To me the biggest difference between substance painter and Mari is learning curve,aside from this channel and flipped normals I have had difficulty finding helpful tutorials for beginners to learn mari,when I started learning substance their was an abundance of tutorials both free and paid and the paid courses really went in depth.I am trying learning mari and unlike substance their isnt really much readily available information both free and paid for introductory lessons to learn mari(as even the tutorials on their website are old) making it chore to try to learn,so unless you are really willing to search hard,try and fail on your own,follow along long someone texturing an asset and take notes while they do it, substance does become the default for many people learning texturing on on their own.
actually there are quite a lot of mari tuts, check tom newbury, meshmen studio, zak boxall, Michael Wilde and countless more. you dont even need to go deep at all. its there everywhere
For me the node graph in Mari is what still gives it an edge over SP. Especially when you have a complicated asset, then Layers and anchor points in SP becomes hell. Node graph in Mari just makes it so much easier.
That is very true, love the node graph. When substance came to DD to demo their new version I actually asked them if they ever gonna make node graph for Substance Painter, seems like they don't have plans like that anytime soon.
@@ArtruismDigital One thing that bugs me about painter, even in it's newer version is the lack of performance when dealing with UDIMS, even with low amounts like 10,15 udims in 4K, i find that gets extremily heavy... even with udim disable feature.
You are truly a god send. These videos are concise and well thought out. I'm very comfortable with Painter but not so much as Mari. I'm starting to understand why each app has its pros and cons and this will definitely be helpful in the future. Thank you!!!
It's quite breathtaking hearing the level of detail and number of UDIMs artists use in high level production. I'm curious, what is an average UDIM count for a hero asset? You mentioned a spaceship could have 100 tiles, but I fee like that's an extreme case. I realise that's a pretty loaded question with a lot of variables. Just looking for a really rough estimate. Like maybe a digi-double, or a dinosaur from Jurassic World?
those high UDIM count is quite common is VFX, every transformer is prob around 200 to 400 UDIM in a transformer movie. Digi-double maybe 20 to 40, a dinosaur is prob 20 -30.
Chun chun Mam ...well said...skill and the will to make what you visualize and breathe life into it is important ....softwares are just the mediums not restrictions ....peace love respect
The other main difference is the cost. Mari is significantly more expensive as far as I can see. If one is working on a budget and spreading work across so much different software it can all get very expensive very quickly.
Mari subscription is 599 a year. For students it is free one year and there is a fully free non-commercial. So technically it is even a bit cheaper than substance for a fully pro license but of course substance offers a lot more in its suite (both software amount as well as substance source credits)
@@MariExtensionPack A perpetual license for Substance Painter on steam is around $50. That's cheaper than 1 month Mari subscription. And a perpetual Mari license is $2168
@@Spyro_2076 A ok. I wasn't aware there were still perpetuals for substance. Thought it was all subs now. Anyway, not trying to go all hail Mari. SUbstance on Subs offers a lot more than just 'texture painting'. But I also am a bit allergic to the general Mari bashing, which I find comes a lot from not being properly informed.
@@MariExtensionPack No prob, I meant no offence.. I got the $50 price because it cost me R819 (south african rand) which is $49.. Guess I'm lucky. Also. I am def not bashing Mari, I've never used it so I am in no position to even have an opinion on it.. Cheers
Why did the red mask textures look so different? The one in Mari looks way more detailed. It makes me think Mark captures more detail somehow, or is that from an old version of substance painter? Thanks for the great video btw!
well one reason is the animation was composited and rendered to film. I believe what you are looking at is a scan of the film not the digital animation file. So once it gets put on film the film chemistry and grain does a lot to normalize the whole picture. Basically add more grain to your shots if you want it too look reall...LOL!!!! of cours if the motion wasn't good and model wasn't good it would look like crap and has nothing to do with textures.. keeping these shots really short is also a key point.
I use zBrush polypaint and SP. From zBrush to other softwares to organize the UVs and then to painter. Not sure how zBrush compares to Mari assuming you can make UDIMs elsewhere cuz we know zBrush is not the best at that. Never used Mari myself but zBrush should be able to handle high poly models and has an okay projection tool I guess (spotlight).
Comparing ZBrush and Mari is like comparing pencil and brick. Even that you can use brick to draw, it don't mean it is good at that, brick is used to build house not to draw. Zbrush is sculpting software and painting in it is just additional option for concept artists, that was never improve to much, and there for it don't have real value in professional production. It is to simple and also, it paint on polygons and not on pixels (in export it blur image when transferring poly color to pixel base image). It dont have layers or nodes that are more than necessary, it paint only color, so you cant make other maps as bump, gloss, masks, it dont have advance adjustment tools... And list go on and on.
@@ArtruismDigital okay but I have to bake again and again right because its projection. and like in substance the brush change the direction according to the surface, dose mari have this future?
Hope to see one that you works on facial make _up material workflow , is that possible to create one pbr material that can cover a lot of variation of lipstick?
The difference make how good is rhe photo you have and the color and to have good pictures tou mast have good canon DSLR and higher so the croma and the megapixel is unique now i use Mari from the beginning for the Reason it use ptex texturing and have beter colors when you painting 🎨 the textures plus is bpr so is amazing
I am using substance and mari for texturing as well and there is one thing I found that is lacking in Substance. If one geometry, es. a flower that has a geometry that overlaps or that is on top of another geometry and I need to hide those polygons in substance, I the geometry is not priorly divided per material in may, is impossible to hide it in substance. Or is there a way to do it?
I think for me its the speed. SP is much faster when it comes to achieving certain results comparing to Mari. In Mari you spend too much time arranging node graph and making procedural mask. I find it such a waste of time. But as you said Mari is good for projection. Apart from that SP is the real winner. Hope they will improve the speed of the software when it come to viewport updates.
What is the best way to project texture to high ploy to low poly. I tried substance painter, substance designer, mari, x normal, but texture is projecting blur.. I stuck there .. I don't know how to project texture whitout getting blur
its going to be low and mid poly stuff for a while. Don't let the unreal 5 demo fool you. You still have to hit frame rates on multiple platforms and a wide range of graphics cards.
@@3DSgeek Also, have you seen the size of a 30mil zbrush sculpt along with 3 of 8k texture maps? Yea, not happening cause you will require 1TB ssd for only one game. We will still use lowpolys, the good thing is that they wont have to be as low which is cool.
Polycount is not only thing that game studio need to wary about. Everything in game have cost, including textures, shaders, is model breakable or not, does it have animation/rig in it. Than effects, render staff (ray trace, global illumination), how far player will see in distance, is it open world or arena style game, how fast is character move, how optimized are shades and other staffs.... For example, if you have boxing game, everything can be made like in movie those days, characters of 250k polygons with decimation turn on and few udims... If it is fps/rpg open world, characters are 30k - 120k, environment is where people start to save polygons, but there is room for more edges... And if it is car race game, o man, it is saving on every polygon, on every teksture map..... On end what is important is not polycount per asset or character but total poly count that camera see and system calculate at any time, and also, how fast new staff will go in and out of memory. (FF 15 hold in memory 11m polygons in every time, and star citizen break record of 30m polygons in any time, on free demo release, but it had to much problems with frame per second, it could drop even to 8 frame in some moments). So, even we will have more freedom, many studios will still use old methods with small upgrade. And for using high poly, it will take probably decades. For 20 years movies didnt learn to use High poly models and game will not win that battle (i dont even know if there is battle, noone even trying to do it yet :) ) Movie use some mid poly in scene, and thay subdivide it for render and for small details, it use displacement map and bump maps to get higher realism in render. One advice is to always believe demo only 60% until you get playable demo and see what is real and what is prerendered. Even if it is not prerendered, it can be scripted (no player that move character, like in games, but instead it is scripted. Something like infiltrator demo on UE4. Also any modern game have examples, when you play, you can see some lag, but when in game cut scene start everything look better and usually it have less lag, because player dont add random factor in calculation at that moment - for example, if you play, and sudenly you deside to turn camera to 180 degres, sistem need to calculate everything that was behind in a secend, and whan it is programmed to happen, system will start render all behind you even before camera turn around)
