DOF resolution and perspective

[lct]

Advertisement (gone after registration)

Whatever CoC values one chooses for film (0.03 or 0.02mm below, for instance, hence 0.023 or 0.015mm for the M8), there is hardly more than one full stop difference between them but if 2 stops work better for you it is your prerogative of course.

 

 

[ho_co]

In one sense, we're really all on the same page here.

 

It may not look that way because some of us are referring to CoC as if we were still dealing with film, others (like me) enthusiastically assume the Karbe rule of thumb, Stephen takes the time to study and relate the possibilities into Excel files, Osuna and García (Do Sensors "Outresolve" Lenses?) restart from scratch.

 

The fact is--and that's what Jaap said several pages back--with a two-dimensional sensor, the rules of DoF have changed.

 

It's because of that fact that we're all following this thread.

 

Karbe's approach to DoF is revolutionary to my mind; lct rejected his approach in his and my earlier exchange on the topic because he didn't follow the traditional way of deriving the idea (which fact is what I like about it); Stephen uses the standard computations to set limits to Karbe's rule of thumb. The LuLa digital focus link I posted above shows--and very effectively since Ferguson uses a larger-than-FF sensor--pragmatically that digital DoF is different from what we're used to.

 

Summary: Digital is new territory, and we are all making that discovery. This is a tiny part of what Puts means when he speaks of a paradigm shift.

 

So this thread and its (sometimes major) differences of opinion offers a major opportunity to broaden our understanding of these things that are changing under our feet.

 

Keep it up! Stephen, thanks for starting it, and thanks for acting so well as moderator!

[Advertisement]

The Leica Q3 is available for pre-order at B&H Photo Video and Adorama

[SJP]

@Howard, thanks for the positive support it makes doing the sums worthwile for me.

 

@lct, also thanks for the comments & criticism, and for keeping us (me) focussed, I am convinced that we are actually on the same track.

 

@Lars & Michael also thanks for your remarks.

 

In fact I am pleased if anyone reponds to all of this stuff because it is so technical & may seem confusing (or uninteresting). Also I hope I can help with understanding the tools of the trade from the optics/physics point of view. The stuff Leica makes deserves such attention to detail - it is not only about reputation or advertising.

[pgk]

Can I just add in some practical thoughts?

 

Firstly the CoC is IMHO defined by the output requirements of the file - so inherently is a variable as not all images are destined for identical usages. And add in such factors as uprezzing in PS and everything is thrown into turmoil.

 

Secondly, I have my suspicions that the 'bokeh' of a lens, which I believe is created by both the lens and the disphragm design, must also impinge on the detailed structure of the image - precisely how I do not know, but this does make sense to me - and consequently this must throw in some effect on the image structure/definition.

 

Thirdly, photography is at its very heart an very practical process. Its relatively easy to test lenses out for given requirements by varying apertures and actually looking at the resulting files. This is also true of different systems and I am fortunate enough to run two currently - FF Canon and M8, I also use some Leica R glass on the Canons.

 

I AM intrigued by the differences in appearance between the various lenses and cameras. Whatever some may say, IMO they produce distinctly different files and I would say that this even extends to their depth of field 'characteristics' - of course we are talking nuances here but nevertheless they do image differently and as with most things in life viewing lots of files educates the viewer until these nuances are often quite distinct. I am sure that these differences could be expressed mathematically, although whether their complexity makes this at all worthwhile or even possible (try adding in the results of residual abberations!?) is something else.

 

So what I suppose I am getting at is that whilst I agree that the mathematics of depth of field has its place (very good guidance), simple maths almost certainly provides an indication of expected outputs, not an absolute description of what precisely, actually, happens in the real photographic world.

[SJP]

pgk, thanks for the comment, indeed this is true. There is more to lenses than DoF and although you can model them quite well this will not tell you how a lens 'feels', or what the lens signature is.

 

I think that Leica's approach is to make a lens with as few abberations as possible - this is a horribly complicated task & also requires considerably more mathematics than I used in my scribblings. Anyway the less errors the lens makes the better the images becomes in terms of 'pop', contrast, distortion, color & edge rendition, flatness of field. Also indeed the shape of the diafragm is very important for the bokeh, with the present Leica lenses the diaphragm is pretty close to a perfect circle.

 

They are not designed to 'look' good, they are designed to be as technically (mathematically) perfect as possible, the 'look' follows from that to a large extent.

 

Nevertheless the look is what we buy not the mathematics.