Animated 3D rendering
rmpsrpms
Posts: 1,928 ✭✭✭✭✭
Here's the mintmark of a 1954-S RPM #4 Lincoln Cent, shot with 20x objective.
35
rmpsrpms
Posts: 1,928 ✭✭✭✭✭
Comments
Ray your stuff is off the charts cool! Curious how many hours it took to make this.
WOW! Let me be the first to say mighty impressive! Awesome, even.
Can you explain how you did it?
THX
That. Is. Awesome.
This looks like some of the new technology coming out for imaging lands and grooves in forensic firearms examinations. Good stuff.
It's already been said, but WOW, AWESOME, and AMAZING. For something as numismatically simplistic as a mintmark, this display is quite mesmerizing.
How are you creating this image?
I always click on your threads. Amazing
Digital dynamite! Fraudsters beware. Peace Roy
BST: endeavor1967, synchr, kliao, Outhaul, Donttellthewife, U1Chicago, ajaan, mCarney1173, SurfinHi, MWallace, Sandman70gt, mustanggt, Pittstate03, Lazybones, Walkerguy21D, coinandcurrency242 , thebigeng, Collectorcoins, JimTyler, USMarine6, Elkevvo, Coll3ctor, Yorkshireman, CUKevin, ranshdow, CoinHunter4, bennybravo, Centsearcher, braddick, Windycity, ZoidMeister, mirabela, JJM, RichURich, Bullsitter, jmski52, LukeMarshall, coinsarefun, MichaelDixon, NickPatton, ProfLiz, Twobitcollector,Jesbroken oih82w8, DCW
Wow, very impressive
Latin American Collection
Wow, thanks everyone for the kind words. Most of the work is done by software and mechanical automation, so not a huge amount of time invested. In this case I was testing a new 20x microscope objective and it was pretty sharp so I decided to publish the result.
The process is:
Shoot 25 images with the 20x objective, with 5um steps between each image: 3 minutes
Pre-process the images for levels, etc: 2 min
Stack the images for 2D rendering: 3 min
Adjust the parameters for 3D rendering and save 90 perspective images: 5 min
Upload 90 images to animated gif website, adjust animation parameters, and run animation: 5 min
Save image: 3 min
Write post and upload image to CU: 2 min
Total time spent: 25 minutes. A fairly long time, but it is faster when you do things more in batch mode.
http://macrocoins.com
It's amazing! Very cool! Thanks for sharing
HAPPY COLLECTING
Fantastic! Can you do a couple to show doubled die vs. machine doubling?
And I was thinking PCGS could be interested in you and your process. It's an obvious for special situations and educational presentations.
Wow, "sweet" !!!
Cool
POST NUBILA PHOEBUS / AFTER CLOUDS, SUN
Love for Music / Collector of Dreck
coming in for a landing on the ridge inside the top serif... retros firing......
That's wild. Thanks for sharing.
I've done some individual renderings, not animations, mainly because of limitations in being able to upload and view the fairly large animated gif files to PB. The last forum upgrade enabled much larger ani-gifs to be uploaded directly to the forum than PB allows, so now it's practical to do these animations and share with the forum. This particular file is over 15MB, which is not really that large but PB blocks ani-gifs over 4MB. You can't do much with 4MB. Most other forums have not come out of the 20th Century yet and block anything over 300k. The new CU is great for images!
Take a look here for some static machine doubling images:
http://www.macrocoins.com/1957d-mdd.html
Perhaps I'll re-shoot one of these and do an animation on it. I think the most dramatic Doubled Die I own is a 1995P Lincoln, but I guess that would be good to show die doubling.
Sure, that would be interesting and fun.
http://macrocoins.com
You're sitting on a million dollar idea here (well maybe thousands). I hope you can patent something and monetize it. It would help collectors understand the minting process and their coins.
Cool! Have you looked into using Imagemagick to automate generating GIFs without having to upload them elsewhere? That's what I do for my stuff.
Keeper of the VAM Catalog • Professional Coin Imaging • Prime Number Set • World Coins in Early America • British Trade Dollars • Variety Attribution
Well, I've been doing this for years and so far haven't really seen much interest. I did static images for a Coin World article on Trail dies, and am working on another article now, but other than that the interest levels have been tepid. One problem is that print articles are static, so no animations can be shown, so only web-published articles can use them. Doesn't seem like it would be such a big limitation these days but it actually is.
Yeah, I've tried a few packages but I don't do it very often, and the cloud software works pretty well. If I end up doing more of them I might go local.
http://macrocoins.com
This looks like a great tool for authentication, it would be hard to hide tooling with images like this. Thanks for posting.
10-4,
My Instagram picturesErik
My registry sets
That is very impressive, superb detail and the animation covers all the angles. How far we have come.... I remember when the first picture of a coin was posted on the forum... wow, everyone was amazed.... and then pictures became common - differing of course, in quality. Now we animated 3D... Fantastic. Cheers, RickO
Have you ever tried taking coin images using a 3D camera. I actually acquired a 3D camera designed to take macro images with the intent of photographing my coin collection. Even got so far as to purchase the recommended Ektachrome film but didn't have the time to follow through. Still have the camera and the rolls of film but not sure if the film would still be useable or even developable today.
That is awesome, well done.
Nope, never tried an actual 3D camera. It would not work at these magnifications anyway. The depth of field is only ~5um so you need to take a lot of images to get the whole thing in sharp focus.
Now, it's possible to render 3D images from the data taken for focus stacking. See below...this is a "crosseyed" 3D image, which seems easier for most folks to make work:
http://macrocoins.com
Or here is a 3D rendering looking from an angle:
http://macrocoins.com
This thread is great! I never bothered to open it until today due to its title.
Do you have any coins with added numerals or letters? It should be really informative to see the "undercut" and missing "metal flow" up the side of this type of alteration. Another SUPER image would be to see a common 1969-S 1c with strike doubling!
PM me if you need an added MM and I'll find one for you to keep for your trouble.
A 3D camera isn't necessary. For a stereoscopic view, you can simply take two images, one for each eye, with the camera moved a little between shots. This is fairly easy for full-coin shots.
For stereo microscopy, you'd need to either re-create the stacked frames at each of the desired view points less than 1 mm apart, or use a 3D reconstruction from a single stack. Keep in mind that the surfaces of the 3D reconstruction are most accurate where they are orthogonal to the the camera.
An added mint mark may look OK from the camera's angle, but when you look at how it meets the coin surface, you may see a seam when viewed at a different angle. This seam's invisibility to the camera would be carried forward into the 3D reconstruction from the stacked images.
Keeper of the VAM Catalog • Professional Coin Imaging • Prime Number Set • World Coins in Early America • British Trade Dollars • Variety Attribution
@messydesk said: "An added mint mark may look OK from the camera's angle, but when you look at how it meets the coin surface, you may see a seam when viewed at a different angle. This seam's invisibility to the camera would be carried forward into the 3D reconstruction from the stacked images."
I'm going to give you a
rather than a BIG 100% disagree as it appears you are trying to explain why what I requested will not work. 
So, from my limited experience, an added numeral or mintmark MAY look OK in a straight down image but only if the shape of the addition is correct and the surrounding field is not all messed up by the alteration. Furthermore, when one of these additions is viewed from the side, there is usually a seam (as you stated) that is visible to the naked eye. I have images of these seams taken with a stereo microscope and electron microscope. Therefore, I see no way a 3-D image would not produce superb educational images of the seam and side. They are very hard to reproduce effectively with a stereo scope due to the "depth of field problem" and the need for a halogen light pipe system. That's why I requested the OP work his magic for all of us here.
In the 1970's we detected some extremely deceptive added mintmarks. It was obvious to us that the person (we named him/her "The Master") employed a stereo scope and some type of miniature "tool" (cut-down razorblade?) to scratch the sides of the added letters to imitate the usual metal flow as seen in the OP's Post. Very deceptive mintmarked Standing Liberty quarters are still seen - possibly from the same person.
Nice 3D image. Really enhances the 3D quality of the coin.
There is for sure less detail rendered on the edges of the features. The more vertical the edge, the less detail available. This shows up especially clearly when viewing the feature edge-on. Higher magnification, with smaller step size, helps with this. The feature edges don't actually go vertical, so some amount of detail is always present, but it ends up getting "stretched" in order to fill in any gaps in the data. I generally use telecentric objectives to mitigate this effect, but it can never be eliminated with 2D imaging.
It's true you would need to view the seam from an angle to see it clearly, but what should show up is the discontinuity in flow lines at the interface compared with other features on the coin. This still goes back to there being no vertical edges. As long as there is some slope, the discontinuity can be imaged. Now, right at the seam, there is likely some retrograde slope on the coin with added feature. The 2D image won't see this, but the lack of data in that area (given sufficiently small step size) will create an artificially-stretched void. The 3D depth map will be expecting to find in-focus pixels in that region, and should fill-in with grey pixels to indicate the gap.
http://macrocoins.com
awesome job rpms, i like that a lot
PM sent to take you up on the offer. Should be an interesting test of the technique.
http://macrocoins.com
A true, 3D image would, indeed produce an excellent view of the seam that undercut an added mint mark, but this technique is not one that makes a complete 3D map of the surface of the coin. It is a 2D image with depth information, where that depth information is capable of reproducing most of the 3D surface. You'd need additional camera projections to get a real 3D map, some of which would be able to see into such a seam, some wouldn't. Ray mentions that there would be an artifact in the image filling in a vertical surface with gray, but you'd see the same thing on a strike-doubled mint mark.
One way Ray could illustrate this (here I am offering him more work to do) is by photographing a 45° tilted coin with his setup, then creating a 3D surface. This will show what would happen where there was a retrograde slope on the surface behind other surface detail.
Keeper of the VAM Catalog • Professional Coin Imaging • Prime Number Set • World Coins in Early America • British Trade Dollars • Variety Attribution
If he can duplicate the side of a mintmark as he has done in the OP. If there is a seam present, IT SHOULD BE very obvious. That's why only a
for you until we see the results with the coins I'm sending to the OP. 
WOW. WOW.
Let's look at this from the perspective of the die used to strike the coin. With the depth map information, I can "invert" the surface of the coin, effectively giving a detailed view of what the die would have looked like that struck the coin. With the "coin" view, tall raised features are more visible, while in the "die" view the field becomes the focus. In this view, the transition from the field to the feature is much easier to analyze, and any lack of smoothness in this area due to tooling of the coin will be very obvious.
Here is a "die" view inverted rendering to show what I'm talking about:
http://macrocoins.com
Now that is cool!
One last rendering...this one shows what the die looks like around the "ridge" of the secondary mintmark. The "notch" in the die near center of the image forms the sharp ridge. It's easier to see the topography of the ridge looking at the die view versus the coin view.
http://macrocoins.com
Very attractive, interesting presentation, and images.
Now the tough questions: What can we learn about the coin that cannot be known without the large investment in time to make the animation?
Can we learn as much or more with less work?
I don't think that matters. When we are doing something we like, the time & money take a back seat.
Sure it matters - if only to the OP who created it.
But, is there more to it than the obvious?
??????????????????? Master, I never had any Philosophy classes so I'm not even at the "Grasshopper" level.
Maybe the question should be whether or not we can demonstrate something to someone else with less work. For example, would this technique showing the difference between die doubling and strike doubling save huge amounts of work when compared with explaining it to people over and over again using strictly 2D images?
Keeper of the VAM Catalog • Professional Coin Imaging • Prime Number Set • World Coins in Early America • British Trade Dollars • Variety Attribution
It's not possible to share the view on a stereo microscope with others (unless in the same room), so this is the best way I know to be able to show folks what these varieties look like at these magnifications. Perspective renderings give more information than does a static 2D image, and the animation of multiple renderings gives even more information. For sure you can get this info (and more) viewing on a stereo microscope, but the classic way to share is through photographing what you see.
http://macrocoins.com
@messydesk said: "Maybe the question should be whether or not we can demonstrate something to someone else with less work. For example, would this technique showing the difference between die doubling and strike doubling save huge amounts of work when compared with explaining it to people over and over again using strictly 2D images?"
Much clearer. I'll agree with it this way: "Maybe the question should be whether or not we can demonstrate something to someone else with less work. For example, would this technique showing the difference between die doubling and strike doubling using strictly 2D images save huge amounts of work when compared with explaining it to people over and over again?"
Creating the 3D renderings is somewhat trivial once the shots are taken and composited for high quality 2D output. So I guess the question is if we really need to create sharp, clear detail images? For decades the standard for publishing has been shots taken through the eyepieces of stereo microscopes. Most variety publications use this technique. I have never been satisfied with the clarity or detail of those shots so I put together a system to do better, both for my own use in identifying and documenting the varieties that I find and for sharing online with others. A growing number of folks are successfully trying this technique, and although it is quite a lot more work than snapping a low resolution image from a stereo microscope, or an even lower resolution one from a cheap USB "microscope", the results are indeed useful for showing details that are otherwise hard to document.
http://macrocoins.com
Ahhh! Sehr Gut!
Now we are looking below the surface and beginning to explore potential for rmpsrpms's technique. (Mark Goodman posted some similar stacked images about 4 years ago, also.) Could this - or analogous processes - also improve numismatic education in the ANA classes?
To insider - Clarity comes first, then efficiency. The first transistor was hard to do; now we sprinkle them on morning cereal.
We did something like this when we were doing receptor localization studies on mice brains while I was in college. We used topographic mapping software to collate our scans... was pretty awesome!
Your images/work is impressive...
I found a nicely strike-doubled (Machine Damage Doubling) 1958-D Cent, and shot it at 5x to show the whole date and MM. Here are the 2D image and the 3D animated rendering of the 5x shots:
http://macrocoins.com