Well, I wouldn't classify it as an imaging microscope. However, given that it's capable of allowing users to view and take very minute measurements, I think I'll categorise it as a profiling microscope. I guess that still makes it a microscope 😁🤷🏾♂️🤷🏾♂️
Hey, at 3:05, I'd recommend you edit that part out, putting out your fingerprint like that might not be safe. Edit: Yeah, we don't know who it's from or which finger it might be, so it can still be safe. Just wanted to let you know though.
I'd love to know more about the gel they are using, It seems absolutely incredible that a gel would conform to such small details so perfectly and yet have no memory in that it returns to it's original shape.
You can see on the lego that it has some issues with cornering. Around the base of the lego nubs it has some issues. Stuff like that may get worked out though in the future.
@@PS-bc5qt I think that tracks with what the guy was saying about the limitations - it won't perfectly show sharp angles but for the application it's not a huge issue as you're not looking at that directly
This would be awesome to generate height maps for making smart materials in substance. Just noticed the height map at 2:41 and wanted to plug it into substance. Mistook height map and normal map as the same thing by mistake here. Thankfully people seem to get what I meant
This is _somewhat_ doable. I make a basic version of it in my tutorial series on convolution in Unreal Engine. The tutorials are quite dry but the example project is free to download. It's not very straight forward to use as a tool though.
this is how normal maps are created, the RGB values on a normal map correspond to "lighting" the object from 3 different direction (on top, sideways, and from the bottom)
I could probably watch about 12 hours of Steve pressing various objects into the gel and just describing what they are before moving onto the next one.
The wierd neon-colored image you can see used to create 3D model is called a normal map and it's meant to hold information about height of the object using different colours for different light angles. It is used widely in video games to create detail in low poly objects by changing how light reflects off of the surface. I find it fascinating they used normal map and not a regular height map but given the technology of 6 light sources it makes all the sense
Height maps contain less information than normal maps. Each pixel's color channel in a normal map contains information on the normal (orientation) of the surface, whereas a height map only contains direct vertical information.
@@BiffGheek Paradoxically heightmaps can be said to have "more" information - you cannot recover height from normal, but you can generate normals from height - a 90 degree drop would not appear in a normal map.
@@BiffGheek It's possible to obtain the normal vector from a heightmap. It was a common trick in computer graphics to use the depth buffer for things like ambient occlusion, an algorithm which needed normal vectors but didn't need to be 100% accurate.
This has WILD implications as a mobile highly accurate surface measurer. Damn, I want one just to have. I bet they're expensive. Industry really needs to have this. Feels revolutionary.
As a 3D artist myself this was one of the most fascinating things I have ever seen in a long while. If produced massively across the globe. It could be really helpful in 3D Film Industry.
Considering the human brain's ability to extract a sense of 3D from watching 2D movies, it might explain the abysmal failure of of 3D movies and TV sets every time there's been a push in that direction. If this can help 3D movies improve on a human's innate ability (and be anything more than a gimmick) I'm all for it.
You can make normal maps of objects by taking 4 pictures with light from up, down, left and right and then composing them manually. Like this thing but of course without the whole "immune to transparency/reflection" thing. The composing would be inverting right and blending with left 50%, put in red channel, invert down and blend with up 50% and put in green. For blue, Nvidia's old normal map plugin for Photoshop could get close with its "normalize" feature from the red and green channel.
I know. The moment I saw that coin being lit from different sides, I immediately knew it was compiling a normalmap. I'd love this ability too, especially because I like making normalmaps of coins.
there's already plenty of software now that can use multiple images to "scan" and object and make a 3d model of it. i'd be surprised if there wasn't similar software that makes normal maps as well
That zigzag motion of the print head is "boustrophedonic". It is a literal reference to the motion of an ox ploughing a field, and you still usually see it in patents to describe print heads, scanners, etc.
We could actually use that almost every day at work to measure engravings on the toolings for chocolate moulds. Would safe us a lot of time. Super interesting 🧐👌🏼
For those wondering, the tool he is using is called GelSight. We use them often in aerospace to measure the depth of damage on metal surfaces. Very cool piece of gear, and not as expensive as you would imagine.
Actually it is very expensive the smallest one called the gelsight mini is about the size of a quarter and costs usd$499 and you have to request a quote to get the gelsight mobile and gelsight max so they are at least usd$1500+
@@bombexplosion69420 in the big scheme of things, for how capable this device is, that’s very cheap. This is an invaluable tool for many industries, including aerospace.
@@bombexplosion69420 Tbf, I would've assumed the cost to be much higher than that. This looks the type of stuff you find in some high tech lab testing prototypes
@@bombexplosion69420i mean, OP just said ‘not as expensive as you’d imagine’ and i honestly thought this would be like £5000 at least, so to hear that it’s £500 instead makes a huge difference
@05:38 I see a mountain in both cases. The shading in the detail around the perimeter comes from bulging shapes, not depressions. Also, if cast shadows are present in such cases, they clarify the situation.
I agree. I think I've seen something like this before that actually flips between being a depression and a bulge, but in this specific photo, it's both a bulge for me. I can't make my brain think that it's a crater.
I work with lots of 3D LIDAR scans and one huge issue for us is the reflection off of reflective surfaces such as mirrors. Interseting how this gets rid of that, at least on a micro scale.
Would training an algorithm be easier using this technique as a source of truth? Small images, and maybe large images later, could be developed accurately with this, but would that information from this technique help you train a model faster? And further, would information about small things like the quarter or matchhead be useful for larger scale applications of LIDAR such as autonomous vehicles?
With the crater illusion, I saw it as a mountain the whole time, even after seeing the shadows falling on it like a crater - the look of the inside of the crater looks so plateu like. Same with the indented text, in fact I think the top light source makes it stronger, and I did eventually recognise the crater, but I had to look closely at the shadows, and looking at the light side makes it look like a mountain no matter what.
I had to work hard to see it as a crater, including switching to only one eye or looking away for a while. Even then, it would pop back up to mountain pretty quickly. I think it's because I saw it as mountain before rotation.
Yeah- I was able to resolve the crater in the second position and not the first, but even after resolving a crater, it was still more natural to see a mountain in both. And that's ignoring that my brain thought it was something small in the gel at the first glance.
4:37 That's called monocular depth estimation and it very much already exists. There are quite a few open-source models that can estimate depth fairly well from a single flat image, and if you have multiple images from different angles there are neural networks that can build up an accurate model of the entire scene, sufficient to move a virtual "camera" around and through the scene and create new views in the process.
Nearly every mundane object under that type of imaging looks quite mesmerizing! I actually thought your stubble was one of the more interesting ones. It really shows just how cleanly the blades cut the hairs at the ends. The draping problem was the first thing that came to mind as someone who has done a lot of vacuum forming. But that’s very interesting that it’s not much of an issue as long as they can get the depth value. And as long as it’s not deeper than it is wide.
As a mechanic this would be really useful to check for wear on parts quickly because most parts that we check for normal wear that occurs during the engine's lifetime is layered and sometimes you may be searching for a deeper scratch in the middle of thousands and it would eliminate the need to send parts for measurements in some cases.
Was really wondering about how easily it was contaminated and how frequently you would need to replace the gel, how expensive is the replacement and how easy is the process? It looked like everything you had on there had some dust or hairs, it seemed like it wasn't like you were constantly scanning dirty pieces but instead that the microscope itself had issue. Incredibly cool technology, I even want one but the dirt issue seems hard to deal with.
what a *fascinating* tool. I personally don't have a lot of use for it, but I can just imagine machinists being absolutely delighted at the details and measurements you can get from it.
I was just thinking about that! I believe it would be amazing for understanding markings, but, if the resolution is enough, even some features of the fabrics of materials
@ well I can imagine a lot of situation in which force it's not a problem. There are some artifacts, like terra sigillata (I don't remember now the english name, sorry) which may not suffer. Or stoneware
@ I was mainly thinking about markings on rustued metal, but on the other hand if you take (for example) a coin, you just scan both sides and you don't have to handle the coin nearly as much, which would make it far less likely to crumble.
This would be an incredible medium to make some kind of animated movie, even just a short little thing. What i'm imagining is in the same vein as the "A Boy And His Atom" animation made with individual atoms.
This is insane. We are getting to watch the first steps of an entire new tool that will change entire fields of study. It is amazing. I'm probably way more hyped about this then I should
I just got sent to Sempre on a training course and got trained to use the GelSight. Lovely bunch of guys! Very nice of you to boost them with this video.
0:15 Miyota 2035 quartz watch movement, as used in countless cheap and cheerful plastic watches. To the left is the stepper coil, to the upper right is the quartz crystal.
I work with 3D making games for a living. Seeing those 3D models being created was so cool. Essentially using the same information that's used to generate and render normal and displacement maps.
Ok, Steve, I'm so glad that youtube thought to put your video in my path. That's an odd microscope indeed, and the gel is strangely sensitive. Removing color does indeed leave us with form, texture, and shadow. Photometric Stereo? Ok then. I wonder what insects would look like with this system? Does that gel wear out? So it has a quality control inspection use, also interesting. Jane Street Academy... sounds excellent. Thank you for all of this information. Well done. But then you already know that.
I saw this (specifically the tiny GelSight mini) recently and thought "what would anyone use this for?". Now that I actually see the accuracy and resolution it has it makes sense
I think my reaction to your example at 6:00 brings up an interesting point. See, I DIDNT experience the crater illusion when you used the touch microscope, and I think it's because humans are VERY good at context clues and learning; Because you'd shown me other visuals of the touch microscope, I was able to tell from the patterning around the letters that a flat, intended surface was pressed against the microscope, and that the letters were smooth because nothing was touching the gel surface. If they really had been protruding from the block, it would be the letters that had texturing, not the background.
Hey man I’m an aircraft tech and just wanted to give some input on other ways we can measure damage nondestructively. We have ultrasonic testers that work much like ultrasound. We also have eddy current which is extremely interesting. We have blanks of different materials; aluminum, titanium, magnesium, steel, and others. We use a calibrated head on the tool that uses electromagnetic eddy currents that are then disrupted by the damage. Using those methods we can actually measure and find cracks and scratches which may not even be visible!
5:45 I can’t really see the crater. I understand why the human mind SHOULD think that it’s a crater, but I can’t see the crater. Sometimes I get the slightest illusion it might be dipping down, but it doesn’t hold.
4:23 "Your brain doesn't think 'Oh, This region is dark because it's facing away from the light source.' You know that it's:" PAINT "The eyeball!" FUCK
This multi-light setup that generates the depth and normal maps is the same method that is used to generate many textures and material used in games and film. You can do it yourself with a camera, a light and either free software or Adobe Substance. You can also remove the reflection from the images using cross polarisation by putting a polarising gel on your light source and in a different orientation on your camera.
This is honestly amazing! You're so lucky to be able to play with. Is there any information on whether this is going to be sold commercially? I'd love to get my hands on this.
In the intro he shows a picture of the company 'The Sempre Group' and their website. The device is there, no price listed but "contact us" so probably very expensive.
@TigerGold 59 It depend... What's the resolution of the camera ? What's the focal point of the camera ? What's the material of the pad ? How many time did the software got developped ? Without thoses anwsers, you cannot estimate the price. It's like saying that VR headset do not look so expensive... Yeah but it's a small 2k screen, a very narrow focal point, and a very difficulte to make tracking device. I'm pretty sure that given the software and the regular change of the pad, it's not sold but got offer at a service.
@TigerGold 59 It's a device made for industrial and commercial use, not consumer use. It will be expensive. I'd estimate around at least 15K USD. Especially with a new technology. My father works in the industrial sector, you'd be surprised by how much the industrial version of something costs vs the consumer version. They're usually 10x more expensive. For example, he once showed me a thermal camera his company was using, it looked like a gun and had a small screen at the back. The price? 5K USD. For comparison you can get a cheap consumer thermal camera for 400 USD. And the 5K USD thermal cam was only the entry-level version, the next better model was 7K USD. The reason these things are so expensive is because industrial tools are made with precision, consistency, and durability in mind. I mean, you wouldn't care much if your toaster shows up as 10°C colder than it is in your consumer thermal camera (I've checked, my toaster goes up to 700+°C), but if an engineer thinks that an electrical cable is 20°C colder than it is....the company can lose millions of $ from the resulting damage. That's why 5K$ to a serious company seems so small in the long run. My advice is to wait for a consumer version of this touch microscope to come up. That might take 5-15 years but unless you got a thick wallet, it's the only option.
I tried diverging my eyes on the octopus images to see them in 3d (Magic Eye style) and it worked! Except the images are in the wrong order so the depth was reversed making the octopus more distant than the background which looks trippy. This means that if you cross your eyes the 3d image would look correct, but I have more trouble crossing my eyes vs diverging, so I wasn’t able to do it. Maybe try it out for yourself at 5:00
The reason why it was reversed for you was because you crossed your eyes the wrong way, outward, rather than inward, which is very difficult to do, impressive! Go the other way and it'll be normal though
@@coreblaster6809 that's exactly what he said. You can either cross your eyes (inward) or diverge them (outward). But it's more difficult for him to cross them.
The Crater illusion doesn't work for me. Neither in the crater image nor with the letters. Both times it just looks like it's sticking out. Never do I perceive an indentation
After it is rotated, the light is actually coming from the top-left. You can clearly see shadows being cast into the crater from the lip of the crater in the top-left.
@@JustOneAsbesto my brain almost always interprets it as an extrusion, even when skipping straight to the crater image, except only once I managed to see it as a crater. I have much better luck with the preview thumbnail though.
Never knew such thing existed. But this is really a valuable tool for a lot of people who do precision work. So thank you for brining it to my attention it deserves that.
02:00 - no lie, my first guess before the reveal was that it was some sort of sonic/acoustic imaging... it'd produce about the same results with no color data
If I'm not wrong, there already exists an AI tool that can scan transparent/reflective objects quite well. They're called Neural Radiance Fields (NeRFs) and the Corridor Crew did an amazing video on it.
NeRFs arent *accurate,* they're what I would call "true." They can guess at the shape and make it *look* right, but due to the nature of the input they're not going to be useful for any precise measurement.
Weird, at 5:23 I got a sudden urge to throw my computer off a high counter or something. This is such a cool device! I was thinking from the thumbnail and title that we’d be looking at an AFM or something. What a fun surprise!
Dear Steve, I work with a microscope which is called a scanning ion conductance microscope (SICM). We can measure objects on the sub um scale. Importantly, its non contact and doesn't need sample labelling. It works particularly well with cardiomyocytes. Let me know if you want to have a go! Ben
We can se because we remember that we had seen the object some time ago. But with optical illusion we can say that this system do not work perfect. This microscope does not have this problem with optical illusion
Complete side note: The free software “Materialize” does an okay job at taking a single color image and extracting depth, reflectivity, and other properties. It was used to remaster the Uncharted game series.
Weirdly, the lettering at 6:13 looks like it's sticking out to me, and at first the lettering before it looked like it was indented. Afterwards the first case is ambiguous (I can switch it in my head), but I still can't see the second case as indented.
There seems to be some limitations when used for surface features with high aspect ratio. For example, it wasn't able to discern the vertical fall off from the edges of the Lego brick pegs But none the less very interesting and could be very useful in a lot of scenarios By the way, what's the MTBF of the gel? (i.e. how many times can it be used)
When creating CG(3D) Objects like for video games. You use something called a normal map. A normal map creates small surface details like little cracks in a wall. It takes 3 different light sources and puts them into the Red, Green, and Blue channels of an image, in the end comes out in this purplish looking image. It is essentially what that microscope does, but that uses 6 light sources instead of 3. Very Cool gadget.
For regular consumers, but people in manufacturing, engineering, even digital 3D artists could make use of this. He explained an example application with how instead of taking molds of scratches to measure they can use this and get the data instantly.
@@ZephyrinSkies No, even for engineering, between making a mold once every blue moon or using a reusable-but-not-really gel pad of fixed size, I feel it is not a significant advantage (esp since cracks have to be more wide than deep which is rare) The only point I'd give would be the way it isn't affected by shine and transparency, but I'm sure some clever signal processing can manage that as well
@@76Arfa If it's just one measurement you can use a micrometer (the tool), if you want full 3D there are microtomography scanners, and in-between there are probably several tools I don't know to make cheaper specific measurements Maybe this gel-scope fits in-between, but I'd be very surprised if we don't already have several tools doing similar things without the drawbacks
I'd call it a PMD (precision measuring device). Have you checked out the technology that uses smoke and a laser as the sense element for an audio microphone? It is unreal.
Would *you* call it a microscope?
The sponsor is Jane Street. Find out about their Academy of Math and Programing here: janestreet.com/amp
have used janestreet before
Well, I wouldn't classify it as an imaging microscope. However, given that it's capable of allowing users to view and take very minute measurements, I think I'll categorise it as a profiling microscope.
I guess that still makes it a microscope 😁🤷🏾♂️🤷🏾♂️
Micro thigmoscope
Hey, at 3:05, I'd recommend you edit that part out, putting out your fingerprint like that might not be safe.
Edit: Yeah, we don't know who it's from or which finger it might be, so it can still be safe. Just wanted to let you know though.
Well microscopes touch objects with light or electrons so why not?
I'd love to know more about the gel they are using, It seems absolutely incredible that a gel would conform to such small details so perfectly and yet have no memory in that it returns to it's original shape.
I'll bet you there's some nice patents on the gel.
Either the video codec has an echo or the gel does have some memory.
You can see on the lego that it has some issues with cornering. Around the base of the lego nubs it has some issues. Stuff like that may get worked out though in the future.
@@PS-bc5qt I think that tracks with what the guy was saying about the limitations - it won't perfectly show sharp angles but for the application it's not a huge issue as you're not looking at that directly
My guess is nanoparticles, and the size of the nanoparticles determines the resolution.
This would be awesome to generate height maps for making smart materials in substance.
Just noticed the height map at 2:41 and wanted to plug it into substance.
Mistook height map and normal map as the same thing by mistake here. Thankfully people seem to get what I meant
This is _somewhat_ doable. I make a basic version of it in my tutorial series on convolution in Unreal Engine. The tutorials are quite dry but the example project is free to download.
It's not very straight forward to use as a tool though.
this is how normal maps are created, the RGB values on a normal map correspond to "lighting" the object from 3 different direction (on top, sideways, and from the bottom)
I wonder if they could make one that's 4x4m in size 🤔
@@Noclip420 The point of a microscope dissappears slowly xD
i got the same idea...
so i think that could really do a thing or two
I could probably watch about 12 hours of Steve pressing various objects into the gel and just describing what they are before moving onto the next one.
Right? I wonder what various kinds of feathers would look like - flight feathers vs down, etc
Then do it.
The wierd neon-colored image you can see used to create 3D model is called a normal map and it's meant to hold information about height of the object using different colours for different light angles. It is used widely in video games to create detail in low poly objects by changing how light reflects off of the surface.
I find it fascinating they used normal map and not a regular height map but given the technology of 6 light sources it makes all the sense
Height maps contain less information than normal maps. Each pixel's color channel in a normal map contains information on the normal (orientation) of the surface, whereas a height map only contains direct vertical information.
@@BiffGheek Paradoxically heightmaps can be said to have "more" information - you cannot recover height from normal, but you can generate normals from height - a 90 degree drop would not appear in a normal map.
@@BiffGheek It's possible to obtain the normal vector from a heightmap. It was a common trick in computer graphics to use the depth buffer for things like ambient occlusion, an algorithm which needed normal vectors but didn't need to be 100% accurate.
@@0x1EGENall I know about Ambient Occlusion is how it lags my computer more thany other graphics setting
Normal map doesn't hold information about the height, it holds information about the angle :)
This has WILD implications as a mobile highly accurate surface measurer. Damn, I want one just to have. I bet they're expensive. Industry really needs to have this. Feels revolutionary.
As a 3D artist myself this was one of the most fascinating things I have ever seen in a long while. If produced massively across the globe. It could be really helpful in 3D Film Industry.
Could you elaborate on why you think this microscope could be helpful to 3D artists?
@@Spudcosmiccc maybe creating normal or roughness materials? its a bit too small though I think
Considering the human brain's ability to extract a sense of 3D from watching 2D movies, it might explain the abysmal failure of of 3D movies and TV sets every time there's been a push in that direction. If this can help 3D movies improve on a human's innate ability (and be anything more than a gimmick) I'm all for it.
this is litterally comercially available since 2015 ua-cam.com/video/VN9ICs02vcY/v-deo.html
with full on 3d measurements since 2019
@@Alan_Berger It's $500, which is actually pretty reasonable (by that I mean somebody who has a use for something like this can actually get one)
im a 3d artist and i want one of these to make instant normalmaps from objects ,these look great
Yes! Inagine having super accurate normal maps in games. It would look amazing.
You can make normal maps of objects by taking 4 pictures with light from up, down, left and right and then composing them manually. Like this thing but of course without the whole "immune to transparency/reflection" thing. The composing would be inverting right and blending with left 50%, put in red channel, invert down and blend with up 50% and put in green. For blue, Nvidia's old normal map plugin for Photoshop could get close with its "normalize" feature from the red and green channel.
I know. The moment I saw that coin being lit from different sides, I immediately knew it was compiling a normalmap. I'd love this ability too, especially because I like making normalmaps of coins.
it instantly reminded me of normal maps as well xD
there's already plenty of software now that can use multiple images to "scan" and object and make a 3d model of it. i'd be surprised if there wasn't similar software that makes normal maps as well
That zigzag motion of the print head is "boustrophedonic". It is a literal reference to the motion of an ox ploughing a field, and you still usually see it in patents to describe print heads, scanners, etc.
I'm here to learn words like boustrophedonic
This concludes my night, I've learnt enough. Thank you for that nugget of information.
I'm here to learn where you learned this
Space filling curves
I learned what that word meant in the context of writing systems.
We could actually use that almost every day at work to measure engravings on the toolings for chocolate moulds. Would safe us a lot of time.
Super interesting 🧐👌🏼
what
For those wondering, the tool he is using is called GelSight. We use them often in aerospace to measure the depth of damage on metal surfaces. Very cool piece of gear, and not as expensive as you would imagine.
Cool Ai comment, for the people that couldn't watch the video for literally one minute.
Actually it is very expensive the smallest one called the gelsight mini is about the size of a quarter and costs usd$499 and you have to request a quote to get the gelsight mobile and gelsight max so they are at least usd$1500+
@@bombexplosion69420 in the big scheme of things, for how capable this device is, that’s very cheap. This is an invaluable tool for many industries, including aerospace.
@@bombexplosion69420 Tbf, I would've assumed the cost to be much higher than that.
This looks the type of stuff you find in some high tech lab testing prototypes
@@bombexplosion69420i mean, OP just said ‘not as expensive as you’d imagine’ and i honestly thought this would be like £5000 at least, so to hear that it’s £500 instead makes a huge difference
@05:38 I see a mountain in both cases. The shading in the detail around the perimeter comes from bulging shapes, not depressions. Also, if cast shadows are present in such cases, they clarify the situation.
Uh, no, I clearly see a crater
I agree. I think I've seen something like this before that actually flips between being a depression and a bulge, but in this specific photo, it's both a bulge for me. I can't make my brain think that it's a crater.
@@hamuelagulto796 same
Maybe it's because you already saw it as a mountain first?
Same here.
I work with lots of 3D LIDAR scans and one huge issue for us is the reflection off of reflective surfaces such as mirrors. Interseting how this gets rid of that, at least on a micro scale.
Just tape a gel cube to the sensor and drive into things!
Would training an algorithm be easier using this technique as a source of truth?
Small images, and maybe large images later, could be developed accurately with this, but would that information from this technique help you train a model faster? And further, would information about small things like the quarter or matchhead be useful for larger scale applications of LIDAR such as autonomous vehicles?
Mirrors look really interesting in a 3D picture. They're kind of like a hole.
With the crater illusion, I saw it as a mountain the whole time, even after seeing the shadows falling on it like a crater - the look of the inside of the crater looks so plateu like. Same with the indented text, in fact I think the top light source makes it stronger, and I did eventually recognise the crater, but I had to look closely at the shadows, and looking at the light side makes it look like a mountain no matter what.
I otoh saw the concave text immediately (after rotation), yet can't see the crater at all. It's always a mountain | plateau for me.
Same for me.
same
I had to work hard to see it as a crater, including switching to only one eye or looking away for a while. Even then, it would pop back up to mountain pretty quickly. I think it's because I saw it as mountain before rotation.
Yeah- I was able to resolve the crater in the second position and not the first, but even after resolving a crater, it was still more natural to see a mountain in both. And that's ignoring that my brain thought it was something small in the gel at the first glance.
5:00 Thanks for making the left/right arrangement compatible with the cross-eye technique! 😸👍
The gel pad is a hexagon, and that fact just makes my brain light up in excitement and happiness all on its own. Hexagons are the best!
The bestagons indeed!
4:37 That's called monocular depth estimation and it very much already exists. There are quite a few open-source models that can estimate depth fairly well from a single flat image, and if you have multiple images from different angles there are neural networks that can build up an accurate model of the entire scene, sufficient to move a virtual "camera" around and through the scene and create new views in the process.
But how reliable and accurate are they? Probably not good enough for the purposes these camera gel deforming sensing things are bought for?
Nearly every mundane object under that type of imaging looks quite mesmerizing! I actually thought your stubble was one of the more interesting ones. It really shows just how cleanly the blades cut the hairs at the ends. The draping problem was the first thing that came to mind as someone who has done a lot of vacuum forming. But that’s very interesting that it’s not much of an issue as long as they can get the depth value. And as long as it’s not deeper than it is wide.
Its incredible how well that gel conforms to objects
I think it is the 80% of that technology.
I bet the gell have "working life time" ..... The gell Will be the thing Will replace regulary ..... 🤔🤔🤔
@@memesfromdeepspace1075 He actually did say that in the video, on the part with the stickiness. He said that it becomes sticky and needs replacing :)
So cool. Its like making a normal map of a object 😍
As a mechanic this would be really useful to check for wear on parts quickly because most parts that we check for normal wear that occurs during the engine's lifetime is layered and sometimes you may be searching for a deeper scratch in the middle of thousands and it would eliminate the need to send parts for measurements in some cases.
Steve: this needs a sound effect
Me: schlorp
Steve: schlorp
I feel it shoud soun like a ruber balloon crackling while it deflates beside an object
Being animals, some noises are built into the experience.
Every one of your videos is a treasure. Thank you
I was expecting a video on an Atomic Force Microscope, but this is far more fascinating! What a novel way to collect 3d data!
You beat me, I am also expect an AFM.
yes exactly!
I also thought it was AFM.
I was thinking STM, but AFM makes more sense.
Yeah same here. Well played on his part.
Was really wondering about how easily it was contaminated and how frequently you would need to replace the gel, how expensive is the replacement and how easy is the process? It looked like everything you had on there had some dust or hairs, it seemed like it wasn't like you were constantly scanning dirty pieces but instead that the microscope itself had issue. Incredibly cool technology, I even want one but the dirt issue seems hard to deal with.
Also if the subject's too fragile, the scope itself might impact and disrupt it. This'd be useless for microscopic stuff
what a *fascinating* tool. I personally don't have a lot of use for it, but I can just imagine machinists being absolutely delighted at the details and measurements you can get from it.
This would be great for artefact studies in archaeology
I was just thinking about that! I believe it would be amazing for understanding markings, but, if the resolution is enough, even some features of the fabrics of materials
So much easier to look for tooling marks and scratches to see how things were made
I'm doubtful about the willingness of the archaeologists to press the artifacts into a gel. Things could easily break from the force.
@ well I can imagine a lot of situation in which force it's not a problem. There are some artifacts, like terra sigillata (I don't remember now the english name, sorry) which may not suffer. Or stoneware
@ I was mainly thinking about markings on rustued metal, but on the other hand if you take (for example) a coin, you just scan both sides and you don't have to handle the coin nearly as much, which would make it far less likely to crumble.
Mould is on another level with his science videos. He always finds the most interesting and obscure stuff.
This would be an incredible medium to make some kind of animated movie, even just a short little thing. What i'm imagining is in the same vein as the "A Boy And His Atom" animation made with individual atoms.
This is nowhere, anywhere, even remotely close to having the resolution to see atoms.
@@JustOneAsbesto I was not suggesting that it did, thanks though.
‘A Boy and His Micron’ maybe?
@@nickcoleman2765 yo thats perfect!
@@JustOneAsbesto I mean isn't technically everything that it sees made of atoms?
This is insane. We are getting to watch the first steps of an entire new tool that will change entire fields of study. It is amazing. I'm probably way more hyped about this then I should
Not at all! We're curious beings and seeing new innovations inspires our mindset!
I just got sent to Sempre on a training course and got trained to use the GelSight. Lovely bunch of guys! Very nice of you to boost them with this video.
0:15 Miyota 2035 quartz watch movement, as used in countless cheap and cheerful plastic watches. To the left is the stepper coil, to the upper right is the quartz crystal.
uploaded 57 seconds ago, gotta binge this up
Really cool that it showed the depth of your mouse cursor in so many shots!
I work with 3D making games for a living. Seeing those 3D models being created was so cool. Essentially using the same information that's used to generate and render normal and displacement maps.
Ok, Steve, I'm so glad that youtube thought to put your video in my path. That's an odd microscope indeed, and the gel is strangely sensitive. Removing color does indeed leave us with form, texture, and shadow. Photometric Stereo? Ok then. I wonder what insects would look like with this system? Does that gel wear out? So it has a quality control inspection use, also interesting. Jane Street Academy... sounds excellent. Thank you for all of this information. Well done. But then you already know that.
I saw this (specifically the tiny GelSight mini) recently and thought "what would anyone use this for?". Now that I actually see the accuracy and resolution it has it makes sense
8:54
Love Steve for moments like these
6:03 "Yeah, it does look like it's sticking out!"
6:12 "Yeah, it does look like it's sticking out!"
lol, same!😁
It's kind of the same for me. Except I see it as indented for a few seconds and then it starts to look like it's sticking out again.
@@That_Awesome_Guy1 "Just when I thought I was in, they pull me back out"
I would love a larger scale of this, it would be amazing for modelling small electronics where the manufacturer doesn't provide 3d files
I would sit on it
I think my reaction to your example at 6:00 brings up an interesting point. See, I DIDNT experience the crater illusion when you used the touch microscope, and I think it's because humans are VERY good at context clues and learning; Because you'd shown me other visuals of the touch microscope, I was able to tell from the patterning around the letters that a flat, intended surface was pressed against the microscope, and that the letters were smooth because nothing was touching the gel surface. If they really had been protruding from the block, it would be the letters that had texturing, not the background.
I work at a jet engine company and we use these all the time! So cool to see!
to what end? wear inspection?
8:55 I was not ready for this...
0:02 ah yes! The polo mint!
Hey man I’m an aircraft tech and just wanted to give some input on other ways we can measure damage nondestructively.
We have ultrasonic testers that work much like ultrasound.
We also have eddy current which is extremely interesting. We have blanks of different materials; aluminum, titanium, magnesium, steel, and others. We use a calibrated head on the tool that uses electromagnetic eddy currents that are then disrupted by the damage.
Using those methods we can actually measure and find cracks and scratches which may not even be visible!
It's a watch!
I've had to replace those little buggers one times too many.
5:45 I can’t really see the crater. I understand why the human mind SHOULD think that it’s a crater, but I can’t see the crater. Sometimes I get the slightest illusion it might be dipping down, but it doesn’t hold.
Same
I think it's because the perimeter of the crater is all lit
I neither see the crater nor understand why would anyone assume it’s a crater!!!
4:23
"Your brain doesn't think 'Oh, This region is dark because it's facing away from the light source.' You know that it's:"
PAINT
"The eyeball!"
FUCK
wow, what an awesome tool :)
thanks Semper Group for letting Steve play with it, maybe send him more stuff to test eh 👍
This multi-light setup that generates the depth and normal maps is the same method that is used to generate many textures and material used in games and film. You can do it yourself with a camera, a light and either free software or Adobe Substance. You can also remove the reflection from the images using cross polarisation by putting a polarising gel on your light source and in a different orientation on your camera.
now we have your fingerprint and can unlock your phone
😂
I had a stroke trying to read this
8:20 you should do all the numbers on the front and back next time too that'd be fun
Funni
What if you press two of these against each other? Will the universe achieve singularity?
This is honestly amazing! You're so lucky to be able to play with. Is there any information on whether this is going to be sold commercially? I'd love to get my hands on this.
In the intro he shows a picture of the company 'The Sempre Group' and their website. The device is there, no price listed but "contact us" so probably very expensive.
@TigerGold 59 It depend...
What's the resolution of the camera ?
What's the focal point of the camera ?
What's the material of the pad ?
How many time did the software got developped ?
Without thoses anwsers, you cannot estimate the price.
It's like saying that VR headset do not look so expensive... Yeah but it's a small 2k screen, a very narrow focal point, and a very difficulte to make tracking device.
I'm pretty sure that given the software and the regular change of the pad, it's not sold but got offer at a service.
@TigerGold 59 Probably isn't expensive to make but they will want to recoup all their research and software development.
TigerGold59, I definitely think you should make one! Maybe sell it at a lower price then this one.
@TigerGold 59 It's a device made for industrial and commercial use, not consumer use. It will be expensive. I'd estimate around at least 15K USD. Especially with a new technology.
My father works in the industrial sector, you'd be surprised by how much the industrial version of something costs vs the consumer version. They're usually 10x more expensive. For example, he once showed me a thermal camera his company was using, it looked like a gun and had a small screen at the back. The price? 5K USD. For comparison you can get a cheap consumer thermal camera for 400 USD. And the 5K USD thermal cam was only the entry-level version, the next better model was 7K USD.
The reason these things are so expensive is because industrial tools are made with precision, consistency, and durability in mind. I mean, you wouldn't care much if your toaster shows up as 10°C colder than it is in your consumer thermal camera (I've checked, my toaster goes up to 700+°C), but if an engineer thinks that an electrical cable is 20°C colder than it is....the company can lose millions of $ from the resulting damage. That's why 5K$ to a serious company seems so small in the long run.
My advice is to wait for a consumer version of this touch microscope to come up. That might take 5-15 years but unless you got a thick wallet, it's the only option.
Now we have your fingerprints
Lol
I imagine this could be incredibly useful for making normal maps for added detail in 3d modeling
"No one would see this face as concave" 6:27
Steve how dare you put this curse on me.
You just made a bowl out of a human
I tried diverging my eyes on the octopus images to see them in 3d (Magic Eye style) and it worked! Except the images are in the wrong order so the depth was reversed making the octopus more distant than the background which looks trippy. This means that if you cross your eyes the 3d image would look correct, but I have more trouble crossing my eyes vs diverging, so I wasn’t able to do it. Maybe try it out for yourself at 5:00
Yes - it's a cross view stereo 3d image (video). Works very well. Not an octopus though.
I managed it with crossed eyes. Very cool 🤩
The reason why it was reversed for you was because you crossed your eyes the wrong way, outward, rather than inward, which is very difficult to do, impressive! Go the other way and it'll be normal though
@@coreblaster6809 that's exactly what he said. You can either cross your eyes (inward) or diverge them (outward). But it's more difficult for him to cross them.
For me crossing its always easier, so it's very nice.
This is the type of rediscovery we need on current technology!
The Crater illusion doesn't work for me. Neither in the crater image nor with the letters. Both times it just looks like it's sticking out. Never do I perceive an indentation
After it is rotated, the light is actually coming from the top-left. You can clearly see shadows being cast into the crater from the lip of the crater in the top-left.
@@JustOneAsbesto my brain almost always interprets it as an extrusion, even when skipping straight to the crater image, except only once I managed to see it as a crater. I have much better luck with the preview thumbnail though.
The fabric you showed was not woven, but in fact knit. Woven fabrics have parallel lines.
**yes** thank you.
5:50 I still see a mountain :()
Same,
Me too
Had to rewatch this, I could watch Steve put items on this and guess what they are for hours, I love this little machine
Never knew such thing existed.
But this is really a valuable tool for a lot of people who do precision work.
So thank you for brining it to my attention it deserves that.
5:38 I can’t see a crater…
@0:20 I am going to guess a wristwatch!
0:27 I didn’t watch the video but I know that this is a watch mechanism
5:03 I cross eyed for that section. Super cool!
02:00 - no lie, my first guess before the reveal was that it was some sort of sonic/acoustic imaging... it'd produce about the same results with no color data
Okay at 6:13 im seeing it as opposite of what you said and its confusing me as to which is real now
If I'm not wrong, there already exists an AI tool that can scan transparent/reflective objects quite well. They're called Neural Radiance Fields (NeRFs) and the Corridor Crew did an amazing video on it.
NeRFs arent *accurate,* they're what I would call "true." They can guess at the shape and make it *look* right, but due to the nature of the input they're not going to be useful for any precise measurement.
That was a very interesting video. For those who haven't see it: ua-cam.com/video/YX5AoaWrowY/v-deo.html
Not at this scale and level of accuracy (yet)
@@testhandle1537 Yea definitely not. But its cool to see that we're making advancements in that direction.
Weird, at 5:23 I got a sudden urge to throw my computer off a high counter or something. This is such a cool device! I was thinking from the thumbnail and title that we’d be looking at an AFM or something. What a fun surprise!
Yeah, *somebody* has a bead-chain fixation!
Dear Steve, I work with a microscope which is called a scanning ion conductance microscope (SICM). We can measure objects on the sub um scale. Importantly, its non contact and doesn't need sample labelling. It works particularly well with cardiomyocytes. Let me know if you want to have a go! Ben
The same thing happens when I press my face into a big bowl of Jello.
Does the gel leave a residue behind to the point to where you have to refill the gel over time?
5:45 I can not see the crater RIP Q_Q
8:36 YO HOLD UP
Why is literally no one else talking about this !?😂
Tes5icle
💀💀💀💀💀
We can se because we remember that we had seen the object some time ago. But with optical illusion we can say that this system do not work perfect. This microscope does not have this problem with optical illusion
Oh you said it,
Complete side note: The free software “Materialize” does an okay job at taking a single color image and extracting depth, reflectivity, and other properties. It was used to remaster the Uncharted game series.
Weirdly, the lettering at 6:13 looks like it's sticking out to me, and at first the lettering before it looked like it was indented. Afterwards the first case is ambiguous (I can switch it in my head), but I still can't see the second case as indented.
0:17 is it inside a watch??
3:00 exact 3d copies of objects are really useful for artifact collections. the leading way this is done is with hundreds of small cameras
There seems to be some limitations when used for surface features with high aspect ratio. For example, it wasn't able to discern the vertical fall off from the edges of the Lego brick pegs
But none the less very interesting and could be very useful in a lot of scenarios
By the way, what's the MTBF of the gel? (i.e. how many times can it be used)
Oh... you mentioned it towards the end of the video... hahaha
I'd probably call it a profilometer more than a microscope. But then again it can be one
does the gel have any memory and would that effect tolerances?
Thats amazing. It's like a digital terrain model, but for tiny things
My past 5+ years of EM experience using various electron detectors approves this video on a fundamental level.
When creating CG(3D) Objects like for video games. You use something called a normal map. A normal map creates small surface details like little cracks in a wall. It takes 3 different light sources and puts them into the Red, Green, and Blue channels of an image, in the end comes out in this purplish looking image. It is essentially what that microscope does, but that uses 6 light sources instead of 3. Very Cool gadget.
I wonder if 1:40 has enough data to unlock Steve's phone...
Probably not. It’s only a partial fongerprint
3:07
The gel specifically is really impressive, but I really feel like it's a solution to a problem no one has
For regular consumers, but people in manufacturing, engineering, even digital 3D artists could make use of this. He explained an example application with how instead of taking molds of scratches to measure they can use this and get the data instantly.
@@ZephyrinSkies No, even for engineering, between making a mold once every blue moon or using a reusable-but-not-really gel pad of fixed size, I feel it is not a significant advantage (esp since cracks have to be more wide than deep which is rare)
The only point I'd give would be the way it isn't affected by shine and transparency, but I'm sure some clever signal processing can manage that as well
better to have it and not need it than need it and not have it...
@@pastek957 and what about checking for tolerances on CNC biomed parts which can have tolerances in the microns sometimes?
@@76Arfa If it's just one measurement you can use a micrometer (the tool), if you want full 3D there are microtomography scanners, and in-between there are probably several tools I don't know to make cheaper specific measurements
Maybe this gel-scope fits in-between, but I'd be very surprised if we don't already have several tools doing similar things without the drawbacks
Whats the gel pad made of?
Omg this is a genius way to finish a step of creating ultra realistic 3D models and CGI renders, that’s very epic
Coolest technology I've seen in years. Makes you think about what else might be out there and kept from the public.
No money in keeping it from the public.
0:47 well that's in interesting fleshlight...
5:05 i crosseyed to see this in 3D, I recommend everyone to learn this technique, its so fun, however hard in the beginning
Such a timely repost of the Micrography youtube channel to Breaing Taps. They use an eletronic force microscope as well.
5:08 you can cross your vision here and get a full 3D view
This could be used to make the coolest music video
4:35 It actually is something that is getting quite good! they are called Neural radience fields and there are being very rapid improvements!
It will be good for applications such as ballistics
I'd call it a PMD (precision measuring device). Have you checked out the technology that uses smoke and a laser as the sense element for an audio microphone? It is unreal.