pixel
« 
pixel
«  
  »
pixel
S M T W T F S
1 2 3 4 5 6 7
8 9 10 11 12 13 14
15 16 17 18 19 20 21
22 23 24 25 26 27 28
29 30 31
 »
pixel

Testing a lens

2009.03.30
pixel
This is the same capture taken at different apertures with the same lens and camera. Note how the blow-ups reveal the lens' deficiencies. See also how various aperture settings affect colours. None of these images were post-processed except for cropping and resizing.
pixel
Above is an example of vignetting. All nine images were taken of the same thing, a wall, in quick succession and at different apertures.
================
pixel
If you have a camera with interchangeable lenses, you may succumb to “lens lust”, the idea that a faster lens with a longer or wider focal length than the POS kit lens will change your photography. That’s when you’re screwed. You will be on first-name terms with every camera shop employee in your city. Your banker will probably call you something else.

This post is not about which lens to buy – that’s for you to decide – but how to test it once purchased. It’s a sad fact that people make mistakes and a brand new lens you’ve saved up for is defective, which is why it is strongly recommended that you buy from a reputable dealer which has a return policy. Another sadder fact is that when you buy cheap, you get nasty. You’ve checked the specifications and know that the lens is compatible. For example, EF-S lenses won’t work on xD Canon cameras and there are all kinds of limitations with Nikons. Perhaps more importantly, you should know that third-party lenses (e.g. Sigma, Tamron, Tokina) may not work with another Canon, Nikon, Pentax, etc., body you decide to buy later. This “tutorial” post is about how to put a lens through its paces after the purchase, but before your period of retraction expires.

1. Check for any obvious specks inside the lens body or scratches on the front or rear lens elements. If there are any that cannot be carefully removed, return it immediately (obvious).

2. Another obvious: check that moving parts – focus grip, switches, etc. – operate smoothly. Note that some, especially cheaper and third-party lenses, may not have FTM, the ability to autofocus and then manually adjust, so check the specs. Use the DOF button on your camera with the lens set at minimum aperture (highest Av number) to make sure the diaphragm blades work correctly.

3. Check the autofocus by, er… focusing on something in good lighting conditions. Short and back-focusing issues (i.e., image slightly out of focus in relation to main subject) may also arise from your camera, so check using another lens

4. Check for vignetting (a.k.a. light fall-off): this means the corners are darker than the centre of the image and is more visible at maximum apertures. If really noticeable at any aperture, return the lens. Some vignetting necessarily occurs, even with good quality lenses that work well (see above)

5. Check for distortion: straight vertical and horizontal lines in an image will curve out (barrel) or in (pin-cushion) at various focal lengths in a zoom lens, especially at its shortest focal length. If too extreme, return. Note: wide-angle and fish-eye lenses are supposed to distort. All zoom lenses will distort at either end of the focal length and especially with extreme zooms.

6. Chromatic aberration (a.k.a. ‘purple fringing’): bright lines will acquire a coloured fringe in high contrast images. Although this can be corrected with post-processing software, if too extreme, return the lens. Note: again, wide-angle lenses are notoriously prone to this problem anyway.

7. Flaring: a strong light source such as the sun will cause artefacts (polychromatic hexagonal halos, etc.). Not much can be done about this. Some lenses are just flare prone. You can attenuate it with a hood or shielding with your hand. Buyers of non-L Canon lenses know that the hood is an "optional" extra.

8. Corner and centre sharpness: corners and centre parts of an image may look soft at different apertures and focal lengths. This will be more noticeable on a full-frame camera or if you’re printing / displaying larger images, say A4 or more.


Test shots should ideally be taken in RAW format (cameras process JPG’s and therefore invalidate the test) and, if possible, of the same subject every time with the same lighting conditions. If you look at lens reviews, you’ll always see the same combinations of colours and shapes to make comparisons easier. I don’t have a cuddly toy or Martini bottle – dislike both – so I always shoot a bookcase.
You can also do the above with a compact – if it shoots in RAW mode – but be prepared for the disappointing result. Compacts or Point-and-Shoots simply cannot rival SLR’s for image quality. The above tests were shot using a Canon 5D mk II full-frame body, my Canon 24-70 L workhorse zoom lens and a Sigma 50 prime (non-zoom) lens.

BTW, if you don’t know what a full-frame camera is, it’s almost certain you don’t own one, but here’s a link to a definition. Basically, the sensor in most digital SLR’s is smaller than a standard 36x24 film exposure, which increases the focal length of the lens by a ‘crop factor’. Thus, a 50 mm lens on a Nikon will behave like a 75 mm lens on a film or full-frame camera. This is great for telephoto and macro lenses because it gives you a longer reach. Inversely, you have to buy special wide-angle lenses.

2010 edit


Adobe and DX Optics, to name but two, each offer lens correction features (primarily for geometric distortion and vignetting). Some cameras also feature 'peripheral illumination' functions. These should be deactivated for the test.



I hope this helps. And if you've found this utterly tedious, all I can say is that it's in your neurons now.
6 Comments
dontblink After reading this post very carefully and pondering it's wealth of information I have one important question to ask. How much of your PB traffic do you think you get from people who search for the word 'sex' while browsing by tags? 20, 30%?
dontblink · 2009-03-30: 15:33
tomie très intéressant ce que je ne comprends pas...c'est pourquoi la photo à f1,4 est sous-exposée par rapport aux deux autres?
autrement pour le vignettage avec lightroom on peut tout compenser...et je trouve que le vignettage a un certains charme sur certaines photos.
tomie · 2009-03-30: 16:31
Robertthebob Great set of info this - I hope TerryL finds it he will love this!
As for me - well Im just a point and shoot person! and I never look at tags either! Boring aint I!
Robertthebob · 2009-03-30: 16:36
Memphis Thank you for this. I feel you should get some of the Licence Fee money that I pay to the BBC for your dedication to public service broadcasting. Don't stop :)
Memphis · 2009-03-30: 18:45
Jarvo Thanks very much for the lesson. Perhaps we have the makings of a new request to the administrators - a section for tutorials (at least a page of dedicated links).
Jarvo · 2009-03-30: 20:21
TIM927 <i will come back to this post when my headache is over,,, and when I buy a new lens for sure, thanks!
TIM927 · 2009-03-31: 14:54
You must be logged in to comment!
Views: 722
Category: Undecided
pixel
Tagged: lens sex testing vignetting halation distortion
« 
pixel
 »
pixel