UV stereo Panasonic 12.5mm lens test

[dabateman]

After seeing Bernard's interesting stereo Images, I thought I would try my own twist on this theme.

 

Panasonic released a 3D lens for M43rds cameras which is 12.5mm f12. But its field of view is closer to 60mm equivalent. It has 4 lens elements, which sounds good for UV work. Also I was able to get a good deal on a brand new one.

 

The Olympus EM5mk2 and EM1mk1 both accept this lens as purchased. However the newer Olympus cameras including the EM1mk2 doesn't. But when you place this lens on the cameras that accept it, the camera forces you to switch to the 3D scene mode, locks the ISO at 800 and locks up most settings. So adding tape to the electrical contacts allows the camera to actually use the lens to its full potential.

 

Bernard if you reading this, you could get a cheap copy and add a M43rds to E-mount adapter and use this lens directly on your Sony. As the electrical contacts only hurt you and its best at manual. There is no focus ring either, its fixed with the aperture.

 

The minimum focus distance is 24 inches (60 cm), which is not good for its stereo ratio. However, you can take the back mount off and shim the lens with 1mm washers, this allows the lens to focus from 6 inches to 9 inches for Macro.

 

Testing its range I found that it was only one stop slower than my UAT using a 370bp15 filter and 2 365nm LED lights. Which is quite good. However I could not get any signal using a 335bp10 filter. So it may be a UV-A lens only.

 

This is a straight out of camera image to show the visible flowers, at a focus distance of 6 inches. The lens mount has been shimmed with 1mm spacer:

 

This is a straight out of camera image using Baader Venus U filter mounted to the front of the lens using two 365nm LED lights. The exposure was 4 seconds, ISO 200 at its fixed f12 aperture.

 

You can cross your eyes to see these images directly in stereo, which I like. Interestingly I need to back away from the flowers using the Baader venus filter, to get a sharper image. I will have to crop, save, and try the software Bernard recommended to see if I can get them into cyan/red. But I think I like the crossed eyed 3D look better. Let me know if anyone can see this in 3D.

[dabateman]

Interesting, on my phone the above image pops perfectly for me. I see a great 3D image. However, when I click on the image its harder to cross my eyes to get the same stereo view. I think I may stick with posting at 800 pixels than.

[ins13]

Very cool!) stereo :) I look at it with PC on a distance of about 1 meter

I can't see Bernard's stereo Images. i dont have red-cyan glasses....

[dabateman]

Thanks Eka,

Its becoming a fun lens to play with. I was imaging flowers yesterday and works well. On the Em5mk2, I can zoom in focus 7x on left side to get the flower and then move the camera forward or backward to trap focus. I think I like a 1mm shimm and may not go more. The standard focus range is ok, but in macro its much more fun.

If you get one, a brass #2 washer would be the best to shim it with. They are 0.5mm thick, so you can adjust the macro distance.

[bvf]

Hi, Da. Interesting different approach to taking 3D. I find the cross-eye approach somewhat uncomfortable, so I created a red/cyan anaglyph of your UV shot.

 

Not entirely sure how this lens works. The lenses are only 1 cm apart, which is very little. So should only get a good stereo effect very close us (say less than 1 foot). But the reviews say you get a poor stereo experience at less than 3 feet. Perhaps it works on a toe-in system.

 

It will be interesting to see what you can get out of it.

 

[dabateman]

Bernard,

Yes this is an interesting lens, that Panasonic may have shot them selves in the foot with. You can't really read reviews on it, as the off the shelf lens isn't really useful. The lenses are really close together, and its minimum focus distance is terrible.

How it really works is only through modification. It needs to be adjusted to focus close, this can be done by taking the lens mount off and adding spacers. Thus moving the lens to macro only.

I have found macro stereo (3D) really hard for me, as the distance is hard for me to predict. However, with this lens its easier as the elements are where I need them, its just finding the correct focus distance. If a ring could be added to scale in the macro adjustment, then this would be perfect. Sadly, I don't think there is enough space for that, as I would like to shift the lens from 0.8mm to 2.4mm into macro. Thats a very small adjustment.

As of now I need to play around more. But 1mm seems to be a happy compromise.

[dabateman]

Update,

I have been playing more with this lens. Using my 1mm shimms I slightly more accurately measured the focus distance. When using my 720nm IR filter and a 75W hallogen bulb the focus is 9 to 10cm, from front of lens to subject.

In visible with a BW486 filter and same hallogen bulb, the focus range is 10 to 11cm.

Using 2 365nm led bulbs and either a baader venus filter or a UG1 2mm filter the focus distance is 14.5 cm.

 

This lens also works for UVA, visible and IR on my stock Em5mk2. Whats fun is it lets through enough light, that the high resolution mode can be used to really squeeze every thing good out. There is a significant improvement using high resolution mode. So even though its an f12 fixed aperture lens, it still works out well in high resolution mode.

I also bought some #2 brass washers, so I may play around with different shims to see the limit to maintain a good 3D image and close UV/visible distance. I am thinking 1.2 or 1.5mm may be best.

I got my washers direct from mcmaster.

https://www.mcmaster.com/92916A310

 

 

[bvf]

Are you gong to post any of your results?

[dabateman]

Bernard here are some photos from my full spectrum EM1 of a test flower:

 

Visible using 486 filter:

 

IR using 720nm Long pass filter:

 

Visible (mostly blue) and UV, using no filter and two 365nm LED lights:

 

UV using Baader venus U filter with the two 365nm LED lights:

 

Here is full High res EM5mk2 image resized. Yes I know this is pointless to load, but you get the Idea with the Baader venus U fitler on a stock camera:

 

Here is a 1:1 crop from the full resolution Jpg from the EM5mk2. I am not sure how this will turn out.

 

To best show the HighResolution files, I will need to learn how to use the software you recommended and then you can zoom in in Blue/Red.

[bvf]

Thanks for that.

 

As usual, I had to convert them to anaglyphs because I can't do the cross-eye thing well. Be interesting to see how good they are when working with the full-res images.

 

In the software I use (Stereo Photo Maker) I can't see a way to import a side-by-side pair in a single image, so I think you'll have to split your images into pairs and import the pairs. That's do-able (and how I created anaglyphs from your images), but it's a bit more time-consuming.

 

I guess you could also print your images and use a Victorian-era stereo viewer for side-by-side images - but you won't be able to zoom in, of course.

[Ultrapurplepix]

Some time ago I did some experiments with 3D X-Ray, 3D visible and 3D thermal, in a mix of cross-eyed and anaglyph. I wasn't overly impressed with my (admittedly fledgling) results. The visible ones were probably the best, which were made with a Loreo 3D lens that was unfortunately rather poorly aligned due to a manufacturing defect. The IR experiments were the traditional move-the-camera type, while the X-Ray was a crude move-the-object.

 

A multi-spectrum 3D image would be an interesting challenge: a mix of see-through (X-Ray), visible (probably from greyscale) and temperature (from LWIR). Any takers? :)

[Andy Perrin]

How did you come by an X-ray machine is the question uppermost in my mind!

[dabateman]

I have not attempted x-rays. The best I have gone is transmission UV. You need thin flowers for that. I should try that in 3D.

[Ultrapurplepix]

How did you come by an X-ray machine is the question uppermost in my mind!

 

Oddly enough, I first encountered my X-ray machine via eBay. They crop up often enough here in the UK and I believe they're more commonly available in places like the US.

 

Mine started life as a mailroom scanner (it's about the size of a 4-drawer filing cabinet). It's intended for office use so it has all the necessary shielding permanently in place, unlike the completely open hospital types.

 

I have been obliged to modify the imaging system substantially. Originally it used a (very) high brightness, low-resolution fluorescent screen, which was observed by a TV-resolution CCD camera and displayed, via a framestore, on an analogue monitor.

 

Incremental improvements have included

- replacing the screen with a (smaller) Kodak Lanex Fine screen robbed from a re-usable hospital dry imaging cassette

- replacing the 45° imaging mirror (it was a regular bathroom mirror tile; now it's a first-surface type)

- replacing the low res CCD camera with a remote-controlled Nikon D700 and 50mm f/1.4 lens

 

All told, this has increased the resolution from about 1 line per mm to 10 lines per mm (x100 better resolution), giving something that can, with care, see the bond wires inside an IC package:

 

 

The bond wires are the hair-thin wires between the tips of the spider-like "lead frame" and the actual chip, which is mounted on the central pad. They're moch more visible if you click-through and access the original via my Flickr page.

 

Bond wire sizes vary with purpose, but the ones seen here are probably in the region of 50µm diameter (typical range is ~10µm to ~100µm, though bond wires up to 500µm or more can be used in special, high-current applications.

 

 

Combining X-Ray and thermal has proved interesting, though there remains much to be done to improve the images (this was made using an earlier, lower-res X-Ray setup):

 

 

All good fun. You can click on each image to access its 'home' page, where you'll find much more information.

 

I got X-Ray images like this from the unmodified machine:

 

 

After the mods I was seeing things like this:

 

 

and things have improved, in smaller steps, from there onwards.

 

There are much better X-Ray imaging systems out there (eg a 'microfocus' X-Ray tube and a solid state image sensor, which is a cross between a large-screen TV and a camera sensor), and gives far better resolution and dynamic range than I've ever achieved with my setup.

[Andy Perrin]

This is amazing stuff. Your modification was certainly very very effective!

 

(it's about the size of a 4-drawer filing cabinet)

Sigh, sounds like I won't be getting one any time soon. I have very little space, living in a tiny studio apartment in Boston!

[Cadmium]

Yes, this is amazing. I really like the X-ray + Thermal composite shot.

[Ultrapurplepix]

This is amazing stuff. Your modification was certainly very very effective!

Thank you.

igh, sounds like I won't be getting one any time soon. I have very little space, living in a tiny studio apartment in Boston!

I have had numerous exchanges with a 'Professor of Magic Moonbeams' who specialises in radiation safety, particularly X-Ray. Using my figures from a very crude measurement method, he calculated that if I X-Rayed my hand, once, using my normal settings, I would receive an X-Ray dose that exceeded two years' civilian whole-body limit. Early-on I had decided not to try X-Raying anything living (except e.g. plants); I am glad I stuck to my decision.

 

I further learned that X-Rays can cause all sorts of secondary effects such as causing materials to re-emit energy at lower, but still dangerous, wavelengths (think UV exciting visible fluorescence) and there's no telling what directions that might fly off at, or its intensity. This would be of particular concern with the 'open' style (unshielded) devices.

 

The lesson I take from this is, if you understand how dangerous short wave UV is, imagine what it's like when the photons are several orders of magnitude more energetic!

 

I'm not trying to stop anyone experimenting but please look after yourself well enough that you'll be able to write up what you did!

[dabateman]

Wow, the improvement you achieved in the resolution is amazing.

Yes the safety would be a major concern. I hope you also have body sheilding that you wear or at least stay out of the room when imaginging for 15 minutes post max dose. Remote timers, wifi cameras or even USB tethering to a distant computer can help for safety.

There are supposed to be some limits on the things you can order, even off ebay. The one I find a little silly is centrifuges. Unless you have a qualified lab purchase number. But I know things get out all the time.

You should create a new post with your work. These are really interesting images

 

[Ultrapurplepix]

@Da Bateman - thanks. As mentioned, the device I have is fully shielded, being intended for use in a mailroom by untrained persons. And I've left almost all of the interlocks in place ;) with the exception of the keyswitch, because I don't have a key so I had to jumper it out. There's no-one under 40 in the house: I would take a very different view of the safety implications if I had children living here, or even visiting.

 

Perhaps I will create a new post looking at ultra-wide multispectral imaging. The trouble is finding a suitable subject. Flowers aren't terribly interesting at LWIR (or, usually, NIR), and can be hard to do much with in my X-Ray setup. I think the X-Ray source is too 'hard' - high energy - to work well with botanicals: the magic moonbeams go straight through. I did image a cabbage leaf a while ago, with difficulty:

 

 

...but the original contrast was extremely low and if you look closely at the original you'll see it was achieved more by noise-density-modulation (my term, which I've just coined) rather than grayscale.

 

You can find more of my images on my business card website or my Flickr account. Most of the latter I regard as 'engineering' photos, rather than high art!

 

(At the time of writing there appears to be a widespread DNS problem that, among other things, causes some browsers to display an old, very broken page referring to Flickr instead of my business card: it's beyond my control but I hope the internet imps will be able to fix it).