Friday, April 24, 2009

Diffraction and Macro Motion Blur

This is, I hope, going to be a discussion about diffraction and how the effects of motion blur can be mistaken for it. But before I dive in I need to make a few things clear:

Please do not give me links to your examples of diffraction limited photos -I know that diffraction exists so there is no need to prove it...

Please read this entire post carefully, and with an open mind, before you post a comment...

When I first got into macro three years ago there were two prevailing themes: You had to use a tripod to get sharp images, and diffraction is the primary reason why you can't stop a lens down and get sharp images. I wanted to let the subject dictate how I shot it, and to go after moving targets, so I came up with the Left Hand Brace Technique for taking control of the motion in a scene so I could get the compositions and sharpness that I was looking for without using a tripod. Along the way I started to realize that what a lot of people were calling diffraction was really nothing more than a form of motion blur that I call macro motion blur. To explain how motion blur can mimic diffraction you first have to really understand what diffraction is.

The classic definition of diffraction is light bending as it passes through a small opening, and that's true. But there's a little more to it than that: The rays of light actually expand, like a cone, so that light rays that should hit a single pixel are now bleeding over into adjacent pixels. When that bleeding over reaches half way into those adjacent pixels you'll see a loss in image quality -diffraction has taken your lunch money.

But if you can wrap your head around expanding light rays causing a loss of detail then couldn't motion blur, as little as half a pixel of movement during the exposure, look like diffraction? My experience has taught me that it can.

Right about now some of you are thinking "But wait a minute Mr. Macro Motion Blur, why do my images get sharper when I decrease my Fstop?!" To answer that question we have to first determine what your real shutter speed is. At macro magnification, high Fstops, and low ISOs, the light that your flash is producing is the only significant light source in a scene. To prove it just set your camera to manual mode, F11, 1/200 of a second (or whatever your maximum flash sync speed is for your camera), ISO 100, and go out and take a photo on a bright sunny day with your macro lens set to its minimum focusing distance (life size magnification). You'll have to shoot something white or yellow to get anything to show up in the resulting image. Then turn on your flash and take the same shot -there will be a major difference in what you can see in the photo. At higher magnifications, with the flash turned off, you'll just get a completely black image. So your flash duration is really your shutter speed, since it's the major, or only, light source -the actual speed of the shutter doesn't matter...

So you set your camera to manual mode, F11, 1/200, and ISO 100. You've got a rig that will let you shoot at 2x, you take a shot hand held, and the resulting photo isn't very sharp. You then set your camera to F8, take the same shot, and this time the image is a lot sharper. See, no more diffraction! Well there's only one little problem -when you went from F11 to F8 you opened up the aperture in your lens to let twice as much light into the camera. Through The Lens (TTL) metering compensated for the increase in aperture (decreasing Fstop) by turning the flash on half as long -so your shutter speed (the flash) doubled. Did you defeat diffraction or macro motion blur?...

For me the answer is simple, since I can get sharp images at life size to three times life size at F13, and four to five times life size at F10 with Canon's MPE-65mm macro lens. The key is to get the flash as close to the subject as possible, use a diffuser that really diffuses the light (not one that just blocks the light), and look for ways to take control of the motion in the scene. Diffraction does exist, but it's not the bogyman that most people make it out to be...

6 comments:

Dave A said...

You are correct to an extent, but I took careful, tripod-stabilized photos using a Canon 40D and Canon 100mm macro lens and flashes awhile back at exactly the same focus, but with different apertures. The results were that to about f8/f10, there was no real difference in sharpness in the sharpest part of the image, but that after that, sharpness degraded. At f16, things were very bad compared to f8, but with other tests, f16 was the only way to get a reasonable dof and sharpness was not horrible. On the other hand, I try not to shoot over f14, as this appears to be a reasonable compromise as far as getting sharpness and dof. Anyway, my point is that I think that you are correct that much fuzziness is due to lack of light, or stabilization during the moment, but that fuzziness due to aperture is also definitely something to consider.

Dave A said...

Oh, and I realize that you were not talking about anything over f13 for any reason, but my tripod tests were at less than 1:1 and there was still slight degradation at f10 or so. I should also qualify this by saying that I've read a lot of your other tips and such and that you are dead on in general, as far a espousing getting as much light as possible where it needs to be to make up for movement of the camera or subject, and your tips have helped me take better and better photos, so thank you very much.

Dalantech said...

Hey Dave,
Again, I never said that diffraction doesn't exist. It does. But I know some photographers who won't shoot over F8 at 2x because they think they're loosing detail to diffraction. Everyone can argue with me, but it's a little tough to argue with the sharpness of my images ;)

Also keep in mind that most lenses start to lose image sharpness past F11 no matter what magnification you shoot at and it has nothing to do with diffraction and more to due with the lens not being able to produce a sharp image circle.

David Röll said...

It actually is diffraction what makes lenses >f/11 bad.
And that's only true for APS-C sensor size and smaller.
Larger sensors or film have a higher diffraction "limit".

Large format photographers often use higher f-stops than what DSLR-photographers even dream about.
There are for example APO-Rodagons with 360mm/F9, and they're very sharp stopped down. They often use F/22 to F/64. But you wouldn't want use them on a small sensor.

A lens that offers very high resolution will deliver better results at the same f-stops than a bad lens, but still, diffraction will always take away sharpness at the same rates.

Diffraction is a difficult subject. Because it matters how large you want to view the photograph later. If you want to crop or get really close to a large print, diffraction will matter earlier than when you look at it from a "normal" distance.

That's why images which are actually past the diffraction limit look good when you resize them to something significiantly smaller. And with that, it should be obvious that sensor resolution (pixel pitch etc) is the real thing for diffraction. If there's too much blur between the pixels, you'll not be able to enlarge or crop the picture very much.

David Röll said...

It actually is diffraction what makes lenses >f/11 bad.
And that's only true for APS-C sensor size and smaller.
Larger sensors or film have a higher diffraction "limit".

Large format photographers often use higher f-stops than what DSLR-photographers even dream about.
There are for example APO-Rodagons with 360mm/F9, and they're very sharp stopped down. They often use F/22 to F/64. But you wouldn't want use them on a small sensor.

A lens that offers very high resolution will deliver better results at the same f-stops than a bad lens, but still, diffraction will always take away sharpness at the same rates.

Diffraction is a difficult subject. Because it matters how large you want to view the photograph later. If you want to crop or get really close to a large print, diffraction will matter earlier than when you look at it from a "normal" distance.

That's why images which are actually past the diffraction limit look good when you resize them to something significiantly smaller. And with that, it should be obvious that sensor resolution (pixel pitch etc) is the real thing for diffraction. If there's too much blur between the pixels, you'll not be able to enlarge or crop the picture very much.

John Kimbler said...

Thanks David Roll. I know what diffraction is, but I also know that there is a synergistic effect between diffraction, motion blur, and lens sharpness (and that's what the article is about). Diffraction, by itself, isn't going to ruin an image.