my first resolution measurement (1 Viewer)

[ronh]

Guys,
I hate to be a blabbemouth, but this little paragraph is imbedded (post #41) in the current thread on the Zeiss 8x42 FL that has turned into discussions of global economy and stuff. I was afraid you'd miss it, and would appreciate any feedback criticism and advice.

I saith:
"Yesterday, I did a centerfield resolution test comparing it to my Trinovid BA, booted to 64x. Weather limited my setup space, so the period of my finest target was a rather fat 3.4 arcsec. The target was alternating black and white stripes, running in four directions. Both barrels of both binoculars could discern which way all the lines ran, but there were differences in the quality of the view. The left barrel was better in both binoculars, and handled the target easily, but the Leica was a tad better. The right barrel of the Zeiss was a little bit worse on three directions, but still handled those three directions easily, but the fourth direction was worse, and I could tell which way the lines ran, but barely, so, some astigmatism there. The right barrel of the Leica was also worse than the left, but not much, and without astigmatism. Overall, the Leica was the close winner. By only barely getting the worst direction in the right barrel, I believe that barrel of the Zeiss could be said to be "resolve" 3.4 arcsec, and all the other barrels were better. Although not well quantified, I believe this is about medium resolution quality for an alpha binocular, the Leica a bit better than average, comments please optics buffs."

Thank you,
Ron

 

[rivergazer]

This is a very interesting thread to me, as it seems few manufacturers show resolution in their specs. I've heard that Nikon SE is the all time champ for resolution. Can anyone confirm or deny? Also curious about Zen ED(2)

marty

 

[shaocaholica]

Can you describe the test in detail for the layman here(me).

 

[ronh]

shao,
I'm new at it, but I read how other people did it, and thought about it a lot before I tried it, and here's what I did.

First, why do it? The low magnification of a binocular is insufficient to clearly reveal optical deficiencies that lurk at the fine scale around the ability of the eyesight to discern. This level of optical error, however, can have a large effect on the overall impression of sharpness, and the finest detail shown. This is the level where "something" about the view separates a cheapie from an alpha, but it's hard to put your finger on, the alpha just looks sharper somehow. The difference in view quality between two binos is influenced as much by contrast and brightness and who knows what all, as resolution, but resolution is an important component of it, and easy to measure compared to the other qualities.

So, how to. A second low power telescope, another binocular for example, placed behind the first, magnifies the image by another factor of x8 or so, and makes those imperfections or "aberrations" big enough to be easy to observe. For reasons that I won't go into, neither the quality of the second telescope, nor the acuity of your eyesight, is important to this test, and that is a great simplification.

So you need some easily adjustable mechanical arrangement to hold tested optic pointing at a target and the "boosting" optic fixed behind the one to be tested.

Then you need a target, typically one made expressly for testing optics, consisting of patterns of alternating white and black stripes of various sizes, each size having four targets with lines running in four directions, 45 degrees apart, 0-180 degrees.

Then you have to either follow the directions of how distant to set the target, or, like me, take a target of unknown dimensions, and measure it, and then set it so that the angle subtended by adjacent white stripes is a few arcsec.

(stripe separation/distance to target) x 57.3 X 60 x 60 = arcsec between stripes

So if you have a fancy target with finely graded sizes of targets, you just focus up and see what's the finest one that you can see which way the stripes in all four targets run, and that's the resolution, the target maker has done all the calculations for you.. Or if you only have one target, you adjust it's distance to where you can barely "resolve" it, and calculate the angle it subtends at that distance, and that's the resolution.

Finally, you need to interpret the result. The average eye can resolve black stripes at around 60 arcsec. Say the bino has a resolution of 3 arcsec, and a magnification of x8, so its resolution, or the size of its optical error, appears to the eye as 24 arcsec. This is so much better than the eye's ability to see, that the binocular will appear as sharp as can be seen on high contrast edges, like twigs against the sunset sky. If, however, the bino's resolution is 6 arcsec, this is increased by x8 to 48arcsec, and the edges of those twigs will start to look just a tad soft and blurry.
Ron

 

[mooreorless]

Here is an USAF 1951 test target that can be used for this. I mount it on a box using tape and set it on a lawn chair to a measured distance of about 75ft./90ft
Steve

 

[edz]

Finally, you need to interpret the result. The average eye can resolve black stripes at around 60 arcsec. Say the bino has a resolution of 3 arcsec, and a magnification of x8, so its resolution, or the size of its optical error, appears to the eye as 24 arcsec. This is so much better than the eye's ability to see, that the binocular will appear as sharp as can be seen on high contrast edges, like twigs against the sunset sky. If, however, the bino's resolution is 6 arcsec, this is increased by x8 to 48arcsec, and the edges of those twigs will start to look just a tad soft and blurry.
Ron

Just a few comments I would make about this.

few people are capable of seeing 60 arcseconds on line pairs. Some can get close. Most will probably fall nearer to 75-80 arcseconds.

When you test you binoculars at boosted power, if you are seeing 3 arcseconds, you are not seeing 3x8=24 arcseconds apparent. If you were seeing 3 arcseconds without the booster, then 24 arcsec apparent would be correct, but I know that is not the case. You must also multiply that by the power of the booster optics. You are probably seeing about 3 x 8 x 8 = ~180 arcseconds apparent. As power is increased towards maximum, minimum resolution (approx 3 arcsec) will go lower, but apparent resolution almost always will increase way beyond minimum. It always takes much greater apparent resolution to see smallest actual resolution near the limit.

You probably should test without the booster to find out what you can see. I'd suspect something near 10 arcseconds with an 8x binocular. I doubt you'll see 9 arcsec, and 12 would be rather poor. In fact you should probably test several binoculars without the booster to establish what your low limit is. You can then weigh normal res vs best achievable normal res, not to be overlooked as a useful measure. Then you can debate, if you wish, which reading is most useful to you, normal or boosted, keeping in mind, you will never ever use boosted magnification in normal practice. This does not say the boosted res reading should be ignored, only weighted as to its usefulness. If you've measured a binocular to have a boosted resolution of 3.8 arcsec, (and perhaps a dozen others are better), however, at normal power it shows the finest resolution of any binocular you've ever looked thru, you need to consider each. There are a whole group of discussions that can take place about potential errors and intended use. For instance if an error puts light outside the 3 arcsec disk at 60x, but at 8x the smallest disk you can perceive is 9 arcsec and all the error is within that disk, then it may not be an error at the intended use power.

If you intend to use a booster to test binoculars, and report max/min resolution, then first you need to test the booster lens for true power. The range of power in a 6x monocular or an 8x monocular can vary easily by 5%. So, do you really know what combined power you are using for this test? Not unless you've tested your booster lens.

I've recorded more than 10 (out of 60) binoculars ranging from 7x50s to 15x70s that all could see resolution of 3.4 arcseconds down to 2.4 arcseconds.

My four best, all 2.7 arcsec or better, are:
Fujinon FMT-SX 16x70
Oberwerk Ultra 15x70
Nikon SE 12x50
Nikon Action Extreme 12x50

Surprising, the $80 Nikon Action VII 10x50 porro measured 3.0 arcsec and the $299 Nikon Monarch ATB10x42 Roof measured 5.7 arcsec.

edz

 

[mooreorless]

"I've recorded more than 10 (out of 60) binoculars ranging from 7x50s to 15x70s that all could see resolution of 3.4 arcseconds down to 2.4 arcseconds."

Hi Edz, The binocular that you say you could see down to 2.4 arcseconds, was this Nikon AE 12x50 binocular you used to have? If so it sounds like it beat out 70mm binoculars on your list. I wish I would of bought it off you.
Regards,Steve

Edit: using Google Cloudy Nights search I did find what you said.
"Using the same Nikon AE 12x50, with a 6x monocular for a net 72x, I could see 3.4 arcsec easily, 3.04 arcsec clearly, 2.7 arcsec barely and 2.4 arcsec suspected. Apparent resolution values are 245, 219, 195, and 174. "

So this binocular was very close to the Rayleigh limit for 50mm, that is astounding:t: Get a good 2 1/2 x extender and use this one for a spotter at 30x

 

[edz]

Hi Steve,

those four best are in this order, although I would say that really indicates they are all very good. These results include readings barely achieved, but not suspected.

Fujinon FMT-SX 16x70 2.4arcsec
Nikon SE 12x50 2.56
Oberwerk Ultra 15x70 2.57
Nikon Action Extreme 12x50 2.7

edz

 

[mooreorless]

Hi Edz, I would say the two 50mm binoculars have to be excellent. Something you said on CN says these 50mm binoculars with the boosters exceeded Rayleigh limit and the 12SE by a wide margin.

"Sure a 50mm lens has a rayleigh Limit of 138/50 = 2.76 arcseconds"

but of course this is line pairs and not point sources here, but all the same these have to be super duper binoculars
Regards,Steve

 

[edz]

Yes those are excellent binoculars. In fact, in one article, I referred to the Nikon Action Extreme 12x50 as the poor man's Nikon SE.

However,

Can't compare Rayleigh resolution limit and line pairs resolution as equals. Rayleigh limit is measured on point sources.

Line pairs resolution will almost always be on the order of about only 60%-70% as wide as Rayleigh.

For example, I can resolve a best with my top dozen binoculars, at normal power, of 80 to 85 ( and two under 80) arcseconds apparent on line pairs. Those same dozen binoculars at best allow me to see 135-150 arcseconds apparent on star pairs.

Another example. I have two very fine 80mm scopes, one a premium Apo and the other a "semi-premium" semi-apo. Both have a Rayleigh limit of 1.7 arcseconds. Both have resolved star pairs between 1.65-1.72 arcseconds. BUT both have resolved line pairs down to 1.15-1.2 arcseconds. We can see by this both scopes can considerably exceed the Rayleigh limit when observing line pairs. FWIW, I've tested the Raleigh limit of a dozen scopes, of which 5 are premium. None have exceeded the Rayleigh limit by more than a few percent.

Seeing a 50mm binocular that resolves line pairs to 2.6-2.7 arcseconds would seem to indicate that at best, with the proper boost, that same binocular would be able to resolve star pairs no closer than 3.7 arcseconds, not anywhere near the diffraction limit.

For a 50mm binocular to be considered as having been able to reach the Rayleigh limit, it would have to resolve line pairs down to about 1.8 to 1.9 arcseconds. Nothing has ever even come close.

edz

 

[ronh]

Ed,
You said "You probably should test without the booster to find out what you can see. I'd suspect something near 10 arcseconds with an 8x binocular. I doubt you'll see 9 arcsec, and 12 would be rather poor."

In fact, with the 8x Zeiss, I can see 12 arcsec easily, but the next finest target, 8.8 arcsec, I could not see, just as you say. I was wrong about 60 arcsec being reached by the naked eye. Pardon me, I thought I had read that.
Ron

 

[edz]

In fact, with the 8x Zeiss, I can see 12 arcsec easily, but the next finest target, 8.8 arcsec, I could not see, just as you say. I was wrong about 60 arcsec being reached by the naked eye. Pardon me, I thought I had read that.
Ron

You may have read that Ron, it is written in many places. In fact it is commonly referred to as the normal resolution of the eye and it is associated with 20/20 vision. However that is for single line resolution, not line "pairs". For instance the capital letter E on a vision chart, sized for 20/20, has lines that are all 1 arcmin thick, or 60 arcseconds thick. However, the entire letter E is 5 arcminutes tall, and observing from center to center of any line pair is 2 arcmin. Although there are a number of other tests, most not generally referred to, common instrument resolution testing in daylight, as you've done, uses line pairs.

edz

 

[mooreorless]

Ed,
You said "You probably should test without the booster to find out what you can see. I'd suspect something near 10 arcseconds with an 8x binocular. I doubt you'll see 9 arcsec, and 12 would be rather poor."

In fact, with the 8x Zeiss, I can see 12 arcsec easily, but the next finest target, 8.8 arcsec, I could not see, just as you say. I was wrong about 60 arcsec being reached by the naked eye. Pardon me, I thought I had read that.
Ron

Hi Ron, Did you try moving the target further away and still be able to see the same line as in seeing between 12 arcsec and the next at the original distance not being able to see 8.8 arcsec? In other words it might be closer to 11/10/9 instead of the 12 arcsec.

I don't think anybody actually answered Shaocaholica question completely. It seems like I am on Cloudy Nights and IMO there is too much information that new people would not understand and give up. I probably didn't help either.

Edz, Maybe explaining how you come up with this [below]would help new people. The procedure you would of used [moving the test target]. The use of apparent resolution on here would add to the confusion of people that are not used to it. Thank you for any more information.:t:

"Using the same Nikon AE 12x50, with a 6x monocular for a net 72x, I could see 3.4 arcsec easily, 3.04 arcsec clearly, 2.7 arcsec barely and 2.4 arcsec suspected. Apparent resolution values are 245, 219, 195, and 174. "

Ron, Thanks for taking the time to actually measure and post on here.:t:

Regards,Steve

 

[edz]

Actually Steve, there are reams of posts on this forum regarding resolution. Look for discussions by Kimmo Absetz. Also search for apparent resolution, as it's all been explained here before, in far more detail than I'd care to get into here. With a bit of search, all the pertinent information is here to be found.

FWIW, apparent resolution, the actual resolution reading multiplied by the power at which it is obtained, gives a method to compare binoculars of different powers. You cannot compare resolution from one power binocular to another directly since power has the greatest impact on the resolution reading obtained. If you use two binoculars both raised to 60x, then they can be compared. If you test one at 48x and the other at 72x, they cannot be compared, or for that matter you can't compare normal resolution readings at 7x to 8x to 10x to 12x. Apparent resolution puts all of them on the same plane, and then they can be compared.

If you ever get the opportunity to gather test data on several dozen binoculars, you will find that normal power apparent resolutions range from your best (probably limited by your eyes), which may be 70 to 80 arcseconds, all the way up to 95-100 arcseconds for the poor binoculars. I've seen a best of 76 arcsec appar., but have only seen a half dozen or so binoculars better than 82 arcseconds appar. I've seen maybe 15-20 that could not see resolution any better than 90 arcseconds.

Also FWIW, don't ever expect high powered resolution readings to come close to best apparent resolution. As you move towards the limits of resolution, power needed to see it increases rapidly. While a best apparent resolution at low power might be say 10 arcseconds at 8x for 80 arcseconds apparent, at ultimate resolution of say 2.5 arcseconds, appar. resolution is very likely to be approaching 3 arcminutes. In a telescope, when attempting to achieve diffraction limited resolution, it is not uncommon to need powers that might result in 5 arcminutes apparent resolution.

My 16x70 Fujinon did not achieve maximum resolution of 2.4 arcseconds until I reach 98x. That's an apparent resolution of nearly 4 arcminutes.

Ron, I'm guessing you are not using the USAF resolution bar chart, since you described a jump from 12 arcsec to 8.8 arcsec. The USAF resolution charts have bars in 12% size increments, so the progression would be from 12 to 10.6 to 9.3 to 8.2

edz

 

[mooreorless]

Well Edz I didn't want this information for myself

"The right barrel of the Zeiss was a little bit worse on three directions, but still handled those three directions easily, but the fourth direction was worse, and I could tell which way the lines ran, but barely, so, some astigmatism there."

I figured Ron wasn't using USAF resolution chart from this post

 

[ronh]

Nope, Steve, I didn't push it to see what I could actually see, I only know it's somewhere between 8.8 and 12 arcsec with 8x

Thanks for the info, Ed, now what I recall reading makes sense. Right, I am using a target from Henry Paul's 1965 booklet "Binoculars and All Purpose Telescopes", which I measured with a Bausch and Lomb 7x comparator loupe. In determining the target periods, I measured across several gaps and divided by the number of gaps for best precision. Four targets, angled 45 degrees apart, is pretty nice, I see that Steve's only has horizontal and vertical stripes. If not for the angled stripes, I would not have learned that my new Zeiss has a tad of astigmatism. I really needed to know that, didn't I?

By the way, that comparator is a fine little instrument, a Hasting's Triplet, showing a beautiful flat field appearing 40 degrees wide. A variety of reticles are available, but mine has a simple 10mm scale marked in 0.1mm increments. It is perfect for measuring exit pupils etc. No optical household should be without one!
Ron

 

[mooreorless]

Hi Ron, A star test would show astigmatism and it would be possible to mount "tape" the USAF target at an angle. I am sure you know about star test etc. and am just posting this for other people that might not know. Your target sounds interesting, is there just the one size spacing? I have seen pictures of a target that Zeiss uses that sort of sounds like the one you used. BTW I bought my lens as a lens testing kit off Ebay, came with 7 lens to test camera lens. I don't have 7 left, I gave some away to some friends as far away as UK.
Regards,Steve

 

[ronh]

Steve,
Or hold the bino crooked. Good guess, as Dr. Paul credits Carl Zeiss for the chart. Unfortunately, the equipment I used to support the binoculars in my boosted magnification test was borrowed, now returned, and I am now sad to say down to judging the new Zeiss in "use". It's the best I've ever seen, man what a clear view.
Ron

 

[CLRobles]

Now Ron we just have to put them along side my SE's and see with one resolves better

 

[mooreorless]

It's the best I've ever seen, man what a clear view.
Ron

I am glad you are happy with the Zeiss. I thought that you were not happy with it for a while.
Regards,Steve