Hack Chat Transcript, Page 2

Thomas Shaddack 3:53 PM

Could polaroid-class instant camera film work for xrays (or autoradiography)?

bruniss 3:54 PM

diagnosis: crack

Ahron Wayne 3:54 PM

Lol. Hey Bruniss

Dan Maloney 3:54 PM

They actually use both Ilford 4x5 film and Instax instant film in the project. Both work pretty well

CLamb 3:54 PM

@Dan Mahoney Photographic film? Are you imaging the film directly or is there a luminescent intermediary.

CLamb 3:56 PM

@Dan Maloney I mean "exposing" not "imaging".

Dan Maloney 3:57 PM

Not me, writing it up as a daily for the blog. Couple of guys presented the work at Schmoocon last week or so. Pretty sure they used some kind of intensifying screen, I'll have to check

Dan Maloney 3:57 PM

Nope, I take that back -- no screens, direct exposure. Long exposure times, like 120-180 seconds

Ahron Wayne 3:58 PM

Sheesh.

Dan Maloney 3:58 PM

Don't want to derail the chat, though -- just a side note. Article will be out in the next day or two

anfractuosity 3:58 PM

wow, didn't know that'd work

Ahron Wayne 3:58 PM

Any questions on the act of making pretend scrolls?

anfractuosity 3:58 PM

haha cool, what material is that

Dan Maloney 3:59 PM

Oh yeah, meant to ask about your roasted scrolls. Where does one get papyrus these days?

Will Fitzgerald 3:59 PM

What kind of ink did you use?

hannes.hoffmann 3:59 PM

Things to scan: spring onions. Preferably dried/carbonized.

Ahron Wayne 3:59 PM

Thomas Shaddack 3:59 PM

Pieces of charcoal. Also internal structure remaining from the wood.

Ahron Wayne 4:00 PM

I tried to get papyrus from a few different sources --- I mainly found small postcard sized sheets, thicker than the ancient stuff, and more realistic swathes that came directly from egypt (it literally had nile river in the return address). Amazon and Ebay were the actual sellers

Dan Maloney 4:01 PM

TIL that Egypt still has a papyrus industry...

Ahron Wayne 4:01 PM

@Will Fitzgerald , I got ink from standard fountain pens and also different India inks, which do have modern binders but are still mostly just carbon... hopefully... And I also tried to mix up my own dense inks to act as reference points, which is surprisingly harder than I thought.

Ahron Wayne 4:02 PM

CLamb 4:02 PM

Was it the carbon in the ink the component which was most radio opaque?

Thomas Shaddack 4:02 PM

barium sulfate wouldn't work?

Ahron Wayne 4:02 PM

I used a 3D printer holding a pen to try and standardize the writing process somewhat.

Thomas Shaddack 4:02 PM

it's available as pigment grade, I think.

Ahron Wayne 4:03 PM

@thomas, if I could find it as a wax stick or something, sure. I did try barium sulfate, but it's hard to write with --- insoluble.

CLamb 4:03 PM

Did you ever try to duplicate the ink on the real scrolls?

Dan Maloney 4:03 PM

Is the chemical composition of Roman ink known? There has to be some kind of record -- seems like they wrote down everything

Thomas Shaddack 4:04 PM

Pigments are insoluble (if soluble it's a dye, I think). Carbon itself. It's then about the size/shape of particles, and keeping them dispersed.

anfractuosity 4:04 PM

Have you heard of sumi ink, that uses soot too

Ahron Wayne 4:04 PM

The basic recipe for carbon ink is --- burn some fat underneath a tarp. Scrape the black stuff that forms. Profit. (in other words, india ink)

SimonAllen 4:06 PM

I wonder if carbon or charcoal ground with a pestle and mortar would work

CLamb 4:07 PM

In modern times lampblack is the carbon that is usually used as a pigment.

Thomas Shaddack 4:07 PM

Soot is better. Nanotech.

hannes.hoffmann 4:07 PM

grinding charoal down to ink particle size is a pain. You bascally need romen slave labour to grind on long enough. That, or floatation.

Thomas Shaddack 4:08 PM

or a postdoc.

wilkij1 4:08 PM

Calcium? Carbonates are pretty soluble, I think. That would triple your atomic number, which would be a big help.

Ahron Wayne 4:09 PM

There are so many factors when created simulations --- ink, papyrus, carbonization, the fact that it's not being aged 2000 years. My thoughts are that if you get enough sheer variation the ML models would figure it out. Though maybe that's why I wasn't in the running for the grand prize

Thomas Shaddack 4:09 PM

calcium carbonate is rather insoluble. if you dissolve it, the white pigment property gets gone.

Ahron Wayne 4:09 PM

@wilkij1 Chalk is probably in fact the easiest way to mark boundaries of your papyrus yeah.

Thomas Shaddack 4:10 PM

but you can precipitate it carefully to fine enough particles, and if you coat them with some surfactant to get them all electrostatically charged so they repulse each other and don't coagulate so easily, you can get a good enough ink.

hannes.hoffmann 4:10 PM

but why calcium? if what you want is just any attenuation by an ink, dissolve metal salts in it. There are tungsten salts that are less toxic than the lead stuff.

wilkij1 4:11 PM

Don't need the pigment property, just the Calcium atoms. I thinking about his need for a reference ink. Take any ink and dope it with something with a high atomic number to get it to be more RO.

Ahron Wayne 4:12 PM

The issue is ---you don't actually want to dope the simulated carbon ink itself, because it would wash out any signal that's on the real scrolls. The dense ink would be just to mark the boundaries of where the real stuff is

Dan Maloney 4:12 PM

Depleted uranium ink ftw

Ahron Wayne 4:12 PM

CLamb 4:13 PM

Ask for things to scan. I have some paint chips from the interior of a colonial era building I'd love to know more about. I was planning on using an optical microscope and a microtome to analyze them but now I wonder if they should also be scanned.

wilkij1 4:13 PM

If the material is easily available as a fine powder then you are all set. I just suggested calcium to get something soluble.

Ahron Wayne 4:13 PM

It's crude, but the white dots are dense ink and the actual carbon ink (after being in the oven) is the much, much harder to see stuff in between

Thomas Shaddack 4:14 PM

@Dan Maloney uranyl salts are a thing.

Ahron Wayne 4:14 PM

It shows up much more easily in infrared light. Hence why THZ imaging will be the way to do this in 5-20 years

anfractuosity 4:15 PM

Would THz imaging, not require a synchrotron source?

wilkij1 4:16 PM

Have to run. Thanks for taking your time to share the experience. Amazing work.

Thomas Shaddack 4:16 PM

Terahertz, the no man's land between the optics and the radio...

Ahron Wayne 4:16 PM

There's a pretty wide band of wavelengths, but it's what the airports use when you hold your arms up above you. it's like halfway between visible light technology and halfway between radio and microwaves. So it's also weird to emit and detect

Ahron Wayne 4:16 PM

Thanks! Enjoyed having you around. Feel free to reach out at any time

hannes.hoffmann 4:17 PM

Thz does not need a syynchrotron, that would be too easy. It needs its own source.

Dan Maloney 4:17 PM

So, yeah -- we usually wrap up after an hour, but I was having too much fun to notice the time. So if Ahron and everyone need to get back to the real world, we can call it here. Or, everyone can keep going.

Ahron Wayne 4:17 PM

Thanks Dan. Happy to stick around for a bit

hannes.hoffmann 4:17 PM

same

CLamb 4:17 PM

@Ahron Wayne Thank you quite and bit and everyone else who helped me gain knowledge.

anfractuosity 4:17 PM

have you see this quote on wikipedia "Most vacuum electronic devices that are used for microwave generation can be modified to operate at terahertz frequencies, including the magnetron" which sounds interesting

Will Fitzgerald 4:17 PM

Do you have an suggestions of best tools for 3d visualizations of the images created?

Dan Maloney 4:18 PM

Cool, we can wait to pull the transcript.

Ahron Wayne 4:18 PM

@Will Fitzgerald --- depends. How much money do you have?

Ahron Wayne 4:18 PM

At work, I primarily use Volume Graphics (20-50 K USD a license)

Will Fitzgerald 4:18 PM

Heh. Open source tools.

Ahron Wayne 4:19 PM

Well then. Invesalius?

Thomas Shaddack 4:19 PM

and python libraries?

Ahron Wayne 4:19 PM

Fiji has lots of extensions

Will Fitzgerald 4:19 PM

Ouch, I see a `K` there.

Ahron Wayne 4:19 PM

Yeah, industry everything is crazy. X-ray microscope I use there is 1.5 million; I can only get one in my dreams

Ahron Wayne 4:20 PM

It's not open source, but Dragonfly is probably your best bet for something you can use for free in academics. Extremely powerful.

Ahron Wayne 4:21 PM

There were lots of utilities made to do this for the Vesuvius Challenge, which might be worth checking out

Will Fitzgerald 4:22 PM

That's helpful. I an ML guy, and have a colleague who worked with the actual Herculaneum scrolls on site, and we are just dipping our toes into the Vesuvius project.

Ahron Wayne 4:22 PM

The visualization of tiff stacks and meshes and such.

Ahron Wayne 4:22 PM

Oh cool! are you on the discord server yet?

Will Fitzgerald 4:23 PM

Yes, I am, as `willf`.

Ahron Wayne 4:24 PM

Well then, cool.

Will Fitzgerald 4:24 PM

I've used the various tools from the project, but honestly haven't done a lot in the image domain, so it's still all quite new to me.

Will Fitzgerald 4:25 PM

by 'the image domain' I just mean, um, images.

Thomas Shaddack 4:27 PM

I did some experiments with 3d visualisations (eg. gcode toolpaths) in browser using Babylon.js framework. Can support even various 3d display standards.

Ahron Wayne 4:28 PM

The silence was me and will connecting on discord ;)

Ahron Wayne 4:29 PM

I think we're just about towards the end of this

Will Fitzgerald 4:29 PM

Thanks again, so much. Very much appreciated.

Thomas Shaddack 4:29 PM

For the 3d displaying, could game engines be leveraged?

Ahron Wayne 4:29 PM

Totally going to put some new things in the machine. Thanks for the discussion and ideas everyone. Always enjoy talking about this. Feel free to shoot me a message at any time

Dan Maloney 4:29 PM

Seems so. Very much appreciate your time today, this was fascinating!

Thomas Shaddack 4:29 PM

In-de-ed!

Dan Maloney 4:29 PM

Thanks all! Transcript coming up

hannes.hoffmann 4:30 PM

Thanks!

Hack Chat Transcript, Page 1

Discussions

Become a Hackaday.io Member