|
Odddzy posted:True Art is when your art director that can't draw sends you a bunch of samurai pics he's found on google image search and asks you to make a character based on his vision. It's always the first page of Google image search too. Being a "creative" at an ad agency must be hard work.
|
#
?
Aug 20, 2019 22:17
|
|
|
|
| # ? Jun 2, 2024 10:04 |
|
|
Just because you don't (can't) actually do the work, doesn't mean you don't have a vision and that you can't direct people. Granted, it often ends up feeling like a process of elimination. We've been doing a lot of 'realistic magic' (hah) recently and there simply isn't any references and it's a quite long process to develop...but then again, it's one of the things I like about the film industry...that you do often spend a long time developing a look...sure, sometimes too long and sometimes on irrelevant details, but I do still enjoy that process. These are usually directed by people who can't do what we do. Often they can't draw either. They do, however, have the ability to roughly tell people, in different departments roughly what they want, refine it over months and then assemble it all at the end into a cool shot.
|
|
#
?
Aug 21, 2019 00:55
|
|
I'm working on a project that has just gone into full production about 2 months ago and already the leads in charge keep mentioning crunch. I almost had a full blown argument about them continually using the term crunch over and over, it's like they've got stockholm syndrome and it's expected at this studio. I haven't crunched in over 6 years and I'm not gonna start now. The art lead (essentially art director) today I overheard him saying we should just turn off specular entirely (we're in a roughness/metalness workflow, so specular should be unadjusted and not hosed with too much, it's default value is 0.5 so roughness/metalness actually WORKS). I also keep making assets in Substance since that's our pipeline now, or so they say/intend, but I keep seeing people taking exported maps out into photoshop and adjusting them in there and then not submitting them to source control or back into the source substance file.  [ /rant]
|
|
|
#
?
Aug 21, 2019 02:21
|
|
Also, you should not be in charge of vision if you don't think you could execute it yourself or have a clear vision of what you want and how to get there. I fought with that in pre-production on my current project and it was torture.ImplicitAssembler posted:These are usually directed by people who can't do what we do. Often they can't draw either. They do, however, have the ability to roughly tell people, in different departments roughly what they want, refine it over months and then assemble it all at the end into a cool shot. "Refine over months" yeah sorry that sounds awful to me. Nothnx. Even one concept artist can convey a clear idea out in a fraction of the time it takes a director or whoever to photobash and collage their intent. ceebee fucked around with this message at 02:26 on Aug 21, 2019 |
|
|
#
?
Aug 21, 2019 02:23
|
|
|
I disagree generally with that. Just because an art director/creative director can't draw doesn't necessarily make them a bad AD. They may be a bad AD but being "unable to execute on it themselves" isn't the reason. One can bitch and argue with other departments all you want but that method isn't going to change anyone's mind and it just makes the person look like an rear end in a top hat (which means others are even LESS likely to listen to their probably good advice). gently caress crunch, though.
|
|
#
?
Aug 21, 2019 02:46
|
|
|
I don't think the people that brief me should be able to do what I do. I just expect them to do their job and properly explore the concept or look they're trying to convey past grabbing the first thing that comes up in google. It's OK to use a bunch of images in a brief but they should be properly explained and maybe you might have to draw a crude shape or two or a diagram to clarify, it should be in their interest to get what they want.
|
|
#
?
Aug 21, 2019 03:06
|
|
|
I think we should kill them all.
|
|
#
?
Aug 21, 2019 03:08
|
|
|
ceebee posted:"Refine over months" yeah sorry that sounds awful to me. Nothnx. Even one concept artist can convey a clear idea out in a fraction of the time it takes a director or whoever to photobash and collage their intent. Concepts are fine (and often help), but when you are doing things in motion, they don't always work. You do a version, you try to pick up the bits from that version that works and develop them. It's a process and most good design goes through that process. Even when you have the general concept nailed, you still end up doing a lot of work to refine it, even if it includes throwing away a lot of stuff. The amount of time we spent on the FX on Joi in Blade Runner 2049 was insane and it probably passed through every FX artist on the show...but the end result was worth it. I also think it's something that the individual artist often don't see. Just finishing a show, where I had 4 different artist work on the same task through the timeline of the show. The first 3 all think that they worked for nothing, but in reality, each of them established a new baseline, for the next one to start with and the end result is really a group effort, even if it feels like 3 failed individual efforts and one successful one. (And I've tried my best to emphasize that it was a group effort).
|
|
#
?
Aug 21, 2019 03:13
|
|
|
ceebee posted:I'm working on a project that has just gone into full production about 2 months ago and already the leads in charge keep mentioning crunch. I almost had a full blown argument about them continually using the term crunch over and over, it's like they've got stockholm syndrome and it's expected at this studio. I haven't crunched in over 6 years and I'm not gonna start now. The art lead (essentially art director) today I overheard him saying we should just turn off specular entirely (we're in a roughness/metalness workflow, so specular should be unadjusted and not hosed with too much, it's default value is 0.5 so roughness/metalness actually WORKS). I also keep making assets in Substance since that's our pipeline now, or so they say/intend, but I keep seeing people taking exported maps out into photoshop and adjusting them in there and then not submitting them to source control or back into the source substance file. wait, why would you export and do adjustments in photoshop when substance pretty much has all the same nodes for that already? that doesnt make any sense?
|
|
#
?
Aug 21, 2019 03:23
|
|
|
Kanine posted:wait, why would you export and do adjustments in photoshop when substance pretty much has all the same nodes for that already? that doesnt make any sense? Because his workplace doesn't have people that can use substance painter and have inneficient workflows to compensate.
|
|
#
?
Aug 21, 2019 03:27
|
|
|
ImplicitAssembler posted:Concepts are fine (and often help), but when you are doing things in motion, they don't always work. To be fair though, the shops you've worked at and the experience you've got is not the norm I've experienced. The vfx guys you work with are probably all very talented and could be cg supervisors in places I've worked at. The norm for the places I've seen is crunch for the grunts and meetings for the DA's while the leads have to crack the whip. It sucks frankly but I've seen it happen often.
|
|
#
?
Aug 21, 2019 03:41
|
|
|
ive always really had issues with "designing" poo poo i think because i never spent enough time with art fundamentals. ive been looking up a lot of stuff online now that im trying to get better at that and this tutorial series by alex senechal is working out pretty well in applying that stuff to 3d hard surface https://www.artstation.com/acms/store/1ad/visual-design-basics-intro-to-design
|
|
#
?
Aug 21, 2019 04:15
|
|
|
ceebee posted:Also, you should not be in charge of vision if you don't think you could execute it yourself or have a clear vision of what you want and how to get there. I fought with that in pre-production on my current project and it was torture. No Hollywood director except for maybe James Cameron really understands the vfx process that well. Heck, I don’t know what kind of notes I’d give to a lighting or Cfx team and they sit down the hall. It’s the client and facility vfx supes job to wrangle the complicated gibberish that comes out of a directors mouth and translate it into a coherent visual, whilst massaging the Hollywoodites egos so the facility gets to bid on the next job. The vfx supe should know what they want though, which is why most of them seem to come from disciplines near the end of the pipeline where they see how it all fits together. Ccs fucked around with this message at 04:32 on Aug 21, 2019 |
|
#
?
Aug 21, 2019 04:26
|
|
|
I guess this is where I (kinda) disagree. The only directors I'd want to work for if I was in VFX would be Guillermo Del Toro and maybe Peter Jackson. I feel like any director worth their salt is willing to get in the trenches with their vision and doesn't have an issue explaining (or executing) where and how they want to take it. Then again I don't work in VFX, I work in Games where most Art Directors should have experience in concepting, character art, environment art, or level design. Ideally a great art director is a great generalist. You know why Sony Santa Monica continually make great looking games? Because they employ really good artists as their art directors or leads. The people who lead visually amazing projects are usually people who have a great understanding of art and execution. mutata posted:I disagree generally with that. Just because an art director/creative director can't draw doesn't necessarily make them a bad AD. In my experience in games the worst art directors are the ones you mentioned. Nobody likes working in vagueness. If they can't draw they should at least be able to photoshop some poo poo together or explain their idea and vision clearly enough for a concept artist to attempt to execute it for the sake of their team. But seriously if you're an art director in games you should have even a basic level of drawing. Odddzy posted:Because his workplace doesn't have people that can use substance painter and have inneficient workflows to compensate. You're mostly right, the workflow is definitely inefficient. But they do know how to use Painter they just refuse to stay in it. ceebee fucked around with this message at 13:25 on Aug 21, 2019 |
|
|
#
?
Aug 21, 2019 12:53
|
|
|
ceebee posted:You're mostly right, the workflow is definitely inefficient. But they do know how to use Painter they just refuse to stay in it. all my experience in the games industry the last 4 years has been working as a contractor for art outsourcing companies. i havent actually worked in an onsite studio environment yet. is it very common for team members to drag their feet on adopting new tools/workflows even when those things would make their jobs a lot more efficient/easier?
|
|
#
?
Aug 21, 2019 14:22
|
|
|
I think that's just a human animal brain thing in general..
|
|
#
?
Aug 21, 2019 14:31
|
|
Kanine posted:is it very common for team members to drag their feet on adopting new tools/workflows even when those things would make their jobs a lot more efficient/easier? Yes, to me it's a symptom of complacency and lazyness. Although to be fair most game studios should set aside some time for their employees to gain new skills and learn new software outside of production. Most people just dont have free time outside of work to dedicate to learning new tools.
|
|
|
#
?
Aug 21, 2019 16:43
|
|
|
It’s also the fact that building a new workflow into a pipeline can create a lot of unintended complications. But it’s also why new studios starting from scratch can be more nimble because they’re creating their pipeline to the needs of today instead of retrofitting something that’s been sitting and rotting for ages.
|
|
#
?
Aug 22, 2019 00:17
|
|
|
while i was in college i worked on and off as a contractor for game art outsourcing places and now im trying to get an actual steady job in the industry. im looking at doing an internship and im pleasantly suprised that it seems like 95% of these are paid
Kanine fucked around with this message at 08:04 on Aug 23, 2019 |
|
#
?
Aug 23, 2019 02:45
|
|
|
I had this coworker who'd stick with her old pipeline when working with things and she'd frequently have trouble exporting artwork and ask me for suggestions on how to get it working. Finally sitting down at her computer I was aghast to find that she was working with a 12 year old version of Lightwave for her 3d work.
|
|
#
?
Aug 23, 2019 06:16
|
|
|
While I get equally annoyed at people who break workflow for entrenched reasons (don't want to learn new tools), I don't think it's bad to use different tools when the existing pipeline can't nail it easily. To use the Painter example, there are reasons an artist might use photoshop alongside. Making masks/alphas/stamps, quickly batch-modifying textures when something like a minor grade or level adjustment is needed, doing more complex handpainted work than Painter can handle live (ie very layer-heavy high res meshes). To avoid the contrarian nature of the above, any outside work should be re-integrated into Painter if possible though. Grading textures after the fact? Write those values down so if the asset is revisited, it can be cloned easily. Handpainted work? Keep on its own layers and bring back into Painter. Designer has replaced a chunk of my PS use though. I used it to generate masks based on handpainted grout lines from Painter, for tinting individual blocks and having control over grout thickness on a massive (100+m tall) hero asset. It was messy to integrate, but leagues quicker than either Painter or PS, and let the supervisor art direct the blocks and grout while still giving quick iteration on broad notes.
|
|
#
?
Aug 23, 2019 16:54
|
|
|
cubicle gangster posted:Finished that project! this came out nice, man! i haven't posted poo poo here in a long time but here's some new sculpture stuff from me     
|
|
#
?
Aug 23, 2019 21:51
|
|
|
Related to the talk about workflows/pipelines/etc. I'm assuming everyone has already seen this GDC talk? https://www.youtube.com/watch?v=072HkIJ_DMs
|
|
#
?
Aug 24, 2019 04:18
|
|
|
Does anyone know how possible it is to return the centre of a face within a shader? I effectively want every pixel on a face be indentical to what the verymost centre pixel would be.. Wondering if this possible to achieve
|
|
#
?
Aug 25, 2019 07:20
|
|
|
echinopsis posted:Does anyone know how possible it is to return the centre of a face within a shader? I effectively want every pixel on a face be indentical to what the verymost centre pixel would be.. Wondering if this possible to achieve I've not done shader programming but am interested in it. Maybe you could bind the normal in the vertex fragment part of the shader to the color of the points of each polygon?
|
|
#
?
Aug 26, 2019 02:30
|
|
|
echinopsis posted:Does anyone know how possible it is to return the centre of a face within a shader? I effectively want every pixel on a face be indentical to what the verymost centre pixel would be.. Wondering if this possible to achieve Can you add attributes to the mesh before you import it into your scene? You could loop through the faces, and for each face get the averaged uv coords of all the connected vertices of that face (which should be the uv coords at the center of the face). Then, take those coordinates and store them as a vertex attribute called something like 'facecenteruv'. That will allow you to sample those uvs at runtime and get the pixel color at the center of the face. You're probably looking for a simple function or something but I'm not sure one exists since a vertex doesn't really have any knowledge of which face it belongs to. I could be wrong though since I only do pretty simple shader stuff myself.
|
|
#
?
Aug 26, 2019 03:03
|
|
|
500 posted:Can you add attributes to the mesh before you import it into your scene? You could loop through the faces, and for each face get the averaged uv coords of all the connected vertices of that face (which should be the uv coords at the center of the face). Then, take those coordinates and store them as a vertex attribute called something like 'facecenteruv'. That will allow you to sample those uvs at runtime and get the pixel color at the center of the face. hmm Odddzy posted:I've not done shader programming but am interested in it. Maybe you could bind the normal in the vertex fragment part of the shader to the color of the points of each polygon? both of you mention vertex but why would a shader ever see a vertex? as opposed to a point on a face? maybe i’m daft
|
|
#
?
Aug 26, 2019 04:10
|
|
|
echinopsis posted:hmm The term "shader" is usually reserved for shader code designed to run on the GPU, that is, vertex and fragment shaders. As opposed to materials which most 3D suites use which represent a whole range of things that are mostly independent from what would normally be considered shaders. Shaders are made up of 2 parts, vertex shaders and fragment shaders. The vertex shader works on each vertex of a shaded geometry, typically to do deformation (when applicable) and the fragment shader works in each visible pixel, typically for coloring and shading. GPUs are massively parallel beasts and so GLSL (shader code) is explicitly designed to work in this context - each chunk of shader code, fragment and vertex, will be running thousands of iterations of different parts of an object simultaneously - as a result they will have essentially no context at all about what's going on with the rest of am object. This is what makes shader work so challenging, and why a seemingly simple thing like doing things to the center of a face can actually be extremely challenging. If in the other hand you mean a material in a specific 3D app the answer is probably a lot simpler and more achievable, but we'll need more context.
|
|
#
?
Aug 26, 2019 04:54
|
|
|
hmmm drat. I was thinking about in blender, idk how much the OSL can access info about the geometry oh well it’s no big deal 🤷♂️
|
|
#
?
Aug 26, 2019 05:37
|
|
|
Yeah sorry, when you said "in a shader" I was thinking you meant inside a shader program. I wrongly assumed you were a developer. If you're curious, a shader program typically looks something like this (see the code on the right and the output on the left). https://www.shadertoy.com/view/ttSSzK With enough complexity, they can generate all sorts of visuals: https://www.shadertoy.com/view/MdXyzX https://www.shadertoy.com/view/Wt2XDm https://www.shadertoy.com/view/tt2XzG Basically, shaders power all the video game graphics you've ever seen. But yeah, if you give more information about the look you want to achieve, there might be another way of doing that.
|
|
#
?
Aug 26, 2019 06:08
|
|
|
500 posted:Yeah sorry, when you said "in a shader" I was thinking you meant inside a shader program. I wrongly assumed you were a developer. nah just a fuckaround I am. I started on https://thebookofshaders.com/ and it really opened my eyes.. but when working in blender, I believe the workflow inside the material (I guess I now realise a shader is just part of the material) would maybe allow me to for example, get an ID of the face of the surface that was hit by the camera ray, then access the geometry of that face, so I could get the centre most point of that face, then generate the colour of the pixel a ray hitting the centre of that face would generate, and apply that across the whole face. I think I could code this, but not inside blender but learn some other image manipulation 🤷♂️
|
|
#
?
Aug 26, 2019 08:47
|
|
|
if you're using a fresnel shader then unsmoothed polygons would return a flat shade. Pure fresnel will just go by the angle of the polygon vs the camera. Hm, if you applied the face normal to that though or mapped a color wheel/gradient to it... you may get something close to what you want. What sort of effect are you after?
|
|
#
?
Aug 26, 2019 08:53
|
|
|
echinopsis posted:nah just a fuckaround I am. I started on https://thebookofshaders.com/ and it really opened my eyes.. but when working in blender, I believe the workflow inside the material (I guess I now realise a shader is just part of the material) would maybe allow me to for example, get an ID of the face of the surface that was hit by the camera ray, then access the geometry of that face, so I could get the centre most point of that face, then generate the colour of the pixel a ray hitting the centre of that face would generate, and apply that across the whole face. You can write GLSL in Blender, but given your description it sounds like you should just use material nodes. There are nodes that'll give you all kinds of info including geometry and normals etc. I can give some better info tomorrow when I'm at a computer. Could probably give some better help if you describe what you're trying to achieve in a little more detail too.
|
|
#
?
Aug 26, 2019 09:20
|
|
|
Synthbuttrange posted:if you're using a fresnel shader then unsmoothed polygons would return a flat shade. Pure fresnel will just go by the angle of the polygon vs the camera. Hm, if you applied the face normal to that though or mapped a color wheel/gradient to it... you may get something close to what you want. What sort of effect are you after? basically, the equivalent of pixelation of what is behind this transparant mesh, but the mesh is a random arrangement of triangles. it'll be an unplanned version of this kind of thing  only there the placement of the triangles are placed, rather than just being a uniform density random bunch of triangles. as in i've taken a heavily subdivided mesh and used the "decimate" modifier on it. then each triangle is one unform colour representing what is behind it
|
|
#
?
Aug 26, 2019 09:24
|
|
|
Synthbuttrange posted:Finally sitting down at her computer I was aghast to find that she was working with a 12 year old version of Lightwave for her 3d work. Ugh, don't get me started on Lightwave. Work just recently forced me to go back to using it and not allowing me to use other DCC apps. My work is probably 90% modeling and it's so primitive compared to the workflow in almost any other program out there. UVing sucks. Texturing sucks. The non-interactive tools, lack of a true 3D gizmo, and OpenGL performance is terrible. Making any complex selection is an exercise in frustration and god forbid if you screw that up and need to start over because the undo system is absolutely the worst. There's been a freeze on technology spending so I'm stuck with it for the foreseeable future. Bleh.
|
|
#
?
Aug 26, 2019 21:27
|
|
|
autojive posted:Ugh, don't get me started on Lightwave. Work just recently forced me to go back to using it and not allowing me to use other DCC apps. So since you put in your 2 weeks when they told you that you'll be out soon, right? Right????
|
|
#
?
Aug 26, 2019 21:44
|
|
|
I just assumed Lightwave was dead by now. Jeez. I started polygonal 3d in ligtwave as a 16 year old, heh. Edit: Tell them Blender is free.
|
|
#
?
Aug 27, 2019 00:25
|
|
|
Lightwave suuuuucks these days. Just convince them to go to Blender, it's free.
|
|
|
#
?
Aug 27, 2019 00:50
|
|
|
Taffer posted:So since you put in your 2 weeks when they told you that you'll be out soon, right? The tech spending freeze essentially ended all subscriptions to the software we rented in our department (except for Creative Cloud) causing everyone to fall back on the only permanent licenses we still have ... Lightwave 2015. We were actually in the middle of finalizing a department-wide transition to C4D with V-Ray so that we could work more easily with the multimedia team but then the freeze came and it was abrupt. It sucks and my manager understands our frustrations and shares them, too, but these are orders from the bean counters and C-level execs to help make numbers for this year. I don't think I'll be leaving any time soon. The company is actually a nice place to work. They really do treat you well and the benefits offered are great so I'll just tough it out until we can get back on to our roadmap. mutata posted:I just assumed Lightwave was dead by now. Jeez. I started polygonal 3d in ligtwave as a 16 year old, heh. They might as well be dead. They are so far behind all the other DCC apps and with the developer resources they have, I don't think they ever will. Newtek was recently bought out by a competitor for their Tricaster product (their real bread-winner) so I won't be surprised if it's discontinued within the next few years. ceebee posted:Lightwave suuuuucks these days. Just convince them to go to Blender, it's free. Tried. Too much resistance from coworkers so it'll never work. :-\
|
|
#
?
Aug 27, 2019 15:10
|
|
|
|
| # ? Jun 2, 2024 10:04 |
|
|
autojive posted:Tried. Too much resistance from coworkers so it'll never work. :-\ I don't think you'll ever convince anybody to switch unless you actually do it first and show them how advanced it is. I'm totally talking out my rear end here because I use c4d, but my 2nd hand understanding of Blender is that it has a really unintuitive interface just for the sake of it. That doesn't seem to be true anymore though, with 2.8 changing things up, watching the one guy at Corridor Crew handle decent shots with it pretty quickly, and seeing those 1 minute tutorials that they have particles, physics, liquids, everything you'd really want. I am REALLY jealous of people who already have a bit of a foothold with Blender, I should probably just start learning it with a little character to model and rig.
|
|
#
?
Aug 27, 2019 15:59
|
|
