When asked about, what's a good software for texturing, I bet many of you would say Substance Painter. I know, it's been longer in the field than Quixel, but I wonder, what is it lacking as from your point of view that if fixed, could make it more attractive for companies to switch to ?
Replies
Now is mixer similar to painter? I don't know its features, nor if its good or not.
I use Substance painter since it early days... its like third arm for me.
Is there any side by side comparison ? Can I try it for free?
Its just how much features you would like to tradeoff.
How much money do you like to spend vs. how many feature do you need.
A good Artist will find a way to create good Art with every tool.
https://quixel.com/blog/2020/2/27/mixer-2020-a-new-texturing-paradigm
https://youtu.be/Eh-lL-iPNe0
https://youtu.be/kRtXIj6ENT4
It does not have to be the best or most efficient (we dont use maya because its a competent modelling package, we use it because you can make tools for it)
Painter is getting there nowadays but the reason it became so ubiquitous in the first place is that it had designer backing it up (which fulfills all the requirements above and then some) .
I've no idea what quixel are planning to do with mixer but honestly it feels like they're too late to the party for it to really matter.
A good and ultra-fast brush engine is a must. UDIMs, projection and normal painting, etc.
If performance were really that important people wouldn't be using maya, mari, solidworks etc.. because they all run like shit.
You can resolve almost all performance issues by throwing money and hardware at the problem - it's not unusual to have £10-15k workstations under every desk with multiple high end GPUs, Hundreds of thousands invested in render farms etc etc. - all of which is cheaper than not being able to make a tool work with your other tools
If you want to compete with painter at a consumer level then everything you've said is valid - it's unrealistic on a £1k consumer level pc at present but still a valid aim.
It will have little to no effect on whether studios adopt it or not though
Mixer is free to use for everyone. You can grab it via https://quixel.com/mixer. We're quite interested in feedback. I'll keep checking this thread for useful suggestions or discussions. Your ideas in particular are fantastic @poopipe - thank you!
I still remember when Photoshop 7 was launched with its "new brush engine", i could handle 12k px height illustrations and huge brush sizes without effort. And the computer i used to use was nothing compared to what i use today (and it was the very best). Right now, if i want to do the same, i need to use a very expensive PC because newer versions always demands a lot. The reasons, are very well known and it's a common issue with old apps like Maya. With each version, they get more and more bloated and pc requirements are higher.
On the hardware side, you must understand that Indie devs or small/medium studios can't waste too much money in those DCC workstations you talked about so casually, nor licenses of apps with nasa computer requirements, and less, upgrade their systems every year. For that reason, an excellent performance on an average workstation is a must for at least 3 years. Performance is the most important thing in computers, both in hardware and software.
The lack of performance is one of the reasons of why i don't use Blender for all, like too many other apps. It doesn't performs like it should be.
Mixer may be free, ok, but it lacks like blender in all its aspects. Performance is horrible being honest, and the worse is that all is pretty basic and simple. The app struggles a lot with common tasks. There are much more competitive options without needing to invest in hardware, and to complicate things much more, substance painter is taking advantage of photoshop's brushes. The race is like lost by 3 laps, and mixer is still in beta. Too late.
And i'd dare to say it's not just about consumer level PC, but standard and average workstations for games development. Take for example an actual Ryzen 3900x or Intel core 9900k /64gb/RTX2080ti with a dual 4K monitor setup and a cintiq pro 32/24 as the most expensive part. In the majority of studios you won't see the latest hardware in all the computers, just plain and common computers with enough ram, a good cpu, a good video card and a dual or triple setup of monitors.
To put the hardware as excuse is not valid for me, and more when actual software don't take advantage of past hardware.
To sum it up, give users performance and you will have a battle won. Take advantage of 5 years old hardware and don't expect users to pay for new hardware every year if they want performance, that's the gold rule.
I never saw such a studio from the inside. Nor one of our client have rigs like that.
2. You can do a lot of this through the Mask Stack by adding a Map layer and placing a Projection modifier on top of it.
3. Projection masks will do this, but we're looking into improving how decals are implemented in Mixer so it may be more streamlined as time goes on.
4. "Blend If" doesn't exist to my knowledge but you can approximate it pretty well using Blend Above/Below layers.
5. Our brush engine is raster right now, not vector. This is an interesting suggestion though, thank you!
6. The brush engine is still in its early development phases so we'll be improving it as we continue developing Mixer.
7. Box Projection is pretty close to what you're looking for.
Also i would really like to see something similar to anchor points on mixer.
Painter also has a finicky UI, strange stacking, and is not fast either
i.e.
Creation/mixing of tileable materials with a simpler workflow than designer. Ideally including robust manual mask painting etc.
Support for UV scaling to support terrain material creation/preview would help also.
This would sit the app at the pointy end of the pipeline-enabling it to take bitmap or (ideally) sbsar data from Designer and give teams a painter equivalent for tileable materials.
As I said above, I've not used mixer- maybe it supports this stuff, maybe not
re: the post above..
It's unlikely that software that does the same things as painter is ever going to be orders of magnitude faster than painter-it's modern, its under constant development. The speed issues it faces come from the architecture which is required for it to support the featureset it has. Obvs it could get faster, but I'd only expect to see incremental improvements until the hardware makes a big leap.
I've yet to run alchemist through a full evaluation in terms of environment art use so can't say whether it's any good or not. we have dismissed it as pointless for specialist material artists given that they use designer.
Fwiw the new SVT feature is sort of a step towards what you're describing but really the big problem is that a gpu with 8gb doesn't have enough capacity to support processing a big layer stack at 2k resolution and there's a limited speed at which stuff can be shuffled back and forth between main and gpu memory.