Zhongyi (Mitakon) Super Macro Lens (1 - 5x)

Have questions about the equipment used for macro- or micro- photography? Post those questions in this forum.

Moderators: Pau, rjlittlefield, ChrisR, Chris S.

NikonUser
Posts: 2623
Joined: Thu Sep 04, 2008 2:03 am
Location: southern New Brunswick, Canada

Post by NikonUser »

1-5x is a convenient range for use on a Nikon camera and it is a range I use a lot. A tempting purchase when they get a Nikon model
Currently I can get 5x with a reversed 28mm Nikon lens on bellows but have to change lenses to get lower mags.
Decided to compare what my reversed 28 mm gives me at 5x with this Chinese lens.
Obviously I had to use a different fly, but a reasonable match.
my 5x image straight out of Zerene was noticeably sharper than the image from the above lens; I'll pass on this lens!
NU.
student of entomology
Quote – Holmes on ‘Entomology’
” I suppose you are an entomologist ? “
” Not quite so ambitious as that, sir. I should like to put my eyes on the individual entitled to that name.
No man can be truly called an entomologist,
sir; the subject is too vast for any single human intelligence to grasp.”
Oliver Wendell Holmes, Sr
The Poet at the Breakfast Table.

Nikon camera, lenses and objectives
Olympus microscope and objectives

ray_parkhurst
Posts: 2752
Joined: Sat Nov 20, 2010 10:40 am
Location: Santa Clara, CA, USA
Contact:

Post by ray_parkhurst »

NikonUser wrote:1-5x is a convenient range for use on a Nikon camera and it is a range I use a lot. A tempting purchase when they get a Nikon model
Currently I can get 5x with a reversed 28mm Nikon lens on bellows but have to change lenses to get lower mags.
Decided to compare what my reversed 28 mm gives me at 5x with this Chinese lens.
Obviously I had to use a different fly, but a reasonable match.
my 5x image straight out of Zerene was noticeably sharper than the image from the above lens; I'll pass on this lens!
I assume the stack was done at f5.6. At 5x mag, this is f33, far into diffraction territory. Unfortunately the lens improves between f4 and f5.6, so probably can not do much better.

rjlittlefield
Site Admin
Posts: 21037
Joined: Tue Aug 01, 2006 8:34 am
Location: Richland, Washington State, USA
Contact:

Post by rjlittlefield »

mjkzz wrote:Note this was stacked from 180 images at 25um step size, total distance is 4.5mm, pretty deep one, but the final scale factor is about 0.98643 from Zerene. I remember this is a cumulative number, if backing up, on average, scale factor between each images is about .98643^(1/180) = 0.999924. I think that can be considered as near telecentric.
That number is pretty good.

We can also calculate the entrance pupil position. Scale change of 1:0.98643 in 4.5 mm of physical movement means that the entrance pupil is located about 4.5/(1-0.98643) = 332 mm in back of the focus point. You should be able to confirm that easily enough by direct observation, looking into the front of the lens when it is set the same way.
ChrisR wrote:It doesn't change size, working distance, iris shape, range - a lot to like.
I don't know what "doesn't change size" means. In the video, between 2:11 and 2:18, it's clear that the lens extends a lot when magnification is changed.

I notice that the aperture is strictly manual with no preset ring. If you've never run into one, a "preset ring" is an adjustable mechanical stop with detents, which in conjunction with a clickless aperture ring lets you quickly change from wide open to stopped down without having to count clicks. They used to be popular on inexpensive telephotos and would be helpful here too, I think.

--Rik

mjkzz
Posts: 1237
Joined: Wed Jul 01, 2015 3:38 pm
Location: California/Shenzhen
Contact:

Post by mjkzz »

Thanks Rik, I will certainly try that. I was surprised to see it is close to near telecentric.

Yes, the lens extends out when changing magnification. At 1x, it is about 120mm long, at 5x, it is about 170mm long.

I think this is an engineering copy (close to production but not there yet), the aperture ring is clickless without detents. I do not think it is calibrated yet.

mjkzz
Posts: 1237
Joined: Wed Jul 01, 2015 3:38 pm
Location: California/Shenzhen
Contact:

Post by mjkzz »

ray_parkhurst wrote:
mjkzz wrote:
Or maybe you can show me how you did it to see difference
I just took small 500x500 crops and lined them up. Here are the crops:

BTW, you state that the 3rd image is a half-stop, so ~f4.7, but I see the name of the file as f5.6, which is a full stop.

f2.8
Image

f4.0
Image

f5.6
Image

Edited to add: here's an animated gif going between f4 and f5.6 showing the change in mag:

Image
Thanks, here is what I did:

1. load 1st one
2. create a new layer on top of the 1st
3. load 3rd and copy it into memory
4. paste the memory into the new layer
5. set layer blending mode to "differene" (using Paint.Net)
6. move the top layer around till you get this:

Image

I do not see too much of variation in terms of size, else, it would have visible border or much more non-black spots in the final image.

ray_parkhurst
Posts: 2752
Joined: Sat Nov 20, 2010 10:40 am
Location: Santa Clara, CA, USA
Contact:

Post by ray_parkhurst »

I don't know how well that method shows differences between the images. The simple overlay/animation I made shows the mag change clearly.

Note that the mag change is not such a big deal except it is an indication of defects in design or manufacturing. I'm sure Rik could relate it back to particular aberrations that the focal length changes vs aperture might cause.

mjkzz
Posts: 1237
Joined: Wed Jul 01, 2015 3:38 pm
Location: California/Shenzhen
Contact:

Post by mjkzz »

Here, I used a Canon 6D + 100mm macro to "look" into Zhongyi Super Macro, set at 5x, from front of it. Though it is not at 332mm (which should be in the middle of 0.31 and 0.35 marks on the 100mm macro lens), it is so close and given the following analysis, I think it matches Rik's theoretical value.

let f(s) = 4.5 / (1 - s) where s is the final scale factor obtained from Zerene. Excuse me for forgetting how to take get 1st order derivative for f, I have been out of school for 20 years :D, but numerical analysis shows df/ds centered around s=0.98643 is 24438.62. This means, a 0.1% (0.001) change in scale factor can result in 24.4mm change.

Anyways, thanks, learnt another thing.

BTW, it is so hard to align optical axes to get clean image of the aperture without proper equipment, if I look into view finder in photo mode (vs video mode), I can clearly see the full aperture, but in video mode, to show it on the LCD, only partial image. I hope it will not distort the experiment.

Image

rjlittlefield
Site Admin
Posts: 21037
Joined: Tue Aug 01, 2006 8:34 am
Location: Richland, Washington State, USA
Contact:

Post by rjlittlefield »

About the scale change between the f40 and f56 images, I agree that there is one but I'm not confident assigning a cause.

For visual analysis, I agree with Ray that the flash-to-compare test is much more sensitive than displaying the difference in pixel values.

The experiment that I did, even more sensitive, was to pull the images into Photoshop, roughly align the two by full pixel shifts, crop to common region, export the cropped images, then pull those into Zerene Stacker and ask it to precisely align them.

A flash-to-compare test on the resulting Aligned images then showed little to no change in scale, and the Scale correction reported by Zerene Stacker was 1.00203 or 0.2%, pretty close to the 0.3% suggested by Ray.

The reason I write "little to no" change in scale is that in the Aligned images, I think I see a bit of scale change along the upper-left-to-lower-right axis, but none along the upper-right-to-lower-left. That difference is difficult for me to understand in terms of a lens aberration, given that the measured crops seem to be positioned in the frame so that those two axes in the crops would be symmetric with respect to the optical axis. I'm considering it a complete mystery for now.

As for what sort of aberrations could cause the major part of the scale change, I'm thinking that nothing more than coma is required. That could also explain the softness at wide aperture. But I haven't worked through the details of even this much, so I'm offering it as nothing more than a possibility.

--Rik

rjlittlefield
Site Admin
Posts: 21037
Joined: Tue Aug 01, 2006 8:34 am
Location: Richland, Washington State, USA
Contact:

Post by rjlittlefield »

mjkzz wrote:let f(s) = 4.5 / (1 - s) where s is the final scale factor obtained from Zerene. Excuse me for forgetting how to take get 1st order derivative for f, I have been out of school for 20 years Very Happy, but numerical analysis shows df/ds centered around s=0.98643 is 24438.62. This means, a 0.1% (0.001) change in scale factor can result in 24.4mm change.
Wolfram Alpha reports that "derivative s 4.5/(1-s)" is 4.5/(1-s)^2, which evaluates to 24437.27738 .

But I never trust symbolic manipulation without confirming against a numerical calculation, so I treat this as confirming both results.

More directly, changing s by +0.001 results in changing f(s) by +26.38 mm, while changing s by -0.001 results in changing f(s) by -22.76 . I think this gives an even better feel for variation in the sensitivity, which I find to be a common result of numerical investigations versus purely symbolic.
mjkzz wrote:if I look into view finder in photo mode (vs video mode), I can clearly see the full aperture, but in video mode, to show it on the LCD, only partial image. I hope it will not distort the experiment.
I agree, this should not affect the results.

--Rik

Macro_Cosmos
Posts: 959
Joined: Mon Jan 15, 2018 9:23 pm
Location: Sydney

Post by Macro_Cosmos »

A friend shared some photos after playing with one at a Beijing show, seems pretty impressive to me. The lack of CA is welcomed, as the laowa 25mm has quite a lot of that.

I'll be looking forward to further tests, the near telecentric performance should make stitching easier and the long WD is great for all kinds of lighting.

mjkzz
Posts: 1237
Joined: Wed Jul 01, 2015 3:38 pm
Location: California/Shenzhen
Contact:

Post by mjkzz »

OK, added one more suggestion IF they are going to make major changes -- maybe it is POSSIBLE to make it even more telecentric by moving the aperture assembly so it would appear close to infinity when viewed from front. But it looks like this is a bit too much of change.

Did a stack and stitch of a coin to test how it performs doing stitching. Microsoft ICE could not stitch it for some reason, but photoshop did it. Have to reduce file size by reducing image quality (81%). Large TIFF file is here: https://www.dropbox.com/s/54zxfo9h3mt7i64/Coin.tif?dl=0

Video showing how it was done (well only one a scan pass): https://youtu.be/CwLrW-Tqa6s
As can be seen in the video, working distance is long enough (3x magnification, WD=120mm), I do not think it would cause much lighting issues where I have 4 150Wx4 continuous lights (total 2400W) inside 4 120x90 softboxes from about 2 feet away around.

Image

Macro_Cosmos
Posts: 959
Joined: Mon Jan 15, 2018 9:23 pm
Location: Sydney

Post by Macro_Cosmos »

That's impressive, even without sharpening, I'm fascinated by the amount of detail. The lack of CA is an absolute winner, now I'm wondering how the 1x and 2x ends compete against my PNs. Hope the retail price is fair.

mjkzz
Posts: 1237
Joined: Wed Jul 01, 2015 3:38 pm
Location: California/Shenzhen
Contact:

Post by mjkzz »

Macro_Cosmos wrote:That's impressive, even without sharpening, I'm fascinated by the amount of detail. The lack of CA is an absolute winner, now I'm wondering how the 1x and 2x ends compete against my PNs. Hope the retail price is fair.
I added some sharpening (25% unmask) after stitching. The reason is I was using a flat profile, 0 contrast, 0 sharpening, 0 saturation (I actually set that to all profiles or using Sony's term "creative styles" vs their true profiles)

The price could be around 600USD or maybe less, not sure, but my personal opinion is that this is a versatile lens.

ray_parkhurst
Posts: 2752
Joined: Sat Nov 20, 2010 10:40 am
Location: Santa Clara, CA, USA
Contact:

Post by ray_parkhurst »

Looks pretty good! Interestingly, there is one tile in upper right corner that did not stitch together well, and that tile is also showing higher CA than the others. Was it processed differently such that the program could not get a good alignment?

mjkzz
Posts: 1237
Joined: Wed Jul 01, 2015 3:38 pm
Location: California/Shenzhen
Contact:

Post by mjkzz »

oh wow, good catch . . . I uploaded wrong file -- it was done with my own software and if the algorithm could not stitch, it simply put a tile at "its place". I did not notice that till I removed background, so I tried Photoshop. New one here

https://www.dropbox.com/s/9lrbgzjmgtn6dbu/Coin.tif?dl=0

In terms of CA, I noticed the lens has more CA outwards (ie, edges, corners), but still much less compared to this SnS:

https://www.easyzoom.com/image/94401/editors

That one was done with a modified Canon 50mm f/1.4 with an acrylic aperture. Back then, Rik actually pointed out (essentially lead me to google how to do it) to remove CA using LR (or PS).

But I did not do anything except cropping off streaking edges (a few pixels wide), stitched it and added 25% unsharp mask (added another 25% this morning). The CA you see could be the part that is at the edge of one tile.

Unfortunately, I am too paranoid about filling up my SSD "buffer drive", I have removed all images. But I am getting something to suspend the coin so I can get cleaner background, will see.

Post Reply Previous topicNext topic