I remember sharing an email correspondence with you regarding this subject several months ago, I'm amazed at how well you were able to do this. I'm so glad 4.3 has a better method for unwrapping now, so hopefully i can preserve the details on my meshes
One clear application of this method of calculating vector displacements is that you could animate the object, create an image sequence of vector displacement maps from the result, and then use them to create an animated vector displacement map.
Man! I've been working in computer graphics for over 15 years, working on movies, animations, and all kinds of projects. For me, most tutorials out there are just meh. But you constantly come up with ideas that are genuinely interesting and conceptually fresh - not just about being good with the software.
long story short: 1️⃣ VDM goes from UV space to 3d positions 2️⃣ with a bit if trickery we can reverse this process 3️⃣ the inverse of this process weve calculated IS THE VDM
Here's what I'm confused by. Now there are MANY ways to unwrap the 3d model onto the UV space right? Yet to generate the VDM we subtract the head's position and UV map vector This means the same VDM can be generated by multiple different forms of UV mapping. How does this work then? If A, B, C are UV maps of the same model but unwrapped differently, this means that Position - A = Position - B = Position - C = Same VDM despite the fact that they should give us different results right?
วันที่ผ่านมา +5
This would be a really cool step in one of those super elaborate hard to solve internet treasure hunts
Thanks a lot! I think when they introduced VDM in blender they added a feature to bake 3D objects to VDM. isn't it better? idk what's under the hood in their setup...
This is technique is used all the time to replicate the details/overhangs of the waves of oceans! Where the amount of geometry is prohibitively large and an animated texture map can contain more details.
I have a question: You know switchlight? The tool that turns an image of a person into a normal map, roughness map, albedo map and such, and allows you to relight the person... could you somehow turn those maps into a 3D model of the face? I have tried this but it just ends up looking really bad, I'm a beginner in terms of blender though. The normal maps looks very good and accurate, would be cool to be able to use it for 3D printing or a digital 3D clone!
The explanation is great, though this method doesn't work great on all meshes (without further modification). I don't know if its just the meshes i was testing on, but if the hole in the geometry is small compared to the total plane size, the mesh will have significant stretching and look terrible. a cartoon head with a small neck for example, will have horrible face shearing and be of low quality.
I remember sharing an email correspondence with you regarding this subject several months ago, I'm amazed at how well you were able to do this. I'm so glad 4.3 has a better method for unwrapping now, so hopefully i can preserve the details on my meshes
One clear application of this method of calculating vector displacements is that you could animate the object, create an image sequence of vector displacement maps from the result, and then use them to create an animated vector displacement map.
Man! I've been working in computer graphics for over 15 years, working on movies, animations, and all kinds of projects. For me, most tutorials out there are just meh. But you constantly come up with ideas that are genuinely interesting and conceptually fresh - not just about being good with the software.
long story short:
1️⃣ VDM goes from UV space to 3d positions
2️⃣ with a bit if trickery we can reverse this process
3️⃣ the inverse of this process weve calculated IS THE VDM
you can check out the .blend either on my website or on patreon
Here's what I'm confused by. Now there are MANY ways to unwrap the 3d model onto the UV space right?
Yet to generate the VDM we subtract the head's position and UV map vector
This means the same VDM can be generated by multiple different forms of UV mapping.
How does this work then?
If A, B, C are UV maps of the same model but unwrapped differently, this means that Position - A = Position - B = Position - C = Same VDM despite the fact that they should give us different results right?
This would be a really cool step in one of those super elaborate hard to solve internet treasure hunts
I can make some truly weird stuff with this now. Your blender knowledge consistently astounds me.
I must admit, I never really understood VDMs until you told me the simple way to do it with a shader. Absolutely cool.
you are one of the coolest Blenderers out there, it's definitely not easy making so much great content so consistently but keep it up please
I didn't even know that this was a thing. Sick.
The tutorial is awesome, but also, you look cool in these glasses 🔥
This truly is sacred 😌
This is some crazy tech
God, I hate CGMatter SO much.
Is it just jealousy?
Yes. And?
Always learn so much from your videos. I appreciate ya. 😁
Incredible. Great work. Thanks for sharing.
Thanks a lot! I think when they introduced VDM in blender they added a feature to bake 3D objects to VDM. isn't it better? idk what's under the hood in their setup...
Can you teach us color space transformation or color science? like in davinci Resolve, and all the stuff is used. Full from basics.
this is a great explanation! thank you!
Holy shit. You are amazing. I love you. This is so clever.
Oooh. Wow. Thankyou😮
WOW, amazing! thank you, can you paint colors, and textures and displace at the same time?
The thing that I am really waiting for is the use-case for this. Where would this be advantageous to use?
This is technique is used all the time to replicate the details/overhangs of the waves of oceans! Where the amount of geometry is prohibitively large and an animated texture map can contain more details.
"No Geometry Nodes"
FINALLY!!!
That's an amazing video! please can you tell me how you did that animation like switching in 2:30
Genius !
I have a question:
You know switchlight? The tool that turns an image of a person into a normal map, roughness map, albedo map and such, and allows you to relight the person... could you somehow turn those maps into a 3D model of the face?
I have tried this but it just ends up looking really bad, I'm a beginner in terms of blender though.
The normal maps looks very good and accurate, would be cool to be able to use it for 3D printing or a digital 3D clone!
The explanation is great, though this method doesn't work great on all meshes (without further modification).
I don't know if its just the meshes i was testing on, but if the hole in the geometry is small compared to the total plane size, the mesh will have significant stretching and look terrible.
a cartoon head with a small neck for example, will have horrible face shearing and be of low quality.
It's very cool but what's the use in production here? I feel like it will always create more geometry than a well-optimized mesh
Maybe mixing them...?
Don't bake it on to just a plane. Bake it on to the model you want to displace.
Train a neural network on VDM and get a Text2VDM model
@@thebigfortuno3329 ew, no AI again
in the first method could you then convert to object and have perfect topo?
I feel like we could use this to make the most accurate map of the globe ever
I might be wrong, but would it be possible that, for the scale, it was a color space problem?
I guess it is time to make an update for the VDM brush baker my boy 😀
Would this animatable in a game engine?
Is there a speedometer to measure how fast this guy's mind work?
How are you not yet being hired from the blender group yet?
Harrier Du Bois thumbnail?!?
ay nice
Wow. ❤
Vector displacement is not as standardized as regular displacement I believe
fun fact VDM in french means FML
Holy smokes
Bro your thumbnails
👑
how cam your chanale only have like 291k
Bro, I'm not in a position to judge anyones handwriting, but what is that last word on your awesome website? bring back "****" ?
I don't know about sacred, but I get scared seeing your knowledge of blender
So I didnt like how you left the head upside down tilted..
It is part of the sacred displacement ceremony, you wouldn't understand
vector displacement is DEAD
You are a CRAZY GENIUS - what the actual!?!? 🤯🤯🤯🫡