Doing a lens comparison

I want to do a side-by-side comparison of IQ between two lenses. I want to compare sharpness, color, and contrast.

What would you recommend I set up to do this in the best 'almost scientific' manner?

As long as the lenses are of a similar focal length. Shoot the same scene with the camera locked on a tripod. Ideally care would be taken to focus properly. Live view zoomed in perhaps. Make sure the scene is in good light. Take as many variables outr of the mix as possible.

Tripod, tripod, tripod.

Yes, tripod, MF, LiveView 5x 10x, remote switch.

As others have said, a Tripod is obviously necessary.

Just take a your time getting the right focal lengths and comparing results. I've done it before and found it does take a bit of time.

Then in Photoshop map the layers on top of each other and do your comparison there.

will the comparison be at the same ISO, focal length and scene (you have to choose the scene wisely so as to look for differences - whether it be flare, high contrast situations, low contrast situations, 2d chart, 3d scene, question of mirror lock up, question of exposure of the picture and same light source....)

I would use someone to put in the frame. Lots of photographers do take photo's of people. They want to see how the lens performs when shooting people (skin tone, sharpness on a person). So, I would take your gear to a location, where you can get a nice background, set your camera on the tripod. focus with manual on the person in the frame. Take the picture and put the other lens on the camera en shoot the exact same shot.

Here is an example:

http://www.youtube.com …Cg3fg&feature=c​4-overview

ceriltheblade wrote in post #16640261
will the comparison be at the same ISO, focal length and scene (you have to choose the scene wisely so as to look for differences - whether it be flare, high contrast situations, low contrast situations, 2d chart, 3d scene, question of mirror lock up, question of exposure of the picture and same light source....)

Yes. Exposure will all be the same, as I'm going to do as much of a controlled environment as possible.

Things like tripod, LV, MF, etc., were a given for me. I'm more or less asking about what 'subjects' would be the best for a comparison. I'll be using an ISO 12233 chart, possibly the Kodak Q-60 chart, and this one (which is more of a printer test chart, but has color and skin tones).

As for including a person makes sense, but couldn't things like subject movement (during lens changes, breathing, natural movement) skew things a bit?

And as for ensure the duplication of FL's, what is the best way? I'm sure the scale(s) on the lenses are a great guide, but I doubt they are exact.

Give us a hint as to which lenses you are comparing. DPreview has a still life studio scene they shoot when they test dynamic range and IQ for camera body tests. I would think a similar scene would be a fair test. Maybe place a few canned goods, some random stuff off your desk, things with some color, and some blacks, whites, etc. The point is to be able to put the shots side by side on the screen to look for differences in the output.

^^^
agreed. I really like the comparison of real things in 3d for many of the concepts you mentioned you wanted to test : sharpness, contrast, and color. a human is usually problematic and won't sit still for as long as you'd like - and they have this annoying thing they do - move! [and some of them speak too! gasp!]

but I would also suggest having more than one scene
one with high contrast objects and low contrast objects as well.

in regards to the focal length - how much of a difference are the focal lengths?

Focal lengths are almost identical.

KirkS518 wrote in post #16640991
Focal lengths are almost identical.

Then dont worry about it too much.

The test procedure using a USAF test chart is to place the chart at either 55 or 27 times the fl of the lens being tested. Then your chart comparisons should be pretty close. There is also an instruction to print 5 charts, one for the center and each corner. This will give you an objective resolution comparison as well as color fringing and falloff. Shoot over a range of f-stops. The metadata will track the info unless you are testing a manual focus legacy lens. Resolution isn't everything, so do some field tests under similar conditions.

i would do it both automatic center and manual center for both lenses.
This way you can eliminate any bad focusing variables... I'd hate to find out I have a really tack sharp lens only for the automatic focus to be way off and need ma adjustment.
Also, eliminate any uv filters.

eddie3dfx wrote in post #16641389
i would do it both automatic center and manual center for both lenses.
This way you can eliminate any bad focusing variables... I'd hate to find out I have a really tack sharp lens only for the automatic focus to be way off and need ma adjustment.
Also, eliminate any uv filters.

I'll be using 10x LV and manual focus only for the test.

and what is this thing you speak of called - uv filters? Never on any of my lenses.