It doesn't get any better than this!

Overread · Nov 5, 2010

http://farm2.static.flickr.com/1259/5150174018_597d82a5db_o.jpg

f13, ISO 200, 1/200sec
Taken with canon 400D, MPE 65mm macro - lighting Canon Twinflash!

Yep this will be a discussion on flash diffusion - specifically with regard to macro photography. First off this is not my own original idea and for a far more in depth discussion on the specifics of the method I am emulating I recomend you have a read here:
Juza Nature Photography Forum • View topic - Concave Flash Diffuser Test

That might look complicated, but the theories make some sense when you think about them. So in the quest to lose the harsh highlights seen on insects when using flash I created the following rough design:

What you see there is incomplete at present, but is sufficing as a basic working model. You have the curved plastic caps for the diffusers secured to the ring of cardboard and held to the flash with sticky tape.
This curves the light, but the diffusion is still incomplete as the actual area of the light source is not made much larger -so we add the curve of paper. This is currently two sheets of regular plain white paper held together and just stick with a few tabs of tap to the bracket holding the flash heads.

Further if you look at the heads themselves you'll see that one is slightly higher than the other - using a kaiser flash mount. This is something that is important when using these flash heads as they make the face of the unit longer and so added height helps to distance them from the setup. For a 1:1 shot its not that bad (as seen above I've only one kaiser at present) however by the time you are at 5:1 things get a lot harder to work with because of the short working distance.*

As you can see even with an incomplete setup (no Vellum paper for diffusion - no breakup of the two layers of paper and no silver foil on the inside of the diffuser sides to help reduce light loss) the amount and nature of the diffusion has certainly reduced the effect of the harshness of the light and you can see for yourself that the shell of the insect does not have the bright highlight marks - there are a few spots, but I hope that slight enhancements to the main method will eliminate these as well.

I'll be interested to hear others thoughts on this and other flash diffusion methods with the twinflash unit as well as other lighting setups of course. Note that this method can be well used with a single speedlite flash and one curved cone of light onto the paper diffuser (infact currently my single 580EX2 tests area little bitter than this at present).

* note that this also requires a different size of paper to be used since the paper sized for 1:1 needs to be physically longer, whlist for 5:1 it has to be shorter.

Garbz · Nov 6, 2010

Why the snoots? In macro would any light not falling on the diffuser possibly be in the picture?

Overread · Nov 6, 2010

The short snoots are so that the light from the flash gets channelled into the curved diffusers - the idea (very simply a better explanation is in the linked thread in my first post) is that a macro photo often has the subject rather near the middle and that subject is 3D in shape. Conventional flash setups have the light coming from a flat diffusion surface, so you get a flat wall of light falling onto the curved 3D shape - the result being that those points closer to the diffuser get far more light than the curved edges of the subject. Thus leading to the often blown highlights you get along the edges and curves of the shell.
So the idea of the concave diffuser is to change the distances and break the light up even more from a curved shape that mimics the curve of the subject so that the light falls more evenly - even after it passes through the small light tent diffusion layer.

As for ambient light most of the time with settings like f13, ISO 100, 1/200sec almost all the light is coming from the flash units, with only the strongest of natural light creeping into the shot. For more natural light you either have to raise the ISO, use a slower shutter speed or sacrifice the depth of field and widen the aperture.

Garbz · Nov 7, 2010

Yeah I get the diffusers, but still not the snoots? Why do you need to specifically channel the light onto the diffuser? Assuming you're going to be taking a photo of something very small and close to the lens I don't see why spill light from the flashes would need to be controlled.

Overread · Nov 7, 2010

Essentially two reasons - the first being to have something to hold the diffusers to the flash themselves - the actual cone of the diffuser is infact touching the front of the flash (they are so close and with a short working distance that they have to be that tight to the flash). The second being that channelling the light helps to minimise spill and thus reduces the amount of light lost - which means one can get away with a slightly lower power rating than if the light were spilling all over the place.

Overread · Nov 7, 2010

A slight modification to the design:

I've added reflective points to the middle of the diffusion cups. Because of how close they are to the flash heads (actually they are touching them) I expect that its more blocking the central harsher light rather than diffusing it around the unit as much, but still it can flash some light back into the flash head and the snoot
Further the diffusion layers have been separated on the holder so that now there is even more break up of the light within the small light tent.

Note that the ring for the light tent is the bottom of a plastic disposable cup with a tab of velcro on the top and bottom which attach to tabs stuck inside the MPE lens (the hood attached without problems).

a 1:1 shot

a 3:1 shot

As you can see the is still some reflection off the diffusion and light tent showing on the shell. I'm not sure if this is a property of using copy paper over the vellum paper others have used or a slight design flaw in the setup I currently have.

Garbz · Nov 8, 2010

Just one thing, if you eliminate all specular highlights you may end up with an image that looks flat and boring. Think model eyes without any catchlights. The diffuse look you have is already stunning

Overread · Nov 8, 2010

Garbz I agree I've already seen the rather flat appearance of the light that one can get when I did these tests with my 580EX2 - for them I had a whole sheet of paper curved over the subject and taped to the table - it gave the max amount of diffusion I've had thus far with not a single reflection on the shell - but it did also look very flat!

I've found that playing around with the shadows and highlights tool if one drags the midtone slider to the right (by quite a lot) you can restore some of the lost depth. That combined with blacks, contrast and clarity in RAW really made a big difference to what starts out as a flatter in camera shot.