StackShot

LSRasmussen · Post by **LSRasmussen** » Sun Mar 28, 2010 8:20 am

Hi All,

This week I received my StackShot from Cognisys Inc. It took a long time to get to Denmark, mainly due to our Customs who wanted their pound of flesh ! I ordered my StackShot with an extra steppermotor as I had a plan to use the controller with my Olympus BHS microscope.
In my view, the stackshot kit is very good quality. The few tests I've made so far certainly indicates that this is a solution that will change my attitude to deep stacks !

I wanted to show you the microscope adaptation (beta !!) and the first result. Please note I'm in the proof of concept phase, so be gentle !

Olympus BHS with StackShot controller - 4x Olympus SPlan - NFK 2.5x projection eyepiece - Nikon D300 + R1C1 Flash kit - 20 images stacked with Zerene

The final stepper mount will be removable so that it works as an attachment.

Is anybody else working with StepShot ?

//Lars

bklein · Post by **bklein** » Sun Mar 28, 2010 9:32 am

How much did they charge for the stepper motor?
(my Stackshot still in its box...)
How did you make up the attachment to the scope knob?

Barry

LSRasmussen · Post by **LSRasmussen** » Sun Mar 28, 2010 10:03 am

Barry,

I paid 50 USD for the extra steppermotor. It comes with the same connector as the one on the StackShot, so it's easy to switch between the setups. Considering the prices here in Denmark and the convenience, I do not think this is too expensive. Regarding the attachment to the BHS, it's made up of bits and pieces I had lying around. The gray discs you can see are actually circular magnets and my reasoning was to attach one to the knob and the other to the motor. In this way it would be easy to snap the stepper motor assembly to the BHS. At the same time they would form a connection that could slip in case something got stuck. However when looking at the fine control knob, I decided to use a plastic tube between the magnet and the knob. It attaches to the knob (friction) without me having to make changes to the BHS. I might rethink this in the final assembly solution.

The whole idea of having a portable controller that works without a PC is in my view great. It means it can be used in various configurations.

BTW, I forgot to mention that the last image shows the tip of a used ballpoint pen.

//Lars

bklein · Post by **bklein** » Sun Mar 28, 2010 7:31 pm

That's a great idea using the magnets. Glad I asked.
How much time did you spend overall taking the pictures and using the stacking software? This is not usually "divulged"...

Barry

LSRasmussen · Post by **LSRasmussen** » Sun Mar 28, 2010 11:55 pm

For the ballpoint pen I used the following parameters in StackShot :
#Pics:1, Tsettle: 4Sec, Toff:500ms, Tpulse:300ms, Tramp:500ms

The Tsettle value was picked so that it ensured that the R1C1 recharged between shots. The Tpulse had to be reduced from 500ms as on my Nikon it sometimes fired the shutter twice. So, 20 images with 4800ms between each image, that's 96 seconds shooting time. Watching this happen automatically for the first time is quite a treat !

I shoot RAW, so I use Lightroom to perform the JPG conversion. Takes maybe 2 to 3 minutes.

I did not time the stacking in Zerene, but on my PC it's less than 2 minutes for 20 images using the PMax algorithm.

In this case I did nothing to try and reduce any vibration. One should really use mirror up by specifying #Pics:2. I'll keep that in mind for next time.

//Lars

rjlittlefield · Post by **rjlittlefield** » Mon Mar 29, 2010 12:04 am

For another reference point / viewpoint, see "A single maple flower" and the accompanying technical description.

The thread was pre-Zerene (Mar 11, 2007), but that doesn't really change anything.

The important point was this snippet in the last post of the image thread:

Turns out to have been just under 10 minutes from first frame in the stack to being done with Helicon Focus. Actually shooting the stack was only 88 seconds for all 30 frames. (They go pretty fast at low magnification where there's plenty of light, large steps turning the screw, and no worries about vibration.) So that's the stacking cost -- 10 minutes.

In contrast, the total time futzing around with specimen prep, lighting, trial images, more specimen prep, wrestling with positioning, still more specimen prep (yes, there's a pattern here...), then post-processing and eventually posting, was more like 3 hours.

If I were faster at all the other stuff, maybe I'd be less fond of stacking.

--Rik

LSRasmussen · Post by **LSRasmussen** » Thu Apr 01, 2010 8:48 am

I've just finalised another test shot with this setup. I had an old butterfly wing (Inachis Io) and decided to try out a scales shot. 30 images with 3.5 micron steps, Mirror Up to reduce vibration and SB800 on manual for lighting. Shutter time was 1/100 of a second. Optics were Olympus 10x SPlan and 2.5x NFK Projection Eyepiece.

So, please have a look at this image & 100% crop.

Is it me or is this image "soft" ? If so, can anybody offer an opinion on why ? I'm thinking about 4 different sources :

1 : Optics - Maybe this is what I can expect from the 10x SPlan and 2.5x NFK combo. Do I need one of Nikons CF objectives so that I can ditch the NFK ? Should I maybe use my Luminars (minus the NFK) instead ?

2 : Diffuser - I'm using a version of the pingpong diffuser. Maybe the light is to soft.

3 : Vibration - I did use Mirror Up with a 1.5 sec pause before firing the shutter. Also between each image I have a 6.5 sec pause in order to make sure that the flash has recharged.

4: 3.5 micron steps - Maybe the step between each image was to large.
I should be able to reduce this further to about 1 - 2 microns if needed.

Any comments more than welcome !

//Lars

rjlittlefield · Post by **rjlittlefield** » Thu Apr 01, 2010 10:45 am

LSRasmussen wrote:Is it me or is this image "soft" ? If so, can anybody offer an opinion on why ?

Shots at this magnification always look soft. One reason is that diffraction reduces the contrast of fine detail even though it is still within the Rayleigh limit. For presentation, my usual practice is to shoot at lowest possible ISO, then apply strong sharpening to bring out detail that would otherwise be hard to see. This can be justified in terms of pushing up the end of the MTF curve so that it is more flat for fine detail, though of course it drops sharply as the absolute limit is approached. I think many other pictures you see posted have been treated in a similar way.

The amount of detail you are showing in the crop looks good.

1 : Optics - Maybe this is what I can expect from the 10x SPlan and 2.5x NFK combo. Do I need one of Nikons CF objectives so that I can ditch the NFK ? Should I maybe use my Luminars (minus the NFK) instead ?

I have not used that combo, but I vaguely recall it is one that Charles Krebs has used successfully. The Nikon CF's are superb. I have tested Luminar 16 mm f/2.5 head-to-head against 10X objectives at NA 0.25 and 0.30; the Luminar is not as sharp. See HERE.

2 : Diffuser - I'm using a version of the pingpong diffuser. Maybe the light is to soft.

This could easily be. Pingpong diffuser plus flash gives very flat lighting because the flash illuminates one whole side of the diffuser. To get less flat lighting, some of us use dual head fiber optic illuminators. As Charles explained a long time ago, this provides great flexibility in controlling the light. At one extreme, you can place bare fiber ends right against the ball to light up an area only a few mm wide.

3 : Vibration - I did use Mirror Up with a 1.5 sec pause before firing the shutter. Also between each image I have a 6.5 sec pause in order to make sure that the flash has recharged.

With flash used like this, vibration is almost certainly not a problem.

4: 3.5 micron steps - Maybe the step between each image was to large.
I should be able to reduce this further to about 1 - 2 microns if needed.

The gain would be very small if any. Having steps that are too large will produce focus banding, but will not degrade sharpness of the best parts. I do not see any focus banding in this image.

--Rik

Charles Krebs · Post by **Charles Krebs** » Thu Apr 01, 2010 11:27 am

Lars,

Two points. Yes, I think the lighting is too flat.

The second is more involved, and has to do with some necessary "compromises". The microscope objective is forming a 20mm diameter image in the eyepiece and trinocular tube. With a 2.5X NFK, this image is enlarged to 50mm diameter at the camera sensor. But the D300 sensor has a diagonal of about 28mm. So in effect you are taking a pretty significant "crop" out of the image formed by the objective. When, at these magnifications, you are already operating at the edge of the diffraction "cliff", this will have an effect on apparent "sharpness". As you know, the 2.5X NFK was primarily meant to be used with the 24x36mm format, which has a diagonal of 43.3mm. Unfortunately, the 1.67X NFK, which provides a much better "fit" for the D300 (and similar sensor sized DSLR's) is very hard to find, and very expensive.

I sometimes try to think of it in terms of... what magnification am I getting, and at what NA? In this case you are getting a 25X magnification, from an optic with a 0.30 NA. If, for example you put a 20/0.40 M Plan on a bellows, you would have a 20X magnification at 0.40 NA. When you look at some "numbers" the effect becomes obvious.

(Note... these numbers involve some simplifications, and "best case" scenarios, but they are usually a pretty good indicator)
With 25X at 0.3 NA, the smallest detail that can be resolved on your subject is 28 microns. With a 20X achieved with a 0.40 NA the smallest resolvable detail would be 16.8 microns. (You could also consider this in more "photographic" terms... in the first case the "effective aperture" is about f39.4, while in the second it would be about f25. )

So a pretty good approach, in general, it that whatever the final magnification, always try to attain it with the highest possible NA. In some cases there may be an argument for a little compromise to attain more DOF, but with image "stacking" this is an issue in fewer instances.

LSRasmussen · Post by **LSRasmussen** » Fri Apr 02, 2010 2:11 am

Rik, Charles,

Thank you very much for your comments. I'll look into building a fiber box for my SB800 in order to fix the lighting issue. I normally do not apply any sharpening aside from the sharpening to screen when exporting the JPEG from Lightroom. I tend to only adjust the levels slightly in order to enhance contrast a bit. So, no doubt I'll have to experiment further with my post processing. I find it interesting that with this subject at this magnification, I'm at or near the limit as to what this lens combo can deliver. I guess I need to see if I can get an APO version of the Olympus 10x or maybe one of the Nikon CF objectives. I do have Leitz APO's on my Ortholux 1 but the working distances are very small, so they are for brightfield observation only.

Once again, thanks for your help,

Back to work,

//Lars

ChrisR · Post by **ChrisR** » Fri Apr 02, 2010 7:14 am

At Rik's suggestion I stopped using JPEGs for stacks. There's more detail coming out using TIFFs. I think I can see JPEG "blocking" in your 100% crop?
My PC is slow, and ot one point ZS seems to freeze forever thinking about 200+ TIFFs, but after that the speed is the same.

LSRasmussen · Post by **LSRasmussen** » Fri Apr 02, 2010 7:34 am

Chris,

Excellent suggestion. I will re-run the stack tonight with TIFFs to see if it makes a difference.

/Lars

Charles Krebs · Post by **Charles Krebs** » Fri Apr 02, 2010 10:28 am

I guess I need to see if I can get an APO version of the Olympus 10x

Well, it is great for diascopic use, but if you need to light from "above" your working distance will drop from 7.5mm to 0.55mm!

I normally do not apply any sharpening...
So, no doubt I'll have to experiment further with my post processing.

Absolutely! Sure, it's possible to overdo sharpening (most of us do from time to time) but done carefully (along with other "corrections") it can make a very significant difference.

LSRasmussen · Post by **LSRasmussen** » Fri Apr 02, 2010 12:13 pm

Charles,

Yes just after posting that reply I checked the working distance of the Olympus 10x APO, yikes, what a difference. I'll keep an eye out for the Nikon instead. I took Chris's advice and ran the stack as 16bit, no compression TIFF files instead. Here is the 100% crop of the same area as before :

On screen, before saving to JPEG (100% crop of the screen output from Zerene) there's no doubt that this retains more detail. Saving the file to JPEG changes the color slightly and some of the detail is lost as far as I can see. It's still better than the previous effort and I guess we cannot totally avoid JPEG's due to the upload file size limit.

I still need to experiment with more aggressive sharpening.

//Lars

Charles Krebs · Post by **Charles Krebs** » Fri Apr 02, 2010 1:01 pm

Lars,

The Nikon CF N Plan Achromat also has a NA of 0.30, so if used with the NFK 2.5X I would not expect to see any improvement in resolution. It can however, be used directly on bellows more successfully, but of course you'll not have the 25X magnification.

The Nikon CF N Plan Apo has a NA of 0.45 and a working distance of 2.75mm. So you just might be able to squeeze a little light in there (wouldn't be fun

) and the resolution should be noticeably better than with a 0.30 NA objective. I don't know how large a problem the correction in the NFK (needed for the Olympus, not needed for the Nikon) would be.