TIR, prism choice and its effect on image quality (3 Viewers)

[[email protected]]

Hi,

while I enjoy the view through some porro bins and well understand the physical disadvantages of roof prisms in general and Schmidt-Pechan specifically, I can say that I personally can not see a difference in image quality between a top notch modern porro and an equally good modern pair of roofs.
It might be my eyes and I really should visit a friend of mine - an ophthalmologist - in his practice some time...
But I fear, that I will come out of that with a prescription for glassess and not be able to use some of my bins any more ;-)

I am also quite convinced, that nobody will be able to see a 2 or 3% difference in transmission, which tends to be the amount that modern roofs are behind the alpha raptor from Swaro...
I am less convinced about the destructive interference and resulting loss in center sharpness due to phase shift on the roof edge. Phase coatings mitigate that quite well nowadays, but I can imagine that a trained observer with exceptional eyesight can see differences.

It might be interesting to design a double blind test - put a few 8x30 or 10x40 bins in black boxes on tripods and intentionally limit the view to only center field (in order to deny the test subjects clues from field of view and the aberrations on the edge).
My candidates would be Habicht, SE, Fujinon FMT-SX if available and whatever current alphas in the correct sizes are at hand...

Joachim

"My candidates would be Habicht, SE, Fujinon FMT-SX if available and whatever current alphas in the correct sizes are at hand..."

You won't see any difference in resolution on-axis. All modern binoculars at the low magnification they are used will have pretty much the same resolution. The only difference you will see is brightness and 3D at a distance. If you ranked your list of binoculars by brightness, they would be Habicht, Fujinon FMT-SX, alpha roofs and SE with similar apertures. Of course, a 50 mm is going to be brighter, especially in low light, than a 32 mm. The porros like the Habicht, SE and Fujinon FMT-SX will have more 3D than the alpha roofs at a distance of about 100 yards.

 

[Scott98]

FWIW...I'm definitely intrigued by the high-throughput of the Zeiss FL HK's and the Habicht, the idea of highest possible transmission and the resulting image quality is what I like.

It sounds like the 10x56 SLC do a little better at the edge than FL, larger sweet spot, that's why I haven't tried to find 10x56 FL's...I'd like to try them someday though...would prefer upgraded 10x54 HT's or 50mm SF's

 

[tenex]

And although I don’t follow the CN forum all that closely, I have seen several references to the contrary
i.e. users do notice a difference in image quality between A-K and S-P prisms.

In post 22 there Erik Bakker said (of Porro or AK prisms equally):
"Total internal reflection means excellent polish of the optical surfaces is always smoother than the slightly more light scattering surface of the (dielectric) mirror coated Schmidt Pechan prisms", so "in comparison the S-P images look slightly smudged and less crystal clear."

Like the "Seil effect" previously referred to by Holger Merlitz, I'd like to believe that I can detect this, but would bet that given all the other variables I wouldn't always be able to distinguish AK vs SP in comparison tests, as no one manufactures the same binocular with different prisms. And I've only been struck by an impression of special "transparency" in some Porro models, like E II. Perhaps overall simplicity of optical design (fewer elements, external focus) has something to do with that also?

 

[hopster]

In post 22 there Erik Bakker said (of Porro or AK prisms equally):
"Total internal reflection means excellent polish of the optical surfaces is always smoother than the slightly more light scattering surface of the (dielectric) mirror coated Schmidt Pechan prisms", so "in comparison the S-P images look slightly smudged and less crystal clear."

Like the "Seil effect" previously referred to by Holger Merlitz, I'd like to believe that I can detect this, but would bet that given all the other variables I wouldn't always be able to distinguish AK vs SP in comparison tests, as no one manufactures the same binocular with different prisms. And I've only been struck by an impression of special "transparency" in some Porro models, like E II. Perhaps overall simplicity of optical design (fewer elements, external focus) has something to do with that also?

Without being a trained optical designer I would agree with the sense in Erik's quoted hypothesis and also yours about simplicity of design being part of the advantage.

Getting closer all the time to getting the Habicht as my 10x, and if I can just live with the stiff focuser I'll have a long-distance daytime optic with both excellent IQ and longevity, and if the Black one it will also be light to carry. My other option is a 2nd hand HT 10x42 with a much nicer focuser and modern eye cups, hmmm...

 

[Binastro]

I am still not convinced that TIR is 100% in the real world.

Surfaces are not totally flat, they are not totally smooth, and there is scatter.

Regards,
B.

 

[Scott98]

another thing to remember when parsing the minute advantages of subtle optics differences is "how clean are your binos?". If you lust after the best image quality you better have spotlessly clean lenses or it's all for naught

 

[Binocollector]

I have a few AK-prism binos and I cannot see much if any difference to SP-prism models. But I have no upper tier model, so I guess, in the end, the quality of the coatings and the glass itself plays a much bigger role than the prisms used.
The ones I have are up to the 500-700€ range, Kite Cervus HD 8x56 (very similar to the GPO), DDoptics Pirschler 10x45 and 15x56, and a few older models, all made in Japan, in 7x50, 8x56 and 9x63. The "Pirschlers" are pretty nice but have no ED glass and both have quite a bit of off-axis CA. My favourite is the 10x45 though because of the compact size, super fast focus and pretty sharp center.
But the models with SP-prisms in the same price range aren't really worse in any way.
From what I read -- AK-prisms had an advantage back in the days when the mirror coatings weren't as advanced as today. Today with improved technology and production processes, a bino with AK-prisms is in fact cheaper to make than one with the same performance and SP prisms -- probably due to the number of coatings applied to the prisms. When it comes to coatings -- a model with SP prisms can even beat a porro in light transmission. At least my Fuji HC 8x42 is brighter in the day than my Fuji FMTR 7x50 -- since both would be stopped down in the day due to pupil constriction, I guess the HC must have better coatings or it wouldn't seem brighter. At night, as soon as the maximum pupil dilation is reached, the roles should be reversed. I'd have to do an extended stargazing session to determine that -- and it's clouds everywhere, looks like summer is over in these parts.

 

[Tringa45]

I am still not convinced that TIR is 100% in the real world.

Surfaces are not totally flat, they are not totally smooth, and there is scatter.

Regards,
B.

Cannot prism surfaces be ground and polished to an accuracy of a fraction of a wavelength, possibly better than the radii of lens surfaces?
A typical wavelength at about one half a µ is after all not that small. Allegedly a finger tip can detect a surface irregularity of about one µ.

As regards scatter, I think that dielectric coatings on S-P prisms in binoculars must be less sensitive than mirror diagonals for astronomical scopes.
In the first case reflections take place inside the glass, whereas inconsistencies in the thickness of around fifty coatings on an erecting mirror could accumulate to quite significant surface irregularity.
Some exacting amateur astronomers now seem to favour enhanced silver erecting mirrors, or 90° prisms, over dielectric mirrors.

Regards,
John

PS:- I suspect that the 1% that doesn't get reflected from a dielectric coating just passes straight through, unless there is a final opaque layer.
Meopta use some sort of enhanced silver coating on the S-P Meostars and I have long suspected that Leica use something similar. Their transmission graphs show a gentle fall-off above 700 nm, whereas dielectric coatings usually show a sharp cut-off above that wavelength.

 

[Lerxst]

It is an interesting question. There is of course no perfectly planar surface, nor any bulk material utterly free from defect, but I would back-of-the-envelope estimate that the deviation from 100% reflectance upon TIR in a quality device would be more than one part in 10,000 (and that is being super-conservative; it is probably well less). I'd expect more loss from the bulk of the prism body and other optical components. Whether that will lead to changes discernable to human eyes under typical viewing conditions is another matter.

 

[Lerxst]

Cannot prism surfaces be ground and polished to an accuracy of a fraction of a wavelength, possibly better than the radii of lens surfaces?

Not sure if optics manufacturers use it, but in the hard drive industry CMP yields sub-nanometer (!) level flatness for various wafer fab steps.

 

[Binastro]

I need to ask an expert when I have time.

All I am asking is whether in practice TIR is 100%.

Binocular prisms can be polished to 1/20th wave, but usually I think they are not that accuracy.

With large flats they are more difficult to polish to high accuracy than curved surfaces.

My 317mm optical window cost more than the primary.

But prisms are usually small.

A friend had a National Physics laboratory test flat about 10 inches diameter.
I think it was expensive and was swapped for the Zeiss 120cm f/7 Cooke triplet 1920 lens, which was very high quality.

I suspect any scatter from say a Swarovski Porro prism binocular might still contribute to high transmission, even though the scatter is not imaging, but I don't know.

I don't think in practice less than 100% makes a difference in a good Porro prism binocular.

I am just interested in any factual data on whether TIR is 100% or not.

This needs someone who actually knows the answer rather than speculation.

Regards,
B.

 

[Tringa45]

I'd expect more loss from the bulk of the prism body and other optical components. Whether that will lead to changes discernable to human eyes under typical viewing conditions is another matter.

Yes, probably several orders of magnitude higher than any possible TIR losses.
Per 25 mm glass path, N-BAK4HT absorbs 1,7% in the blue at 400 nm.

John

 

[tenex]

All I am asking is whether in practice TIR is 100%.

Do we know the typical internal reflective efficacy of the mirrored surface of a Pechan prism, for comparison?

 

[Patudo]

My brother, who doesn't read Birdforum and as a consequence is blissfully untroubled by most of the technical gobbledygook, says he notices a slight difference in contrast with Abbe-Koening prismed binoculars, but, in his words: "I think I can tell the difference, but can't, hand on my heart, guarantee that I could in a blind test".

I myself use porro (Nikon SE and others) and Abbe-Koening binoculars much more than Schmidt-Pechan binoculars. I probably ought to be bigging up the qualities of the totally (or maybe not, lol...) internally reflective prisms, but I cannot honestly say I think they are better optically in any way that I can attribute to the prisms. I've used my brother's 8.5x42 EL SV FP a lot, in a wide variety of conditions, and find it outstanding (every bit the equal, if not better, thanks to a larger sweet spot, than my 8x42 FL), have tried a friend's 10x42 Noctivid which I also thought was very impressive and the NL in both 8x42 and 10x42 formats was also superb. I feel modern Schmidt-Pechans do have that quality of "transparency" - I think because of their dielectric coatings achieving a more neutral colour rendition than the previous generation of silver-coated prisms, which give a warmer colour cast. Winding the clock back some 20-odd years, Zeiss's 7x42 Dialyt did have a better image than the 10x40 - brighter and cleaner - but to my eyes something like the 8x32 FL is just as bright (in decent light anyway) and clean.

For what it's worth I thought the 10x42 Conquest HD was very good indeed when I tried it - very bright and very sharp. I could be very happy using it as my only birding binocular, ditto the Monarch HG or Meostar HD.

 

[Maljunulo]

The reflection may indeed be 100% but there will be some scattering and absorption in the glass, so if you put 1000 photons in, you will not get 1000 photons out at any wavelength.

Prisms are bulky.

 

[Binastro]

Prisms are small and were typically made twenty at a time.

The quality depends on who is making them.

But they are usually good or very good.

Although the TIR can't probably be 100%, it is probably nearly so, because the polish, flatness etc. isn't perfect.

Large optical windows are completely different, and are usually made individually.
There are few people who can actually make them accurately.

I couldn't get an actual efficiency for practical prisms TIR, but in practice it doesn't seem to be a problem.

Regards,
B.

 

[Maljunulo]

Compared to individual lenses, they are bulky, compared to the aggregate of lenses in an optical train, maybe not.

 

[kimmik]

TIR in optical fibre lets light run many kilometers without a boost. That is in the order of billions of TIR reflections. Even then it is probably the glass attenuation that reduces signal, rather than imperfection in the TIR.

Of course even superconductivity is not actually perfection depending on your standard of perfection.

 

[Maljunulo]

TIR in optical fibre lets light run many kilometers without a boost. That is in the order of billions of TIR reflections. Even then it is probably the glass attenuation that reduces signal, rather than imperfection in the TIR.

This.

Are the physical properties of optical fiber glass at all similar to binocular prism glass?

 

[Lerxst]

Another example would be optical ring resonators that use TIR to circulate light and achieve Q factors that are astronomically large, like 10^10.

Optical fiber transmission loss is what I used to get the 1 part in 10^4 estimate I made above. For visible wavelengths, attenuation is quoted at worst ~ 10 dB/km. The ratio of surface area for a 100 um fiber that is 1 km long, to the surface area of a prism is on the order of 1,000. The loss is fibers is almost all due to Rayleigh scattering in the volume, so any contribution from imperfect TIR is going to be in the noise. And this is not factoring in that the planarization processes applied to the few sq cms of a prism faces are going to be far more exacting that what will be done during the manufacture of a such immense lengths of cabling.

I've found nothing in the literature yet that gets at quantification of any measurable deviation for perfect TIR, but I'm still looking. Of all the mechanisms in an optical path that can incur loss, TIR is the last one to worry about.