600 PF w/ and w/o 1.4 TC compared at (really) long distances

[baekgaard]

There are frequent discussions on whether it's better to just crop or to use a TC with a long tele, and I've also had doubts myself on this. I've also seen claims that using a TC may give even worse performance at longer distances compared to cropping, although I'm not entirely sure why that would be.

So in order to somewhat settle the question for myself, I decided to photograph the moon. It isn't exactly abundant in wildlife, so I picked it more as an example of an object at a distance as long as I could ever think of using

I mounted my camera, a Z8 with the latest firmware, to a 600 mm f/6.3 PF which I then mounted to a tripod and aimed it at the moon and took a series of shots with and without the Z TC 1.4x. Each shot was re-focused, and I then picked the two best shots of the bunch (which was a bit hard since many were about equally sharp).

I then cropped the moon so that image was 2000x2000 pixels with the TC and 1428x1428 without the TC (i.e. approximately 2000/1.4). I subsequently resized the images down or up, so that I ended up with two shots at (approximately) 2000x2000 and two at 1428x1428 pixels, where the 600mm was upsized and the 840mm was downsized. I used darktable as my image processing program (which is what I otherwise often use and it's easier to crop to an exact pixel size for me).

There is NO sharpening applied nor any noise reduction; I used the default setup and just increased the exposure about 1 stop for both images and cropped as described.

I'm attaching the images here, so you can judge for yourself. I am also including a shot that is processed more like I would normally do, with a bit of sharpening and some adjustments of dark and white-points and contrast (filmic RGB) and also some noise removal -- just as an example of what the system can do.

Overall I'm fairly impressed by what the 600mm f/6.3 PF can produce. It is a very sharp tele.

I would say that to me, the shot that uses the 1.4x TC show significantly more detail, although the bare 600mm is still pretty good. Apparently the 600mm outresolves the sensor and when using the TC the total resolution increases -- it's visible in the 2000x2000 shots but surprisingly also in the 1428x1428 shots. This is, by the way, consistent with the findings that PL did previously, although the difference was may be a bit larger in real life than what their numbers indicated to me.

There are of course other factors than image acuity or resolution!

Feel free to give critique to the method or to my conclusions, of course. YMMV.

First, here is the native 600mm unscaled (i.e. 1:1 pixels):

You can only see EXIF info for this image if you are logged in.

and then the comparable 840mm shot downsized, so each camera pixel is now about 0.71 output pixels:

You can only see EXIF info for this image if you are logged in.

Here are then the same images, in the same order, but now it's first the 600mm pixels upsized to about 2000x2000 pixels, so each camera pixel now becomes 1.4 output pixels:

You can only see EXIF info for this image if you are logged in.

and then the 840mm image at native resolution (1:1):

You can only see EXIF info for this image if you are logged in.

For reference, I'm also including a sharpened image:

You can only see EXIF info for this image if you are logged in.

Hope this may help some to decide also (at least if you want to take pictures of the moon)!

 

[Tibor Hrnko]

It is only 1 light second

 

[Dinusaur]

Very practical comparison and thanks for doing this. This must be applicable to other telephoto lenses e.g. 400 2.8TC, I'll give it a try with internal TC and also an external one.

 

[Nasos Kosmas]

Shooting moon is not a good practice to check lens sharpness
The air in the atmosphere can produce some haze and ruing your tests
To understand better this phenomenon try to make a video of the moon on a steady tripod and then see it on a good screen, you will notice small sharpness changes on the moon surface during this observation
Because of this if you want to have the best photo of moon and stars you never use one shoot but you get a series of photos and use photo stacking programs to do the job, but this is way to far from just a lens sharpness test

 

[NorthernFocus]

Shooting moon is not a good practice to check lens sharpness
The air in the atmosphere can produce some haze and ruing your tests
To understand better this phenomenon try to make a video of the moon on a steady tripod and then see it on a good screen, you will notice small sharpness changes on the moon surface during this observation
Because of this if you want to have the best photo of moon and stars you never use one shoot but you get a series of photos and use photo stacking programs to do the job, but this is way to far from just a lens sharpness test

What was presented here isn't a test of lens sharpness. It's a comparison of two different lenses under the same conditions. And therefore is relevant data for practical consideration.

 

[StefanSC]

Well, it's good to know that the 1.4x TC will help when shooting Moon Birds

Don't want to diminish your work, it's just that it covers only a very niche situation and there is no way you can draw a definitive conclusion from it.

(As a note: I currently own 4 tc's for 2 mounts that I use with 5 or 6 lenses for various situations and there are so many nuances that go in using a TC in the field that I'm getting a headache just thinking about it...)

 

[baekgaard]

Shooting moon is not a good practice to check lens sharpness
The air in the atmosphere can produce some haze and ruing your tests
To understand better this phenomenon try to make a video of the moon on a steady tripod and then see it on a good screen, you will notice small sharpness changes on the moon surface during this observation
Because of this if you want to have the best photo of moon and stars you never use one shoot but you get a series of photos and use photo stacking programs to do the job, but this is way to far from just a lens sharpness test

What was presented here isn't a test of lens sharpness. It's a comparison of two different lenses under the same conditions. And therefore is relevant data for practical consideration.

Exactly, @NorthernFocus

First, thanks for the feedback both. I am indeed aware of the atmospheric distortions (I also own a couple of telescopes, a reflector and a refractor, and am used to see things wobble around), and the point of the test was indeed not to check lens sharpness -- if so I might have done a slanted-edge test to find the MTF or even a typical "TC" ("Teddybear and Checkerboard") test But I already knew my optics perform well. Nor was it an attempt at taking the best possible moon-shot.

Just to explain the background, what motivated me to do the test was that I've seen claims that 1) a TC will ALWAYS cause a loss of acuity and resolution in the final image and 2) using a TC will make atmospheric distortions worse compared to cropping or shooting DX mode, especially if you mount it on a PF type lens. I could probably argue on a theoretical basis regarding some of these points, but just to validate my assumptions I thought a test through a lot of air could experimentally give some indications on especially the second point (as a scientist I know that it's sometimes good to check up on your theoretical assumptions). And no, this is not a peer-reviewed study, just a test done where I document the method and the results and leave the analysis and discussion to the reader.

I picked the moon exactly to include some level of atmospheric distortion through about as much air as I would ever think of. However, in order to not skew the results by a random outlier, I took a number of pictures and picked the best from each series, although I think I had a good number of just about equally good shots in each series.

I don't think the conditions changed much if at all between the series, so the images are about as representative of a "real world" shot at long distance with this particular combination of camera and optics as I could get.

As for stacking (oversampling and averaging or using some more advanced aggregation function), it can indeed take out some of the random "noise" assuming it is somewhat close to a gaussian distribution. However, there is also be a transfer function involved representing the sampling in the camera and the de-mosaic algorithm, and if your tracking is not pixel (or even sub-pixel) accurate then you may also start to "compensate" for some of that by stacking, similar to when you shift the sensor between images (of course all dependent on a lot of other variables). So even a perfectly aligned EQ mount may not have been sufficient for the type of test I wanted to do if I had to stack also -- and definitely my own EQ mount is far from good enough

 

[baekgaard]

Well, it's good to know that the 1.4x TC will help when shooting Moon Birds

Don't want to diminish your work, it's just that it covers only a very niche situation and there is no way you can draw a definitive conclusion from it.

(As a note: I currently own 4 tc's for 2 mounts that I use with 5 or 6 lenses for various situations and there are so many nuances that go in using a TC in the field that I'm getting a headache just thinking about it...)

Thanks. I did look for the moon birds, but didn't see any this time around...

I don't claim any generalised conclusion at all. I think there are some conclusions that I have drawn for myself, which is that my optics in this concrete case seemed to perform in a similar way to the PL test (i.e. adding the TC 1.4 to the 600mm PF gives a little higher acuity/resolution than cropping) and that the atmospheric distortions (shooting through a lot of air) do not impact the TC shots more than the non-TC shots. Both are in line with what I expected. I would tentatively think this also applies when shooting earthly birds and if I don't get a sharp shot with the TC mounted, I would blame my technique or the lack of light and not the lens. And I'm not going into a discussion on noise and resampling here...

I may also some day do a Teddybear/Checkerboard test at shorter distances, but I agree that even if I did, it wouldn't mean you (or anyone else) could conclude anything with your 4 TCs and 6 lenses, of course. But it seems you do use the TC's also from time to time then and thereby rely on improving the picture compared to cropping? But perhaps there are some with the 600mm PF that have worried about using a TC, and if so, they are welcome to look at the small "test" I did and draw their own conclusions.

 

[dabhand16]

I've also seen claims that it may get worse performance at long distances, although I'm not entirely sure why that would be.

The longer the distance, the more air you shoot through. So more dust, heat haze and other factors that will degrade images.

 

[baekgaard]

The longer the distance, the more air you shoot through. So more dust, heat haze and other factors that will degrade images.

Ouch; I meant to say that some claim the difference between using a TC and cropping gets larger with more air. It's clear, of course, that a lot of air will cause image distortion. I'll fix this above, thanks!

 

[Nimi]

@baekgaard

I'm very interested in your finding that the lens outresolves the sensor, which I agree with for many of the new lenses (not just Nikon). Which also means that this test is specific to the sensor (duh). In a few years, when 100mp sensors are the norm, test results will be different.

Have you examined the edges of the frame? Even with this "MTF-perfect" lens, there is a drop-off in sharpness, ie it does not outresolve the sensor a few mm from the edge. Obviously on the long edge, as the image circle is cropped more significantly on top and bottom.

 

[Nasos Kosmas]

I picked the moon exactly to include some level of atmospheric distortion through about as much air as I would ever think of. However, in order to not skew the results by a random outlier, I took a number of pictures and picked the best from each series, although I think I had a good number of just about equally good shots in each series.

If you already picked the best shots of a series your results are more valid towards 100%
I already own a 500 pf and made the same moon tests with the three F mount TCs latest versions, but as I realized the atmospheric variability factor I just forgot about it.
My general finding in the use of TCs is that you always loose contrast, clarity, AF speed but you gain detail if the lens in front of the TC is capable

 

[baekgaard]

@baekgaard

[...]

Have you examined the edges of the frame? Even with this "MTF-perfect" lens, there is a drop-off in sharpness, ie it does not outresolve the sensor a few mm from the edge. Obviously on the long edge, as the image circle is cropped more significantly on top and bottom.

Unfortunately not; partly because in my use cases reach and the central part is often more important and the periphery of the image usually contains out-of-focus areas anyway.

But it would be interesting to know how it performs "in real life" also, of course!

 

[Krish]

This is very interesting. Thank you.

 

[glsmith9823]

Taken this past Wednesday June 5 Nikon Z-8 (firmware up to date) Nikon Z 600mm F/6.3 with Nikon Z 1.4 TC Distance: 3.5 miles, Temperature 91 F/ Humidity 82
1/1250, F/9, ISO 160 Cape Canaveral Space Force Station Launch Complex 41, ULA Atlas V N-22 Rocket carrying Boeing CST-100 "Calypso" Space Capsule

You can only see EXIF info for this image if you are logged in.

 

[jhallettbc]

Amazingly sharp. I have used that combination with birds and can't notice a decrease in sharpness or contrast. However when I have tried the 600 PF with and without the 1.4 TC using Bob Atkins test chart on white paper, it is quite easy to see the loss of contrast and some sharpness. If it's not really noticeable in real life photography I will use the combo.

 

[macwalter]

I spend a lot of time in my backyard practicing with the 600pf and both TCs. I have found for me as I handhold the main difference is my technique and the greater difficulty to acquire the subject at 840mm vs 600mm. Im shooting wide open at f6.3 and f9 and in good light I have found that the subjects, birds etc tend to be a bit better in terms of detail at longer distances with the 1.4tc than the bare lens while at closer distances the opposite is true. The longer shots are anywhere from 60-150ft while he closer shots are 20-50ft. This is not a scientific study mind you but I do it a lot and the results tens to stay consistent for me. The 1.4Tc works well in decent light while the 2x is a challenge to handhold and you need good light at f13. However when I “nail” the shot even this combo can be quite sharp. In the real world I wouldn’t hesitate to use the 1.4c while the 2x Tcwould be for extreme situations with no other option and as you need good light, atmospheric issues would be a concern.

 

[Neil H.]

Nice tones, but most moon images do not show the colour. An interesting discussion on sharpness but if you want moon colours you will need to stack light and dark images (example below). As previous commentators point out, I do not think the moon is a good topic for this test. I have done enough astro-photography to know that "intermittent light high level clouds" (often not visible to the eye) would play havoc with this experimental design.

You can only see EXIF info for this image if you are logged in.

 

[macwalter]

Nice shots of the moon but not sure I would photograph the moon with a lens like the 600pf. I don’t do Astro photography so can’t relate to these comparisons. I would say photograph wildlife may have more impact.