The 10X42 HT compared to the SF and SV (1 Viewer)

Vespobuteo · Jul 16, 2016

On basis of the discussion we have had I have tried to refine the method for simulating color bias in binoculars.
Please note that this is experimental.

I used Henry's comparison photo posted above and measured color bias (tint) and white balance temperature (WB) for the different patches/bins.
I got more conclusive results this time with bigger differences in WB (warmth/balance between red and blue)
and also in tint (amount of of green-pink bias).
Probably because the light Henry used in his setup was more adequate (i.e. daylight).

Finally applying these binocular specific value, on the portrait test photo below, after some slight transformation.
Tint must be applied in reverse for example.
In lack of a HT test patch the FL gets to represent Zeiss in this example.
Attached is a comparison between the simulated look of (from left to right):

1) Nikon EII
2) Zeiss FL 8x56
3) Reference photo.

Seems that the photo gets resized, but hopefully you will see the difference, even in this small format.
Skin tones are often very good for detecting color casts in photos.

typo · Jul 17, 2016

My memory might be unreliable of course, but Henry, those patches ar looking pretty close to me. Well done! I note you thought Powerpoint might have altered something. Perhaps you could just put up crops of the untouched originals?

Vespobuteo, the differences in those photos definitely look more in line with what I would expect than the previous set, but I still get a sense that the colour balance still feels a little off, particularly on the EII image. Could the camera sensor filter mosaic be different from RGB or something? Probably just my imagination.

I'll try to look at the numbers later.

Cheers,

David

Vespobuteo · Jul 17, 2016

typo said:
My memory might be unreliable of course, but Henry, those patches ar looking pretty close to me. Well done! I note you thought Powerpoint might have altered something. Perhaps you could just put up the original crops?

Vespobuteo, the differences in those photos definitely look more in line with what I would expect than the previous set, but I still get a sense that the colour balance still feels a little off, particularly on the EII image. Could the camera sensor filter mosaic be different from RGB or something? Probably just my imagination.

I'll try to look at the numbers later.

Cheers,

David

Yep, I think it's getting closer.
When you say off, in what direction, too weak reds?

Henry's camera should have a Bayer RGB color filter array, if it's a Nikon, I recall he used a Nikon DSLR in test shots before. Most cameras use a Bayer filter but Fujifilm have used other color filter arrays. Personally I think their cameras are superior in rendering color. Especially skin-tones.
There are some info on the Fujis here:
https://en.wikipedia.org/wiki/Fujifilm_X-Pro1

Btw, the reference photo I used is not perfectly white balanced. Probably better to have a neutral reference.

Attached another series, perhaps a better photo for comparison, same order of bins.

typo · Jul 17, 2016

Vespobuteo said:
....... Digital cameras are not spectrometers obviously.......

But maybe they could be. At least the budget variety of spectrometer uses the same kind of sensor as a camera. Instead of a usual rectangle it's a stip a few pixels wide. A simple diffaction grating splits the light and the image is simply corrected by a reference to a calibrated light source. It's certainly within the realms of the decent hobyist to turn a camera in to a crude spectrophotometer. The filter mosaic will make it low definition of course, and you would need a camera with a removable lens and a decent sized sensor. A lot wizer to buy a cheap spectrophotometer than muck about with an expensive SLR I suspect. The tricky, and potentially expensive bit is sorting out the optics.

David

typo · Jul 17, 2016

Vespobuteo said:
Yep, I think it's getting closer.
When you say off, in what direction, too weak reds?

Curiously, my feeling is that the green is a fraction low, but it's a while since I last tried the EII so probably a comment best ignored. I'll see what the numbers say.

I mentioned the sensor masks because the remapping of the colour separation to a screen display caused some analysis artifacts in a scientific project a few years back but don't remember the details now.. We ended up using an industrial monochrome cameras and custom dichroic filters. May not apply here as you say.

David

Vespobuteo · Jul 17, 2016

typo said:
But maybe they could be. At least the budget variety of spectrometer uses the same kind of sensor as a camera. Instead of a usual rectangle it's a stip a few pixels wide. A simple diffaction grating splits the light and the image is simply corrected by a reference to a calibrated light source. It's certainly within the realms of the decent hobyist to turn a camera in to a crude spectrophotometer. The filter mosaic will make it low definition of course, and you would need a camera with a removable lens and a decent sized sensor. A lot wizer to buy a cheap spectrophotometer than muck about with an expensive SLR I suspect. The tricky, and potentially expensive bit is sorting out the optics.

David

I don't doubt that, things tend to get tricky when colors are involved.
:eek!:

DXO seem to measure color response now in their camera tests. That was new to me.
(They measure a lot of stuff actually)

http://www.dxomark.com/Cameras/Nikon/D810---Measurements

Gijs van Ginkel · Jul 17, 2016

Vespobuteo, post 103,
The picture in the left through the Nikon EII is fully in line with the measured transmission spectra, which show a clear higher plateau in the red (600-675 nm) compared to the blue green (450-550 nm). Also visually the Nikon EII shows a red bias, so we can say with the Romans "quod erat demonstrandum".
Gijs van Ginkel

typo · Jul 17, 2016

I've come up with some numbers. I took a rectangular area within each image and used an old scientific image analysis package to come up with RGB values. Vespobuteo's photos it wasn't easy to come up with exactly the same area, but I was close. I think when you look at the results you will see that small differences aren't going to make much difference to the conclusions.

In order to make the numbers comparable I normalised the red and blue channels against the green channel. the result is how the relative levels of red and blue change between the images. The baseline values will be dictated by the light at the time. It's the difference between the models that's interesting.

Henry's photo comparison #98
Red Blue
background -2.6 -9.53
Habicht -3.3 -12.3
EII 2.7 -13.9
FL -3.3 -17.1
SE 1.45 -17.8

Vespobuteo #101
Original 17.1 -18.8
EII 15.5 -13.1
FL 12.7 -19.1

Vespobuteo #103
Original 17.2 -14.5
EII 15.0 -8.2
FL 13.4 -15.2

Just to pick out a couple of points. In Henry's image, the FL had about 6% less red than the EII which hardly sounds surprising, but it also 3.2% less blue than the EII which might surprise some. It means the blue level in EII shot was higher or the FL value lower than we might have expected from the transmission plots. I think this simply shows up the difficulty of using daylight. A small difference in the density of the cloud cover between the shots might do that unfortunately, but there could be other reasons.

Vespobuteo's #101 photos show the difference in the red between the FL and EII roughly half that of Henry's at 2.7%, but the difference in blue is nearly twice as big at 6%. I #103 the difference in the red is down to 1.6% and the blue 7%.

Something doesn't seem quite right with these images. Need to think about this some more, but I would welcome any suggestions.

David

Vespobuteo · Jul 17, 2016

David: I had to WB correct the background first and then each patch before normalizing WB and tint relative to the background values. Perhaps it may have some impact. Sunlight might not be perfect either of course.

chill6x6 · Jul 17, 2016

SuperDuty said:
From what I saw in my initial viewing session, I REALLY like these, my favorite in comparison to the 10X42 SF, 10X42 and 50 SV, and 8.5 X42 SV. The things I liked were superb CA correction-significantly better than the SVs and noticeably better than the SF, noticeably brighter while maintaining beautiful color and contrast, lightweight and super easy to get a steady hold, excellent focus knob action, very (sharp) or at least very sharp appearing. The things they didn't do as well,( less FOV compared to the SF, more pincusion than the SF and SV, fuzzier edges than the SF and SV), we're far outweighed for my taste by the positives. Another thing of note, as the sun became low in the horizon behind me, I saw none of the yellow hazy effect that I noticed in the SF. These observations are based on one session during almost perfect conditions, I'll see what I think after a little more use.

Robert

Robert...
Today I set up my CA test target. I looked thru 10 binoculars... You know what the best was in this regard? It was the newly acquired Nikon EDG 8X42. Second best went to a newly acquired used Zeiss FL X42 BUT it was kind of back and fourth between it and the SV 8.5X42. NEXT was my HT 10X42.

I have to LOOK for CA. I have to MAKE it appear. Normal viewing, I hardly ever see it...at all. #10 was my Leica Trinovid 10X42 which to me is a FANTASTIC BINOCULAR because in normal use, I hardly ever see it unless I'm TRYING.

I thought this would interest you and those that are CA sensitive. Optically, that EDG....oh MY! :t:

typo · Jul 17, 2016

Vespo,

It might seem very odd to reduce a very attactive subject to a shade of grey but I guess numbers are a bit abstract. I've needed tolook up the HEX codes for the RGB values I generated your #103 images. Try entering the following values in the FFFFFF boxes in the program below.

EII 656058
FL 6D6051
Original 705F51

Seems to have drifted a bit from Henry's original, but I'm afraid have no idea how.
http://www.colortools.net/color_compare_colors.html

David

typo · Jul 18, 2016

Just ignoring the white balance for now, and other variables such as the light, Henry's EII patch appeared redder than the background. It had relatively more red and less blue. The FL patch looked green. It had relatively less blue and less red than the background.

Now in the #103 the EII shot the reds look redder and the blues look bluer than the original and the FL now looks it now just has a fraction less red but no change in the blue. That doesn't appear to reflect Henry's original or what we would expect from the filtering effect of the binocular transmission profile. The EII shot in particular is looking pretty weird to me now. The RGB is too crude a tool to reflect what seems to be acolour by colour non-proportional transformation, but it does show the shift in the blue is massive in the EII shot. It's changed from being relatively negative in the Henry's original to now being significantly positive.

I might be being dumb, but however you shift the white balance (or colour temperature) the EII's shouldn't the relative red content should always be higher than the original and the blue lower? For, the FL both red and blue should be lower by comparison. Somehow that seems to have got lost in the processing. Don't really understand what you've done, but at a guess fewer variables not more?

Do you think we might have gone as far as we can with this?

typo · Jul 18, 2016

I don't know what the tint tool does but in the FL shot it looks like it increased green rather than decreased red and particularly the blue. With the EII it looks like it increased red without reducing the blue, but I'm viewing these on a tablet which doesn't have the best colour rendition. I'II try to look at the numbers later.

David

Vespobuteo · Jul 18, 2016

typo said:
I don't know what the tint tool does but in the FL shot it looks like it increased green rather than decreased red and particularly the blue. With the EII it looks like it increased red without reducing the blue, but I'm viewing these on a tablet which doesn't have the best colour rendition. I'II try to look at the numbers later.

David

The white balance adjustment is similar with combining two glass filters. Tint and Temperature. Tint corrects the green/pink cast. Temp corrects the blue/red cast. Attached is the test photo now with both tint and temp applied. Note that the difference from #103 photos is that the temp correction is positive instead of negative. I think that it will work better at least it seems to be more correct looking at it.

Tint:EII: +4, FL: -9
Temp:EII: +380K, FL: +300K.
Order from left to right: EII, FL, reference.

SuperDuty · Jul 18, 2016

Hi Chuck

That's interesting, tell me more about the EDG and FL compared to the SV and HT. Do you see allot of distortion -(bowed vertical and horizontal edges) with the FL ?

Robert

chill6x6 said:
Robert...
Today I set up my CA test target. I looked thru 10 binoculars... You know what the best was in this regard? It was the newly acquired Nikon EDG 8X42. Second best went to a newly acquired used Zeiss FL X42 BUT it was kind of back and fourth between it and the SV 8.5X42. NEXT was my HT 10X42.

I have to LOOK for CA. I have to MAKE it appear. Normal viewing, I hardly ever see it...at all. #10 was my Leica Trinovid 10X42 which to me is a FANTASTIC BINOCULAR because in normal use, I hardly ever see it unless I'm TRYING.

I thought this would interest you and those that are CA sensitive. Optically, that EDG....oh MY! :t:

typo · Jul 18, 2016

Vespo,

I'd acknowledge the images cosmetically 'look' the best fit so far, but sorry I was unaware how far the exercise had drifted from a quantitative scientific adjustment. To my eyes the various previous images had been obviously wrong and I wanted to uderstand how it could possibly happen. Didn't you find it obvious? It sounds to me like tint is just an aesthetic tool for adjusting skin tone and temperature correction is based on something much more simplistic than the colour temperature algorithm. I don't think they are capable of performing this job with any accuracy. I think you need to go back to the original RGB and mathematically correct the illustrative images. That may not be possible with your software.

Time for me to stop on this one I think. Good luck.

Chuck and Robert apologies for the distraction.

David

SuperDuty · Jul 18, 2016

No problem, learning new things is what it's all about. I'm having fun in just looking at the best available before settling on one binocular, not sure when I'll be done, maybe never.:-O

typo said:
Vespo,

I'd acknowledge the images cosmetically 'look' the best fit so far, but sorry I was unaware how far the exercise had drifted from a quantitative scientific adjustment. To my eyes the various previous images had been obviously wrong and I wanted to uderstand how it could possibly happen. Didn't you find it obvious? It sounds to me like tint is just an aesthetic tool for adjusting skin tone and temperature correction is based on something much more simplistic than the colour temperature algorithm. I don't think they are capable of performing this job with any accuracy. I think you need to go back to the original RGB and mathematically correct the illustrative images. That may not be possible with your software.

Time for me to stop on this one I think. Good luck.

Chuck and Robert apologies for the distraction.

David

james holdsworth · Jul 18, 2016

SuperDuty said:
Hi Chuck

That's interesting, tell me more about the EDG and FL compared to the SV and HT. Do you see allot of distortion -(bowed vertical and horizontal edges) with the FL ?

Robert

I own both the FL and HT and both show about the same level of pincushion - which is about the same as most of my BGAT's, which is about the same as most of my other bins.

It may jump out at you when looking at poles and buildings but it is almost never an issue when birding.

SuperDuty · Jul 19, 2016

Thanks

chill6x6 · Jul 19, 2016

SuperDuty said:
Hi Chuck

That's interesting, tell me more about the EDG and FL compared to the SV and HT. Do you see allot of distortion -(bowed vertical and horizontal edges) with the FL ?

Robert

FL>EDG but only slightly so. I have only had both binoculars a VERY short time...FL<wk and EDG<2 weeks...so I haven't spent all that much time with them.

This is the first 42mm FL I've owned or even handled. I can't help but ask myself, "Why the HT?" The FL is lighter, smaller, and really nothing to complain about so far. Too early to tell but it may very well be my favorite Zeiss binocular...

The 10X42 HT compared to the SF and SV (1 Viewer)

Well-known member

Attachments

Well-known member

Well-known member

Attachments

Well-known member

Well-known member

Well-known member

Well-known member

Well-known member

Well-known member

Registered User

Well-known member

Well-known member

Well-known member

Well-known member

Attachments

Well-known member

Well-known member

Well-known member

Consulting Biologist

Well-known member

Registered User

Attachments

Similar threads

Users who are viewing this thread