and less chromatic aberration.
Speaking of IS, I wonder why Zeiss restricted her IS technology to a very big and expensive 20x60, which is mainly an astronomy binocular and impractical for birding, tourism etc.
Daniel Vukobratovich's classic paper, Binocular performance and design, published by the SPIE in 1989.
Ed
Excellent paper! Thank you for sharing it.
-Omid
You're welcome. It's the single most informative article I've ever read about binocular design and performance.
Ed
I've added some notations in red to Fig. 1 from Daniel Vukobratovich's classic paper, Binocular performance and design, published by the SPIE in 1989. As can be seen, theoretical (or ideal) efficiency increases 1:1 with magnification. However, all real binoculars are compromised by various factors, a very important one being hand tremble.
Without being supported, the asymptotic efficiency level of handheld binoculars is approx. 7.0. In other words, no matter how large the magnification the efficiency never exceeds about 7.0 — unless the binocular is supported, i.e., by a tripod. Otherwise stated, handheld binoculars don't perform better than ideal x7 binoculars.
With maximum support, which eliminates tremble, efficiency is still not theoretically perfect, but a significant improvement is realized. For example, a 10x binocular improves from about 5.5 to 8.8, an increase of (8.8-5.5)/5.5= .6 or 60%. My guess is that the support provided by an IS mechanism, like the Canon's, might not be quite as effective, perhaps only improving the binocular by 40-50%, let's say for an efficiency of 7.5.
It should be evident from the graph, therefore, that a 15x or 20x handheld binocular may be expected to perform arguably worse than a stabilized 10x, and in that respect it replaces them with one instrument.
Kimmo seems to be on to something important! :smoke:
Ed
Yes, It is. I learned a lot by reading it. In Section 6 the article mentions research by R. Home that indicates that the optimal angle between a binoculars' barrels may not be parallel but slightly convergent. This is very interesting since I realized this myself a while back and mentioned it in this thread here on Birdforum. Do you happen to have this paper too?
Thanks again,
-Omid
I would say that Home's tenuous implication is based on the so-called "dark focus" or resting focus of the eye, which obtains in the absence of a retinal image, — such as darkness. Without reading the article, however, it's not intuitively clear how any fixed convergence angle of the tubes would allow the "vergence mechanism of the observer's eyes to match their accommodative state," except at some fixed distance, since the accommodative state necessarily varies with working distance. No doubt that's partly why Vukobratovich says "The actual error tolerable in binocular alignment is controversial."The collimation of binocular telescopes appears to be based on the assumption that the user prefers to accommodate to a stimulus at infinity. Recent work favours the view that for normal observers the preferred accommodative state of the eye is approximately 1 dioptre. ...
The conclusion is that the present strict tolerances on the collimation of binoculars could be relaxed if the design of binoculars were modified to take into account the accommodative-vergence state of the eye.
Why the big three (Zeiss, Swarovski, Leica) have not yet made an image-stabilizer binoculars has been a question for me too. A few months back in a discussion with Mr. Stephen Albercht (Leica Business Development Manager for Sports Optics) he mentioned to me that the less-than-perfect optical quality of an stabilized binocular is a concern for Leica. I am guessing the feedback control systems and the electronic actuator mechanisms are too complex for an optics-only company to make. There could also be issues with patent protection by the Japanese companies.
Ed,
Thanks for linking the full Vukobratovich article. It really is excellent and contains a wealth of interesting information.
The section on tremor in hand-held binoculars mentions Schober's measurements of hand tremble with the surprising (to me) finding that people belong to two distinct groups regarding the type of tremble they have. One group has three characteristic tremble frequencies centered between 1-2 Hz, 6-9 Hz, and 10-12 Hz with a sharp maximum between 7-9 Hz. The other group had either no distinct tremble frequencies or a flat maximum between 6-10 Hz. In the studied sample, the two groups of people were nearly equal size.
If this is true, then we all belong into one of two types of tremblers, and may accordingly have different experiences of hand-held viewing. The type of trembler you are may also influence your experience with stabilized binoculars. I have tried to subjectively evaluate which type I may be, but cannot say that it would have been successful yet.
In an article by Nolting and Kiesel, Verwackelt? Bestimmung der Sichtlinienstabilität stabilisierten Ferngläser (Fachhohschule Aalen), there are measurements that show the effect of IS on image shake with different stabilized binoculars. The Canon in this test, an 18x50 IS, damps less than half of the shake amplitude at under 3 Hz, but above about 4 Hz the damping is very good and consistent, keeping the shake amplitude consistently below 0.05 degrees. Fujinon is better between 2-3 Hz, but worse at all other frequencies. Zeiss 20x60 has significant damping of shake between 3-6 Hz, but not very much at other shake frequencies.
It is quite possible that the widely varying experiences people report with image-stabilized binoculars are at least party attributable to viewers' different characteristic tremble frequencies and amplitudes. Testing this would be quite a task, and trying out for yourself is the easiest way to go.
The other comment I'd like to make about Vukobratovich's article concerns the efficiency estimates. Based on my testing, I firmly believe that top-level modern binoculars when supported come significantly closer to their theoretical maximum efficiency than what was suggested in the article. Instead of an 8.8 efficiency rating for a 10x binocular, I'm quite sure the best ones get over 9.5 or 95% when well supported.
Also, at least in my tests, with a hand-held but image-stablized Canon 10x42, I get well over 90% (about 95%) of the efficiency I get with the same binocular tripod-mounted. I don't know if I'm a particularly lucky trembler, but that is my mileage and the primary reason I'm such a strong advocate for image stabilization as an essential feature of hand-held binoculars.
Kimmo
Ed,
Thanks for linking the full Vukobratovich article. It really is excellent and contains a wealth of interesting information.
The section on tremor in hand-held binoculars mentions Schober's measurements of hand tremble with the surprising (to me) finding that people belong to two distinct groups regarding the type of tremble they have. One group has three characteristic tremble frequencies centered between 1-2 Hz, 6-9 Hz, and 10-12 Hz with a sharp maximum between 7-9 Hz. The other group had either no distinct tremble frequencies or a flat maximum between 6-10 Hz. In the studied sample, the two groups of people were nearly equal size.
If this is true, then we all belong into one of two types of tremblers, and may accordingly have different experiences of hand-held viewing. The type of trembler you are may also influence your experience with stabilized binoculars. I have tried to subjectively evaluate which type I may be, but cannot say that it would have been successful yet.
In an article by Nolting and Kiesel, Verwackelt? Bestimmung der Sichtlinienstabilität stabilisierten Ferngläser (Fachhohschule Aalen), there are measurements that show the effect of IS on image shake with different stabilized binoculars. The Canon in this test, an 18x50 IS, damps less than half of the shake amplitude at under 3 Hz, but above about 4 Hz the damping is very good and consistent, keeping the shake amplitude consistently below 0.05 degrees. Fujinon is better between 2-3 Hz, but worse at all other frequencies. Zeiss 20x60 has significant damping of shake between 3-6 Hz, but not very much at other shake frequencies.
It is quite possible that the widely varying experiences people report with image-stabilized binoculars are at least party attributable to viewers' different characteristic tremble frequencies and amplitudes. Testing this would be quite a task, and trying out for yourself is the easiest way to go.
The other comment I'd like to make about Vukobratovich's article concerns the efficiency estimates. Based on my testing, I firmly believe that top-level modern binoculars when supported come significantly closer to their theoretical maximum efficiency than what was suggested in the article. Instead of an 8.8 efficiency rating for a 10x binocular, I'm quite sure the best ones get over 9.5 or 95% when well supported.
Also, at least in my tests, with a hand-held but image-stablized Canon 10x42, I get well over 90% (about 95%) of the efficiency I get with the same binocular tripod-mounted. I don't know if I'm a particularly lucky trembler, but that is my mileage and the primary reason I'm such a strong advocate for image stabilization as an essential feature of hand-held binoculars.
Kimmo
Abstract
Muscle tremor limits the resolution of hand-held field glasses. All mechanical or optical compensations of tremor movements depends on its spectral frequency distribution.
In this paper, the spectral frequency distribution was recorded for 22 emmetropic observers between 18 and 36 years of age. Only tremor frequencies less than 20 cps are of practical importance. The frequency spectrum is a characteristic of the single observer. The position of his individual peaks remains constant for a period of more than one year. The most usual frequency maxima are in the ranges 1.3–1.7, 2.7–3.5, and 6–11 cps.
Other factors, such as weight and shape of the instrument, luminance level, and structure of the visual field, are of less significant influence.
Kimmo: ...The other comment I'd like to make about Vukobratovich's article concerns the efficiency estimates. Based on my testing, I firmly believe that top-level modern binoculars when supported come significantly closer to their theoretical maximum efficiency than what was suggested in the article. Instead of an 8.8 efficiency rating for a 10x binocular, I'm quite sure the best ones get over 9.5 or 95% when well supported.
Also, at least in my tests, with a hand-held but image-stablized Canon 10x42, I get well over 90% (about 95%) of the efficiency I get with the same binocular tripod-mounted. I don't know if I'm a particularly lucky trembler, but that is my mileage and the primary reason I'm such a strong advocate for image stabilization as an essential feature of hand-held binoculars.
Chosun:... My takeaway from this experience, was that I had a distinct preference for the plain view with NO IS on. Every time I engaged the IS I could detect visible movement of the viewing picture which was most unpleasant, and even detected 'artifacts' to the view, which I did not like at all. I found this somewhat surprising, given all that has been said, the fact that magnification was 15x, and my own experience with stablised photographic equipment. However, my preference to have IS off was quite clear in this instance.
It is a well known fact that spherical aberration of the EP causes kidney beaning
Hi Ed:
The simple equation in the paper that defines binocular efficiency can be questioned on the grounds that "target detection" is a somewhat vague notion that may not be that easily defined.
Holger Merlitz wrote a more recent (2015) paper on these aspects that you might find interesting---- I can send you his paper, and any other paper you want (by R Home etc), just let me know.
Regarding adding the IS feature to SV, I do not believe that's going to happen. My speculation is that the next big step will be an alpha digital bin set, but it will not come from Swaro but rather from Zeiss or possibly Leica.
Cheers,
Peter.
...Detection is the discrimination of an object from its background and its
assignment to the class of potentially interesting objects. Classification is the
assignment of the detected object to a gross class of objects such as vehicles or
combatant vessels. Recognition is the assignment of the classified object to a specific
subclass such as tanks or destroyers. Identification is the assignment of the
recognized object to an even more specific category such as M-60 tanks or Spruance
class destroyers.
Hi Ed:
Regarding adding the IS feature to SV, I do not believe that's going to happen. My speculation is that the next big step will be an alpha digital bin set, but it will not come from Swaro but rather from Zeiss or possibly Leica.
Cheers,
Peter.