Filters and image quality

[pico]

Advertisement (gone after registration)

The focus shift caused by a filter is about one third of its thickness.

 

For example if the thickness of a filter is 0.6mm, then the focus shift is about 0.2mm. Hence the thinner the filter the better.

 

Sorry, cannot accept that. Too vague and general.

.

[Advertisement]

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

[jaapv]

I damaged the front element of a lens by having a filter break by a bang on the side. As it happened it was a 240/4.0 apo. The glass broke and scratched the front of the protective element badly. And I had it on as I was doing some rough hiking. .

If I hadn’t I would just have dented the hood….

I remember -it was maybe 20 years ago- that replacing the “protective” element on the lens was shockingly expensive.

[jaapv]

The focus shift caused by a filter is about one third of its thickness.

 

For example if the thickness of a filter is 0.6mm, then the focus shift is about 0.2mm. Hence the thinner the filter the better.

A blank filter does not cause focus shift. It causes refraction errors especially out of the centre. As the refraction error is worse the thicker a planparallel piece of glass is, it is made as thin as possible.

 

Focus shift is caused by colour filters on non-APO lenses, mostly by red filters. The amount of shift is not determined by the thickness, but by the spectral transmission of the

filter and the amount of chromatic correction of the lens.

 

Nice explanations here:

The Complete Guide to Lens Filters (Part One) | wolfcrow

[masjah]

A blank filter does not cause focus shift. It causes refraction errors especially out of the centre. As the refraction error is worse the thicker a planparallel piece of glass is, it is made as thin as possible.

 

Focus shift is caused by colour filters on non-APO lenses, mostly by red filters. The amount of shift is not determined by the thickness, but by the spectral transmission of the

filter and the amount of chromatic correction of the lens.

 

Nice explanations here:

The Complete Guide to Lens Filters (Part One) | wolfcrow

 

Agreed. I do though keep UV filters on most of my lenses for protection and keeping clean; people have different views on this. As I said earlier,the 280/4 Apo is an interesting case in point. The lateral displacement of the slot-in filter is explicitly allowed for in the lens computation, so Leica say that actually leaving it out generates a "sin of omission" so to speak.

 

The 28/2.8 PC is another interesting case. I have two lens hoods for this lens, one of which has the unmounted glass disc UV filter screwed in. I prefer not to use it; I think it does degrade performance, and I suspect that, with the lens shifted, some of the incident rays are very oblique, with a consequently higher lateral displacement through the filter.

[pico]

Many of us consider clear protective filters because we go back to the earlier days of fragile coating. Heck, early Zeiss lens coatings could be wiped off with a soft cloth.

 

It took me too long to become wise concerning modern lens coating.

 

Too bad that I still use those old lenses!

 

 

 

 

Sent from my Etcha-sketch.

[gisling]

A colorless filter or color filter, as long as it is made of parallel glass plate, then shifting of image focus cannot be avoided, like it or not, it is a matter of physics.

 

Former head of Kodak Optical design department, Professor Rudolf Kingslake

 

wrote in his book "Lens Design Fundamentals" p 55 -56

 

Section D: Image Shift Caused by a Parallel Plate

 

amount of shift = thickness of glass *(1-1/n), where n is refractive index of glass

 

"Thus, if n=1.5 the shift will be one-third of the thickness of plate"

 

Well known American lens design expert Dr. Warren Smith has a more detailed explanation of this matter in the book" Modern Optical Engineering"

 

see attach facsimile.

Welcome, dear visitor! As registered member you'd see an image here…

Simply register for free here – We are always happy to welcome new members!

[gisling]

Advertisement (gone after registration)

Luckily for Leicaflex and Leica R users, the image shift cause by filter is automatically compensated by the single lens reflex system, as what you see in the viewfinder is what is on the film plan, the user will be able to rotate the lens barrel to

bring image back into sharp focus

[gisling]

The same is true for twin lens reflex cameras, as long as you put identical filters on

the viewing lens and taking lens of Rolleiflex, the image shift cause by the filter is

equally compensated.

 

The focus shift caused by the filter may not matter for autofocus cameras as well.

 

But there is no such compensation mechanism for rangefinder cameras (Leica IIIc, M6 etc), because what the rangefinder see is only matched to the lens without filter,

there is no way the rangefinder can compensate for filter-caused- image shift.

[gisling]

For certain lens, filter is absolutely no no.

 

A case in point is Leitz Elmarit R19 version I. Any filter will cause very visible vignetting

[jaapv]

A blank filter does not cause focus shift. It causes refraction errors especially out of the centre. As the refraction error is worse the thicker a planparallel piece of glass is, it is made as thin as possible.

 

Thanks, I stand corrected

Now let's talk phase shifts and interference

[masjah]

Luckily for Leicaflex and Leica R users, the image shift cause by filter is automatically compensated by the single lens reflex system, as what you see in the viewfinder is what is on the film plan, the user will be able to rotate the lens barrel to

bring image back into sharp focus

 

So why do Leica caution against leaving out the plain glass filter in the slot of the 280/4, saying that it's presence has been allowed for in the computation, if all one has to do is refocus slightly? I don't think a simple focus shift is the whole story. The formulae in many textbooks are based on linearised thin lens theory, and the computations by people like Leica (and all other lens designers) include second and higher order aberrations. I certainly agree though that the thinner the glass, the less the effect of the filter, and probably the effects may be worse for a slot-in filter at the back of the optical system, rather than the front of the lens.

[wizard]

...If I hadn’t I would just have dented the hood….

 

Well, Jaap, but then this was just bad luck on your part. I have a friend who does not like to buckle up since he heard of an accident on a mountain road where the unbuckled driver was thrown out of his car and survived almost unhurt, while the car crashed down into a canyon and was completely destroyed. That was good luck on his part, but chances were 1 in a million.

 

Best regards,

 

Andy

[pop]

It depends a good deal on where the filter is mounted. If it's on the side of the object to be photographed, one third the thickness of the filter will not matter greatly. If it's on the image side, the same third of the thickness will make itself felt. If it's between both sides, i.e. within the lens, leaving out the glass element ought to amount to changing the distance between the elements on both sides of the filter and can be understood only within the context of the optical construction of that lens.

[jaapv]

Well, Jaap, but then this was just bad luck on your part. I have a friend who does not like to buckle up since he heard of an accident on a mountain road where the unbuckled driver was thrown out of his car and survived almost unhurt, while the car crashed down into a canyon and was completely destroyed. That was good luck on his part, but chances were 1 in a million.

 

Best regards,

 

Andy

Of course it was - and I still use filters from time to time But I do not rely on them to protect my lens in all circumstances. A nice solid metal hood is often more effective.

[erl]

You guys must be anal about focussing your lenses, and always on a tripod, to even notice such things you talk about, none of which I doubt. But, it sounds like the sailor I competed against who spent who knows how long polishing the hull of his yacht so it would move through the water faster. Meanwhile, I spent the same time honing my skippering skills in a rough old boat. Yes, I won!

 

back to filters. I originally bought IR cut filters for all my M lenses when I first had my M8. I now leave them on for my M9 simply because some colours, mainly greens, in daylight reproduce better, IMO. Under tungsten light the improvement is even more noticeable. Do I lose focus accuracy? Yes, because I am clumsy! Got nothing to do with the filter, in my case. Just real world stuff. No theory involved.

[lct]

Same on M8.2 and M240. Never got focussing problems due to IR-cut filters so far.

[pico]

A colorless filter or color filter, as long as it is made of parallel glass plate, then shifting of image focus cannot be avoided, like it or not, it is a matter of physics.

 

The example shows the plate behind the lens. Does it behave differently in front of the lens?

[pop]

The example shows the plate behind the lens. Does it behave differently in front of the lens?

 

No, it does not. The distance of the object which will be in focus is increased by one third the thickness of the glass plate. Difficult to compensate for when - for instance - taking the picture of a cathedral.

[Advertisement]

Hallo,

Look here: Interesting Leica gear

[masjah]

No, it does not. The distance of the object which will be in focus is increased by one third the thickness of the glass plate. Difficult to compensate for when - for instance - taking the picture of a cathedral.

 

Which confirms that the optical effect of a filter is much more significant when it is between the lens and the film/sensor, thereby explaining Leica's warning with regard to the 240/4. And this effect will also depend on the colour of the light, since even the simple linearised formula involves the refractive index, which depends on wavelength

 

Even so, I'm still a bit bothered by the effect of a front-of-lens filter on very off-axis rays at very large angles of incidence (such as on a 28/2.8 PC at maximum shift). I still think that there's more to the story than just focus shift.

[pico]

No, it does not. The distance of the object which will be in focus is increased by one third the thickness of the glass plate.

 

 

Taken literally that means a picture would suffer no significant absolute defocus at anything but macro distances of which RF use has no concern. One-third the depth of the filter is nothing for any subject from 3-foot to infinity.