Fast FD lens comparison Pictures. Warning, its not pretty...

[dnhkng]

I have uploaded some pictures to flickr, showing the performance of the following lenses on a Lumix G1 with a FD adapter:

canon 24mm f1.4 aspherical chrome ring
canon 55mm f1.2 chrome ring
canon 85mm f1.2 aspherical chrome ring

together with the images from the kit lens.

http://picasaweb.google.com/dnhkng/LensComparison#5409157599728976690

The original images, and crops used for comparison, have been made at both the widest aperture, and at f2.8 for each lens. With the kit lens, I used the fastest aperture at 14mm and 45mm. In each image, I used the zoom and EVF to focus on the patterned plastic near the bright orange LED.

Anyway, the 24mm is TERRIBLE! unbelievably bad when wide open. The 55mm is quite washed out, but the 85mm pics are fantastic. If anyone want the full, 10mb jpegs, let me know.

Below is a sample, at low res.

top row: 85mm f1.2, 55mm f1.2, 24mm f1.4, kit 14mm f3.5
bottom: 85mm f2.8, 55mm f2.8, 24mm f2.8, kit 45mm f5.6

I'm really shocked that I spent so much money for pictures like this...

[trifox]

Hi

this comparison is quite irrelevant ..

but sell the FD 24 f1.4 -- you don't need it ..

tf

[dnhkng]

why is it irrelevant?

[trifox]

[dnhkng]

Sure, but its on a 2x crop camera.

Also, the pictures were all take from the same distance, and I cropped away the outsides of the shorter focal distance images. The examples from the 24mm lens are just the very very center of the image. So the bad-corner argument is also irrelevant

The comparison just shows the same thing (the power supply to my xbox 360), with three different fast lenses, from the same distance. And the 24mm at f1.2 looks absolutely terrible in my opinion.

[trifox]

[dnhkng]

do you want it?

[DSG]

I agree the test is kinda irrelevant as all the shots were taken from the same spot...This means he had to downsize some of the sample pics to keep the size of the subject the same in each sample due to the different focal lengths used and therefore resolution would have been lost from the samples from the longer focal length lenses compared to the samples from the wider focal length lenses.
As it is though, at f2.8, they all look pretty much the same sharpness wise except for individual contrast differences.
If you bump up the contrast of the ones with lower contrast to match the the ones with higher contrast they should all look pretty much indentical.
However, assuming lens to subject distance had been changed to keep the framing identical despite the differences in focal length, as it would have been in a properly conducted test, then there would be no need to resize and I would expect the results to be slightly better with the FD lenses.

[Laurence]

Whether "scientific" or not, I do like to view images from lenses that are
sort of compared. Of course, on the Net I'm aware that there are many
factors and variables, not the least of which is affected by the simple act
of posting the images.

[cooltouch]

The 24mm seems to be exhibiting a severe amount of coma wide open. So much so that I would suspect that the lens is out of collimation -- perhaps it was dismantled or repaired before and not reassembled properly? I've been a Canon user for 27 years, and I've never come across this amount of coma with any of my Canon lenses, including some very fast ones.

When doing a comparison like this, I think it's okay to show images of the same size, as long as the ones of different subject size are upsized. Detail loss is minimized that way. In this case, the 85/1.2 would be the reference image to which all others would be upsized. However, I agree that the comparison does not hold much value since the lenses being compared are so different from one another. Well, maybe the 85 and 55. Because they're both teles on the camera being used, and since they're both f/1.2 lenses, it makes some sense. But the 24 is just too far away, unless I was wanting to compare something very specific, like bokeh.

[dnhkng]

I just noticed something else. Whenever there is a bright object in the picture, such as the LED on the powersupply in the first set of pictures, the glare is always in the same direction. Its always towards the top right corner, and its a lens effect, because when i rotate the camera, the glare follows the orientation of the camera.

Below is a picture of my ipod, taken in the dark. By eye, the back-lit screen edges are all the same sharpness, but in the picture takes at f1.4, the top and right edges are totally blurred by the directional glare.

Maybe this particular lens is just misaligned. Otherwise I would expect the blurring to be circular, not so highly directional.

[F16SUNSHINE]

Micheal is right IMHO about the 24.
You should send in for a proper alignment.
It should not cost you much and the lens will then become relevant. Like this it's below run of the mill.
I suggest you go out and shoot these lenses in the real world.
Making comparisons on a 2x crop camera is no way to show the capability of a lens.
It's like trying to judge a book without reading the beginning or end.

That said you can still make wonderful images with a 2x crop especially with a line up as you have.

[dnhkng]

Well, I bought the 2x crop first. Then I decided to go FD lenses later, partly in the hope that in a few years time a full-frame mirrorless will be released, so I can take full advantage of these lenses. Also, if such a full frame is released, the prices for the remaining fungus/scratch free lenses might go up beyond my budget, like what has happened with the fast cine lenses.

anyway, three things then:

Firstly, if you have a range of lenses of various lengths, and you want to test them for sharpness etc, whats the fairest way to test? I am happy to redo the lens test with some advice

Second, I am pretty new to Munich, Germany, and I have no idea who can do a lens alignment on a 30 year old lens... I would be very happy to hear some suggestions!

Lastly, the lens just arrived by post today after I bought it on ebay. Externally I would say it is in very good condition, with barely any wear, and the lenses are clean and clear. I paid 480 euro for the lens, in the original box. I have no idea what an alignment would cost, so, would it be cheaper return the lens to the seller and buy another one, or get it serviced?

BTW, there are hardly any 24mm f1.4's on ebay. I have only seen 2 in the last few months (not including the ridiculously overpriced priced buy-it-now ones), compared to maybe twenty 85mm f1.2's (also not so common) in the same time.

[egidio]

this lens is probably misaligned for a kilometer

[cooltouch]

I'm thinking it would be cheaper to return the lens than have it serviced.

But I can understand the appeal of having a fast 24mm for a 2x crop body, though. Basically it's the equivalent of a FF 50/1.4 -- just costs a whole lot more.

As for tests, I don't think it's necessary to do comparisons with other lenses when evaluating a lens's sharpness, contrast, and CA, etc. If you have two or more that have the same focal length or close to it, then, sure -- do a comparison. But as for testing, whether it's a group or an individual lens, what I prefer to do is find subjects with lots of detail, one for close-focusing, and one for infinity focus. A currency note (dollar, euro, pound note, whatever) is good for close-focus because it will usually have lots of fine detail. As for infinity focus, it could be most anything with good detail. Sometimes I'll use distant signs or structures that have some sort of small detail -- distant trees with good bark texture.

A stout tripod is a must to eliminate shake. A remote release or using the self-timer also reduces the possibility of shake. If you can trip the mirror ahead of the exposure, try that too and see if it makes any difference.

I'll shoot the test images at successive apertures. Sometimes I'll go through them all, sometimes maybe just wide open, f/8, and f/16.

If your camera has some way to magnify the image, like Canon's Live View, be sure to use it for focusing accuracy.

Then load the images into your favorite image processing software, and take a look at them, especially at 100%. You can usually tell very quickly which of your lenses are good performers and which aren't.

[dnhkng]

thanks for the advice.

The images I posted were taken on a tripod, and I was very careful taking the shots to avoid any camera shake. The 24mm lens picture wide open is just so bad its not worth using the tripod though. I took a few shots at dusk, and my hand-held half second exposures at a higher aperture were sharper than wide open on a tripod. And I'm really not that steady...

I think I will return the lens on Monday. Its been a very disappointing weekend

My 55mm aspherical arrives, and the breech mount ring is dented and broken, and now the 24mm is misaligned.

about the price, yeah, I know a 24mm on a 2x crop is crazy but a lumix G1 and a 24mm f1.4 is still cheaper than a 5d and an EF 50mm f1.4

[pirius]

I only have Canon FD 85/1.2 from this list and in my tests on Canon 5D II it is sharper at f/1.2 than some lenses at f/2.0. To say "it's perfectly usable at f/1.2" wouldn't do it justice. It is sharp at f/1.2!

Oh, I also have Canon EF 24/1.4L. Things must have changed quite a bit since the FD days in this range. The EF version is awesome wide open.

[cooltouch]

[dnhkng]

pirius, could you post a few pics taken with you 24mm f1.4 wide open? I would love to see what I am missing

[dnhkng]

Hey guys,

Just thought I would post a warning!

I sent that bad lens back last week, along with a letter describing the problem with the CA, coma defect, and links to the example pictures. Anyway, the assholes have put it straight back on ebay as "Das Objektiv wurde von unserem Techniker geprüft und arbeitet einwandfrei", ie, that its been tested by their technician, and is working properly!

So, if you bid on this auction:
http://cgi.ebay.de/ws/eBayISAPI.dll?ViewItem&item=110466312878&ssPageName=STRK:MEWAX:IT
at least you know you are getting a dodgy lens now. I'm pretty pissed off about this, but there's not much I can do. There will always be assholes who lie in their ads.