Canon 200mm f/2.8 FD vs EF comparative lens test using a Sony A7ii

[Mfcrisis]

Canon 200mm f/2.8 FD vs EF comparative lens test with a Sony A7ii

I was somewhat disappointed with the purple fringing of the Canon nFD 200mm f/2.8 v2 lens in high contrast situations(eg photos with dark trees/birds against the bright sky) at wide apertures. Unfortunately Canon never made an L version of this lens, although the ordinary version is beautifully made.

I decided that I couldn't live with the purple fringing and got hold of a Canon EF 200mm F/2.8L USM II lens to compare against the FD. The FD was released in 1982 and the EF in 1991 with a minor facelift to a version 2 in 1996.

Key similarities and differences...

Both lens have a 72mm filter thread, an 8-blade aperture and a minimum focus distance of 1.5m. Both lens are well constructed, despite not being an L lens, the FD feels equally well built, compared with the FD.

Other physical differences:

a) The FD has 7 elements in 6 groups(none special); the EF has 9 elements in 7 groups (including two ultra-low distortion elements).

b) The FD combo is about 18% lighter at 840g vs 1012g for the EF (including adapters).

c)) Fitted with adapters (cheap Fotga for the FD, Metabones IV for the EF), both lens are a similar size as shown in the pic below, though the EF's hood (pictured next to the lens) adds considerable bulk.

d) The clever screw-out lens metal hood of the FD is a more convenient design than the reversing plastic hood of the EF.

f) The FD is purely manual focus, whereas the EF is notionally autofocus with the Metabones, but the AF is too slow and too inaccurate to be of any use on a Sony camera

g) The FD has a larger focus ring and is designed for manual focusing, whereas manual focusing is secondary with the EF

h) The FD plus dumb adapter does not convey EXIF info to the camera, whereas the EF and Metabones passes both aperture and focal length to the camera and the A7ii uses this info for the IBIS stabilisation.

Testing...

For convenience, I took pictures of the TV screen at about 3m with both lens at F2.8. The dot pitch of the TV is about 0.8mm (I can make the pixels out with the naked eye at distances up to about 1.5 m. A few scanning artefacts are visible.

Here are the results, hand-held with IBIS on (all processed from RAW via Capture One, with no adjustments):


Above are the side-by-side photos from each lens (low res screen-print). At this resolution The FD (on the left) is noticeably warmer, has better contrast and is generally more appealing. My (non-photographer) wife confirmed this view, without hesitation.

Surprisingly, with all else being equal, the EF needed a shutter speed that was about 35% longer than the FD: 1/13s to 1/15s for the EF vs 1/20s for the FD - quite a large difference. Note that the EF had a (stuck) Kenko UV filter on it, whereas the FD was bare - but that shouldn't make 35% difference.

I zoomed in on the left eye to 100% magnification:


At this level of magnification, the TV's pixels are clearly visible in both pics ... BUT, the FD is clearly sharper. I took six or seven photos with each lens and these results were consistent.

 

[Mfcrisis]

I've found the purple fringing / LoCA is a signature of most of the FD line of lenses. Makes them unsuitable for subjects where parts of the image are out of focus. That being said I'm not sure what you are seeing is this as the area is in focus. This could be just sensor bloom.

What is 'sensor bloom'? And how does one get rid of it? Since the EF is much better in this respect, the lens clearly plays a part.

 

[Alupang]

Well, it's the worst FD prime at the long end ever made, versus a modern zoom in the range that's easiest to make stopped down a bit.

Well it could be that Canon designed the nFD 200mm f2.8 for maximum sharpness at the expense of CA at max aperture. A design choice that obviously did not go over well with many photographers given its reputation as "the worst FD prime at the long end". But what if the nFD at f5.6 is actually sharper (+ CA well under control) than the $1500 FE70200 and EF 200mm f2.8L?

Should it be judged a great lens then? I think so. I try to shoot my lenses at their optimum aperture (in most cases=f5.6). The nFD 200mm f2.8 could be one of the very best 200mms at f5.6 for all we know.

The Fe is resolving nigh on 4000 lp/ph on the a7r, and not much worse in zone B, which an older designed mon premium prime will find it hard to achieve, and it was certainly true of my pair when I first got the FD ...

Well then aren't you ignoring the test in this thread then? I've seen some pretty convincing tests of the Canon EF 70-200mm f4 L being equal or outperforming the FE70200. I would guess the Canon EF 200mm f2.8 L prime outperforms the EF 70-200 L zoom...right? And here we see the nFD sharper at f2.8 than the EF 200 f2.8 prime. Interesting.

So I suspect at f5.6, the nFD 200mm f2.8 would be at least on par with the EF and FE in terms of sharpness, contrast and CA.

Anyhow I'll try to do some tests again in the next couple of days to make sure my memory isn't playing tricks on me!

That would be great. I'm very interested in seeing your results.

 

[Timbukto]

This is a test of two samples. My second EF 200mm was quite a bit better than my first. In addition I would not buy a 'filter' stuck lens without significant discount. You can be naive and give the seller the benefit of the doubt that its just a problem with the 'filter' and filter threads. Or you can be on the safe side and assume there was impact damage to cause the filter to be stuck on the lens.

 

[dbm305]

A lot of things get called sensor bloom; the most common is probably a kind of CA that involves near infra red. Digital sensors are much more sensitive in this region than film, and most lenses are not well corrected. The sensor stacks filtering characteristics can mitigate it.

but that's not under your control.

Yes the lens can play a part: the closer the lens is to true APO the better (I doubt you would get with the Coastal Optics lens which is corrected for IR and UV).

You could also mitigate it with a filter that takes out the near IR on the front of the lens (I think b+w makes a filter that removes moth uv an ir and near ir)

 

[ProfHankD]

I've found the purple fringing / LoCA is a signature of most of the FD line of lenses. Makes them unsuitable for subjects where parts of the image are out of focus. That being said I'm not sure what you are seeing is this as the area is in focus. This could be just sensor bloom.

What is 'sensor bloom'? And how does one get rid of it? Since the EF is much better in this respect, the lens clearly plays a part.

Sensor bloom is a saturation effect in which charge leaks from pixels to adjacent pixels.

This is NOT what we see here.

 

[dbm305]

In a previous post I said that I had the FD for when I really wanted f2.8, but I thought based on previous experience that the FE zoom was better at f5.6

Some doubted it do I did a quick test on the A7r.

Manual focus, focus bracketed three times and the best chosen in each case. Shot at 3 metres. Tripod mounted using shutter speeds outside the shock zone of the A7r.

For the FD I have used remove CA in Lightroom (which does relatively little for this lens, but improves it a tiny bit), but not the Sony FE which didn't need it.

In camera corrections all turned off.

Tests for the centre and the bottom right (as you can see bottom right I moved in a tiny bit relative to centre, but it's consistent between lenses of course)

I did the crops in Lightroom individually; I should have done one and then just synced for uniformity but there you go.

So: odd colour effect on the FD which I recall; it has a magenta tinge even when in focus. A bit less contrast and sharpness than the Sony zoom. Still, not a bad performance for a legacy tele with no fancy glass.



Centre focussed "Vueillez"; Bottom right focussed on "Merci"


FD centrally


Sony FE centrally






FD bottom right


Sony FE bottom right

 

[Alupang]

Yeah could be sample variation in the initial tests above, or your test for that matter. You are showing me what I did not expect--the nFD should be as nice as the FE zoom at f5.6.

Seems to be resolving fine but the color and contrast is off.

Does an autocolor adjustment fix the nFD's weird color and washed contrast?

 

[dbm305]

No; I've already corrected for CA. There's not much you can do for the weird colour (I think it's SLOCA) that won't make the colour overall worse.

Of course you can use the clarity slider or equivalent, and up the contrast, and sharpen, and get OK results (Plenty of lenses, including most consumer tele zooms resolve less in the corners). But it still doesn't quite match my copy of the Sony G Zoom (but that is a very good zoom, designed to challenge prime lenses and priced accordingly - Jim Kasson has a test showing it going toe to toe with the Leica R APO Tele Tessar 3.4 in the centre although not up to that lofty and expensive standard in the edges -- but not much is)

Of course something could be wrong with my copy of the FD, but I a bit doubt it precisely because the corner resolution is fairly OK and importantly symmetrically so -- most effects that make a copy go out will introduce some asymmetries)

 

[ProfHankD]

So: odd colour effect on the FD which I recall; it has a magenta tinge even when in focus. A bit less contrast and sharpness than the Sony zoom. Still, not a bad performance for a legacy tele with no fancy glass.

That looks like either SA with "bokeh CA" or PF. Agreed; it's quite usable. For either cause, it should get better stopping down more. I have several 200mm f/3.5, and they all have fairly severe "bokeh CA" issues, so maybe this is that + SA rather than PF?

 

[dbm305]

Yep I think likely SA plus LoCA;

but the real Achilles heel of this lens is the SLoCA in the bokeh which is very hard to clean up in many circumstances, even at mid apertures.

 

[Alupang]

No; I've already corrected for CA. There's not much you can do for the weird colour (I think it's SLOCA) that won't make the colour overall worse.

Of course you can use the clarity slider or equivalent, and up the contrast, and sharpen, and get OK results (Plenty of lenses, including most consumer tele zooms resolve less in the corners). But it still doesn't quite match my copy of the Sony G Zoom (but that is a very good zoom,

Well I'm no PP expert (just use freeware Irfanview) and a little autocolor adjustment + a little blue results in this:




Canon nFD 200mm f2.8 @ f5.6 PP w/ Irfanview autocolor + 6 blue

For under $200 and if you actually like manual focusing, AND if you don't mind a little PP, the nFD can hang in there with the FE.

I think that comes down to about $650 saved per "if".

Of course something could be wrong with my copy of the FD, but I a bit doubt it precisely because the corner resolution is fairly OK and importantly symmetrically so -- most effects that make a copy go out will introduce some asymmetries)

 

[Alupang]

nFD 200mm f2.8 Center




nFD 200mm f2.8 Corner

 

[dbm305]

Much better!

But very much harder to do with a detailed scene, I'm afraid, because the colour errors aren't consistent the way they are with black and white text.

 

[Alupang]

Much better!

But very much harder to do with a detailed scene, I'm afraid, because the colour errors aren't consistent the way they are with black and white text.

OK, well, give me a real world shot @ f5.6 then and I'll give it a try--let me see how hard it may be.

Like a real world shot with the FE VS the nFD. I bet I can fix it up very quick and easy...??

Just autocolor and add a little blue (canon is too warm for me). No big deal, 2 clicks.

 

[Alupang]

The nFD 200mm f2.8 was not a cheap lens--in 1992 it was B&H priced about the same as a nFD 50mm f1.2L. $479.

Adjusted for inflation, it would be worth more than the FE all day long.

 

[dbm305]

You've got one haven't you? Just take some shots with busy bokeh and try to sort out the colour.

It's the SLOCA that is the issue; you can fix it tolerably but I have sometimes slave over it with the defringeing tool and the LOCA sliders in Lightroom. One problem is that by defringeing extremely, which you have to do, you desaturate edges that shouldn't be.

All camera gear is cheaper than it used to be, adjusted for inflation!

I'm not dissing the lens; it's useful enough and can be had at a low price. It's a good example of an attempt to build a moderately fast tele without using high RI glass. Still, it was a fast tele and that cost. These designs performed much better at shorter focal lengths.

 

[Alupang]

You've got one haven't you? Just take some shots with busy bokeh and try to sort out the colour.

No, I never owned a copy but, after seeing this thread, I am considering giving one a try. I could use a $200 200mm beater lens.

It's the SLOCA that is the issue; you can fix it tolerably but I have sometimes slave over it with the defringeing tool and the LOCA sliders in Lightroom. One problem is that by defringeing extremely, which you have to do, you desaturate edges that shouldn't be.

At f2.8 yeah, but at f4-5.6 this isn't as much an issue I suspect. I say this based on your samples I fixed with 2 clicks in freeware Irfanview. I could be wrong--I would love to see a real world shot at f5.6 that is so difficult to color fix. Let's see a side by side nFD VS FE70200 real world comparison at f5.6 please.

It is my opinion that at f4 & f5.6, after a little color adjustment, the nFD performs about equal to the FE70-200. That's a nice $200 option for those like me that like manual focus. As a nice bonus, the f2.8 speed is there for low light B&W and color shots where CA isn't an issue.

All camera gear is cheaper than it used to be, adjusted for inflation!

Depends on the lens.

I'm not dissing the lens; it's useful enough and can be had at a low price. It's a good example of an attempt to build a moderately fast tele without using high RI glass. Still, it was a fast tele and that cost. These designs performed much better at shorter focal lengths.

I agree and if it performs as well (after color adjustment) at same apertures as the FE70200, it's a great buy.

You want to sell your copy to me? PM me if so.