Any experiences with the Thorlabs 2x superapochromat?

Justwalking · Post by **Justwalking** » Tue Jun 12, 2018 11:33 am

Lou Jost wrote:Justwalking, what if those resolution figures refer to resolution on the subject rather than on the sensor?

I bet your camera sensor have lower resolution.

https://www.edmundoptics.com/resources/ ... esolution/

Lou Jost · Post by **Lou Jost** » Tue Jun 12, 2018 12:36 pm

If the "120 line pairs per mm" refers to the resolution in image space, then according to the table in your link, my 20Mp MFT sensor slightly out-resolves this lens when not using pixel-shift, and greatly out-resolves it when using pixel-shift (effectively 50Mp, so 2.1micron pixel width, Nyquist limit 227 lp/mm).

Nevertheless this lens has a higher NA than my other options, so it is interesting.

Justwalking · Post by **Justwalking** » Tue Jun 12, 2018 12:53 pm

Lou Jost wrote:If the "120 line pairs per mm" refers to the resolution in image space, then according to the table in your link, my 20Mp MFT sensor slightly out-resolves this lens when not using pixel-shift, and greatly out-resolves it when using pixel-shift (effectively 50Mp, so 2.1micron pixel width, Nyquist limit 227 lp/mm).
.

Oh! I did not know you camera. Thought there is less MP. Then it is problem to find outperfoming optics to take maximum. Do you prepare 50?P pic for some since magazin?

Justwalking · Post by **Justwalking** » Tue Jun 12, 2018 12:57 pm

Oh! I did not know you camera,Lou. Thought there is less MP. Than it is problem to find outperfoming optics to take maximum. Do you prepare 50?P pic for some since magazin?
(#####, i have again general error from board engine during repla?)

Lou Jost · Post by **Lou Jost** » Tue Jun 12, 2018 1:05 pm

Yes, it is challenging to out-resolve it. I publish scientific photos in my articles, and I sometimes make larger-size images for exhibition, so I try hard to get the best resolution possible.

Justwalking · Post by **Justwalking** » Tue Jun 12, 2018 1:13 pm

But as i understand pixel-shift can not give your real more optics resolution.
It's help for color clarity and redusing noise. Think 20 MP lens for 4/3 probably is maximum today on free market.

Lou Jost · Post by **Lou Jost** » Tue Jun 12, 2018 1:17 pm

No, there are two kinds of pixel-shift implementations. You are correct for Sony and Pentax (whole-pixel shifts, four shifts per image) but I have an Olympus which does half-pixel shifts, 8 photos per image, and this does produce real increased resolution. I have made many photos with sub-pixel resolution.

Lou Jost · Post by **Lou Jost** » Tue Jun 12, 2018 1:18 pm

I have asked Kowa if their lens resolution figures refer to object-side or image-side. Will report back here when they tell me.

Justwalking · Post by **Justwalking** » Tue Jun 12, 2018 2:09 pm

Lou, anyway each time your sensor needs only 20MP from optics and then you can shift and combine separate frames as you need in memory. But yes, finish results will be more in MP.

Keep in touch about Kowa's answer.

Lou Jost · Post by **Lou Jost** » Tue Jun 12, 2018 3:08 pm

Yes, I think you are right about that.

ray_parkhurst · Post by **ray_parkhurst** » Tue Jun 12, 2018 3:29 pm

Even though each sub-image is only 20MP, for the composite to have higher resolution the lens must support the resolution of the composite. The sub-pixel information must be there for the lens to sample with the sub-pixel shifts.

Lou Jost · Post by **Lou Jost** » Tue Jun 12, 2018 3:47 pm

I'm not sure. Can you give an argument in favor of that?

I do know that the Oly algorithm stops working at f/8. That number is more like the aperture at which a 20MP MFT sensor becomes diffraction-limited, not the aperture at which a 50Mp or 80MP MFT sensor becomes diffraction-limited.

rjlittlefield · Post by **rjlittlefield** » Tue Jun 12, 2018 3:50 pm

Justwalking wrote:(#####, i have again general error from board engine during repla?)

This is caused by a character set problem.

From http://www.photomacrography.net/forum/v ... 207#224207 :

rjlittlefield wrote:The place to look for problems is where that "?" appears in the list of words. In this case it's at '?lens' and '?na'.

The underlying issue is that for historical reasons the forum database uses a character set named "latin1" (description: https://en.wikipedia.org/wiki/ISO/IEC_8859-1). latin1 is only a subset of Unicode. If a post uses any Unicode characters that are not contained in latin1, then the post will preview OK but produce the message about "Illegal mix of collations" when it gets Submit'ed and stored in the database.

It's a very annoying problem that was caused by a change in the PHP specification several years ago. If you're curious, see the 3-page discussion at http://www.photomacrography.net/forum/v ... hp?t=25154 .

--Rik

ray_parkhurst · Post by **ray_parkhurst** » Tue Jun 12, 2018 4:07 pm

Lou Jost wrote:I'm not sure. Can you give an argument in favor of that?

I do know that the Oly algorithm stops working at f/8. That number is more like the aperture at which a 20MP MFT sensor becomes diffraction-limited, not the aperture at which a 50Mp or 80MP MFT sensor becomes diffraction-limited.

At best, the up-sampling algos simply mimic a higher resolution sensor. A low resolution lens on high resolution sensor will produce a low resolution image. If it were not so, where is the limit? If the lens was not the limit, then you could continue to sub-sample forever and achieve infinitely small resolution.

Lou Jost · Post by **Lou Jost** » Tue Jun 12, 2018 6:02 pm

Looking at their other spec sheets, the resolution is always 120 lp/mm regardless of lens or magnification. That means it must be the resolution on sensor, otherwise it should be roughly proportional to magnification.

But I am concerned that the spec sheet says that the resolution is appropriate for a 4.5 micron sensor. The Oly sensor is 3.3 microns.