Photographing nymphs

[stevekale]

I'm interested in entomology as it relates to fly fishing and that's how I found my way to this forum some time ago. Over the last couple of days I have been photographing some nymphs I collected from the River Frome in Dorset while fishing on Friday. (I had to rent a Canon camera in order to prove to Canon that an issue I have been experiencing with my own camera wasn't caused by my Stackshot but that's a long story...)

I'm processing the images now but during the shooting I experienced a couple of issues that I wanted to ask about.

First, let me explain the basic technique I used which was rather simple. The nymphs were backlit with a slide viewing light box. (Unfortunately I couldn't control the intensity of this and so the background isn't fully white.) I had two flash units on either side with linear polarising film over the end of the snoots (both oriented the same way). The camera and lens were suspended in a vertical orientation, looking down at the light box. A circularising polariser was on the end of a Canon f2.8L 100mm macro lens and a full set of Kenko extension tubes. So about 1.7x (?) magnification. I place a small "pool" of water on the light box, simply spooning it on, and placed the deceased nymph within it. I could then orient the nymph to suit and position legs etc.

My first question is whether there is a better fluid to be using rather than water? I'm thinking here about two factors - refraction/optical clarity and while as I noted above that I could position legs etc occasionally they would drift back towards their former orientation.

Here's an image of what is probably a Heptagenia sulphurea - it seemed to work ok given the inherent lack of rigidity in my setup.




An actual pixels version can be found here.

The second question relates to cross polarisation and an effect I witnessed doing these shots. I found that by rotating the circular polariser I could eliminate almost all specular reflection except for small bits that would turn purple. In most cases this wasn't too annoying but in others it was a bit more so. Here's an example from a set that I intend to reshoot in any case. A true Mayfly (Ephemera) nymph. The highlight is on the head of the nymph. What is causing this?



PS: With Zerene, is there a way to rename the output images after they've been prepared? If one does a DMAP and PMAP and then a few "stack selected files only" it becomes rather easy to forget what's in the latter, especially if a project is saved and returned to a day or so later.

[Craig Gerard]

An actual pixels version can be found here.

Hmm - what's wrong with the code immediately above?

Steve, it's the quotation marks "" at either end of the URL, remove those and you are good to go.

Here.



Craig

[stevekale]

Thanks Craig. Every forum seems to use different HTML! Corrected above now.

[rjlittlefield]

My first question is whether there is a better fluid to be using rather than water? I'm thinking here about two factors - refraction/optical clarity and while as I noted above that I could position legs etc occasionally they would drift back towards their former orientation.

You might try 100% isopropyl alcohol. In my area it is easy to get in pharmacies. With the specimens I have tried it is very good at getting rid of bubbles. Alcohol may also stiffen specimens by pulling water out of the tissues, but I have no idea what that will do with mayfly nymphs.

Other than bubbles and meniscus I would not expect much difference in clarity between various fluids. A thin layer of fluid is better than thick, as long as the surfaces are equally flat.

At the comparatively low magnification and small NA of the setup you're using here, I wouldn't expect any degradation from a few mm of fluid. The full size image you have posted looks great at actual pixels.

The second question relates to cross polarisation and an effect I witnessed doing these shots. I found that by rotating the circular polariser I could eliminate almost all specular reflection except for small bits that would turn purple. In most cases this wasn't too annoying but in others it was a bit more so. Here's an example from a set that I intend to reshoot in any case. A true Mayfly (Ephemera) nymph. The highlight is on the head of the nymph. What is causing this?

Have you tried crossing your polarizers and looking directly through them? Many photo-grade polarizers develop a strong blue color when crossed. If that's the case with yours, then you'll also get a blue cast on what would be very bright reflections without the polarizers. If the polarizers are neutral when crossed, then I guess the blue must be due to some surface characteristic of the subject, but I don't know what that would be.

With Zerene, is there a way to rename the output images after they've been prepared? If one does a DMAP and PMAP and then a few "stack selected files only" it becomes rather easy to forget what's in the latter, especially if a project is saved and returned to a day or so later.

The GUI doesn't provide any way to change output image names after they have been created. For the case that you've mentioned, I use an output naming template that automatically incorporates the method name and source frame numbers into the output image names at the time they are created. See Options > Preferences > Image Saving, the tags named method, mininp, and maxinp.

--Rik

[stevekale]

Have you tried crossing your polarizers and looking directly through them? Many photo-grade polarizers develop a strong blue color when crossed. If that's the case with yours, then you'll also get a blue cast on what would be very bright reflections without the polarizers.

That was it. Hmm...I am going to have to rethink my white balance also. The circular polariser is a very good quality one - B&W, F-Pro - but clearly not intended to be crossed. (The flash polarisers are sections of Techspec film.)

The GUI doesn't provide any way to change output image names after they have been created. For the case that you've mentioned, I use an output naming template that automatically incorporates the method name and source frame numbers into the output image names at the time they are created. See Options > Preferences > Image Saving, the tags named method, mininp, and maxinp.

--Rik

Thanks!

[stevekale]

I don't suppose there is a way to align frames based on a section of the frame? That would help greatly with managing movement issues. For example, when 'stacking selected' for just a leg I don't care about the alignment of the rest of the image. Sounds like wishful thinking....

[rjlittlefield]

I don't suppose there is a way to align frames based on a section of the frame? That would help greatly with managing movement issues. For example, when 'stacking selected' for just a leg I don't care about the alignment of the rest of the image. Sounds like wishful thinking....

Sorry, no such capability.

--Rik

[stevekale]

Sorry, no such capability.

--Rik

It would be a cracking addition though

The GUI doesn't provide any way to change output image names after they have been created. For the case that you've mentioned, I use an output naming template that automatically incorporates the method name and source frame numbers into the output image names at the time they are created. See Options > Preferences > Image Saving, the tags named method, mininp, and maxinp.

I guess I don't yet think in terms of frame numbers, just in parts of the image (e.g. leg)

[rjlittlefield]

I guess I don't yet think in terms of frame numbers, just in parts of the image (e.g. leg)

Renaming after image creation is "on the list". But with some planning ahead, you can use the template field to do that too. Just set the template to be say "LeftFrontLeg" before running that Stack Selected, and that's what the image will end up getting named. Of course if you forget to change it before the next image, that right rear leg will get the same name!

--Rik

[ChrisR]

I don't suppose there is a way to align frames based on a section of the frame? That would help greatly with managing movement issues. For example, when 'stacking selected' for just a leg I don't care about the alignment of the rest of the image. Sounds like wishful thinking....

Take them all (the sub-stack) into Photoshop and really blur the bits you aren't interested in (use a macro or upper layer of fog), save them out then run a new stack on those. It won't quite align with the rest of course, but it never could have, in the sense that the rest of the stack is pre-alighned. So you'll have to 'shop it in to the final image separately.
Have you tried running a stack with scaling and rotation turned off?

[stevekale]

Thanks for the suggestion. I've not yet tried scaling and rotation off although I am not sure how that would help. The current alignment method works well if the entire image is moving but obviously doesn't like it if one part of the image moves for whatever reason. If we could select a subset of the frames and tell ZS to ignore the rest when aligning that would be an awesome feature. I'd rank that way ahead of the ability to rename output files.

[stevekale]

BTW here is another example of what I've been doing lately. Nothing near as fancy as what you guys do! I had trouble with the two top, rear legs, tail and left antenna moving while capturing the frames. The images aren't as sharp as I would like but I'll get there.



(This guy had lost his uppermost tail in a scrap in the collection jar and so I did a quick fix in PS.)

[stevekale]

When I select

{method} {mininp:4}-{maxinp}

from the list in preferences for output image names I get a filename with (e.g.) DMap, then the last 5 digits in the first filename, - , then the last 5 digits in the 9th filename.

So if my filenames are Nymph-2-x.tif where x goes from 1 to 47, my output names are:

PMAX h-2-1-h-2-9
DMAP h-2-1-h-2-9

when aligning and stacking all 47....

[stevekale]

Absolutely LOVING DMap! I've not had reason to use PMax or any part of a PMax output yet. Bubbles moving around the frame? DMap fixes that. Need a clean background? Run DMap at 98%, obliterating any annoying stuff, use as base for editing then paint in main image from conventional DMap. Great stuff Rik.

[rjlittlefield]

The general form of the name is correct. The :4 says that at most 4 constant characters should be truncated from the front of the file names. So, "Nymph-2-1" and chop off the first 4 characters gives "h-2-1".

The reason you're getting "h-2-9" as an ending value is because the minimum and maximum values are identified using a simple string-based comparison, with no attempt to parse variable length numbers. "h-2-9" comes after "h-2-47", so "h-2-9" is what gets chosen.

Generate your file names with leading zeroes so that they're all the same length, and everything will be fine. In that case the comparison will be that "47" comes after "09".

I have updated the documentation to clarify this point.

--Rik