34 replies to this topic

[Andy Perrin]

(Borrowed from the lime thread.)

Mark, on 09 February 2017 - 21:01, said:

Andy, how was your SWIR image photographed? My understanding is that CMOS and CCD image sensors are photo blind above 1200 nM.

You are correct. The cutoff is actually closer to 1125nm.

So the SWIR is photographed using a fluorescent screen that upconverts SWIR photons into NIR photons. Physically, it is the anti-Stokes effect. The converter tube and screen are the CamIR Adapter, which is the same as the Edmund Optics version, which I obtained on eBay. I paid about 12% of Edmund's price!




(I believe this to be the cheapest SWIR camera setup in existence! ...and it was still too damn expensive. )

This is with the Cine Velostigmat 25mm/F1.5 (20", F/1.5, ISO1600)


With a Thorlabs 50mm SM1-mount lens (F/2)

[Andrea B.]

That certainly is an impressive looking rig!

What investigations do you hope to make with this set-up??

[Andy Perrin]

Just exploring! I have started already with the limes, as you saw. Anything with water in it seems to go dark, including humans. That's the overriding conclusion so far. I have no idea what else I'll find as I proceed.

Edited by Andy Perrin, 10 February 2017 - 21:57.

[Hornblende]

You could try to demonstrate the transparency of a piece of sillicon metal in IR!

[Andy Perrin]

Heh, yes, if I had a piece of silicon!

[Andy Perrin]

Glass of water

[Hornblende]

This is so cool.
We are drinking Coca-Cola without noticing it.

Edited by Hornblende, 15 February 2017 - 23:38.

[Cadmium]

You can also use RG1000 stacked with S8612. I know this sounds crazy, but I have seen it done.
I would not have thought it possible, because the sensor is limited to about 1200nm, and the stack transmission is very low,
but with a long exposure I have seen this work the same way. Water is black (dark) with the stack.
I have not tried this myself yet, but I will.
Here is a graph I made showing 3mm thick RG1000, the test I saw used 2mm thick.

[Andy Perrin]

That's not possible (at least for the same reasons). Show it to me and I will try to find an explanation, but it's not shortwave infrared! Silicon goes transparent at 1125nm. Water darkens at 1400nm or so.

Water has a number of dark regions other than SWIR, however.

Edited by Andy Perrin, 16 February 2017 - 03:02.

[Hornblende]

With my Zwo astro-camera and an IR 850nm filter, water looks darker than usual.
I will post a picture tomorow.

[Andy Perrin]

It does get a bit darker even in the NIR. This graph is from Wiki:
https://upload.wikim...iquid_water.png

But note that the graph is log-log, so it jumps by several orders of magnitude in SWIR.

Uploader gave a link to a list of references, and it matches other plots I googled. (By the way, it is important to google LIQUID water...the vapor spectrum is not the same.)

Edit: This link has an example with some "dark water" near the very bottom of the page. It does not go as dark as in SWIR, but it is certainly noticeable!
http://www.randombio...90infrared.html

Edited by Andy Perrin, 16 February 2017 - 05:13.

[Cadmium]

Andy, you are complicating the issue by showing a false color IR image as an example of dark water. That always happens with false color IR water, you have to go higher to remove color from the water.

Edited by Cadmium, 16 February 2017 - 05:34.

[Andy Perrin]

Cadmium you are looking at the wrong pic. See the pic that is two above that.

Edited by Andy Perrin, 16 February 2017 - 06:09.

[Cadmium]

Yes, I see the correct pic now.

Well, I never said it was possible, I only said I had seen it done. ;-)
Here is my test of the same, using the same glass thicknesses as I had seen done, but I don't know the ISO and Aperture used on the test I saw.
Suffice it to say it requires a lot of exposure time and ISO. Not much fun for me. But this shows what I had seen.
This is a 1200nm+/- range, pushed to extremes, nothing more. I used RG1000 2mm thick + S8612 1mm thick (similar to the graph I posted above).
What you are seeing here is upper transmission of the S8612, as much as I can push the sensor. The RG1000 is cutting off the lower transmission of the S8612. At least that's the idea.
I am not endorsing it, I am just testing the idea.
Shot using Kuribayashi 35mm lens, D7000 full spectrum camera, incandescent light, RO/DI water.

Edited by Cadmium, 16 February 2017 - 08:11.

[colinbm]

What about S8612 by its self Steve ?

[Cadmium]

Col, S8612 alone would deliver a visual range photo. 300nm to 740nm using 1mm.

Edited by Cadmium, 16 February 2017 - 09:45.

[colinbm]

What in the world was I thinking of ??? :-(

[Andy Perrin]

Well the spectra I found show the absorption does go up quickly past 1000nm, so it's not surprising in that light (no pun?) that it looks darker. But I don't believe that's really 1200, or if so, it's just BARELY. Like the last two photons in the bucket, so to speak!

[Cadmium]

Yep, and my hand is not dark when photographed with the RG1000 + S8612 stack.

[Andy Perrin]

Miso bottle

Visible (iPhone 6S Plus), 1000nm LP filter (absorption glass from eBay, not sure which glass)

SWIR using a modification of the above setup. I have now got an iris, a lens hood lined with Acktar Metal Velvet, and an SWIR BBAR-coated 50mm lens from Thorlabs. The latter increases the transmission quite a lot, enough that I can stop down to F4 now.

Main difference seems to be that the miso bottle cap goes from light to dark.

Edited by Andy Perrin, 18 February 2017 - 07:58.