If you have a tutorial covering the whole workflow from Blender to Substance painter then I will buy it. That was a solid tutorial man thanks.

Thanks Chipp, that means a lot!

What he said. That was super well explained. I finally got it.

Agreed, I finally understood it, that's how a tutorial is made.

Glad that it helped!

I since found out that it's UVPackMaster and it's a paid one but it's totally worth it :)

devtalk.blender.org/t/call-for-content-themes/3174/377?u=jonlampel

@@cgarcane6653 aaah man! thanks a lot!

Thanks, beat me to it!

Ah yes, thanks for the catch!

I think it was an affix but it's the same as the image sequence mechanics by the looks.

@@cg_cookie , now I'm confused: prefix or suffix?

Sure thing! devtalk.blender.org/t/call-for-content-themes/3174/377?u=jonlampel

I just want to know if you paid the troll toll yet.

Essentially yes.UDIMs allow texture artists to upscale specific parts of a mesh while still keeping the same texel density. This could of been explained better with an organic model instead of inorganic model made from separate parts.

No. That is just visualization, UDIM isn't a merged texture, it's a way to make a program use many textures in the same mesh. This way you can avoid working with ultra large texture (which are hard to edit and use due to performance) and you no longer need to slit a mesh in many pieces to get the same effect. This is mostly useful for movies because you need a lot of texture detail for realism particularly in close up.

"Variable-resolution tiling” seems to make more sense to me.

@@lawrencedoliveiro9104 But why be so complicated? Imagine Blender calling " 2D coordinates space mapping" instead UV map. Just to be the weird software that doesn't follow the industry naming? It's a good name your idea, but way off the industry workflow, like, for example, when you search for something and found information in another software tutorial but you understand the concept in Blender because it is the same thing, you know? Like IPO Curves to Graph Editor, etc. These are a stupid examples but make sense when you are a professional and are trying to solve something for a project. Cheers, mate!

@@MoraesJr my guess is they aren't calling it UDim because it doesn't completely comply with the spec (yet) so it technically isn't UDim (yet)

@@sinom Yeah, make sense. I hope it's the case. Cheers.

Good idea. I encourage you to make this suggestion on RCS -- sometimes those proposals actually get implemented! But yeah: what's up with Blender's wonky terminology? Is it a Dutch thing?

Not really it's mostly useful for texture painting and node trees but effectively it's like having different UV maps for different components

I could be wrong, but if it's anything like sprite images in app/web/web hybrid development, you're saving resources (thus speeding up prep time and render time) by skipping the act of the render engine requesting more than a single file depending on how complex the code needs to be. If they're using threading and each individual texture requires a separate thread then they might be holding up other operations while that's going on (depending on your available threads of course). Also if this is for a game the world be a benefit in load times as well as increased performance. No idea how much of a difference that makes though. If you've got a really elaborate scene with a ton of actors (like a large battle in a strategy game) the effect would add up quickly. That's my best guess as someone who works in a different area of software development.

Yeah, but you do it with one shader instead of one shader for each textureset. Fast and more efficient to work with.

Its faster and doesnt require as much computational power. Its all about efficiency and portability these days.

I don't know about the performance effect, but not having to have multiple materials for different textures sounds like a good enough reason to want this by itself, especially if doing anything more than the basics with nodes.

Exactly, I'm confused. I've been doing this in max / maya since the 90s. I dont think I've ever used software that couldnt do this. Its like if everyone suddenly started talking about an amazing new technology that allowed you to assign different smoothing groups to different face and they were calling it "UltraNormals" or something :P

Separate chunks! It's just separate images that are read by Blender as one texture.

@@cg_cookie - ah gotcha, thanks

In every other software I've ever used it can be multiple objects.

Udims give the highest texels per part

You should use UVs when your UV unwrapping (and modelling) skills are stronger. After the basic unwrap, you can set up your texture map so that areas with high detail get more space, and your seams are all hidden away where no one sees the pixelation from stretching a texture across vertices. OTOH, you should use UDIMS when your texturing and painting skills are stronger, and you want to spend more time painting and less time reseaming and rearranging your texture maps. UDIMS are also much more useful in small groups, where multiple persons can work on texturing the model without file sharing snafus. All of this involves a workflow where you want a single material to cover a whole complex object, which is useful in game engines, but not so much in various artistic renderings. Also, your final engine might not support UDIMs, but instead have a different solution (I see you, UNITY).

Yes, with UDIMs you could then apply single UDIM image texture, to single shader, on single object. The texture set design in Substance Painter is, and has always been, flawed. It was intended to only be used for splitting structurally separate objects but because visibility control in Substance tools is very mediocre people immediately started using texture sets to split models based on visibility requirements rather than shaders. UDIMs actually make this somewhat better, but ironically UDIMs are very unused for games while Painter is predominantly used in games. But oh well. UDIM is essentially just standard for packing multiple UV textures into one file. At the end of it the main consideration for UDIMs is: - Can you afford the performance penalty of having multiple maps per asset? (Compared to a single map) - Does every app in your authoring pipeline support UDIM's (If not it might not be worth using them) - Does your end application support UDIMs (Most game engines still don't out of the box far as I know) If you want to generalize more the common opinion is use UDIMs for VFX and use UVs for games. When you combine the penalty from multiple maps, with in general not having enough resources for more than 2-5 maps per asset anyhow and most game engines not supporting the files out of the box, it becomes pretty obvious why they aren't widely used for in-game. On the other side VFX often has no texture nor compute limits so UDIMs are widely used. UDIMs offer workflow benefits, too. If you are working on single UV increasing the texel density of a part of the model means you have to scale up some UV islands and move around others. With UDIMs you'd just split the model into tiles structurally and then if one part seems like it needs more resolution you'd just up the image resolution of that tile. (Especially with something like Substance which is procedural so you can just click from 540x540 to 8k that is very powerful) Incidentally GPU memory has really exploded recently I think. Consumer enthusiast cards are starting to get close to 10GB in memory, and professional cards like RTX Titan have over 20GB It could be that UDIMs will become mainstream in games before too long, too. That is purely speculation, though.

i suggest, watch it again. and if you don't, Basically, when you join your meshes into one object and you want your textures in high resolution but not heavy. then you break it into different images or udim tiles. it is a very important tool in the film industry.

@@spiritofmax Not in video games. UDIMs in games is unnecessary and inefficient. Movies on the other hand, live and die by UDIMs for hero assets.

Unauthorized Expression It's not uncommon to split a continuos mesh for example a human head into several udims to have an insane amount of resolution for close ups, using a robot with independent parts is not the best way of explaining the use of udims. In fact painting across multiple udims is key to get the most benefit in cases like the one I mention. Some time ago I made a character that had 8 4K udims just for the head, using a single map with the equivalent resolution would be very difficult to manage and even worse if not impossible using different shaders, it's not only the albedo map...specular, roughness, normal, subsurface scattering, displacement and everything you need needs to be taken into account.

@@anurandev7337 yeah you're right

Anuran Dev "not for video games".. so far, this allows a future proofing pipeline

I'm not able to make a followup video in the coming weeks, but if we do cover the topic again I'll be sure to use a better example like that. Thanks!