1:20 to 1:5 (0.05x to 0.2x)

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

Moderators: rjlittlefield, ChrisR, Chris S., Pau

jvanhuys
Posts: 49
Joined: Sun Jan 22, 2023 9:24 pm
Location: Seoul, South Korea

Re: 1:20 to 1:5 (0.05x to 0.2x)

Post by jvanhuys »

chris_ma wrote:
Tue Jan 24, 2023 6:45 pm
jvanhuys wrote:
Tue Jan 24, 2023 5:59 pm
So I had a look at that 'd.fine HR-M 2.8/80 0.09x' and it seems literally perfect for my needs. I contacted the distributor I'm hoping it'll be around $1000USD or less. The Rodenstock you mentioned is out of the question due to the price though. Thanks again.
great, that should be an excellent fit. Let us know if you get any pricing info.
from the predecessor lenses in their line up I'd expect something in the 2500-3000USD range.
the 120 float is about 5000USD here in europe.
the xenon sapphires are also in the 4000-5000USD range new if I recall properly (I got one second hand for a bit over 1000).
Thanks Chris for the headsup. I was definitely not expecting those price ranges... clearly I need to manage my expectations here. Will let you know once I get the quote on the 'd.fine.'

lothman
Posts: 957
Joined: Sat Feb 14, 2009 7:00 am
Location: Stuttgart/Germany

Re: 1:20 to 1:5 (0.05x to 0.2x)

Post by lothman »

I assume for aquiring the images to be stitched you move the camera sensor in a big image circle of your lens. Therefore you need the best available lens. Correct?

But why can't you move the camere together with a usual macro lens an stitch those images together. This would give you more resolution with a much cheaper lens like the mentioned new Sigma 105mm.

Lou Jost
Posts: 5944
Joined: Fri Sep 04, 2015 7:03 am
Location: Ecuador
Contact:

Re: 1:20 to 1:5 (0.05x to 0.2x)

Post by Lou Jost »

First, that's the wrong lens, the macro lens is f/2.8. It is for full frame. But that image circle number is wrong even for the mentioned 105 1.4 lens, which is also for full frame.

The right lens is this one:
https://www.lenstip.com/595.4-Lens_revi ... ution.html

https://www.youtube.com/watch?v=yPJ1uTfKJDY

https://www.bhphotovideo.com/c/product/ ... ltEALw_wcB

The field is quite flat but perhaps enlarger lenses are flatter. Stacking fixes any small deviations from flatness, if any.

This is widely considered to be among the best commercially made macro lenses ever. Pixel shifting can make up for the FF image circle; it gives you medium format resolution on a FF camera.

patta
Posts: 86
Joined: Mon Apr 20, 2020 9:51 am
Location: Stavanger Norge
Contact:

Re: 1:20 to 1:5 (0.05x to 0.2x)

Post by patta »

A bit of a silly observation - hope not to offend...
- the main complaint of current lenses seems to be Chromatic Aberration;
- the artworks are black and white.
--> can't you go monochromatic?? No more CA! Either get illumination with blue LEDs; or a narrowband color filter (see Baader); or from the image, just keep one color channel & dump the others.
It's cheating but works; maybe also (more work) add a layer with color but lower resolution.

Other almost censored ideas:
- flatbed scanner?
- An X Y Z automated movement, to take many small photos efficiently, may cost less than some above mentioned lenses

patta
Posts: 86
Joined: Mon Apr 20, 2020 9:51 am
Location: Stavanger Norge
Contact:

Re: 1:20 to 1:5 (0.05x to 0.2x)

Post by patta »

I see that you're based in Korea...
The world's best lenses for your job are already (all) there, in the factories of LG and Samsung to inspect TV screens. Line-scan for large format, price new is a couple of cars.
But, there is an urban legend saying that they change those lenses pretty often and sell the old ones for scrap... I'd give it a try.

jvanhuys
Posts: 49
Joined: Sun Jan 22, 2023 9:24 pm
Location: Seoul, South Korea

Re: 1:20 to 1:5 (0.05x to 0.2x)

Post by jvanhuys »

patta wrote:
Wed Jan 25, 2023 7:46 am
I see that you're based in Korea...
The world's best lenses for your job are already (all) there, in the factories of LG and Samsung to inspect TV screens. Line-scan for large format, price new is a couple of cars.
But, there is an urban legend saying that they change those lenses pretty often and sell the old ones for scrap... I'd give it a try.
It's a great idea. Regarding black and white, I think I did mention it has some accents of color here and there, but primarily B&W. Great idea though. Thank you. I will look into the Korean 2nd hand market for industrial items. There are some sites I know of, just never thought about the line inspection lenses from Korean factories.

jvanhuys
Posts: 49
Joined: Sun Jan 22, 2023 9:24 pm
Location: Seoul, South Korea

Re: 1:20 to 1:5 (0.05x to 0.2x)

Post by jvanhuys »

Lou Jost wrote:
Wed Jan 25, 2023 4:49 am
First, that's the wrong lens, the macro lens is f/2.8. It is for full frame. But that image circle number is wrong even for the mentioned 105 1.4 lens, which is also for full frame.

The right lens is this one:
https://www.lenstip.com/595.4-Lens_revi ... ution.html

https://www.youtube.com/watch?v=yPJ1uTfKJDY

https://www.bhphotovideo.com/c/product/ ... ltEALw_wcB

The field is quite flat but perhaps enlarger lenses are flatter. Stacking fixes any small deviations from flatness, if any.

This is widely considered to be among the best commercially made macro lenses ever. Pixel shifting can make up for the FF image circle; it gives you medium format resolution on a FF camera.
Hi Lou,

Thanks for pitching in.

That is no doubt a good lens, however I mentioned earlier that I see my monorail at the bottom of frame when photographiing larger artworks. I can't drop my front standard lower than the tripod head, and I definitely won't rotate the camera for a stitch, so the only option is to rise the back standard. For this I need at least 60mm IC, which I don't think the Sigma will cover.

Look forward to hearing from you again.

chris_ma
Posts: 570
Joined: Fri Mar 22, 2019 2:23 pm
Location: Germany

Re: 1:20 to 1:5 (0.05x to 0.2x)

Post by chris_ma »

jvanhuys wrote:
Wed Jan 25, 2023 8:57 am
That is no doubt a good lens, however I mentioned earlier that I see my monorail at the bottom of frame when photographiing larger artworks. I can't drop my front standard lower than the tripod head, and I definitely won't rotate the camera for a stitch, so the only option is to rise the back standard. For this I need at least 60mm IC, which I don't think the Sigma will cover.
what I don't quite understand is why you need a monorail, specially with the sigma which has a native sony E-mount available?
even with an industrial lens, it seems better to use a bellows or helicoid system for variable magnification then a monorail since it's much easier to keep parallelism.

I don't think you'll find a lens with top image quality and an uniformly sharp image circle larger than 60mm for sub 1000usd, specially with that magnification range.
there are some lenses in the second hand market which are quite affordable which are really good, but they are usually at fixed magnification and more in the 0.3x to 1x range.
chris

rjlittlefield
Site Admin
Posts: 23561
Joined: Tue Aug 01, 2006 8:34 am
Location: Richland, Washington State, USA
Contact:

Re: 1:20 to 1:5 (0.05x to 0.2x)

Post by rjlittlefield »

jvanhuys wrote:
Tue Jan 24, 2023 5:48 pm
I slightly disagree over CA and geometry correction though, as there is definitely a filtering hit when either of those are applied.
I agree that this happens, but whether it limits the final result is a different issue.

The ideal situation would be to have a lens with a perfectly flat field, and no CA, and crisp over a 600 MP area, then physically scan your sensor over that area so precisely that you don't ever need to interpolate pixel values. Barring that, you'll have to do at least one step of sub-pixel interpolation. In the pano stitching software that I know, standard practice is to compose all the steps of geometry calculations into a single overall coordinate transformation. That way only a single stage of pixel interpolation is required to get from source images to output. But I am not aware of any packages that bundle CA correction into that pipeline. Likewise for focus stacking.

So, using off-the-shelf software to do each of those steps separately, you'll be looking at a chain of three pixel interpolations instead of just one. That may sound pretty painful, or as chris_ma notes, it may be OK in practice.

In any event, there's nothing that says you have to start with only as many pixels as you want in the final product. If you want a super high quality 600 MP result, then one option is to set up your optics and workflow to produce a 2400 MP stitch, followed by 2X downsampling to get 600 MP. That way essentially all the softening from multiple pixel interpolations happens at a level that gets averaged away in the end.

I will happily grant that this approach seems rather inefficient at first glance. Whether it actually is depends on all the tradeoffs combined. The original gigapixel panoramas, which really did have resolvable detail at all those pixels, were made using lenses that would have struggled to deliver 50 megapixels in their optical image. The trick, of course, was to use a long lens that made that optical image cover only a tiny part of the subject, and scan that high quality area over the subject to accumulate the desired result. It's a way of trading off time against equipment availability.

lothman makes the same point in fewer words:
lothman wrote:
Wed Jan 25, 2023 4:27 am
But why can't you move the camere together with a usual macro lens an stitch those images together. This would give you more resolution with a much cheaper lens like the mentioned new Sigma 105mm.
--Rik

jvanhuys
Posts: 49
Joined: Sun Jan 22, 2023 9:24 pm
Location: Seoul, South Korea

Re: 1:20 to 1:5 (0.05x to 0.2x)

Post by jvanhuys »

chris_ma wrote:
Wed Jan 25, 2023 9:54 am
jvanhuys wrote:
Wed Jan 25, 2023 8:57 am
That is no doubt a good lens, however I mentioned earlier that I see my monorail at the bottom of frame when photographiing larger artworks. I can't drop my front standard lower than the tripod head, and I definitely won't rotate the camera for a stitch, so the only option is to rise the back standard. For this I need at least 60mm IC, which I don't think the Sigma will cover.
what I don't quite understand is why you need a monorail, specially with the sigma which has a native sony E-mount available?
even with an industrial lens, it seems better to use a bellows or helicoid system for variable magnification then a monorail since it's much easier to keep parallelism.

I don't think you'll find a lens with top image quality and an uniformly sharp image circle larger than 60mm for sub 1000usd, specially with that magnification range.
there are some lenses in the second hand market which are quite affordable which are really good, but they are usually at fixed magnification and more in the 0.3x to 1x range.
Hi Chris, I really enjoy the monorail for the incredibly precise focus adjustments. I got a bit bored rotating my variable length extension tube countless times for stacking etc. So a monorail was a renaissance for me. It's so precise and everything is lockable. I should add that I have geared movements, on rise, fall, shift, tilt and swing, so it's a lot of fun. Totally agree with you on the lens prices. I'm asking for too much. Cheers.
Last edited by jvanhuys on Wed Jan 25, 2023 10:42 pm, edited 1 time in total.

jvanhuys
Posts: 49
Joined: Sun Jan 22, 2023 9:24 pm
Location: Seoul, South Korea

Re: 1:20 to 1:5 (0.05x to 0.2x)

Post by jvanhuys »

rjlittlefield wrote:
Wed Jan 25, 2023 4:59 pm
jvanhuys wrote:
Tue Jan 24, 2023 5:48 pm

In any event, there's nothing that says you have to start with only as many pixels as you want in the final product. If you want a super high quality 600 MP result, then one option is to set up your optics and workflow to produce a 2400 MP stitch, followed by 2X downsampling to get 600 MP. That way essentially all the softening from multiple pixel interpolations happens at a level that gets averaged away in the end.
--Rik
My gosh this part is interesting. I was sort of subconsciously sticking to my Symmar-S 180mm (1970s, single-coated) for closer shots and more stitching... not really thinking about why my prints look sharper with it, than more modern, smaller IC lenses with lower magnification.

Could you elaborate at which point the oversampling cancels out the gain in sharpness? Like, at which point does it average out? Is there a rule or a formula to follow? I can of course Google this, but I always prefer speaking to an informed person first.

chris_ma
Posts: 570
Joined: Fri Mar 22, 2019 2:23 pm
Location: Germany

Re: 1:20 to 1:5 (0.05x to 0.2x)

Post by chris_ma »

jvanhuys wrote:
Wed Jan 25, 2023 5:21 pm
Could you elaborate at which point the oversampling cancels out the gain in sharpness? Like, at which point does it average out? Is there a rule or a formula to follow? I can of course Google this, but I always prefer speaking to an informed person first.
that depends on the resolution of the full optical imaging system.

let's say as an argument you're using a 40MP camera, but use a lens that can only resolve 10MP.

in this case you need at least 4 captures (more with overlap) for a 160MP image (containing 40MP of real resolution) and then downsample to 25%. you'll also want to account for the debayer resolution loss but in this example this is so small that it will be a vanishing factor.

it's simple to do a test:
capture a sample image with your setup, downsample by a factor of X with a lanczos filter, break concatenation of the scaling, upscale by a factor of 1/X.
see at which point the result still looks identical to the original image (ignoring debayer artefacts and noise)
chris

jvanhuys
Posts: 49
Joined: Sun Jan 22, 2023 9:24 pm
Location: Seoul, South Korea

Re: 1:20 to 1:5 (0.05x to 0.2x)

Post by jvanhuys »

chris_ma wrote:
Thu Jan 26, 2023 12:17 am
jvanhuys wrote:
Wed Jan 25, 2023 5:21 pm
Could you elaborate at which point the oversampling cancels out the gain in sharpness? Like, at which point does it average out? Is there a rule or a formula to follow? I can of course Google this, but I always prefer speaking to an informed person first.
that depends on the resolution of the full optical imaging system.

let's say as an argument you're using a 40MP camera, but use a lens that can only resolve 10MP.

in this case you need at least 4 captures (more with overlap) for a 160MP image (containing 40MP of real resolution) and then downsample to 25%. you'll also want to account for the debayer resolution loss but in this example this is so small that it will be a vanishing factor.

it's simple to do a test:
capture a sample image with your setup, downsample by a factor of X with a lanczos filter, break concatenation of the scaling, upscale by a factor of 1/X.
see at which point the result still looks identical to the original image (ignoring debayer artefacts and noise)
Hey! How do access the filtering algorithm in Photoshop? I can only access it in Nuke, which is totally overkill for still images. Also Nuke is procedural, not exactly fun to work with layers. I've attached the filtering algorithms available to me in Nuke. Interesting that you mentioned Lanczos, as it's generally what we use in high-end VFX, Lanczos6 to be specific.

Why would you break concatenation though? That incurs a filtering hit... that is not good. The point of concatenation is to give a net transform with a single filter hit instead of multiple ones. Or are you saying you need to break concatenation to enable the effect of oversampling? If so, how would that work?
Attachments
filtering.jpg

patta
Posts: 86
Joined: Mon Apr 20, 2020 9:51 am
Location: Stavanger Norge
Contact:

Re: 1:20 to 1:5 (0.05x to 0.2x)

Post by patta »

I couldn't resist, the observation by Rik that, in practice, we are looking for an objective lens with a native resolution of 600 MP, prompted the urge to make this graph: How much a lens cost vs its resolution. It is a quick ballpark; Hubble and the scanner are in just for fun
Attachments
Lens cost vs resolution - ballpark Jan2023 (5).png
Last edited by patta on Thu Jan 26, 2023 3:42 am, edited 3 times in total.

jvanhuys
Posts: 49
Joined: Sun Jan 22, 2023 9:24 pm
Location: Seoul, South Korea

Re: 1:20 to 1:5 (0.05x to 0.2x)

Post by jvanhuys »

patta wrote:
Thu Jan 26, 2023 2:53 am
I couldn't resist, the observation by Rik that, in practice, we are looking for an objective lens with a native resolution of 600 MP, prompted the urge to make this graph: How much a lens cost vs its resolution. It is just a ballpark...
hahahahahaha thanks for that. Just finished work. That was what I needed to combat the blues!

Post Reply Previous topicNext topic