,
Neat Image does not use the file name of a profile nor its device mode comments to do matching profile-image. Instead, it uses the EXIF values.
When you build a new profile, the EXIF values are received from the image you analyze.
When you match profiles against a specific input image, the above profile EXIF values are compared with EXIF values of the input image.
So, to have Neat Image distinguish profiles for images with and without camera built-in noise reduction, you have to build two profiles from two test images that were shot with and without noise reduction.
After that, Neat Image will automatically select the right profile from these two based on exposure time. No additional manual tweaking of device mode comments or file names are necessary.
Please try that.
Also, please make sure the profile matching priorities are set according to your needs in Neat Image Options. Try to increase the priority of exposure.
Vlad
P.S. Thinking about the exposure:
- 1/250 is 125 times shorter than 1/2
- 1/2000 is 8 times shorter than 1/250;
Given that, which one, 1/2 or 1/2000 is closer to 1/250?
Neat Image does not use the file name of a profile nor its device mode comments to do matching profile-image. Instead, it uses the EXIF values.
When you build a new profile, the EXIF values are received from the image you analyze.
When you match profiles against a specific input image, the above profile EXIF values are compared with EXIF values of the input image.
So, to have Neat Image distinguish profiles for images with and without camera built-in noise reduction, you have to build two profiles from two test images that were shot with and without noise reduction.
After that, Neat Image will automatically select the right profile from these two based on exposure time. No additional manual tweaking of device mode comments or file names are necessary.
Please try that.
Also, please make sure the profile matching priorities are set according to your needs in Neat Image Options. Try to increase the priority of exposure.
Vlad
P.S. Thinking about the exposure:
- 1/250 is 125 times shorter than 1/2
- 1/2000 is 8 times shorter than 1/250;
Given that, which one, 1/2 or 1/2000 is closer to 1/250?
Vlad,
Many thanks for your quick reply!
I have tried what you describe (two version of the images, with and without NR, and to change of priorities). The problem is what you describe, 1/2000 vs 1/2 sec, but also that there is no EXIF information in the images that in-camera NR have been used, at least not what I can find. So NI just have to select the closest one in exposure time but NI has no idea where the exact break point is (between 1/25 and 1/30) I guess?
My idea is, I hope I am no completely wrong; what I need to do is to make new images of the calibration target at all ISO setting at 1/30 and at 1/25 sec, respectively. That is relatively easy at 1/30 sec using the flash as it is just to prolong the opening time of the shutter, but it should not make any big difference from the 1/2000 images I think as it is the flash that should give most of the light anyway. I feel much more reluctant to use the flash with the long exposures, in combination with the in-camera NR, as I am afraid that will be too much cheating. But, without the flash it is much more difficult to make equivalent exposures at all ISO settings (64-800) but all using 1/25 sec...
(The nice thing with the flash is that I can receive very equivalent exposed images for the different ISO values just by stepping up or down the power of my flash, a Nikon SB-28.)
An alternative, as it looks to me, would be to edit/tweak the EXIF data of my original images and then produce new profiles, or even better edit/tweak the EXIF data in the already prepared profiles. I am, however, not sure which program to use for such a procedure.
Another alternative would of course be to make it manually, put the profiles into different folders, NR and non-NR, and change the search patch for images taken with the in-camera NR, but it would be a pity. I really like to have a fully automatic selection of profiles!
Best wishes
Anders
Many thanks for your quick reply!
I have tried what you describe (two version of the images, with and without NR, and to change of priorities). The problem is what you describe, 1/2000 vs 1/2 sec, but also that there is no EXIF information in the images that in-camera NR have been used, at least not what I can find. So NI just have to select the closest one in exposure time but NI has no idea where the exact break point is (between 1/25 and 1/30) I guess?
My idea is, I hope I am no completely wrong; what I need to do is to make new images of the calibration target at all ISO setting at 1/30 and at 1/25 sec, respectively. That is relatively easy at 1/30 sec using the flash as it is just to prolong the opening time of the shutter, but it should not make any big difference from the 1/2000 images I think as it is the flash that should give most of the light anyway. I feel much more reluctant to use the flash with the long exposures, in combination with the in-camera NR, as I am afraid that will be too much cheating. But, without the flash it is much more difficult to make equivalent exposures at all ISO settings (64-800) but all using 1/25 sec...
(The nice thing with the flash is that I can receive very equivalent exposed images for the different ISO values just by stepping up or down the power of my flash, a Nikon SB-28.)
An alternative, as it looks to me, would be to edit/tweak the EXIF data of my original images and then produce new profiles, or even better edit/tweak the EXIF data in the already prepared profiles. I am, however, not sure which program to use for such a procedure.
Another alternative would of course be to make it manually, put the profiles into different folders, NR and non-NR, and change the search patch for images taken with the in-camera NR, but it would be a pity. I really like to have a fully automatic selection of profiles!
Best wishes
Anders
Vlad,
OK, I will try that.
I thought I found the way to do what I wanted to do. I made two new images with 1/25 and 1/30 sec exposure time at each ISO value with the purpose of just using their EXIF data. I then opened those images and used the command "Profile/Copy EXIF Data to Profile" to the corresponding profile. Unfortunately that did not work as "Profile/Profile Viewer" shows that the old EXIF data from the original calibration image are retained. The only thing that change are the "Device Mode:" data.
Anders
OK, I will try that.
I thought I found the way to do what I wanted to do. I made two new images with 1/25 and 1/30 sec exposure time at each ISO value with the purpose of just using their EXIF data. I then opened those images and used the command "Profile/Copy EXIF Data to Profile" to the corresponding profile. Unfortunately that did not work as "Profile/Profile Viewer" shows that the old EXIF data from the original calibration image are retained. The only thing that change are the "Device Mode:" data.
Anders