Looking for some feedback on images and setup

ChrisJones · Post by **ChrisJones** » Fri Apr 05, 2013 6:08 pm

Hi,

I work on Drosophila melanogaster and sometimes need to take pictures. After being unimpressed by a 5MP Nikon DS-Fi1 CCD camera attached to a Nikon SMZ800 scope, I began looking at ways to take pictures with a DSLR instead. I decided to buy a Nikon D5200 as they are not horrifically expensive, and I have my own D3200 (and D3100) so they are familiar to me. After reading around on this forum and elsewhere, I came across three options for connecting the D5200 to the SMZ800 (which already has a beam splitter). The options were a c-mount to f-mount adapter, a DD20NLT adapter and a VariMag II adapter. I settled on the VariMag II as it seemed to me to be the most complete system.

I have had everything set up for a few weeks now and after some learning I have produced a number of practice pictures, which I’ll link below. I would be interested to hear any suggestions about how I can improve the images.

To produce the images I took several shots with different focus (typically ~3-7) at between 1/100 and 1/25 shutter speed and ISO around 640. I sorted out the exposure in PS CS6 Camera Raw then stacked, aligned and blended them in PS.

I have come across a number of issues while doing this, though none has been insurmountable, I would be interested to know how others deal with similar problems:

- The design of the beam splitter for the SMZ800 means that the field of view pans when focusing. This proved to be a problem initially, but was solved completely when I realised how easy it is to align images in PS.

- There is a lot of chromatic aberration (purple/green fringing) visible in the images taken by the camera. There is often so much that the maximum correction for this in Camera Raw is not strong enough and I am having to do it manually after them images have been blended together.

- The images seem to be quite noisy to me. I don’t know if this is par for the course or if I should be aiming for less. I didn’t apply much noise reduction in Camera Raw and I think I should have done more, but I’m always scared of losing sharpness.

- It’s hard to get enough light onto the camera sensor. I’m using a fibre optic LED light source. Set to full, it produces a light so strong I can’t look at the white CO2 pad it’s shining on, but I still have to use what feels to me like a rather high ISO and low shutter speed. Ideally I would like to diffuse the light using tracing paper.

- I have not been able to produce any nice images at higher magnification than the ones shown below. This is due to lack of light and movement when I press the shutter release.

- A DSLR screen, even rotating one, is insufficient for focusing accurately. I’m going to connect the camera to a monitor.

Anyway, here are some examples of what I have managed to produce so far:
(+ some more examples here: http://www.flickr.com/photos/jonesci/se ... 025629963/)

Drosophila melanogaster (y- B- cn-, male) by Chris_Jones., on Flickr

Drosophila melanogaster (w[1118], female) by Chris_Jones., on Flickr

Drosophila melanogaster (w[1118], male) by Chris_Jones., on Flickr

Drosophila melanogaster Serrate (Ser) phenotype by Chris_Jones., on Flickr

Drosophila melanogaster (w[1118], female, dorsal view) by Chris_Jones., on Flickr

This one is not a fly.

Biro by Chris_Jones., on Flickr

Overall I'm quite pleased with how these images have come out, they're certainly an improvement over the CCD camera, but I'm sure they can be improved further. Any suggestions?

rjlittlefield · Post by **rjlittlefield** » Fri Apr 05, 2013 8:13 pm

Chris, welcome aboard!

One of our less pleasant tasks here at photomacrography.net is explaining to people that their expensive top end stereo microscope is not a particularly good tool for photographing small subjects.

Some of the major problems you've mentioned -- not very sharp, lots of chromatic aberration, hard to adjust for focus stacking -- are very common and are essentially embedded in the design of the instrument.

It's not difficult to make large improvements in those areas, but it does require a significant change of direction. Basically you swap out the stereomicroscope and swap in different optics and a different focusing mechanism that are better suited for photography despite being pitifully bad for direct observation.

As an example of what can be done for a pretty modest investment, see the blowfly face HERE and the fruit fly face HERE. Both of these were shot through 10X objectives designed for use on compound microscopes. The fruit fly face actually was shot at 10X on sensor; the blowfly was pushed down to 5X by tweaking the lens configuration. The optics do not need to be very expensive for this approach. Both those photos could be reproduced with lenses costing a total of about $600 new.

If you're interested in pursuing that direction, then probably the best approach is to go read our FAQ: How can I hook a microscope objective to my camera? and skim through some of the focusing rigs shown in the links at FAQ: What's the best way to focus when stacking?.

In your current setup, probably the most helpful thing you could do is switch to flash illumination. Inexpensive consumer flashes, placed close to the subject and diffused by paper or plastic, will still produce plenty of light to set your camera on base ISO while also giving you a short exposure time that would solve movement issues. Those flashes would be usable with any sort of optics and focusing mechanism, so they're probably the first thing to consider.

--Rik

ChrisJones · Post by **ChrisJones** » Sat Apr 06, 2013 6:56 am

Thank you for the advice Rik, you've given me a few things to think about.

I'm tempted by the attaching an objective to a telephoto lens, but most of this is new to me and it looks like it would take me a long time to do the research and create a working set up.

I have a Nikon 70-300mm lens. I may be able to find a suitable objective in the lab, or perhaps a Nikon CFI Plan Achromat 4x would be suitable (although I haven't found a price for that as I've only had a brief look). I'd need something with a magnification low enough to be able to photograph a whole fly.

Unfortunately I have to consider whether such a set up would be something I actually need... rather than something I want. I'd also need it to be usable by students...

We have several SMZ800s in the lab and their primary uses are sorting flies and dissection. We have two beam splitters, 1x and 0.7x c-mounts and two CCD cameras, so setting one up with a DSLR was pretty easy. The images produced are not perfect, and ease of use is not ideal (more post-processing that I'd like), but at least it wasn't too complicated.

I'll certainly give the flash suggestion a go.

rjlittlefield · Post by **rjlittlefield** » Sat Apr 06, 2013 8:10 am

Chris,

Thank you for the additional information. For your situation, there are some aspects that make the SMZ800s look better.

The big one that comes to mind is "need it to be usable by students". A lot of the setups that you'll see here at photomacrography.net are designed, assembled, and used by a single individual who knows how they work and has a lot invested in keeping them in good shape. They tend to have lots of parts that are easily removed and misplaced. The simplest of objective-on-telephoto setups is not too bad: just a couple of adapter rings. But as the systems become more capable, they also tend to accumulate more parts that have to get rearranged in use. Switching between 10X and 4X objectives, for example, is often done by unscrewing one objective and screwing in another one, giving plenty of opportunity for loss or damage by inexperienced hands.

To get high quality images in a more integrated system in your environment, you might consider mounting your DSLR on a traditional compound microscope to take advantage of its sharper lenses, finer focusing mechanisms, and built-in turret. HERE are a couple of homebrew variations. I'm wondering if you already have in your lab a suitable scope that has gotten overlooked for this application because you're used to the SMZ800s for direct observation.

The other aspect is a tradeoff between sharpness and the time needed for acquisition and processing. There is an unavoidable tradeoff between sharpness and depth of field (DOF). The optics in stereo scopes are typically designed to favor larger DOF but sacrifice sharpness in the bargain. The objectives for compound scopes go the other way, favoring sharpness but sacrificing DOF. If you switch optics to get more sharpness, you'll also need to shoot more images in each of your source stacks.

One other tradeoff I should mention relates to the post-processing. I see that you're currently using Photoshop CS6 for focus stacking. That has the advantage of being integrated with other tasks such as raw conversion and CA correction, but it also can have some disadvantages in terms of missed detail. This becomes more apparent with higher quality source images and deeper stacks; see for example the discussion and illustrations HERE.

--Rik

Charles Krebs · Post by **Charles Krebs** » Sat Apr 06, 2013 12:31 pm

Rik said:

One of our less pleasant tasks here at photomacrography.net is explaining to people that their expensive top end stereo microscope is not a particularly good tool for photographing small subjects.

So true, and it seems to come up at least once or twice a year. Some info in this post might be of interest:
http://www.photomacrography.net/forum/v ... hp?t=16878

I can see why in your teaching setting a stereo is a good tool (albeit not the best for photography). If photography is the primary goal a different equipment solution will give better results. But if you want to keep the highly desirable viewing properties of a stereo, but also try to improve the images over what you are now getting there are some things to try.

I decided to buy a Nikon D5200 as they are not horrifically expensive, and I have my own D3200 (and D3100) so they are familiar to me. After reading around on this forum and elsewhere, I came across three options for connecting the D5200 to the SMZ800 (which already has a beam splitter). The options were a c-mount to f-mount adapter, a DD20NLT adapter and a VariMag II adapter. I settled on the VariMag II as it seemed to me to be the most complete system.

at between 1/100 and 1/25 shutter speed and ISO around 640

A DSLR with a mechanical shutter used between 1/25 to 1/100 mounted high on a stereo microscope is a recipe for vibration. Short duration electronic flash is likely the best solution if you are to continue to use the D5200.

Don't be tempted to increase the magnification in the camera by increasing the magnification in your VariMag adapter. One of the photographic problems with a stereo is typically the very low to modest NA (numerical aperture). Increasing the magnification into the camera via the adapter makes this problem even worse. Use the lowest adapter magnification you can while still getting an image to the corners in the camera body.

- There is a lot of chromatic aberration (purple/green fringing) visible in the images taken by the camera.

You did not mention which lens you have on the SMZ-800. Based on this quote I would venture that is is a Plan Achromat. This aspect will be greatly improved with one of the Plan Apo lenses Nikon makes for this microscope. But the cost of one of those could probably set up a complete basic "imaging station" using the methods Rik mentioned (and which are ubiquitous on this site). When I say a Plan Apo will "greatly improve" the chromatic correction I know that is the case since I have tried it. What I don't know about your set-up is whether or not the VariMag adapter is contributing to the chromatic errors you are seeing. I have never used one.

A DSLR screen, even rotating one, is insufficient for focusing accurately. I’m going to connect the camera to a monitor.

Magnified live view is very effective for focusing. But in a teaching environment I would think that an HDMI connection to a screen could be beneficial on a variety of levels. Very simple to do. I would also explore the possibility of establishing "parfocality" between the viewing eyepieces and the camera body. This way when the eyepieces are in focus the camera is as well. (This can be tricky to maintain with a multi-user environment. Once set up, the eyepiece diopter adjustment should not be moved. But... if a single user were to setup for a "session" where he/she will be doing a great deal of photography it can typically be "re-calibrated" in about 30 seconds thanks to the live-view feature).

I would forget about using Photoshop and go with a dedicated stacking program like Zerene Stacker or Helicon Focus

ChrisJones · Post by **ChrisJones** » Sat Apr 06, 2013 1:41 pm

Thanks for the additional info Rik and Charles.

Having never used anything else, I never realised how inferior the stacking is in Photoshop compared to specialised software. I may well have a look at something else. Initially I did have a look at Helicon focus and it is not expensive, but I thought "oh, I'll just do it in Photoshop".

I suspect the VariMag is responsible for some of the chromatic aberration, but I haven't used it enough to be sure.

I have the VariMag set to its lowest magnification (1.5x) along with a 1x c-mount, which has no problem covering the sensor. I also have a 0.7x c-mount and I think the image still covers the sensor with this. Do you think there would be any benefit to using this c-mount over the 1x?
I did try higher magnifications with the VariMag, but the loss of sharpness is significant. Combined with the low shutter speed, image quality is very poor.

I also never realised that you could get different lenses for the SMZ800. Might have a look into this.

As for parfocality, the VariMag has a thread to adjust the focus, but I have not been able to get it to match the focus in the eye pieces at any magnification. The CCD cameras I have are not parfocal either (they can't be adjusted though). It's not ideal, but I'm pretty used to living with this anyway.

I didn't even think of zooming the live view display to focus more accurately. I've ordered a cable to attach the camera to a monitor anyway though. An additional minor problem I've found with the D5200 live view screen is that I have to use it upside down. Although it rotates reasonably freely, the orientation of the live view image is determined by the screen's position relative to the camera body, not by an accelerometer. To have it on the side I want, the image is upside down. But a monitor connection will make this point moot.