Diffraction and DSLR's?

[revi]

No, as has been said above, the 1.5 is the (approximate) crop factor.
Have a look here for specs: http://www.sony.fr/support/fr/produc...specifications:
lens 14,3 - 71,5mm (35mm eq.: 24 - 120), F2,8 - 4,8
sensor size: 21,5 x 14,4 mm
So that lets us calculate a crop factor of 1.66-1.68 (by focal length) or 1.67 (from sensor size)

That said, the lens can get much closer to the sensor than in a DSLR, as there's no mirror to be taken into account. That might make it a lot easier to build a distortion-
free lens with a nice zoom range. 1.5 or even 6 mm lens-sensor distance would be pushing it with a sensor of 21×14 mm, though...

[pjbw]

....seems to be impossible to replicate these days (2012, I think I know why!))

There are EVF top-end cameras from Fujifilm (X-E1) and Sony (NEX-6) available at the moment. However they both have interchangeable lenses with only 3X zoom and the EVFs are not 'live' as with the R1. Their histograms would help. I am only guessing but I could be sure that the zooms at wide angle would have barrel distortion because the lenses would have to be spaced further away from the sensors because of the lens mounts.

[ajohnw]

Surely that is not what the tutorial says ... as the distance increases so does the diameter of the Airey disc at the sensor after the fashion of circle of confusion.[ CoC also mentioned in the tutorial ].

I haven't read all of this thread but the distance the light travels has nothing what so ever to do with the Airey disc. Where is can have an effect is when there are optical errors in the system of lenses but general effect of errors is less contrast. Talking what are relatively small sub wavelength of light error these reduce the intensity of the actual Airey disc and deflect more light into the rings that surround it. More often in camera optics light is just scattered and that also reduces contrast. It's also safe to say that a camera lens wont be diffraction limited in any case so is highly unlikely to produce a spot of the size of its theoretical Airey disc anyway.

The classical and easiest to understand cause of diffraction disc is to visualise waves of light being directed to an exact point of light by a lens. Unlike a laser the light isn't coherent so the peaks and valleys of these waves do not coincide and where a peak coincides with a valley they interfere and cancel out. It's then necessary to consider that a circular aperture will have captured the light and concentrated it into a spot. More mathematical factual information is about on the web. In very simplistic terms this explains why faster F numbers produce smaller diffraction spots, the waves have a more obtuse angle. This can cause people to think that this means that the closer the rear element is to the sensor the better. Not so as this all relates to what is called the nodal point of the lens - this is where it can theoretically be replaced with single simple lens. This can be visualise by imaging a lens is focused on something and and producing an image. If the lens is swung around at it's nodal point the image doesn't move.

There are distance effect and it's better to think in terms of telescopes to under stand what happens. First point is that the effective diameter of the lens sets it's resolution eg a 50 mm F2 lens has an effective diameter of 25mm and that is what sets it's maximum resolution. The F number sets the size of the diffractions spot and the focal length sets the scale of the image. So a 50mm F2 lens has exactly the same resolution as a 100mm F4 as it's effective aperture is still 25mm. The diffraction spot size has doubled but so has the image scale so the net effect is the same.

As to why this is all a bit dubious really in camera optics term well diffraction limited these days usually mean a 1/4 wave error in the wave front produced by the lens. In distance terms that means an error of about 0.0014mm in the "positioning" of the light waves. Positioning isn't a good terms to use but will do. Say a lens has a 6 elements, those have 12 surfaces so for axial images the maximum error on the surface of the glass rather loosely speaking is 0.00012mm In actual practice the refractive index of the glass will be higher than air so the maximum allowable error will be even less than that. Throw in off axis images and these distances where the light it travelling in glass all go to pot. A well designed optic of this type will produce suitably sized circles of confusion rather than diffraction spots. The front lens will also be larger than these number suggest so that the light fall off will be acceptable over the coverage angle of the lens. This can also help with off axis optical path length distances.

There is a good section on photographic lenses on the wiki - note the symmetrical designs to get round that - the comment about are new lenses better than the old are interesting too. Having seen some of the astro photo's some of the old plate lens cameras could produce I have my doubts. Some of them are exceedingly hard to make. Most of the improvements really have been less glass due to synthetic fluorites of one sort or another,

http://en.wikipedia.org/wiki/Photographic_lens

John
-

[ajohnw]

That laser image is a little different Remco. It's the pin hole that produces the pattern. The waves are truly coherent so the wave interference is perfect leading to many black minima.. A better idea of an optically generated pattern is shown above that. For practical highish quality optics with a total error of 1/4 wave on the wave front, the intensity of the central spot is reduced by 20% and the fist minimum is no longer zero. More lights also finishes up in the rings. Things get worse as the errors increase. This is why the image on the wiki is computer generated - perfection is rather hard to achieve. There are better examples about on the web along with how the light is distributed.

It's easy to see a diffraction pattern that is similar to the laser one. Just make a small pin hole in a piece of card and look at clear light bulb through it. It's even possible to make the hole slightly larger and see the central spot size reduce. The explanation for this is that only a small portion of the filament is viewed so the light tends to be coherent and maxima and minima interfere.

Reading again about laser spot sizes it seems I may be wrong. My understanding was that they could be focused to a spot size that was largely dependent on the accuracy of the lens rather than diffraction and that is why they can cut or burn etc. Producing a parallel beam is an entirely different kettle of fish and is a function of source size and lens complexity. The wiki entries may relate to that.

I do take you point though but was trying to keep it simple. Light is rather strange stuff. For another rather interesting take on it in relationship to optics take a look at this . It relates to microsocopes where things aren't naturally illuminated with diffuse light - ie hit the object at all angles. I asked Peter some time ago if he minded this link being posted. He doesn't but requests that people don't plague him with questions about it.

John
-

[ajohnw]

A post seems to have disappeared ? Odd it;s still in my latest posts page,

John
-

[Donald]

I've noticed there have been a couple of glitches on the forum today. I see some posts getting posted twice and it's sometimes taking a very long time for the message to go up once you press the 'Post quick reply' button. This happens from time-to-time. Will keep an eye on it and ask Sean (McQ) to have a look under the hood if it persists.