Magnification, sharpness, diffraction, aperture and the MPE65mm macro

[Overread]

I did this test to see the effects of the differing apertures and magnifications just to see how this lens does perform before I really got a chance to test it in the field and some others might find the results interesting

Setup:
Camera and lens:
Canon 400D
Remote cable release
Canon MPE65mm macro
Teleconverter used for last tests - Sigma 2*teleconverter
Hoodman anglefinder - set to 2.5* magnification to aid focusing

Lighting:
Canon 580EX2
Offcamera flash cord
Lumiquest softbox

Support:
Manfrotto 055XPROB tripod
Manfrotto Junior geared head
Ebay focusing rail (single section) (note apparently also the same as the Adorama focusing rail)

Subject setup:
Single 2 pence UK coin
Single LED bulb torch with flexi arm

Camera settings used for each shot:
1/200sec shutter speed
ISO 100
Aperture (varies)
Manual flash power output (varies)



The full setup:

Camera and lens mounted using the lenses tripod collar (sold with the lens) onto the focusing rail; the rail being attached to the quick release plate of the geared head and the whole setup mounted on the tripod. The geared head was set to an angle (as shown on its head markings) of just over 45 degrees - this tilted angle was delibate as I wished to know the performance of the focusing rail whilst at a tilted angle with the whole setup. In addition a tilted angle lets the varying depths of field be displayed on the coins. Due to limitations in gear the flash had to be handheld by the photographer.

Each shot was focused by using the LED torch as a point of bright light, the light shining onto the coin surface. This proved to be essential at the greater magnifications as the image through the viewfinder would become very dark, coupled with a very small depth of field this further increases the difficulty of accurate focusing. The 2.5* setting on the Hoodman angle finder proved to be very beneficial for the aid of focusing at such fine magnifiations - though those with liveview can use up to 10* magnifcation on the preview image for even further fine focusing.
Focusing for each setting was achived by moving the focsing rail and not adjusting the focus on the camera, for this purpose in the tests. In this the focusing rail performed very well, allowing easy movement of the rail whilst the pressure screw was applied and not requiring any tightening further once the hand is removed from the rail controls. In the field further finer focusing could be done by adjusting the focus on the lens itself, unless preserving a fixed magnification was needed.
Test shots were taken and the flash power adjusted accordingly based on the output of the camera histogram - note that this along with the handholding of the flash itself did lead to some inconsistant results - especailly where the angle of the flash changed between shots. AT this small distances even small changes in the flash distance and angle could lead to larger changes in the resulting image. To counter this a better test would be to have the flash mounted on a fixed stand

Once each series of shots was taken the magnification was adjusted to the next level (as written on the lens barrel) and the focusing rail repositioned to aquire the new focus.

Test shots were then uploaded to the computer and the RAWs were processed - however only the white balance was adjusted for these shots (and the brightness on a very few). Even when output as JPEGs the sharpness, contrast and other settings have not been adjusted from the cameras base settings. This results in a softer image than would be possible by using tools such as the clarity slider in RAW processing and the unsharpen mask in the final stages of image editing = however it gives a level field that all the shots can be viewed on, rather than show up slight differences in the amount of sharpening applied (since whilst the amounts could be fixed each shot would have required different amounts in normal processing). In addition the major dustspots were also removed (least I think I got most of the major ones). More minor dustspots which would normally be removed from proper images were left in.

Link to the image set on flickr:
http://www.flickr.com/photos/2453447...7623350445656/
The title of each image denotes the aperture used and the magnification factor of the shot. Note that fullsized versions of each shot are availible for review

Test results - after a quick review of the presented images at fullsize (you know its quicker to view them on flickr than it is to wait for windows image viewer....)
1:1

Here we can see that f8 is clearly the sharpest aperture to use of those tested. However results from both f10 and f13 show little difference between each other and would still be very usable when sharpened in editing. The results from f16 however are showing a far more noticable effect of diffraction (espeacially when compared to the f8 results)

2:1

Here we can see that things have got harder still, f13 is now more comparable to that of f16 from before. F10 is fairly usable still however, whilst f8 remains the sharpest aperture from those displayed.

3:1

f16 is now becoming a very soft aperture to use indeed and f13 is quickly following it. f10 would be the highest suitable aperture at this magnification now. Also note how dustspots are becoming far more of a problem in these shots (remember only major spots were removed the rest remain)

4:1

f10 might just be usable, but its really softened a lot since the smaller magnifications. In all honestly f8 would be the new ideal aperture to use for this magnifcation.

5:1

The first thing to notice is that f5.6 has appeared on the scale now, this is in responce to the constant softening of results and whereby f8 might no longer remain the most suitable aperture to use. Secondly looking at the results its clear that whilst f16 gives a massive depth of field for this range the results are very soft indeed. Even opening all the way up to f8 we still have a very noticably soft image - f5.6 might be the most desirable aperture for this magnification - even though it makes things a lot harder with its far finer depth of field.

10:1

Ok not as strict a test here since lower apertures are missing, but still at this high magnification one might be more forced to pick the smaller f8 over f5.6 just because of the improvement in depth of field that the aperture brings.

So to summaries the results from the aperture tests:
magnification - optimal aperture
1:1 f13
2:1 f13
3:1 f10
4:1 f8
5:1 f5.6 however in practice f8 might be a more workable aperture due to the small depths of field

remembering that this is as tested on a 400D and that other camera bodies might give slightly varying results.

Test improvements:
select a single point on the coin to be the point of focus for all the shots.A left and bottom line on the image would also allow easier compartive results in the difference of subject area shown at each different magnification. Further a fixed flash position would have allowed for more consistant results with the flash head to have been recorded - lessening differences in sharpness due to contrast changes.

[Overread]

A comparison between the Canon MPE65mm macro, the Sigma 70mm macro and the Sigma 150mm macro.

Images from the test can be found here: http://www.flickr.com/photos/2453447...7623359678512/

An early assessment of this test shows that whilst there is a little difference betwen the frames covered at 1:1 (as marked onthe lenses) the frame area covered is basically the same and for a field shot the difference would not be noticable.
Further whilst an above view gives a good idea of frame area it does appear to suffer the limitation that a focus error during shooting can cause possible erroring results - a repeat test might be performed with the setup at an angle so as to reduce the chances of a focus error giving incorrect results.

Preliminary results from this test though I would say:

At f8 65mm clearly the sharpest, followed by the 150mm. The 70mm is showing considerable softness and this might be a missfocus.

At f10 65mm is clearly the sharpest again, followed by the 150mm and 70mm which show very similar levels of performance (further suggesting that the f8 results for the 70mm might be flawed.

At f13 the 65mm and 70mm give very similar performance whilst the 150mm appears to be more noticably softer.

As in previous tests these results are unedited save for white balance setting in the RAW editing process (all other settings were kept as standard). Further even though results from the 70mm and 150mm might be in question the obvious evidence is that the 65mm performs to a fantastic level of sharpness at each of the tested apertures

[Overread]

IF anyone is interested John Hallmen has now completed a comparison test between the MPE65mm and a selection of his other highmagnification setups.

http://www.flickr.com/photos/johnhallmen/4320502042/

A very interesting read which shows that there are other options out there for the macro shooter and that several of them can infact beat the MPE65mm macro, especailly when starting to work at the much higher magnifications