A demo of depth of field, CoC and diffraction

[ErikKaffehr]

Hi

http://echophoto.dnsalias.net/ekr/index.php/photoarticles/49-dof-in-digital-pictures?start=1

The article is based on a series of mages taken at 3.0 m with a 100/2.8 macro lens on a Sony Alpha 55 (16 MP APS-C) camera. The camera was shifted 3, 6 and 9 centimeters to cause defocus, and a "Circle of Confusion" was calculated for each defocus at each aperture. Figures included show:

- Image sharpness
- Sensor pitch
- Airy ring diameter (due to diffusion)
- Calculated CoC

Best regards
Erik

[LKaven]

Generous of you to serve up your findings.  The Topaz looks great to my eyes, much better than Smart Sharpen.  Do you have any guidelines on setting parameters for Topaz Focus?

[walter.sk]

Thanks for posting your research.  I have been playing with Topaz Refocus, and am also a longtime Focus Magic user.  Of course, FM has not yet made it to 64-bit processing.  Your results with Topaz are impressive.

[NikoJorj]

Great info, many thanks!
What strikes me most is the difference in blur between diffraction and defocusing, at a same size of Airy disc/CoC (eg between 30mm shift @f/1.4 and no shift @f/8).

PS just a quibble, but wouldn't it be more exact to write simply f/8 rather than f=1:8?

[ErikKaffehr]

Hi,

It's not f/1.4 but f/4.

I limited apertures to f/4 - f/16 because I wanted to use adjustment on studio flash only, and it has a five steep range.

At large apertures aberrations dominate at small apertures diffraction will have a tendency to dominate.

I'll change to f/# convention on next update.

Best regards
Erik

Great info, many thanks!
What strikes me most is the difference in blur between diffraction and defocusing, at a same size of Airy disc/CoC (eg between 30mm shift @f/1.4 and no shift @f/8).

PS just a quibble, but wouldn't it be more exact to write simply f/8 rather than f=1:8?

[ErikKaffehr]

Hi,

I don't really have any advise on using Topaz Focus. The way I used it was with "estimate blur" and probably one pixel blur. In this case it would use blind deconvolution and calculate the "Point Spread Function" itself. A problem is that PSF may vary across the field so what is good at center would be awful in the corners. I may look a bit more into this.

Best regards
Erik

Generous of you to serve up your findings.  The Topaz looks great to my eyes, much better than Smart Sharpen.  Do you have any guidelines on setting parameters for Topaz Focus?