Good first experience with (used) Mitutoyo 20x
Moderators: rjlittlefield, ChrisR, Chris S., Pau
Good first experience with (used) Mitutoyo 20x
My lens adapter arrived yesterday, so I was finally able to test my recently purchased used Mitutoyo 20x/0.42 (#378-804-2) for the first time. As suggested here, I did a direct comparison with the Nikon CFI 10x/0.25 (#MRL00102) which I purchased new some months ago. The results were as expected (the Mitu outresolving the Nikon), but given the risk factor when buying used objectives it was still nice to see this confirmed so I wanted to share the results here.
For this test I used the wing of a moth (Noctua sp.). Here are some side by side comparisons:
Animated GIFs:
https://c1.staticflickr.com/5/4897/4508 ... f7c4_o.gif
https://c1.staticflickr.com/5/4911/3194 ... 5e82_o.gif
https://c1.staticflickr.com/5/4892/4576 ... 888f_o.gif
Full-size stack with the Mitutoyo:
https://c1.staticflickr.com/5/4835/4489 ... eece_o.jpg
Full-size stack with the Nikon:
https://c1.staticflickr.com/5/4854/3087 ... 79fa_o.jpg
The Nikon stack was enlarged to match the Mitutoyo stack (I didn't want to lose any information by down-sampling the 20x image). I used the same tube assembly as shown here (when using the Mitutoyo the only difference is the M52 - M26 lens adapter). The images were stacked with Zerene (DMap to preserve information which was needed for this test), 40% unsharp mask 1px radius. Canon EOS 5D Mark II, ISO 100, diffused flash. The step size was 3um (should have been 2um).
My final conclusion is that I'm happy with the Mitu (thanks Javier ) and that this is a good place to buy used equipment !
Any feedback is welcome.
For this test I used the wing of a moth (Noctua sp.). Here are some side by side comparisons:
Animated GIFs:
https://c1.staticflickr.com/5/4897/4508 ... f7c4_o.gif
https://c1.staticflickr.com/5/4911/3194 ... 5e82_o.gif
https://c1.staticflickr.com/5/4892/4576 ... 888f_o.gif
Full-size stack with the Mitutoyo:
https://c1.staticflickr.com/5/4835/4489 ... eece_o.jpg
Full-size stack with the Nikon:
https://c1.staticflickr.com/5/4854/3087 ... 79fa_o.jpg
The Nikon stack was enlarged to match the Mitutoyo stack (I didn't want to lose any information by down-sampling the 20x image). I used the same tube assembly as shown here (when using the Mitutoyo the only difference is the M52 - M26 lens adapter). The images were stacked with Zerene (DMap to preserve information which was needed for this test), 40% unsharp mask 1px radius. Canon EOS 5D Mark II, ISO 100, diffused flash. The step size was 3um (should have been 2um).
My final conclusion is that I'm happy with the Mitu (thanks Javier ) and that this is a good place to buy used equipment !
Any feedback is welcome.
Santiago
Flickr
Flickr
-
- Posts: 2627
- Joined: Thu Jan 17, 2013 9:34 pm
- Location: United States
- Contact:
Thank you Robert. Yes, the Nikon stack has been upscaled from 5616 px to 11178 px (width), that's 1.99x.RobertOToole wrote:At first glance the The 10x results look very soft but the image was upscaled to match the 20x right? Otherwise the 10x sharpness looks like something you find out of a defective 10x objective.
Here are 3 100% crops from a Nikon stack (raw DMaps with artifacts)... are these looking better?
Thanks for your comment. I'm always impressed by your tests btw .
Santiago
Flickr
Flickr
-
- Posts: 1511
- Joined: Mon Jan 15, 2018 9:23 pm
- Contact:
Thanks Macro_Cosmos. What could be the reason? The reversed Raynox DCR-150? The objective was purchased new some months ago from SEO Enterprises.Macro_Cosmos wrote:Still looks a bit too soft to be honest from my experience.
(Edit) To me these last 3 Nikon crops look crisper than the Mitutoyo crops (as expected). What do you think of the Mitutoyo crops, do those also look too soft to you?
Thanks in advance
Santiago
Flickr
Flickr
-
- Posts: 2627
- Joined: Thu Jan 17, 2013 9:34 pm
- Location: United States
- Contact:
Hi Santiago,santiago wrote:Thank you Robert. Yes, the Nikon stack has been upscaled from 5616 px to 11178 px (width), that's 1.99x.RobertOToole wrote:At first glance the The 10x results look very soft but the image was upscaled to match the 20x right? Otherwise the 10x sharpness looks like something you find out of a defective 10x objective.
Here are 3 100% crops from a Nikon stack (raw DMaps with artifacts)... are these looking better?
Thanks for your comment. I'm always impressed by your tests btw .
This looks about right for that objective at 100%. I'm pretty familiar with CFI Plan 10x since I was just shooting with one last month for a 10x test.
The 10x/0.25 CFI Plan is not as sharp as the Mitutoyo 10s or other higher NA Nikon 10x objectives. I think this objective's strength is coverage not sharpness anyway.
Best,
Robert
It's good to hear that!RobertOToole wrote:This looks about right for that objective at 100%. I'm pretty familiar with CFI Plan 10x since I was just shooting with one last month for a 10x test.
I have a beginner's question: to calculate the step size for stacks I use the formula 0.5/(NA^2). For NA 0.42 this gives 2.834um. But does magnification play any role at all? If I were to push down the magnification of the Mitu from 20x to 18x (using bellows), will the DOF increase at all? I wouldn't mind (ocasionally) trading some magnification for more DOF, so that I can use a step size of 3 instead of 2.
Thanks in advance!
Santiago
Flickr
Flickr
Not really, you're oversampling anyway, so changing mag would result only in different FOV.
That being said, you certainly can use 3um step or even bigger, without obvious banding, just decreased resolution.
If you want to experiment, check your most recent 2um stack with software stacking every 2nd or 3rd frame and compare output.
Im pretty sure you'll be hard pressed to point any difference between 2um or 3um step.
That being said, you certainly can use 3um step or even bigger, without obvious banding, just decreased resolution.
If you want to experiment, check your most recent 2um stack with software stacking every 2nd or 3rd frame and compare output.
Im pretty sure you'll be hard pressed to point any difference between 2um or 3um step.
Ok, good to know.JohnyM wrote:Not really, you're oversampling anyway, so changing mag would result only in different FOV.
That is a great way to simulate different step sizes, will definitely try that!If you want to experiment, check your most recent 2um stack with software stacking every 2nd or 3rd frame and compare output.
Thank you JohnyM
Santiago
Flickr
Flickr
-
- Posts: 1511
- Joined: Mon Jan 15, 2018 9:23 pm
- Contact:
Oh my bad, I didn't read carefully. I thought it was a mitutoyo 10x. Sorry about that.
For step size, I use 3um for 10x and would use 1-2 um for 20x and 50x magnifications. Another contributor to softness is lighting, which is the hardest aspect IMO. I've seen some photos with old finite objectives that are just amazing because of extremely well done lighting. It's an area everyone would always be able to improve in.
MC
For step size, I use 3um for 10x and would use 1-2 um for 20x and 50x magnifications. Another contributor to softness is lighting, which is the hardest aspect IMO. I've seen some photos with old finite objectives that are just amazing because of extremely well done lighting. It's an area everyone would always be able to improve in.
MC
- rjlittlefield
- Site Admin
- Posts: 23561
- Joined: Tue Aug 01, 2006 8:34 am
- Location: Richland, Washington State, USA
- Contact:
Another thing to remember is that images like these can and should be aggressively sharpened to get the most out of them. Just from diffraction, MTF drops to only around 40% at half the cutoff frequency. But you can recover that to near 100% with simple sharpening. See http://www.photomacrography.net/forum/v ... hp?t=33724 and http://www.photomacrography.net/forum/v ... 198#203198 for some illustrations and discussion.
--Rik
--Rik
That was very interesting material on sharpening and deconvolution, thank you. I tend to use little sharpening for fear of ruining the images but these posts have made me reconsider that a bit!rjlittlefield wrote:Another thing to remember is that images like these can and should be aggressively sharpened to get the most out of them. Just from diffraction, MTF drops to only around 40% at half the cutoff frequency. But you can recover that to near 100% with simple sharpening. See http://www.photomacrography.net/forum/v ... hp?t=33724 and http://www.photomacrography.net/forum/v ... 198#203198 for some illustrations and discussion.
I'm also curious to try RawTherapee's (blind) deconvolution sharpening...
Santiago
Flickr
Flickr