Polypaint a Head in ZBrush - Color Theory 

Thanks a lot! Really appreciate it :D

I totally agree 🤝

Fantastic! Glad it's useful.

I think he was just talking about Poly Painting (or vertex colors), and I don't entirely agree. If you are baking to a UV mapped model, then chances are you want to keep editing in a painting app or Photoshop or whatever you choose to use, in which case the level of detail should be based on how far you plan to place the camera. Moreover, if your model has any wasted space in the UV, you'll need higher resolution to get detail for every vertex. I have been working on a full character which I decided to unwrap into a single map (Blender doesn't support UDIMs at this time.) and I have to crank the resolution to get clean detail all over the face.

Hey Danny! For me it really depends on what it is you're doing and how close to camera it is. For a model like this, it's a complete waste to do more than 2k, simply because the polycount isnt high enough to support it. Sure, there might be unwasted parts of the UV, as not all parts of the UV map is being utilised. That said, like we show in this video, the texture is incredibly soft, as it's all hand painted. You can see the same if you export a displacement map; the map is really super soft. If you're doing really stylized work, really dont need a lot of resolution compared to if you're doing photo realism, as you then mostly want areas with softer colors and some patterns, while in realism, you need just a lot higher fidelity in your map. Instead of working up the resolution (above 4k), I prefer to split it into multiple UDIMs. Remember that a 4k map is 4x bigger than a 2k map; 4million pixels vs 16 - and an 8k map has 64 million pixels. To Josephs point, you can always upres your map in Photoshop if needed afterwards. if I was rendering this model directly, you'd never see the difference between a 2k and a 4k map - as the map is just so soft. I usually overshoot my work a bit, just so I know I'm not working at the breaking point always. So in short: - Depends on how close to camera - How complex the texture is. Let me know if this helps! /Henning

Thanks to both of you for the responses. I also spoke to Vince Dromart about it and from what he said I'm realising is that as well as how close up it is, it also depends on the rsize/res of the final image - which seems really obvious and something I definitely should have picked up on by now! It makes a lot of sense now and I think my workflow will improve as a result - Thanks a lot!

Great, thank you!

You bet! No promises as to when, but I'm definitely eager to do Mari tutorials soon. /H

Thank you! :)

Thanks for watching!

Hi i have some questions

Haha! Glad we can help. Best tip for dissertation writing is to do the bare minimum you have to do in order to pass and spend all your time on your reel and portfolio (if you're in CG, that is).

In ZBrush the layer system is pretty old, so no blending modes unfortunately. I really dislike using layers for polypainting. When I do advanced textures, I'll use Painter or Mari. /H

Oh, you bet! We are definitely doing texturing tutorials in the future.

FlippedNormals awesome!

Mari or Substance Painter :)

"pure reference" thats the name of the software.

maybe enable backface mask

Right Click - Mode - Always On Top

Polypainting is painting directly on the verts, which means that every single vertex has a color. This is really useful for concepting, but it would be far too heavy for production, as you'd essentially need one material per polygon. It's also tired to the polycount. In a texture map, youre painting directly on a map, which is completely independent from the polycount. You could have an 8k map on a single polygon for instance.

Painter is a much better tool for texture painting. For pure color map painting, ZBrush is a bit faster and immediate, but you can really only do the base there.

@@FlippedNormals So why you not make video about painting model in Substance. I really need this.

For sure, proper retopo is *mostly* for assets which are deformed. We have a whole video on the topic where we discuss it in-depth :)

I think its a bad idea to fill your scene with hipoly assets if you can get the same effect with lopoly assets, simply because you will run out of RAM and increase render build time. Depends on the render engine of course, but you'll have to wait for all of that data to copy from disk to RAM or VRAM (or maybe both) and then you'll have to wait for the Render BVH to build (for a path tracer, at least). Messy topology can be a problem with tesdelation/deformation at render time, also. If you're using some sort of adaptive subdivision, you'll want to keep your meshes clean.

Joseph Brandenburg sounds like a pain then. But usually you can cache it so it only needs to processed once, obviously with deformations that's not going to work out but on a static object in a scene it usually works out fine. Thanks for the advise (:

Caching is great, just make sure to have an external hard drive to hold the cache files (which can get pretty big) and of course it still uses the RAM. I use Blender, and it saves the cache in the Roaming App Data hidden folder, which means I sometimes run out of disk space and don't know where to look. Be sure to check hidden files. It doesn't really matter how much RAM you use or how much disk space you eat up if you are just doing still renders. The problems arise when doing animation, because little problems get multiplied by thousands. It's a good idea to retopologize and UV unwrap stuff and bake the hi-res data onto it. That way, you can reuse it later. You want to save yourself the trouble later when you have a client and a deadline. Moreover, if you can convince a client that you can do things fast and efficienty, even in animation, it might make it easier to find a job. I don't know, it's still a hobby for me.

Joseph Brandenburg completely agree. I'm only 17 so you probably know better than me. UV mapping is something I'm currently learning in Maya but oh my god it's tiresome. At least with rigging and retypology there's some visible progress on the model itself

Both are good! Polypainting is better for experimenting but Painter is better for final maps

@@FlippedNormals by final maps u mean the end product right? Like the finished model?

just hold shift and click on the Zadd.

You probably have the zadd option enabled when using the smooth brush. So press shift to use the smooth brush but first make sure that the zadd option is turned off(while still pressing the shift button).

Yup! flippednormals.com/products/the-complete-zbrush-to-modo-workflow/

FlippedNormals awesome, thanks!

There's a zplugin for that, in the top ribbon menu

Fair enough - well spotted!

I believe it's because ZBrushes colors are backwards. The secondary color is on the Left, and the primary color is on the Right. So you're doing it right, it just feels backwards.

PureRef

That's a great idea! Thanks.

FlippedNormals thanks for listening!

I'd suggest just observe reference of people with darker skin tones. Really observe it and sample the colors.

flippednormals.com/downloads/introduction-to-zbrush/ :)

Im facing the same issue!

@@noor96498 that's zbrush... I guess you just got to keep practicing to figure it all out. It's worth it though.

What resource are you talking about here?

We'll do mari tutorial in the future, I promise :D

FlippedNormals awesome, I'm already hyped for it :D

Maybe a comparison video follow up to that! See the same model being paint in Mari :-) I know how painful it is but... want to see... am I the only one struggling with Mari's "bake"? ;p~ Nice video btw! Thank you guys!

Hey, you gotta take one of the team and explore some weird areas :D

Stay tuned! We'll soon be releasing something that might help you out

@@FlippedNormals i love you thank you

It's great for studios but for personal use Painter is much more achievable