Recommendations on flash for stacking?

rjlittlefield · Post by **rjlittlefield** » Mon Mar 26, 2012 12:55 am

Well, that's certainly suspicious. But I'm pretty sure you're looking at the histograms including the Saturation=97 adjustment layer. The ones I've been posting are without the adjustment. What do yours look like that way?

--Rik

Craig Gerard · Post by **Craig Gerard** » Mon Mar 26, 2012 1:05 am

Rik,

Yep. You caught me editing my previous post

I've added another screenshot to that post.

Craig

rjlittlefield · Post by **rjlittlefield** » Mon Mar 26, 2012 2:01 am

Those look similar to what I've gotten in this M 1/32 test. The changes in color balance seen here are pretty minor, but the fact that they're occurring at all is interesting. My gut feel is that under some circumstances the changes are a lot bigger, but I don't know what those circumstances are. All I really know is that in some stacks I've shot, there are obvious changes in color balance between adjacent frames when I've played the stack as a filmstrip.

--Rik

Craig Gerard · Post by **Craig Gerard** » Mon Mar 26, 2012 5:28 am

Rik,

I had a look at the RAW files. The K temp is consistent (identical) for all frames in the previous test.
I scrolled through each of the 15 images and cannot see any shift between frames.

A sample screenshot; image unadjusted as imported from the DSLR. PhaseOne do not use ACR.

Craig

rjlittlefield · Post by **rjlittlefield** » Mon Mar 26, 2012 8:59 am

I'm not familiar with Phase One.

Is that reported K temp a value that it determined by looking at the image, or is it just echoing the nominal K temp that it used when converting the raw?

In ACR, the reported temp would be the one used for conversion unless I specifically went to the trouble to reset gray point on each image. Without resetting the gray point, of course all the values would be the same because all the images were processed the same. The fact that you're getting identical numbers down to the 4th digit, despite clear differences in the histograms, suggests that Phase One is doing the same thing.

--Rik

ChrisLilley · Post by **ChrisLilley** » Mon Mar 26, 2012 12:39 pm

DQE wrote:Sounds like we're approaching a conclusion that one needs to dump one's Canon stuff in favor of an all-Nikon system!

I assume this is humour but just in case, way to little data to draw any such sweeping conclusion.

Craig Gerard · Post by **Craig Gerard** » Mon Mar 26, 2012 3:57 pm

Is that reported K temp a value that it determined by looking at the image, or is it just echoing the nominal K temp that it used when converting the raw?

The latter.

Looking at the images in PhaseOne, prior to any adjustments, there are some inconsistencies in the histogram when comparing images.
When I open multiple images in 'multi view'/'grid view' my eye can see subtle variations between frames resembling those concerns mentioned at the beginning of this thread.

Craig

rjlittlefield · Post by **rjlittlefield** » Fri Mar 30, 2012 12:11 am

I played with this a bit more this evening.

It seems that the flash power is important. Color variation with the 580EX II was much worse at 1/128 power than at 1/32 power. Here's a set of thumbnails at 1/128 power, as seen by a Nikon D5000 with saturation=97 added:

I shot both Canon T1i raw and Nikon D5000 raw while I was at it. Adobe Camera Raw 6.6.0.261 in Photoshop CS5 then gets gray point on the corresponding raw files as follows:

Nikon DSC_0119 5550K +14 tint (Canon IMG_5963 5550K +22 tint)
Nikon DSC_0106 5150K +04 tint (Canon IMG_5950 5200K +14 tint)

This increased variation at low power also makes sense in terms of the older situations that I've been able to check. Turns out that the ones I remember as being the worst were also low power, notably cases where I shot live subjects by setting the camera on continuous exposure and used the "lean in" technique. When doing those, I set flash on the lowest power I can get away with, figuring that then it won't run out of juice part way through the sequence.

--Rik

Craig Gerard · Post by **Craig Gerard** » Fri Mar 30, 2012 3:27 am

Rik,

Thanks for the additional report.

Kinda of makes you want to try using an Nikon SB800 speedlight.....anyone up for it? The onboard D5000 pop up flash performed admirably.

I wonder if the latest flagship Canon Speedlight 600EX-RT would return a different result.

These tests can be rather tedious; but if using E-TTL with the 580EX II does it fair any better? I may have time over the weekend to have a look; but if someone beats me to it I won't be disappointed.

Craig

DQE · Post by **DQE** » Fri Mar 30, 2012 10:09 am

I wonder if the following would help re white balance stability and/or exposure stability, while shooting with flash or other light sources:

1. Place a very small, very uniform white balance card somewhere in the field of view of the camera, just before photographing one's stack. This "micro-white balance card" should be in each frame at the same place, but it need not be in accurate focus throughout the stack AFAIK.

Alternatively, intentionally place the white card at an angle to the direction of travel of the rail. This way, some part of the card would be in focus as the camera moves, and it shouldn't matter too much that one is using a different region of the white balance card for each slice. Using a different location in the card for each slice would complicate writing the Photoshop script. In an ideal universe one would automatically move the white card so as to keep the same spot in focus as the camera rig moves through the stack, but this would probably be significant trouble and cost to keep two focus rails in sync.

Another thought is to mechanically couple the white card to the camera or sample, depending on which one moves during the stacking process. This way, the same location would always be in the same focus and same relative position vs the specimen.

It might not matter too much if the angle of the white balance card changes slightly as one goes through the stack - these things are pretty diffuse reflectors.

2. Write a Photoshop script or manually set the white balance based on sampling the card's image in each frame of the stack. In ACR of course there is a special-purpose white balance "eyedropper" tool just for this purpose.

I suppose this would mean that one would have to do this before feeding the resulting .tif or .jpg files to Zerene. It should mostly be rate-limited by the I/O speed of one's computer rig. Probably yet another reason to invest in a couple of solid state hard drives (SSDs) for photographic computing, at least in principle.

Just thinking out loud - hope it inspires someone to think of something along this line that would be useful.

Having read this thread, I now don't trust *any* of my camera equipment!