MPE focus tests.

[BasAndrews]

It seems I am not the only one who has had issues with the MPE and potentially soft focus.

To help those who also may be affected I suggest posting or a link to a set of standard shots to compare the lens with different settings. If those who do not suffer this issue have the time and inclination to post cropped images at 100%, it would be useful to compare, and much appreciated.

My attempts are in a set:
http://www.flickr.com/photos/basandr...7630406887062/

All the MP-E shots were taken on a crop sensor (600d).

There are examples of the same subject at 1x, 2x and 5x at F13 and F7.1. Each has been posted as a normal shot and also cropped at 100%. There is also a shot at approx 2x using the tamron lens and tubes. That is to compare the softness with the same settings between lenses.

100% crop at 1:1 F13

MPEx1F13-crop by BasAndrews, on Flickr

100% crop at 5x F7.1

MPEx5F7.1-crop by BasAndrews, on Flickr

[Canon Bob]

Having a shot library of a standard subject isn't a bad idea Bas, but I think going for a crop % (100% in your request) will be misleading. The relative pixel density of the different bodies will result in output magnification changes.
A full frame sensor will show a 7.2mm wide frame at maximum magnification. Cropping the image to 0.5x will frame 3.6mm and cropping APS-C to .75x will yeald the same (0.64x for APS-H bodies)

This way we'll see all images at the same final magnification irrespective of pixel count.

I hope that was clear and the reasons understood.

Bob

[alquimista]

first of all thank you Bas for staring this tread, I will also do a series of pictures using the different magnification distance you have used on your images, however I'm interested on what Bob posted about the difference between crop and full sensors. Bob honestly I'm not great at doing the math; is there a way you can give us a number we should be saving our images at? Pixel count at x dpi so that all images equal out to the same percentage, is there a formula such as camera megapixel times or divided by x that will give us the ratio we should save the image at? Thank you in advance for your input

[BasAndrews]

Thanks Bob, but the main aim of this was to see the clarity of the subject, so we can compare how the detail of the image varies at different magnification. If it works out that images are clearer on a FF sensor then that is useful info.

My perception of the pictures above is that the shot at 5x does not give any individual area that is sharp focus, whereas the image at 1:1 is fine even at 100%. OK the D0F will play a large part, but there should be some part of the image that is crisp. If anyone else perceives that differently then please say so, part of the problem could simply be that what I see at 100% is what is expected.

Comparing the two lenses I use at similar mag:
Tamron at 2x

tamronx2f13-crop by BasAndrews, on Flickr

MP-E at 2x

MPEx2F13-crop by BasAndrews, on Flickr

OK the lighting is different, but the image from the Tamron lens looks slightly better to me. I will try a session using the same body and flash and change the lens to really see the difference.

[Martin G.]

Bas, did you use a tripod? What measures did you take to ensure there is absolutely no motion blur?

[Overread]

You might find this of interest:
http://www.flickr.com/photos/2453447...7623350445656/

On the far left is a full description of what I did for those tests. From what I can tell you're not so much testing focus as you are the ultimate possible sharpness of the lens itself. Remember this is a tricky thing, different MP cameras will show different views at 100% (typically the more MP the more enlarged a photo is at 100% view and thus the softer the results might appear to be, even though they are equally if not better when resized for normal use).

In addition I prefer test shots for sharpness to be akin to the kind I shot there; a rough 45 degree angle to a detailed flat surface. That means that you remove the problem of accidentally missing the focus, even if by a touch. You'll always get a proper clear band of in-focus coin in every single shot. Making it easy to perform the test and to see the sharpness.

My biggest test error would probably have been that I didn't have the flash fully fixed in position - a fixed light source, fixed coin and stable tripod for the camera would all help accuracy of the test (in short ensuring an even level of contrast in all photos).

edit -
http://www.flickr.com/photos/2453447...7623359678512/

a 1:1 test between the MPE - Sigma 70mm and 150mm macro lenses

[BasAndrews]

First, thank you all for the assistance and comments.

Martin, these were taken handheld (lens resting on a table with the subject on the same table). I originally tried a set with the tripod but it was taking forever to set up each shot.

As the percieved problem is ay higher mag, I will try the coin shots (if I have any left after buyng the lens ) using mirror lockup on a tripod.

Thanks overread, much appreciated.

[alquimista]

Wow this is getting good! I'm going to be setting up my gear tonight and do the test, will be posting results soon, will do the coin shot.

[AbPho]

Try to keep your effective aperture no smaller than f/22 or your going to get a hugh quality hit due to diffraction.

effective aperture = nominal aperture * (magnification + 1)

so, your 1:1 at f/13 = 13*(1+1) = f/26. Not bad
5:1 at f/7.1 = 7.1*(5+1) = f/43. Not so good.

At 5:1 on the MP-E use f/4. It's a lot sharper than at f/2.8 and you get a (very) little bit more DOF. Your DOF is going to be around 60µm. Not very much.

[Lester Wareham]

What apature are you shooting at Bas, I don't stop down any more that f/8 at 2:1 and above. At 4:1 or 5:1 I might open up to f/5.6.

All this is to avoid difraction, but of corse you are trading dof for peak sharpness.

[Lester Wareham]

Originally Posted by Lester Wareham

What apature are you shooting at Bas, I don't stop down any more that f/8 at 2:1 and above. At 4:1 or 5:1 I might open up to f/5.6.

All this is to avoid difraction, but of corse you are trading dof for peak sharpness.

Also be aware diffent lenses will have different pupilary magnications. This means the effective apature at the same magnification will be different.
See http://www.zen20934.zen.co.uk/photog...ro_Photography

[BasAndrews]

Excellent information Lester and abpho, thank you. I was not aware of the different effective aperture at the same mag. It seems maths gets in the way of everything in photograhy

Bob from the the forum kindly sent me some shots taken at the same mag and aperture (but on a IDMKIV) and the shots were noticebaly clearer. That leaves the camera, lens or technique on my end as being a problem.

I do not believe the camera is an issue, so it is me or the lens. To be sure I may hire a lens and then I can put the blame squarely on my shoulders if there is no difference. BTW this does not stop me using the lens, most of the really naff shots are obviously me and I can get some pleasing results, so I am just attempting to get the best I can from the gear and also fix whatever it is that I do wrong when taking my shots.

[alquimista]

thanks for the Info Lester! and thank you Bas for starting this tread, I had family issues this week and was not able to do any shooting or testing all the information posted here is of great help and Im sure will help me improve my shooting

[Lester Wareham]

No probs.

I did do some relative sharpness tests on various macro solutions using a bank note as a target. Of course getting the same bit of the note and ideal focus is very hard for getting criticle results.I used an angle finder C at the time because the body I was using did not have live view.

These are here but only include life size I think, but do include a range of set f-stop.

[Canon Bob]

Originally Posted by Lester Wareham

....Of course getting the same bit of the note and ideal focus is very hard for getting criticle results......

For higher magnification tests, (5x-20x) I've used an Emery board (disposable nail file), and it becomes quite intuitive. Trying to keep it perpendicular is a challenge but any errors are obvious with a shift of the region in focus to one side or the other,

Bob