Home Adobe Substance

[WiP] A Pilgrim's Mountain Journey: A Material Blending Project

2

Replies

  • zachagreg
    Offline / Send Message
    zachagreg ngon master
    Yea focus on that art-test honestly, whatever internship that is shaping up like that is not worthy of your time. 
  • Flight
    Offline / Send Message
    Flight polycounter lvl 10
    Well, you live in the right place. Plenty of studios around there, that or Malmö. You got some good stuff going, just need to practice and get in the swing of things.
  • Finnn
    Offline / Send Message
    Finnn greentooth
  • radiancef0rge
    Offline / Send Message
    radiancef0rge ngon master
    Finnn said:
    don't quit
    how is this good advice?  the 'experience' isnt worth free work man. 
  • zachagreg
    Offline / Send Message
    zachagreg ngon master
    Finnn said:
    don't quit
    how is this good advice?  the 'experience' isnt worth free work man. 
    Especially if he is the one fixing their stuff. Clearly if that's the case they need him more than he needs them and him not being paid is quite literally not worth it. If it were the case where that was his livelihood then that is a different story, but that isn't the case.
  • defragger
    Offline / Send Message
    defragger sublime tool
    No, he is right! Don't quit! ... at least not right away!
    How to renegotiate is also a valuable part of work experience.

    Your textures are dope by the way! Following this thread for a while now.
  • radiancef0rge
    Offline / Send Message
    radiancef0rge ngon master
    what?  its not relevant work experience if hes working for free and fixing their mistakes - it means hes working at a non reputable company.  based on what experience as a hiring manager are you giving this advice? 
  • defragger
    Offline / Send Message
    defragger sublime tool
    ...hes working for free and fixing their mistakes....

    thats why I said renegotiate

    If there is no agreement you can still leave the place.
  • Daniel_Swing
    Offline / Send Message
    Daniel_Swing ngon master
    Thanks for the encouraging comments guys!
    But yipes, please don't derail my thread!
    I'm definitely not sticking to my current "internship" for much longer. Even if it's experience, it's experience that I won't need in the future, because I hope that I'll never have to work in project like this ever again. What they want me to fix are not some aesthetic details or minor resource sinks. It's much more grave than that and so fundamentally flawed that it would be easier to start their 4 year project from scratch, at least when it comes to the graphics (did I mention that the game is supposed to release in march?).

    It doesn't fill me with confidence when they call me their 'savior', when I'm the one not getting paid.

    The only value I really get out of this is the "Bestiary of Monstrosities":
    - it looks like a simple corner pillar for an energy field, but it's actually a non-instanceable asset with 80+ draw-calls. I call it 'FPS devourer'.
    - colored Metallic maps, known as 'the Butcher of PBR'.
    - the odd 700x700 texture, don't know what to name this yet.
    - the common, oversized, colored textures that are actually just a uniform gray value, 'the swarm of incognito resource leeches'.
    Consistent aesthetics, proper usage of PBR and texture packing is all but unknown to the poor, wretched assets.
    Every so often, one can spot an assets with only one material ID, with responsible color and specular values, reasonable pivot point and with a name more descriptive than "mesh". Some say it doesn't exist and that it's only a mirage in all the madness. A diamond in the rough.
    The highlight of my day is if I find a new, unique monster to add to this list.

    And that's that.
    I'll be getting back to this project some time or after next week (when the art-test is done).
  • zachagreg
    Offline / Send Message
    zachagreg ngon master
    Oh those are just...gross. Reminds me of a royalty project I worked on briefly. HOWEVER!! 

    When going through and editing your height information via histogram range did you use any convention or metric to determine the correct range? I've had similar issues authoring a couple of materials in the past. 

    I imagine a rough visual estimate of the levels would be sufficient but just curious of other ways.
  • Daniel_Swing
    Offline / Send Message
    Daniel_Swing ngon master
    Getting back into things.
    Soon done with the rock material (I hope):


    @Flight I didn't plan on doing any mesh support for my materials. Though I may consider doing if I want to add rock meshes to my scene before I call it done.

    @zachagreg I'm mostly just eyeballing the values, trying to keep it around 0.5 linear (because that's what my shader defines as "middle" with no displacement). I'll probably have to put some more robust thought into my value ranges before I finish up.
  • Daniel_Swing
    Offline / Send Message
    Daniel_Swing ngon master
    Tweaked the color, roughness and specular a bit. Getting really happy with it actually.
    Sad face for the displacement hole in the landscape :(
    A quick sample of the Base Color buffer:

    I think I'm really close to happy with all landscape materials. So I'll be moving on to doing some sort of dead, dry tree?
  • TheGabmeister
    Offline / Send Message
    TheGabmeister interpolator
    Dang. I like where this is going. Keep it up.
  • zachagreg
    Offline / Send Message
    zachagreg ngon master
    I think I'm really close to happy with all landscape materials. So I'll be moving on to doing some sort of dead, dry tree?
    Or Joshua tree's?  ( ͡° ͜ʖ ͡°)
    Image result for joshua tree


  • McGergs
    Offline / Send Message
    McGergs polycounter lvl 9
    Blown away with how far you have come from the last time I looked!
  • Daniel_Swing
    Offline / Send Message
    Daniel_Swing ngon master
    Thanks peeps!
    Done some slight tweaking to the rock material. Some more test-shots:

    I'm thinking I should make it a bit more gray in the tone, to match the rubble rocks.
  • Daniel_Swing
    Offline / Send Message
    Daniel_Swing ngon master
    I can't get happy with these rock shapes... I think I'm doing progress though.

    I've been thinking that I really should get this project done asap and not prolong it by adding more stuff. That way I can get started on something new and fresh...
  • Daniel_Swing
    Offline / Send Message
    Daniel_Swing ngon master
    I  think this is my last attempt.
    I'll move on to investigating the actual blending a bit more, getting some artifacts on the rubble rocks in slopes and the height blend is wrong in certain areas.
  • Daniel_Swing
    Offline / Send Message
    Daniel_Swing ngon master
    Today I've been sitting mostly with my materials/shaders to clean up and improve how the textures blend. I've tried to cobble something together from old tutorials and my own experimentation. I'm not confident in that these are the easiest/best approaches, so feel free to give me some pointers/tips if anyone knows a better way.

    I've made 'Material Functions' for each sub material to make the master landscape material more foreseeable.
    You can see how I've chosen to pack my textures in the graph below.

    Special note: I found that using the same height maps for displacement and height blending didn't really give me a desirable result in the height blend. So I included one height in the Base Color alpha that I input into 'World Displacement' and another more 'mask-like' height map (see the thumbnail in the graph) in the red channel of my packed textures. I output the height mask separately, so that it's easily accessible in my landscape material to hock up to the 'layer blend'.
    I still don't know how the "height blend" actually works and how it uses the information in the texture. Though it seems to be more of a masking function than an actual "height" function?
    I have two input parameters in my material functions that I multiply together, to control the UV/tiling. I assign these values in my landscape material, one functions as a global UV scale and the other one is for individual material tiling.

    I blend my material functions using a regular layer blend with one empty layer (I read somewhere that this was a fix for potential black spots), two 'weight blended' layers and three 'height blended' layers.

    I bring the 'World Displacement' values into -1 to 1 range (so that it displaces both in and out), multiply it with the landscape's vertices normals (to make the displacement tangent the landscape) and then multiply it with an intensity parameter - I did this in another order before today, which resulted in a number of visual artifacts by the displacement in slopes.

    Today I also noticed that the tessellation didn't automatically LOD away in the distance (don't know why I assumed that it would, but it ate a lot of performance), so I added a manual control for this tied to the distance of the camera .


    A major reason that I've been struggling with the Rock/Cliff/Mountain material is because I couldn't get it to look good both up close and from far away. So today I cheated around that, by modifying the material function to lerp into a lower tiling when far away. I do this with the same method I control the tessellation distance, with different constants.

    The hill's shapes lerp in real-time depending in the distance of the camera.

  • Daniel_Swing
    Offline / Send Message
    Daniel_Swing ngon master
    Been tweaking the lighting a bit:
    I think I'm about to call this done.
  • zachagreg
    Offline / Send Message
    zachagreg ngon master
    Just curious, what's the performance like right now on this bad boy? It's looking phenomenal and I'd like to see some of those transitions for tiling and tessellation in action. 
  • Daniel_Swing
    Offline / Send Message
    Daniel_Swing ngon master
    It really depends on how far I pull the tessellation factor and how far into the distance. In the latest screen-shot it's stable above 60fps, though it's above 100fps with more reasonable settings.
  • Daniel_Swing
    Offline / Send Message
    Daniel_Swing ngon master
    It's up on ArtStation: (https://www.artstation.com/artwork/6aRbgW) now.

    I'll be making a break-down of the entire project soon (a more condensed form of this thread + some extras that I haven't been able to mention in here). I'm just not sure if I should make an entirely new thread for it (to keep it focused?) or put it in this thread?
  • 3DKyle
    Offline / Send Message
    3DKyle interpolator
    Well done on this Daniel, love it <3
  • zachagreg
    Offline / Send Message
    zachagreg ngon master
    I'd recommend making a new thread just so that it can be condensed and focused easily.
  • Daniel_Swing
2
Sign In or Register to comment.