Who else is cheating?

[John Russell]

Holger Merlitz recently tested the 12x50 Optolyth Alpins and found the exit pupil to be a mere 3,5mm. This makes them effectively 12x42s.
I have just read a report on a German forum, in which the exit pupil of the Canon 12x42 L IS WP was measured at 3,7mm. As the 10x magnification was found to be correct they are, despite their excellent optical qualities, effectively only 10x37s due to the inadequate diameter of the stabilizing prisms.
Rather than a deliberate attempt to delude their customers, I suspect however that the manufacturers have just not completed their homework - remember the "Elk Test"?

[elkcub]

John,

Where can the Merlitz article be found?

Thanks,

Ed

[kabsetz]

John,

Prompted by a thread on Cloudy Nights, I made some measurements on the Canon and a Nikon 10x42 SE. It turns out that matters are not quite as simple as manufacturers not doing their homework properly. True, the exit pupil of the Canon is not 4.2mm wide, although it is not 3.7mm either. I got 4.0mm horizontal e.p. diameter, 3.9mm diagonal diameter and 3.8mm vertical diameter. For the Nikon, I measured 4.2mm horizontal, almost 4.2mm diagonal and only 37mm vertical diameter. These results, however, do not mean that the entire objective diameter would not be utilized. For the Canon, even with a 43mm mask covering the objective centrally, a sliver of field edge could be seen on either side. Every point of the objective lens is contributing to the image at least partially, but it seems that the very edges of the objective are not contributing to the image center. It would be just as misleading to call the Canon a 10x37 or a 10x39 when masking the aperture to those dimensions would cut out light that in the present design is contributing to the image. I'm quite sure that Canon knows what they are doing, and that the overall design compromise they have reached is a conscious and deliberate one. Anyway, the proof of the pudding is in the eating.

Kimmo

[ThoLa]

Holger Merlitz recently tested the 12x50 Optolyth Alpins and found the exit pupil to be a mere 3,5mm. This makes them effectively 12x42s.
I have just read a report on a German forum, in which the exit pupil of the Canon 12x42 L IS WP was measured at 3,7mm. As the 10x magnification was found to be correct they are, despite their excellent optical qualities, effectively only 10x37s due to the inadequate diameter of the stabilizing prisms.
Rather than a deliberate attempt to delude their customers, I suspect however that the manufacturers have just not completed their homework - remember the "Elk Test"?


Hi!
Cheating is a harsh word, isn't it?
Even the Leica Ultravid 8x32 was MEASURED to have "only" 3,9 mm exit pupil by STIFTUNG WARENTEST, a nationwide consumer organisation that you will know as you are based in Germany.

Swarovski SLC 10x42 = 3,9 mm.

But how relevant is it in real life?

And there are those "understaters" who are actually "larger than life":

Pentax DCF HRII 10x42 = 4,3 mm.
Zeiss Victory FL 10x42 = 4,5 mm (!).


The same amount of variations is found with respect to fields of view (which are often discussed as well).

I do not know who Mr. Merlitz is. But a very sensible general rule is: Beware individuals of unknown affiliations who make statements whose origin cannot reliably traced back. Any serious statements and measurements ought to be made public in a trust-worthy, generally available source.

It does not make very much sense discussing 0,something values if they ooze out of the grapevine.

Tom

[bill lord]

Holger Merlitz recently tested the 12x50 Optolyth Alpins and found the exit pupil to be a mere 3,5mm. This makes them effectively 12x42s.
I have just read a report on a German forum, in which the exit pupil of the Canon 12x42 L IS WP was measured at 3,7mm. As the 10x magnification was found to be correct they are, despite their excellent optical qualities, effectively only 10x37s due to the inadequate diameter of the stabilizing prisms.
Rather than a deliberate attempt to delude their customers, I suspect however that the manufacturers have just not completed their homework - remember the "Elk Test"?

12x50 does not say that the exit pupil is 4.166. it is a measure of the binoculars, it implies that the magnification is 12X and that the objective lens is manufactured with a 50mm diameter. As long as this is so then there is no deception.

[kabsetz]

Tom,

Could you perhaps provide more data from the Stiftung Warentest report? It would also be interesting to know how they measured the exit-pupil.

Kimmo

[Surveyor]

FWIW

I recently obtained ISO 14133-1 Optics and optical Instruments-Specifications for binoculars, monoculars and spotting scopes--Part 1 General Purpose Instruments and (14133-2) Part 2-- High performance Instruments.

I was very surprised (dismayed) at the allowable tolerances i.e. +/-6% for high performance exit pupil, 100/60 arc minutes (gen. purpose/HP) divergence in the horizontal plane for instruments <20x. 5%/4% for power, etc. Note, these are max. allowable from nominal, but if any of my instruments approached these figures I am sure I would dispose of them.

These figures, especially collimation, are greater by a large margin than I had considered acceptable.

Ron

[elkcub]

Hi!
Cheating is a harsh word, isn't it?
Even the Leica Ultravid 8x32 was MEASURED to have "only" 3,9 mm exit pupil by STIFTUNG WARENTEST, a nationwide consumer organisation that you will know as you are based in Germany.

Swarovski SLC 10x42 = 3,9 mm.

But how relevant is it in real life?

And there are those "understaters" who are actually "larger than life":

Pentax DCF HRII 10x42 = 4,3 mm.
Zeiss Victory FL 10x42 = 4,5 mm (!).


The same amount of variations is found with respect to fields of view (which are often discussed as well).

I do not know who Mr. Merlitz is. But a very sensible general rule is: Beware individuals of unknown affiliations who make statements whose origin cannot reliably traced back. Any serious statements and measurements ought to be made public in a trust-worthy, generally available source.

It does not make very much sense discussing 0,something values if they ooze out of the grapevine.

Tom

Tom,

Holger Merlitz is a highly respected binocular collector and evaluator with a Ph.D in physics. Here is his home page: http://www.holgermerlitz.de/. This interesting essay concerning the "rolling ball" effect was recently released. http://www.birdforum.net/showthread.php?t=74727

The reason I asked for the original reference in post #2 is to examine what he found and the methodology he used.

Ed

[ThoLa]

Tom,

Could you perhaps provide more data from the Stiftung Warentest report? It would also be interesting to know how they measured the exit-pupil.

Kimmo


Hi!
Measurements were done according to ISO protocols.
A full list is given at the end of the article.
I don't know how to make an attachment here
but I can send you a pdf of the article if you
contact me at
tlazar 'at' web.de

Kind regards,
Thomas

[elkcub]

Hi!
Measurements were done according to ISO protocols.
A full list is given at the end of the article.
I don't know how to make an attachment here
but I can send you a pdf of the article if you
contact me

Kind regards,
Thomas

Thomas,

To add attachments, scroll down below the text panel to a button labled "Manage Attachments." From there it's self explanatory. ;)
(But there are limits on picture file sizes.)

Ed

[John Russell]

Kimmo,
Here is a link to the Canon test (in German):- http://www.juelich-bonn.com/jForum/read.php?9,17044,17044#msg-17044.
The contributor originally noticed the discrepancy with the Canons mounted on a tripod next to some 8x32s (4mm exit pupil).

John

[John Russell]

Where can the Merlitz article be found?


Ed,
I take it you have now found the Optolyth test on Holger Merlitz' home page.

John

[John Russell]

12x50 does not say that the exit pupil is 4.166. it is a measure of the binoculars, it implies that the magnification is 12X and that the objective lens is manufactured with a 50mm diameter. As long as this is so then there is no deception.

Bill,
If the exit pupil of a 12x50 is only 3,5mm then either the magnification is significantly higher than 12 or the 50mm objectives are being masked in some manner and their light gathering power is not being used to its full potential.

John

[henry link]

Masking of the objective in the 50mm Optolyth Alpins is pretty old news to me. I bought a 10x50 Alpin almost 20 years ago and found that the Porro prism was so undersized that its front aperture effectively stopped down the objective lens to about 45mm. Curiously, in that case the exit pupil was 5mm because the magnification was actually only 9X. Obviously nothing much has been changed since then.

More recently I found that an "8x40" Celestron Outland binocular had an aperture stopdown just behind the objective lens that reduced the true aperture to 32mm.

[elkcub]

Ed,
I take it you have now found the Optolyth test on Holger Merlitz' home page.

John

John,

Yes, I have. As usual it's a pleasure to read his concise analyses, as something is always to be learned and little is left to add. His explanation for the design is (my emphasis):
With 58 degs. apparent field, even the Optolyth delivers a reasonably wide field. How is this possible with small prisms? What matters here is the size 'S' of the intermediate image. For an un-obstructed binocular, and neglecting the distortion, S can be estimated using the focal ratio Fr = F/D (F being the focal length of the objective and D its diameter), the exit pupil diameter 'd' and the apparent angle of field 'W' using S = 2*Fr*d*tan(W/2). Since the intermediate image is located in front of the ocular, the prism must be of proper size such that the light cone, which is growing even larger towards the objective, can be fully passed through the prism. The focal ratio Fr of hand held binoculars is usually restricted around a value of 4 and would not vary much. Therefore, S grows with the angle of field W and with the exit pupil size d. If the prisms are too small to accommodate the light cone, and the angle of field is already fixed, the designer is forced to stop down the exit pupil d instead, as was done with the Alpin. During this process, non-axial principal rays are cut off, leading to a reduction of the effective entrance pupil size, which means that the usable size of the objective is reduced to D = d*m (m being the magnification). As a result, the Optolyth Alpin has got a fairly wide field of view, but stopped down exit pupils.

Since he didn't mention the exit pupil sizes of the other two binoculars, one assumes they didn't differ markedly from what one would expect, i.e. D/P. Later he does suggest the Alpin might have been designated an 12x42, considering its effective aperture, but in my view the company was arguably not 'cheating' since physically there are 50mm objectives and 12x eyepieces. So, I agree with Bill. We are not in a position to know why the extra glass was used, but it might not be nefarious. Is it more deceptive to limit effective aperture with undersized prisms than it is to limit effective FOV with inadequate eye relief? The bottom line is the binocular ranked well against the other two used in the comparision.

Ed

[elkcub]

FWIW

I recently obtained ISO 14133-1 Optics and optical Instruments-Specifications for binoculars, monoculars and spotting scopes--Part 1 General Purpose Instruments and (14133-2) Part 2-- High performance Instruments.

I was very surprised (dismayed) at the allowable tolerances i.e. +/-6% for high performance exit pupil, 100/60 arc minutes (gen. purpose/HP) divergence in the horizontal plane for instruments <20x. 5%/4% for power, etc. Note, these are max. allowable from nominal, but if any of my instruments approached these figures I am sure I would dispose of them.

These figures, especially collimation, are greater by a large margin than I had considered acceptable.

Ron

Ron,

Interesting post. Of course, manufacturer's are not limited to allowable tolerances, and the high-end boys may set a higher standard than the industry as a whole.

I've been concerned for some time that 'nominal' specifications for power are, with rare exception, integer numbers, e.g., 7x, 8x, 10x. etc. Rounding to the nearest integer could, therefore, introduce real differences of as much as ±5% for 10x, ±6.25% for 8x, ±7.14% for 7x, and so forth. For example, anything in the ±5% range of 9.5+ to 10.5- might be called a 10x. Given nontrivial retinal offset differences between designs, the observer's size perception might not even make this apparent.

Assuming this were true, it seems quite possible to compare two nominal 10x42 configurations that differ by as much as 10% in power, reaching the erroneous conclusion that one has much better 'resolution' than the other, or that one is not quite as bright as the other. In the context of this thread, my suspicions are only heightened by Kimmo and Henry's observations about measured exit pupil variation, which could reflect such differences in power. For the case discussed, a 10x42 with -5% error would present an EP of 4.42 compared with an EP of 4.0 with +5% error. The nominal is 4.2.

Of course, these are simply comments, not indictments. :brains:

Ed

[henry link]

This subject has been thrashed about at Cloudy Nights binocular forum for several years. There are really two seperate issues which are easily confused, at least by me. I've seen no binocular that has prisms large enough to allow full illumination of the field edges by the entire surface of the objective. What is surprising is how the gradual loss of brightness toward the field edge goes unnoticed most of the time. In binoculars like that the exit pupil will measure as it should (if the magnification is accurate). However, some binoculars like the 50mm Alpins are worse. Their prisms are so small that even the center of the field receives no light from the edge of the objective. In that situation the objective is effectively stopped down and the exit pupil will be smaller than it should be.

I have to disagree a bit with Ed and Bill's view that there is nothing wrong with a 12x42 being marketed as a 12x50 since the objective lens really is 50mm. Buying decisions are often based on aperture. I don't think anyone will be happy to find they have decided on a 50mm binocular and actually bought a 42mm. In the case of the Alpins I suspect very sloppy design work. As for the Celestron Outlands, an aperture stop placed behind the objective has to be intentional. Why stop down a 40mm objective to 32mm instead of just using a 32mm objective? I can only guess that the targeted consumer for this binocular likes big numbers.

[APSmith]

When I'm looking at basic model information on the outside of the box, I know the information format is the work of Sales/Marketing/Advertising. I expect nominal numbers, or even a little bogus trade jargon. I'm okay if a 9.8x bin says 10x on box. A similar example is a "2x4" piece of lumber (although always smaller - i guess you could say that is "cheating").

However, when I'm reading a spec sheet, I expect that the information is of a technical nature, originating from Engineering/Manufacturing. Therefore,I expect reasonably accurate information. Since models are distinguished by as little as 0.5 power (8.5X, for instance), the actual specs should at least be given in tenths. If they are going to give a nominal number, the spec sheet should say "Nominal magnification".

To be fair, the numbers are typically not formatted with a decimal place, so we could assume it's only as accurate. i.e. The magnification is typically given in integers. So, for an 8x, it should actually be between 7.5 and 8.5. The problem is, other models are given in 0.5 increments (8.5 for example), with the implied accuracy of the nearest tenth. So, if it's not between 8.45 and 8.55, they're misleading. And if it's not between 8.26 and 8.74, they're just plain "lying". It's as if we're to assume it's between 8.0 and 9.0, which would be the the implied range of integer magnification data (+/- 0.5). This, of course, leaves way too much room for "cheating".

Should the "8.5"X42 EL box say 8x or 9x?

I don't suspect all this matters a whole lot. But, it would be nice to know the truth.

[henry link]

Using Kimmo Absetz method for measuring true magnification I haven't found many examples of seriously inaccurate magnification among my binoculars. Of twelve 8x binoculars I measured all but two fell between 7.9 and 8.1. A Zeiss 8x42 FL measured 7.8x. An old Kern 8x30 was 7.6X. My 8.5x42 EL measured about 8.4X.

Another thing to consider is that binoculars that focus by varying the distance between the eyepiece and the objective might show as much as a 4-5% difference in magnification between close focus and infinity focus.

[Surveyor]

Henry;

That pretty much mirrors my results using photo scale for power. Most fall in the 1-2% range but I have 1 Zeiss that comes in at 7.7x and a Trinovid that hit 8.24x.

[APSmith]

Henry/Surveyor,

Thanks for the measurements. With minor exceptions, it looks like we're pretty much getting the truth afterall. I had begun to wonder.

I would still like to see those type of actuals in the tech specs.

[elkcub]

I was moving on after concluding there would be no point discussing "cheating" any further. But in the back of my mind an old memory trace was nibbling away, which probably influenced why I felt comfortable saying in post #15 "...We are not in a position to know why the extra glass was used, but it might not be nefarious."

Steve Ingraham did a comparative analysis of 10x binoculars in 1993, including the 10x40 Optolyth Alpin. http://www.betterviewdesired.com/HighX.html.
The whole report is worth reading, but here are his comments about the Alpin:
Optolyth Alpin 10X40
Multi-coated, heavy ribbed rubber armor, good woven strap

It is difficult to imagine a more compact or lighter 10X40 glass. You simply forget you are wearing binoculars. They offer excellent optical performance, with minimal distortion at the edge of a very flat field, but at this power their "tiny" porroprisms seem to limit the brightness of the view. It is noticeable in full daylight...though it is not evident in their twilight performance (a true mystery...I suspect their multi-coating is exceptionally efficient at the twilight level). If you hate heavy binocs, (or have very small hands) but are not willing to sacrifice optical excellence or the larger image size of 10X glasses, these are the binoculars for you!

Sound familiar, "tiny" porroprisms? Apparently Optolyth has been "cheating" with undersized prisms for 14 years or more! Later on in his article Steve described the "Living-Bird Feather Resolution Test" (which is really why I was re-reading the article.) The results were:
...in order then, from the most apparent feather detail to the least: 1. Optolyth Alpin, Leica Ultra, Pentax PCF, Nikon Lookout, Celestron ED, Swift Ultralite, Bausch &;Lomb Elite, Zeiss, Bausch &;Lomb Custom, Swift Audubon, Nikon Execulite, Swarovski Traditional, Celestron Ultima.

My, my, how odd. There is little doubt that the 10x40 also had less than its nominal 40mm effective aperture, for essentially the same reasons as Holger's 12x50, but this design approach apparently was successful in meeting the ultimate criterion: 'resolving' feather details on a live bird (for Steve, anyway). It's beyond me how the extra glass might have influenced the outcome; maybe it's all part of the same "mystery" as the Alpin's twilight performance. ;)

Blue skies,
Ed