My setup was Canon 80D with the 100mm F2.8L macro lens on a monopod (Manfroto MPMXPROC4). One problem I have found with this is that the lens hood for this lens (Canon ET-73) is very deep (80mm) and, given that the distance from the front of the lens to the subject at 1:1 is only about 90mm, you end up almost touching the subject with it. In full sunshine like yesterday, a lens hood is fairly important, so I bought a VHBW-tec copy of the ET-73 from eBay (since the Canon branded item is ridiculously expensive!) and cut it down to 40mm. I find this modified hood is still sufficient to avoid flare, but provides a more comfortable working distance.
Even when it is this warm, so that they are very active and alert, Sympetrum are reasonably approachable. The knack is to move slowly and steadily and avoid any sudden movements, especially when raising your hand to the shutter release button - since this is a movement made rather close to the dragonfly. One of the things I like to do is to try and get some real close ups, usually of the head. I wanted a high shutter speed and a small aperture to get the necessary depth of field, so I set it to manual mode, dialled in F11 and 1/400s and set the ISO to auto. With plenty of patience, I managed a number of shots like this:
 |
| Sympetrum striolatum, Canon 80D, 100mm F2,8L Macro, F11, 1/400s, ISO 1250 |
I quite like this one because the wing venation is sufficiently out of focus and the body mostly shadowed so that attention is focused very much on the head - and the depth of field is sufficient for the whole head to be in focus. It also shows the striped legs rather nicely - this is where the name "striolatum" comes from!
The other thing I wanted to experiment with was trying to get sufficient depth of field to get the whole insect in focus. Now S. striolatum averages around 40mm long with a wing length of around 28mm. My 80D has an APS-C sensor which measures 22.3 x 14.9mm. So, to fill the frame with a side view of the insect, I will need a reproduction ratio somewhere in the range of half life size (1:2) or a little less - which will require me to be about 30cm away from it. Looking at the depth of field tables, even at its minimum aperture of F32, the 100mm macro will deliver a depth of field of a bit under 10mm at these settings. So I am not going to be able to get even the nearer wing tip in focus as well as the body.
The obvious way around this is to take a sequence of images at different focal points and stack them. There are two ways to do this: either change the focus between shots or keep the focus fixed and move the camera relative to the subject.
Manually changing the focus in controlled steps is difficult. It is possible to automate it using the lens focus motor, for example, using Magic Lantern addon firmware for Canon cameras. Unfortunately, there isn't currently a version of Magic Lantern for the 80D. It seems that porting it to the Digic6 processor has proved difficult and the problems have not yet been ironed out. I have successfully used Magic Lantern with my 60D and I have used this method to capture stacks of fungi - so perhaps I will do another post about that.
Moving the camera backwards and forwards is more straightforward, so I decided to try a simple way of doing this - which is rocking backwards and forward whilst taking burst of pictures using the camera's high speed shutter mode. The maximum burst rate for the 80D is 7fps. Clearly, since the camera is going to be moving, this requires a high shutter speed to avoid movement blur, so I set it to 1/1000s. Since the depth of field is going to be handled by stacking, you can afford to open up the aperture to get the ISO down. I found that F5.6 would give me an ISO of around 1250-1600. The 80D's high ISO performance is very good so I was quite happy with this, but the depth of field of each individual shot would be a bit under 2mm. One thing I should have done was to switch to capturing JPEG images (rather than raw CR2). It takes much longer to store raw images and the camera's buffer fills up at 24 shots, after which the frame rate drops precipitously. This doesn't occur if you capture only JPEGs. The processor can keep up and you can keep shooting until your SD card fills up! Anyway, I found that sequences of 24 shots were sufficient for this purpose.
 |
| Sympetrum striolatum from a stack of 16 images, Canon 80D, 100mm F2.8L, F5.6, 1/1000s, ISO 1600, stacked using Zerene Stacker 1.04 |
This images is produced from 16 shots stacked using Zerene Stacker. I didn't start quite far enough back and the wind tip is not quite fully covered and there is possibly a bit of movement in the wings. I also stopped just too soon and the bristle along the far, rear leg are not quite sharp. However, these are living beasts and they don't keep still - even if they appear to sit still for quite long periods. In particular, they are constantly moving their heads and track every movement you make - like rocking back and forwards near them! Here is an animation of another stack in which the wretched beast decided to clean its eyes with its front legs just as I pressed the shutter button! It is animated at 4fps so, since it was shot at 7fps, it is roughly half real speed.
It is interesting to note how much the subject moves around in the frame. Since I have the camera on a monopod and am moving towards it, the fact that it rises up across the frame is not unexpected due to the pivoting motion of the monpod. But the forward and backward swaying shows I didn't get my forward movement very smooth! The degree to which Zerene managed to correct for these shifts and accurately align the images in the above stack (which suffered similar subject movement!) is quite impressive.
