Camera Labs

I'm currently investigating what the best long tele-lenses could be for my Nikon D80. The goal is to get at least to 400mm, better even to 500mm.
Why not post this for you Canon-owners also, as the third-party lenses are the same for Nikon and Canon! I just switched the original Nikon-lenses for original Canon-one.
So here's my preselection (all zooms in the end!), figures in brackets denote weight, length, price in EUR!
Up to 400mm:
1. Canon EF 100-400mm 4.5-5.6 L IS USM (1360g,189mm,1400€)
2. Tokina AT-XD 80-400mm 4.5-5.6 (1020g,137mm,550€)
3. Sigma AF 80-400mm 4.5-5.6 EX DG APO RF OS (1750g,190mm, 1100€)
4. Sigma AF 135-400mm 4.5-5.6 Asp APO RF (1210g,181mm,520€)
Up to 500mm:
6. Sigma AF 50-500mm 4.0-6.3 EX DG APO HSM RF (1842g,219mm,1050€)
7. Sigma AF 170-500mm 5.0-6.3 DG Asp APO RF (1315g,230mm,680€)
8. Tamron SP AF 200-500mm 5-6.3 Di LD IF (1237g,227mm,770€)

Unfortunately only 2 lenses have IS: (1) and (3) and both are met with mixed reviews: PhotoZone raves about (1) but mentioned it got a bad copy first, ePhotoZine says bad things about it, (3) looks not bad, but got some remarks on its OS
(2) and (4) do not hold up in the IQ-department
(6) and (7) also found mixed reviews, (6) also being quite heavy
the Tamron 200-500 (8 ) looks quite good at PZ and ePZ, with pictures nearly matching an excellent original lens It also focusses to 2.5m giving a 5:1 magnification Just lice the Canon (1) @ 1.8m!

That leaves you with a clear favorite if you're ready to invest in a stabilized lens: (1 ) is on top, followed by (3).
And if that is "a bit" over budget then go (8 ) and enjoy the longest (500mm) reach.

-------------
I didn't mention the Sigma AF 300-800mm 5.6 EX DG APO HSM IF because it's 5880g, 544mm, 6000€ and has no IS

_________________
Thomas (beware: Nikon-fanboy and moderator!) My Lens Reviews, My Pictures, My Photography Blog
D800+assorted lenses

Hi Tom,

Thanks for "Canonising" your original Nikon post. On the face of it the EF 100-400mm f/4.5-5.6L IS USM wins the battle if you have the budget and need IS.

That said, here's a question I really don't know the answer to but which may be equally validly asked in the Nikon camp for the equivalent lenses.

I am fortunate enough to own the EF 70-200mm f/4L IS USM which, according to Photozone's review "may well be the very best tele zoom on the market today". Given how sharp this lens is (click on the the Blur Index plot in the SLR Gear review) is it worthwhile my buying a 100-400mm lens if that lens is less sharp? Mightn't I do almost as well by doing a "digital zoom" with Photoshop etc.?

Difficult to know the answer to this without hands-on testing and side by side comparison of the candidate lenses. I know you want to get VR at the longer focal length but just on the question of resolution have you compared your 500mm mirror lens with your 18-200mm lens at 200mm and with a bit of Photoshopping to zoom it up to 500mm?

Gordon, if you are reading this have you done any such comparisons?

Bob.

_________________
OM-D E-M5 + ED 12-50mm, H-PS14042E, H-F007014E, L-RS014150E, M.ZUIKO DIGITAL 45mm 1:1.8.
Gitzo GT1541T, Arca-Swiss Z1 DP ball-head.
Astrophotography: TEC 140 'scope, FLI ML16803 camera, ASA DDM60 Pro mount.

Dear Professor B. Andersson,

may I kindly remind you to get your calculator out and do a quick math?
If the best resolution of a shorter lens is say 2000 LW/PW and you do some digi-zoom by the factor of 2, what will be the resulting real (not smeared by interpolation) resolution?
Rrrrrrrright: It's 1000 LW/PH! I have never seen such a bad lens

So perhaps this trick is viable to bring an excellent 400mm pic (2000 LW/PH) up to a good 500mm equivalent with 1600 LW/PH, but not more...

_________________
Thomas (beware: Nikon-fanboy and moderator!) My Lens Reviews, My Pictures, My Photography Blog
D800+assorted lenses

I've certainly never deserved the Professor epithet but I'll do my best to live up to it, just for you . Your reply prompted me to have a look at the test data for three lenses reviewed at Photozone, all strapped on an EOS 350D. Apologies for the amount of detail but I believe it is needed to make my point.

Sigma AF 80-400mm f/4.5-5.6 EX APO OS At 400mm and f/5.6 the quoted resolution is around 1570 LW/PH and chromatic aberration is almost 1 pixel across. At 300mm and f/5.6 the figures are 1560 LW/PH and 0.4 pixels.

Canon EF 70-200mm f/4 USM L IS At 200mm and f/5.6 the quoted resolution is around 2120 LW/PH and chromatic aberration is 0.25 of a pixel.

Canon EF 100-400mm f/4.5-5.6 USM L IS At 400mm and f/5.6 the quoted resolution is around 1630 LW/PH and chromatic aberration is 0.64 pixels across. At 300mm and f/5.6 the figures are 1690 LW/PH and 0.47 pixels.

So, having got my calculator out , it would seem that, provided the camera sensor is up to the job, the Canon 70-200mm lens can capture detail at 200mm which requires the Sigma to be zoomed to 270mm (2120/1560 times 200). The Sigma has chromatic aberration at 300mm of 0.4 of a pixel so this may be visible. The Canon 100-400mm isn't much better than the Sigma. The Canon 70-200mm lens carries its significant advantage in resolving power across the entire range of focal lengths common to all three lenses.

Current best prices in the UK are £690/€1035/$1380 for the Canon 70-200mm and £757/€1135/$1515 for the Sigma 80-400mm f/4.5-5.6 EX DG APO OS. I note that this is not the lens tested by Photozone but also note that Photozone say the "tested lens is the older variant without Sigma's new DG (Digital) coating but regarding the unaltered design most findings should be valid for the current lens as well". The Canon 100-400mm figures are £1000/€1500/$2000.

I respectfully submit that the Canon EF 70-200mm f/4 USM L IS is a significantly better lens across the range of focal lengths common to all three lenses and that only after the Canon 70-200mm has been "digitally zoomed" out to beyond 270mm can the Sigma be said to have an advantage in resolving power. Even so, by the time the Sigma has reached 400mm its chromatic aberration may be a significant issue.

I know you are fond of producing A3 prints so the difference between the digitally zoomed Canon 70-200mm resolution at 400mm of 1060 LW/PH and the Sigma's figure of 1570 LW/PH means you wouldn't want the Canon lens even if you had a Canon body!

In my previous post I asked, given I own an EF 70-200mm f/4L IS USM lens, "is it worthwhile my buying a 100-400mm lens if that lens is less sharp?" I didn't know the answer before I composed this post but now I do. For my needs the answer is "No" for the Sigma and "No" for the Canon 100-400mm which buys a slightly improved performance over the Sigma at a much higher price.

Forgive me if I discard my professorial stance before I try the same analysis with Nikkor lenses: I'll leave that to others. I hope I have demonstrated that, within limits, it is not valid to assume that just because one lens has a longer focal length than another it is necessarily a better choice.


I totally agree. The problem is "where do we find an affordable 400mm lens with that resolution and image stabilisation".

Bob.

P.S. The trouble with these numerical analyses is that they can't always tell you the whole story. User perception and feedback is important too. As before, good luck with your own search for a 400mm plus telephoto.

_________________
OM-D E-M5 + ED 12-50mm, H-PS14042E, H-F007014E, L-RS014150E, M.ZUIKO DIGITAL 45mm 1:1.8.
Gitzo GT1541T, Arca-Swiss Z1 DP ball-head.
Astrophotography: TEC 140 'scope, FLI ML16803 camera, ASA DDM60 Pro mount.

Nice analysis, Bob!
The prob is that the stabilized 400mm zooms of Sigma, Canon and Nikon are not the very best at the long end.
The only real alternative that gives significantly better resolution is the Tamron 200-500mm at 500mm with 1600LW/PH. You would need a 200mm with 4000LW/PH resolution (and a FF-body to match) to compete.
But unfortunately this lens is not stabilized. And as I never carry a tripod and only seldom a monopod IQ-loss through shake is a worse adversary than a few lines of resolution that the lens may (not) have. So I'm in a total dilemma here...
---------------
Still dreaming of the Nikon VR 200-400mm/f4

_________________
Thomas (beware: Nikon-fanboy and moderator!) My Lens Reviews, My Pictures, My Photography Blog
D800+assorted lenses

As a collateral question:
Why is it that a 1.7x tele-converter does not necessarily reduce the resolution also by a factor 1.7? See Klaus Schroiff's test here

_________________
Thomas (beware: Nikon-fanboy and moderator!) My Lens Reviews, My Pictures, My Photography Blog
D800+assorted lenses

Like all the best questions this one afforded me hours of entertainment. Not sure if I've got a handle on this or not, as I have been learning as I go, but here goes...

I followed the link to Klaus Schroiff's test of the Nikkor AF-S 300mm f/4D IF-ED (nice lens!) and had a look at the MTF50 results for this lens with and without the two tele-converters. As you point out, the best case center resolution values don't decrease markedly and certainly don't decrease in proportion to the tele-converter's magnification.

That got me wondering exactly what the MTF50 test was and what LW/PH units were. The Imatest Image Sharpness page gave me rather more answer than I wanted. My take on this is that, for a given camera sensor, it is proportional to the number of line pairs that can be resolved per millimetre on the sensor.

So how does the sensor size plays a part in these resolution measurements as the tests at Photozone are done with a camera/lens combination rather than having the lens on an optical test-bed? A term I am familiar with from my interest in astronomy and telescopes is "Airy Disk" and a Google turned up gold with a fascinating page entitled Diffraction and Photography. As an aside, looking at the Tutorials Index Page this site looks like it's worth bookmarking.

Back to the subject in hand. The Diffraction and Photography page has an interactive diagram of Aperture vs. Pixel Size which may be relevant to the issue with tele-converters. It took me a while to understand how to get the best from the interactive diagram but it turns out that you should first place the mouse pointer over the camera model of interest to set the pixel grid size and then (without mousing over a different camera model) move the mouse pointer left so that you can then mouse up and down over the different aperture settings to see the Airy Disk size.

The diagram doesn't have a Nikon D200 in its list but the "Canon EOS 20D/350D" will serve. On those models with a 41.2 µm² pixel area one can see that the size of the Airy Disk is significantly smaller than the pixel at f/5.6 and it is only at f/11 and beyond that diffraction becomes an issue. That accords well with experience.

Back to the Photozone test of the Nikkor AF-S 300mm on a D200 body. I note that to the side of the MTF50 test data the max LW/PH for the D200 is quoted as ~2350. That said, a quick scan of Nikkor primes tested by Photozone shows that the AF-S 300mm MTF50 results are pretty much at the top of the tree. Indeed, it would be surprising to me if the best Nikkor lenses, such as this one, didn't manage to produce Airy Disks smaller than the pixel area when diffraction isn't limiting. With such small Airy Disks it is the pixel size which limits the resolution.

So, I think (but can't be sure) that the reason that the tele-converters don't significantly worsen the test results is that they are magnifying an Airy Disk which started off rather smaller than the pixel area.

This has ramifications for the post I made above regarding the Sigma and the two Canon zooms. My EF 70-200mm f/4 USM L IS lens is producing resolution results which are almost certainly limited by the camera sensor and so, like the Nikkor AF-S 300mm with the Nikon tele-converters, it might well benefit from using a Canon EF 1.4x Converter. However, the other Canon lens and the Sigma lens are already producing MTF50 figures well below the camera sensor resolution and so using an extender/tele-converter with these is likely to prove disappointing.

As I said, I have been learning as I have been composing this post so there might well be a flaw in the argument above. As long as it is done nicely I promise I won't get upset if that flaw is exposed!

Bob.

_________________
OM-D E-M5 + ED 12-50mm, H-PS14042E, H-F007014E, L-RS014150E, M.ZUIKO DIGITAL 45mm 1:1.8.
Gitzo GT1541T, Arca-Swiss Z1 DP ball-head.
Astrophotography: TEC 140 'scope, FLI ML16803 camera, ASA DDM60 Pro mount.

Well, I thought, that when the resolution of a lens is ~15% below the max body-resolution, that the sensor resolution would not be the limiting factor for the combined lens+body system.
Let's do some quick maths:
Let's assume that the resolution is linear with the magnification of the tele-converter (in this case 1.7x). Calculating backward from the center-resolution of 1812 LW/PH w/converter, that would theoretically give a resolution of 3080 LW/PH w/o converter.
The big question is: Why does a 3080-res lens plus a 2350-res body only yield 2050-res combined (and not something closer to 2350)?
And the other question: Let's assume the successor of the test-body (let's call it 40D ) has 14.4MP and thus could theoretically resolve 20% more, or 2820 LW/PH. What will the combined resolution of lens+body (w/o converter) be? Also 20% higher (we're approaching the (theoretical) limit of the lens (3080) now), that is 2050x1.2=2460?
Then I should take back my remark that "todays sensors are already surpassing the resolution of todays lenses"!

_________________
Thomas (beware: Nikon-fanboy and moderator!) My Lens Reviews, My Pictures, My Photography Blog
D800+assorted lenses

Here are some results from the August 10th issue of ColorFoto (Germany) for longer tele-zooms tested on a Canon 20D. The evalutation gives points for 3 focal lengths based mainly on resolution/contrast (the higher, the better; max=150pts). The last figure is the grand total.
Sigma 80-400mm OS: 80mm/69, 180mm/72, 400mm/66.5, total=69
this test-lens was 5% decentred
Canon 100-400mm: 100mm/73.5, 200mm/72, 400mm/69, total=71.5
This test-lens was 14% decentred, meaning that
1. Even a Canon lens can be decentred
2. The Canon could have yielded better resolution-values had it be well-centred
3. Still it's better than the Sigma.
Canon 70-300mm IS: 70mm/73.5, 120mm/73, 300mm/67.5, total=71.5
This test-lens was 11% decentred, and came out clearly better than its "DO" cousin (62.5 total pts.) that costs twice as much.
Sigma 100-300mm: 100mm/71.5, 170mm/72.5, 300mm/67.5, total=70.5
this test-lens was 7% decentred. It costs more than double that of the Canon 70-300 and has no stabilization, which more than compensates the larger aperture of the Sigma.

As an asside for this Canon section, but nonetheless interesting: In combination with a Nikon D200 the results of the Sigma were:
Sigma 80-400mm OS: 80mm/70, 190mm/69, 400mm/59.5, total=66
this test-lens was 15% decentred (!)
That means:
Although the Sigma was mounted to a body with higher resolution capabilities the results came out worse! This should be attributed to the quite heavy decentering, as everything else should be more or less equal.
So again the ugly quality control issue raises it's head: not only with Sigma (again) but also with Canon (max decentering 16% in a fixed focus lens) and Nikon (max decentering 19%!)

What I didn't get though with this test: why didn't ColorFoto take a 10MP-Canon body, to make it a little more comparable to the lenses tested on the Nikon D200 or the Oly E400
And no, there were no lenses tested with 500mm focal length

_________________
Thomas (beware: Nikon-fanboy and moderator!) My Lens Reviews, My Pictures, My Photography Blog
D800+assorted lenses

I don't really know. Could it be a function of the software that converts the Bayer pattern of red, green and blue pixels into the combined RGB pixels that we play with on our PCs? There is a nice illustration of de-mosaiciing and the artefacts which can be introduced here.

Particularly when dealing with quality prime lenses such as the Nikkor AF-S 300mm it would be surprising to me if they were not designed to be diffraction limited at around f/4 or f/5.6 even if doing so for lenses capable of f/1.4 or f/2 is too expensive or difficult. After all, I imagine such lenses still find use on film cameras where the grain size of some black and white films is of the order of 1 µm² rather than the 41.2 µm² in my example above.

I wonder if any of the lenses we have been discussing has test results posted on the net when mounted on a film SLR? A search for such a review with the Canon EF 70-200mm f/4 USM L IS came up empty.

Bob.

_________________
OM-D E-M5 + ED 12-50mm, H-PS14042E, H-F007014E, L-RS014150E, M.ZUIKO DIGITAL 45mm 1:1.8.
Gitzo GT1541T, Arca-Swiss Z1 DP ball-head.
Astrophotography: TEC 140 'scope, FLI ML16803 camera, ASA DDM60 Pro mount.

Couldn't find any tests for modern lenses using fine grained black and white film but I did turn up a page of tests (http://www.paulfvs.dds.nl/lenstest.html) done on various Rokkor lenses from the 1970s.

Several of these ancient lenses achieved central resolutions approaching or achieving 150 line pairs per millimetre.

By comparison, my 400D sensor has a height of 14.8mm over which there are 2592 pixels yielding no better than 2592/(14.8 x 2) = 88 line pairs per millimetre.

The inference (not proven fact) is that if, at optimum apertures, these lenses could provide better resolution than is offered by typical DSLR sensors then the best modern lenses can do likewise.

That might explain why tele-converters/extenders can, when used with the best lenses, bring out more detail than performing the equivalent digital zoom despite the fact that tele-converters/extenders introduce optical defects of their own.

As I mentioned before, I'm learning as I go with this and it may still be that I have missed something.

Bob.

_________________
OM-D E-M5 + ED 12-50mm, H-PS14042E, H-F007014E, L-RS014150E, M.ZUIKO DIGITAL 45mm 1:1.8.
Gitzo GT1541T, Arca-Swiss Z1 DP ball-head.
Astrophotography: TEC 140 'scope, FLI ML16803 camera, ASA DDM60 Pro mount.

Your findings could be the answer: Good lenses having a resolution of 3-4k LW/PH!
Thus with a 2x tele-converter they give you 1.5-2k LW/PH resolution, which results on a body with 2.3k res in 1.3-1.8k combined system resolution...
Still I haven't found the formula that calculates the system-resolution from the res of the lens combined with the res of the body/sensor.
My assumption is that according to Shannon's theorem, if the res of the lens is double that of the res of the sensor, the system has the res of the sensor (i.e. a 4k-res lens plus a 2k-res sensor gives a 2k-res system).
But if both (i.e. lens and sensor) have the same resolution, what is the combined (system-)resolution? I think perhaps it could be 71% (1/square-root(2))of the resolution of the lens/sensor. But this is just a wild guess...

_________________
Thomas (beware: Nikon-fanboy and moderator!) My Lens Reviews, My Pictures, My Photography Blog
D800+assorted lenses

Wow, Shannon's Theorem - that takes me back to a short telecommunications course I did about a quarter of a century ago! Happy days.

I think your analysis is spot on. I hadn't realised until this thread that good but affordable lenses at optimum aperture are only around a factor of 2 or less better in resolving power than the pixel spacing of modern cropped DSLR sensors. Point and shoot cameras already seem to be hitting these diffraction limits with their small pixel sizes, even if fronted with optically perfect glass.

Bob.

_________________
OM-D E-M5 + ED 12-50mm, H-PS14042E, H-F007014E, L-RS014150E, M.ZUIKO DIGITAL 45mm 1:1.8.
Gitzo GT1541T, Arca-Swiss Z1 DP ball-head.
Astrophotography: TEC 140 'scope, FLI ML16803 camera, ASA DDM60 Pro mount.

Yep, this is a problem facing compacts today, but like mobile phone cameras, the manufacturers will keep increasing the resolutions as sadly that's the specification most people believe corresponds to quality.

It's certainly a concern that it's not just increased noise and reduced dyanmic range we have to worry about with higher sensor resolutions, but also whether the actual optics can resolve sufficient detail for them.

It must be frustrating working in sensor development, being told by your bosses that marketing simply demand ever-higher resolutions...

I found this article very interesting: http://www.clarkvision.com/imagedetail/ ... mary1.html
There is also a lot more on Clarke's site.

_________________
Thomas (beware: Nikon-fanboy and moderator!) My Lens Reviews, My Pictures, My Photography Blog
D800+assorted lenses