What determines our color sensitivity as looking through bins? (1 Viewer)

[BobinKy]

Henry--

This is good news on your test methodology.

Now, to be sure, do the four color squares in the above photo reflect the results of your research on the four models you listed?

Or is this just the format you will use and the actual color differences will be posted at a later date?

Thank you for your work and dedication on this project.

--Bob
Kentucky, USA

 

[henry link]

Bob,

The central squares are crops from actual photographs made with the binoculars placed in front of the paper like in the Swarovski example above. The background around the edges is an actual photograph of the paper with no binocular in front of it made at the same time. Exposures are all identical. A perfect binocular with 100% transmission and no color bias would produce a square that looks just like the background. I found I didn't need Photoshop. I just assembled the crops into a Powerpoint slide and saved it as a jpeg.

It's hard to believe that the old binoculars are really that dim, but thats what the camera recorded. The eye is easily thrown off by context. If I look at those squares in isolation and surrounded by black as through a binocular they don't look dark at all.

Henry

 

[Kevin Purcell]

It's hard to believe that the old binoculars are really that dim, but thats what the camera recorded. The eye is easily thrown off by context. If I look at those squares in isolation and surrounded by black as through a binocular they don't look dark at all.

This is an example of the Simultaneous Contrast Illusion or the Contrast Effect

http://en.wikipedia.org/wiki/Contrast_effect
http://en.wikipedia.org/wiki/File:Gradient-optical-illusion.svg
http://en.wikipedia.org/wiki/White's_illusion

Very interesting work, Henry.

Do you have a RAW mode on the camera? Minimizing the amount of processing (especially lossy conversions like JPEG) of the image might help in making measurements.

 

[Pileatus]

Henry,

When I cropped out the 8X30 E and the Zeiss 8X50 the FL and SE were visually indistinguishable from one another. Interesting.

John

Just a quick update on this project for those who are interested.

The photo below is my tentative format for showing the test results. The background color around the edges is sunlight reflecting directly back to the camera from a white piece of paper. The central squares are the reflection from the very same piece of paper after traveling through four different binoculars. The bottom left is a Nikon 8x30 E from 1985 with single layer MgF coating, typical of that generation of coatings. The upper left is a Carl Zeiss/Jena 8x50 Octarem from 1987. It had an early multi-coating called "T3M". The upper right is a Zeiss 8x56 FL and the lower right is a Nikon 8x32 SE.

I know everyone is actually interested in comparisons of the current alpha binoculars. That will have to wait until I can make a trip to a dealer who has generously humored me in things like this before.

 

[BobinKy]

Henry--

Impressive!

I hope you do not mind, but I posted a link on Cloudy Nights to this thread.

--Bob
Kentucky, USA

 

[ronh]

Henry,
That is really an interesting thing you have done, thank you. Don't hurt your back doing flips of joy over your big FL, then you couldn't carry it!

I bet Photo shop or a similar program could be used to analyze these images quantitatively, for transmission vs color, and overall vision-weighted transmission.

Great work!
Ron

 

[Matt_RTH]

Excellent test. I have a couple gems from yesteryear that I want to try this on.

 

[Kevin Purcell]

I bet Photo shop or a similar program could be used to analyze these images quantitatively, for transmission vs color, and overall vision-weighted transmission.

The problem is that you only get three points on the spectrum (RGB) as this isn't a spectrometer it's a camera.

The other issue is knowing exactly where and how wide these three "receptors" (color bands) are. I would presume they would have to be fairly close to the average human positions (so the colors map) but being colorblind (RG deutroanomalous ... so my green receptor color max is shifted to the red) I know how this things can differ.

That will vary from camera to camera so you have a problem with comparing different results from different users.

As a technique as Henry has presented it (for comparisons between bins with the same camera) it is an interesting technique. We just have to send bins to Henry to get them done

 

[edz]

As some have alluded to above, I wonder how much this represents differences in transmission vs how much this represents color cast? It seems a good way to visually detect differences in total transmission. Of course, wouldn't it also seem that the difference in total transmission results in a different perception of color cast?

edz

 

[edz]

To support comments that coatings make a difference in color cast (and obviously total transmission), I'm posting two paragraphs here from an article by Roland Christen, noted optics expert and Owner Astro Physics.

Comparative Coatings and Obstructions
by Roland Christen
The test for a really good coating is to set it outside in the shade with the eye end up so that light from the sky can flood the optics. The other end should have a black end cap attached. Look straight down into the eyepiece to see how much of the light is coming back out at you. This is the light that is reflected back from the various coated surfaces, light that should have gone through and be absorbed by the black end cap. The darker the optics look, the more light gets through to your eye, and the less is reflected back or scattered. The really bad oculars literally glow with reflected light.

Although all the eyepieces in this picture are "multicoated", they are not optimized. The different elements in a typical wide angle ocular are made from different index glass. For each index there exists an optimum coating design, one that will achieve the lowest reflectivity for that element. For optimum efficiency, each element must be coated with a different coating design according to its refractive index. What is often done is to coat all elements with the same formula. This results in some wavelengths getting attenuated more than others, usually in the blue end of the spectrum. These eyepieces will show a distinct yellowing of the image, as well as an overall lower transmission efficiency. This can be seen in the second image where I have compared the 25mm Zeiss Ortho with the 3rd ocular. You can see a definite yellowing of the white wall behind the ocular, as well as a faint reflection of my head and sky behind me, which is completely missing in the Zeiss.

http://geogdata.csun.edu/~voltaire/roland/coating.html

 

[edz]

These excerpts from another RC article might stir some thought on two points; 1)using a camera to detect color cast - the main point here being that the camera detector is much different than the eye as a detector, and 2) color correction (degree and bias of chromatic aberration) of a lens has some influence on the color cast of the output. Considering the very fast optics of binoculars, CA bias may influence color cast to some level.

Color Correction in Refractors
by Roland Christen

In lenses, the key (to color correction) is the color of the in-focus disc. The whiter it appears at focus, the better is the correction. In a lens with lesser color correction, the in-focus disc will appear more and more yellowish

To clarify the situation on color correction a bit more ... I will attempt to explain where achros, EDs, fluorites etc fit into the scheme of things by comparing the ability of each design to produce a reasonably focused image spot diameter over its wavelength range.

Fast 6"F8 Cde achromat: 550 - 650 nm
Fast 6"F9 ED doublet: 450 - 650 nm
Fast 6" fluorite doublet: 420 - 1000 nm
Fast 6" fluorite triplet: 360 - 1000nm

...why would you need correction well into the blue-violet past 480nm? With black and white emulsions, this was necessary because they have considerable sensitivity down to 380nm. Today's new blue sensitive CCD cameras also need good correction in the violet. Also, CCD cameras pick up lots of IR light below 650nm, so correction to 1000nm is a distinct advantage. For pure visual use, it would be quite sufficient if the useable range extended only from 440 to 650 nm.

See post above for link to Astro Physics articles.

 

[falcondude]

edz, thanks for posting your comment and link.

 

[henry link]

Thanks for all the interest. I haven't been able to get back to this thread today. A neighborhood rezoning dispute is on the front burner. Maybe tonight.

Henry

 

[elkcub]

Ronh; ...I bet Photo shop or a similar program could be used to analyze these images quantitatively, for transmission vs color, and overall vision-weighted transmission.

The problem is that you only get three points on the spectrum (RGB) as this isn't a spectrometer it's a camera.

The other issue is knowing exactly where and how wide these three "receptors" (color bands) are. I would presume they would have to be fairly close to the average human positions (so the colors map) but being colorblind (RG deutroanomalous ... so my green receptor color max is shifted to the red) I know how this things can differ.

That will vary from camera to camera so you have a problem with comparing different results from different users.

As a technique as Henry has presented it (for comparisons between bins with the same camera) it is an interesting technique. We just have to send bins to Henry to get them done

It seems to me that ronh makes a valid point with regard to analyzing the color content of the pixelated photos. This is different from analyzing the light emerging from the instrument's exit pupil, which would require a spectrometer. But, if one is going to make such digital recordings in the first place, it's probably a good basis for making an objective comparison; rather than relying on individual color perceptions that are influenced by biological factors, color contrast, and the color bias of the monitor.

Using Adobe ImageReady for a quick evaluation, the RGB color mix of each panel is shown in the attachment. Weightings vary somewhat over each area, but the main result is clear. (I'm too lazy to take an average of several readings, so I picked the center.) The background is slightly weighted towards blue, and brighter than any of the four panels. The average of the three RGB background values is 223, suggesting that exposure was 223/255 = 87.45% of maximum, — since on this scale pure white = 255 and pure black = 0.

On individual panels, the Zeiss 8x50 image is biased green-red, the Nikon E image is biased red-green, the Zeiss FL image is almost neutral but slightly green, and the SE image is neutral. These results represent the binoculars' filtering of the blue-biased background image.

By expressing the average of the three RGB values as a percentage of the average background value (i.e., 223), the transmission efficiency of the four binoculars can be estimated as:

Zeiss 8x50 = 172/223 = 77.1%
Nikon 8x30 = 162/223 = 72.6%
Zeiss 8x56 = 214/223 = 96.0%
Nikon 8x32 = 213/223 = 95.5%

This seems more or less reasonable to me. The reliability of the procedure can be determined by others doing the same thing with their Adobe or similar software.

Ed

 

[Matt_RTH]

Amazing, Ed. I'll post my results if I can get decent results. BTW, on the 8x30, is that a later era ('90s) or early/mid '80s?

Matt

 

[Matt_RTH]

Well I was mistaken! Now that I actually see the results, I see the photos were taken with the binoculars reversed! I thought these were basically digiscopes through the bin, which to me would be more indicative of actual use, but I see why taking the photo through the objective to the ep would be more approachable.

Well I'll have to do it again but just for fun, I'll post my images. I tried to keep everything as consistent as possible, but didn't stabilize the camera.

ISO 1250
1/50 shutter
F5.6 aperture
135mm FL
White balance = incandescent
Manual focus and attempting to focus on the eyepiece of the bin.

I intentionally took shot of a less than bright-white surface as I was attempting to check brightness (as opposed to color which was the original topic here, but the results have gone the way of brightness).

I may replace this photo once I replicate the other previous tests but to clarify, this is through the ep to the objective.

Ignore any odd colors like in the Bushnell photo because it was a reflection.

Matt

 

[elkcub]

Amazing, Ed. I'll post my results if I can get decent results. BTW, on the 8x30, is that a later era ('90s) or early/mid '80s?

Matt

Henry said it was a "...Nikon 8x30 E from 1985 with single layer MgF coating." I own one just like it. It was followed by an E model with multi-coating that was similar to the E2.

Ed

 

[Matt_RTH]

Gotcha. Somehow I thought we were talking about different test specimens. I'm going to have to try in daylight. My histogram looked good but somehow I think my results are still way too dark and probably compressing values.

 

[ronh]

Ed,
The colors breakdowns are consistent with appearance for the darker ones. I saw a German hunting magazine that had an independent measurement of the 8x56 FL transmission as 96%, exactly your value. Your analysis gives a convincing result, and one of the greatest interest to binocular fanciers.

All the color vs transmission curves that I have seen for binoculars are very smooth, no surprising wiggles or dips. If you knew the frequencies of the R, G, and B, a simple spline fit to the transmission at those three points would certainly be a good approximation of the true curve.
Ron

 

[henry link]

Very promising, Ed! I hoped somebody who actually knows what he's doing with cameras and computers would run with this.

I wonder if your software could be used to correct the overall blue bias in the test by dropping the blue channels in the image by 6? (Is channels what you call them?)

Today it's car repair. I hope to make some comments this afternoon or maybe I'll just let everyone else continue to do all the work for me.

Henry