Hairy areas Zerene stacking problem

A forum to ask questions, post setups, and generally discuss anything having to do with photomacrography and photomicroscopy.

Moderators: rjlittlefield, ChrisR, Chris S., Pau

pawelfoto
Posts: 90
Joined: Tue Mar 09, 2021 2:51 pm
Location: Poland

Hairy areas Zerene stacking problem

Post by pawelfoto »

Hi, I'm new to the community and just learning focus stacking. This is one of my first stacks with Nikon 10x CFI + Raynox150, 270 frames in 9um step lenghts, Zerene Pmax, partialy Dmap. I am asking experienced colleagues for advice:
1. What would you improve in this photo?
2. What is the solution to problem areas like crossed hairs under the chin. Please send me your workflow. Can it be improved in Zerene, for example "Stack Selected", or with special retouching brushes? or will I have to work hard at PS with the use of a clone stamp? I am asking for advice and links to places where it is described.
2. I am convinced that this topic was explained many times on the forum. What's the best way to find a solution yourself? what should i type in the search box to refine my search?
Full frame: https://drive.google.com/file/d/1bQEnPI ... sp=sharing
best regards, Pawel from Wroclaw, Poland
Attachments
the tongue of the fly.jpg

Online
Adalbert
Posts: 2426
Joined: Mon Nov 30, 2015 1:09 pm

Re: Hairy areas Zerene stacking problem

Post by Adalbert »

Hi Paweł,
will I have to work hard at PS
yes, if you want to achieve good results, you will have to work hard :-(

Some workflows:
https://www.photomacrography.net/forum/ ... 27&t=43340

The crossed hairs can be retouched e.g. by Zerene (using sub-stacks).

Best,
ADi

rjlittlefield
Site Admin
Posts: 23561
Joined: Tue Aug 01, 2006 8:34 am
Location: Richland, Washington State, USA
Contact:

Re: Hairy areas Zerene stacking problem

Post by rjlittlefield »

For Zerene Stacker, a good place to start is with documentation on the main Tutorials index page, at https://zerenesystems.com/cms/stacker/docs/tutorials/tutorialsindex . I always recommend to study at least the first 7 bullet points, down to and including "Tips for Retouching", including the two video tutorials on retouching. In your case, it would be good to add "Using Stack Selected to Retouch Transparent Foreground", which is linked farther down on the page.

With very hairy subjects, the process called "slabbing" can be helpful. It is another use of Stack Selected, which simply splits the stack into overlapping sections, called "slabs", each containing the same number of frames, typically 10-20 frames with an overlap of 30% or so. Stack Selected is then run on each slab, producing a set of intermediate outputs that can be used either as source for retouching, or as source for subsequent stacking operations such as DMap with varying radii.

I wrote Zerene Stacker, and I agree with Adi. If you want flawless results, you will have to work hard. In general, it is best to retouch inside Zerene Stacker to handle visibility issues like crossed hairs. This is because retouching inside Zerene Stacker uses the same adjustments for image alignment that were used for stacking, so cloning is automatically done from the exactly matching area in the source frame. However, Zerene Stacker does not allow for cloning from arbitrary locations, so if you want to clone out dust or other defects that were on the subject, then you'll have to use Photoshop or another similar tool for that.

Switching subjects, I am intrigued by a bit of anatomy in your image. At several places in the eye, there are elliptical paths where the facets of the eye come together in odd ways, like dislocations in a crystal lattice:
RingInEye.jpg

Do those dislocations exist in the source images, or did something go wrong in the stacking process to create them in the stacked output images? Or are they an unusual form of moiré patterns, resulting from resizing to fit the forum limits?

--Rik

pawelfoto
Posts: 90
Joined: Tue Mar 09, 2021 2:51 pm
Location: Poland

Re: Hairy areas Zerene stackig problem

Post by pawelfoto »

hi Rik, it's an honor for me to hear advice from the master himself. I'm your big fan. I admire your knowledge, achievements and great kindness. As I can see on the forum, no question remains unanswered. Just yesterday I watched Live-chat (2018) with you and Janice Sullivan. So I had the chance to even see you! I also heard a lot of nice words about you on the Allan Walls Photography channel. As for the eyes, I also saw these irregularities. I checked the source images are fine, no distortion on the eyes. So I will try the slabbing method. I know there is still a lot of learning ahead of me. Thanks a lot, Pawel

rjlittlefield
Site Admin
Posts: 23561
Joined: Tue Aug 01, 2006 8:34 am
Location: Richland, Washington State, USA
Contact:

Re: Hairy areas Zerene stackig problem

Post by rjlittlefield »

pawelfoto wrote:
Mon Apr 05, 2021 9:56 am
I checked the source images are fine, no distortion on the eyes.
Thank you for the kind words. I am not on social media very much, so it's good to hear that I am well thought of.

Now, about your stack...

Because the source images are fine, and the pattern looks like it could be on a depth contour, I think there may be some problem with alignment between the images on either side of the dislocation.

To confirm that, you can put a checkmark on "Show as adjusted" in the Input Files panel, then click back and forth between the source images. With a well aligned stack, flashing between two frames that are shown as adjusted will show nothing changing except for focus. But in this case I expect you will see some lateral shift also. If so, then the question becomes "What is causing the shift?"

Shifts like this can result from totally unexpected causes, for example a passing shadow that causes distribution of brightness to change slightly from one corner to another. This is because the alignment algorithm looks at average difference of pixel values across the entire frame, including out-of-focus regions.

I will be interested to hear what it turns out to be in this case, if you can figure it out.

--Rik

pawelfoto
Posts: 90
Joined: Tue Mar 09, 2021 2:51 pm
Location: Poland

Re: Hairy areas Zerene stackig problem

Post by pawelfoto »

I tested everything thoroughly. You were right - the problem is in alignment. If I watch files before alignment, only the focus changes smoothly. After Zerene alignment, a shift between the two frames is visible. If you want to test yourself, here is a link to download a few files concerning problematic layers. https://drive.google.com/drive/folders/ ... sp=sharing
I dealt with this by making the crop and stack of the eye itself separately. Output was OK, then I put it on the "damaged" eye in PS. I am falling in love with Zerene more and more. Yesterday I extended the version to Prosumer and I excercise slabbing.
==Best, Pawel

rjlittlefield
Site Admin
Posts: 23561
Joined: Tue Aug 01, 2006 8:34 am
Location: Richland, Washington State, USA
Contact:

Re: Hairy areas Zerene stackig problem

Post by rjlittlefield »

Thank you for the sample images. In the Google Drive folder, I see files named 2A0A8279.tif through 2A0A8283.tif, which I will call 279 through 283. I wonder if these are the frames you intended to upload, because they do not show any of the eye in focus. However, they do have sharp detail in some of the bright portion of the face near center of frame, and that area does show some interesting behavior.

I loaded these files into Zerene Stacker, processed them with Align All Frames using all default settings for Options, then scrolled through the frames both with and without "Show as adjusted" being check-marked.

What I see is that...

Without a check-mark on "Show as adjusted", there in fact is a significant lateral shift between frames 279 and 280. Then from 280 through 283, there are no lateral shifts. Brightness is consistent for frames 279-281, but then a little brighter for 282 and much darker for 283.

With a check-mark on "Show as adjusted", the original lateral shift between 279 and 280 has been removed, but a slight lateral shift has been introduced between 282 and 283.

Closer examination of frames 282 and 283 shows that the pattern of illumination shifts between those two frames. Frame 283 is overall darker, but the darkening is especially strong on the front of the face, some in the background, and not much at all in the gray area that is anatomically below the eye. So, it looks like the scene was illuminated by multiple sources, and some but not all of them lost brightness in this frame.

Here are a couple of animations that toggle between frames 282 and 283, first without brightness adjustment, then with brightness adjustment. These animations are oriented as they appeared in the original source frames, where anatomically "below the eye" is visually above-right of the eye. Notice that before adjustment, frame 283 is darker over most of its area but almost unchanged in the upper right corner as shown, while after adjustment, there is no overall difference in brightness between the frames, but that's because some parts that used to be the same got brighter, while some parts that used to be darker are still darker, just not by as much.

Without brightness adjustment, frames 282 and 283
BeforeBrightnessAdjustment.gif

With brightness adjustment, frames 282 and 283
AfterBrightnessAdjustment.gif


In any event, this shift of overall pattern of illumination is what's causing a lateral shift of details to be introduced.

So, this is a vexing stack. Even for just these 5 frames, turning off alignment altogether leaves some degradation because of the lateral shift between 279 and 280 in the original frames. Turning on shift correction fixes that problem, but introduces a new one between 282 and 283 because of the illumination shift. In the particular area that I looked closely at here, there were fewer artifacts with alignment turned on, but in the full stack it is apparently the other way around.

I'm glad to hear that you were able to get a better result by turning off alignment and combining the best bits of two approaches. However, in the long run you will be better served by getting your illumination stable so that you can leave alignment turned on and have it do the right thing.

I hope this helps!

--Rik

pawelfoto
Posts: 90
Joined: Tue Mar 09, 2021 2:51 pm
Location: Poland

Re: Hairy areas Zerene stackig problem

Post by pawelfoto »

My fault, I got the photo numbers wrong. These are the proper ones: https://drive.google.com/drive/folders/ ... sp=sharing
You can clearly see what you write about. Thank you very much for careful analysis - I'm surprised by the differences in lighting. I use speedlite canon MT-24EX with additional power supply, flash interval 6 sec, 1/32 power, so the flash should be stable. Both lights are reflected from the diffuser and directed at the specimen. Maybe I need to try a 580EX or find another. Have you ever encountered such instability in lighting? What will be the best solution?

rjlittlefield
Site Admin
Posts: 23561
Joined: Tue Aug 01, 2006 8:34 am
Location: Richland, Washington State, USA
Contact:

Re: Hairy areas Zerene stackig problem

Post by rjlittlefield »

Thank you for the new images. Yes, I agree it is the same effect -- alignment being messed up by changing pattern of illumination.
Have you ever encountered such instability in lighting?
Yes, it is a common problem. One of the examples in my "How To Use DMap" tutorial is a small moth, shot with Speedlite 580 EX II by "leaning in" while shooting continuously. The stack looks awful in DMap unless I turn on brightness correction. The awfulness in that case consists of obvious banding in the depth contours, where DMap transitions between bright and dark frames. Brightness correction makes that issue almost disappear.

The most stable illumination is continuous. I expect that pulsed LED is second best, though I have not seen any data on that. Flashes from gas discharge tubes always vary by some amount. The best are professional studio strobes, specifically designed for uniform flashes that are relatively slow -- roughly 1 ms duration, regardless of power. The worst seem to be consumer speedlights that are turned down to low power. Speedlights work by triggering the discharge and then cutting off the current a short time later, so low "power" really means short duration. Unfortunately there seems to be some random variation in either starting the flash or cutting it off. That variation turns into variable exposure, and for shorter flashes the variation is relatively more important.

I have never tested an MT-24EX for variation, so I do not really know how much is "normal". But based on experience with other flashes, I would not say that you have a "defective" unit.

You might try increasing the flash power, and absorbing any extra light with more layers of diffuser.

If you have a 580EX handy, that would be worth a try also. Bearing in mind my experience with the moth, I am not optimistic it would give much less variation in total exposure. However, the 580EX has only one head, and in this case having only one head is a big advantage because it means that whatever variation there is affects the entire frame to the same amount. Zerene Stacker is pretty good at handling that sort of overall variation, versus side-to-side variation like the MT-24 is giving you. Having only one head is less convenient to set up balanced illumination, because dividing the illumination has to be done with reflectors, diffusers, and gobos. But having the pattern of illumination be consistent may be worth the extra trouble of dealing with just one head.

--Rik

Post Reply Previous topicNext topic