Jump to content

Cerbera

Community Staff
  • Posts

    17,954
  • Joined

  • Days Won

    733

Everything posted by Cerbera

  1. Another hilarious (extended) episode of the much appreciated Rollin' Wild Series, notable for the humour AND the immaculate 3D work / dynamics... So enjoy that... CBR
  2. Yeah, pretty damn good, but concur with Vector about skin too shiny, and phone and cables are suspiciously scratch-free... CBR
  3. Yeah, I run out of enthusiasm sometimes long before finishing some stuff, but it's rare that I don't get back to things eventually. But it does happen. I started modelling Hogwarts a few years back, and got defeated by the sheer volume of work it was gonna take to complete. I still give myself gargantuan tasks, but have learned over time that I stand more chance of finishing them if they are slightly smaller scale - you know; build a Tie Fighter instead of an Executor class Star Destroyer... that seems to work for me. CBR
  4. Cerbera

    Frist base mesh

    Your ear mesh is very nice, and the face is going the same way if you continue to be careful about where your loops run. Here's the ideal facial loops for reference. Drawing-wise, I don't think it matters if you can't properly draw, if a) you get your reference material / character proportions from somewhere else, or b) just do the most basic of sketches to give you a general guide for modelling or c) you can model new characters without any reference at all. So while there's no doubt that being able to draw is a definite advantage at the design stage, it's not necessary IMO... It's also worth thinking about using sculpting to rough out forms very quickly, which can act as a sort of 3D sketch pad with helpers like steady stroke and symmetry or using simple Doodle strokes to plan your edge loops when poly modelling... CBR
  5. Cerbera

    Lightsaber

    No, we can't criticize the triangles because it's a game asset ;) Except if those triangles are not helping reduce the polycount ! So what started off as a nice almost all-quad mesh seems to have been unnecessarily triangulated for no reason I can immediately see - surely this has just doubled the amount of polygons hasn't it, which is very much not the point of using triangles in game assets, surely ?! But despite all the tri's, the render seems good, although the damage looks weird because of its unevenness (none on some parts, loads on others), and because of where it appears, which is not necessarily where it would do so in real life. Exposed edge damage is good for example, but don't also have it in the deep grooves, because those grooves actually protect their inner surfaces from external damage, so should remain largely clean and scratch-free, although they will still contain some buildup of a different kind of 'gunk' of the sort that usually collects in crevices and corners of devices that are in regular human contact :) CBR
  6. I have a feeling you won't like my answers, but honest is always better, so here they are... 1. Yes, 4K is generally enough, except for 360 degree HDRI's which could do with being 8 or even 16K, if they are to be used as backdrops. Blurrier versions used for GI can be smaller. 2. Almost never. Most of my textures come from original photos, or downloaded photos online, or noise based stuff I tend to just make myself. When I do buy textures, I tend to buy very high quality photo-based images from a company like Arroway if I haven't got the time or the locations to go get them myself. As an example of just how often that is, I have bought 2 texture packs in the last 7 years, both of those for specific projects, and only then because I was really struggling to find the quite specific type of images I needed in an online search. But 90% of the time I can find everything I need with a just a quick look on Google. The sheer amount of free reasonable res textures online means that if you are selling a texture pack your textures need to be pretty damn special and quite reasonably priced. Your samples do look good, but are they better than what I could make given a 5 minute web search, and 3 minutes with crazybump for example, which would make all my maps for me from one original photo ? 3. Displacement, every time, because it holds up from all angles in a way that parallax doesn't. Also, you'd be mad to make textures that only worked with R18. To be honest, Bump I would be expecting to be dealing with smaller details only, so no parallax required. 4. I'd want scale variation only - ie a patch of grass 1m x 1m, then 5m x 5m, then 50m x 50m for example... 5. Depends entirely on the project. In general I would say I have more call for 'distant' textures than close-up ones. 6. I would like to see more textures that have their dimensions in the title, so that users know what physical area the texture is designed for. Lastly I would emphasise the importance of tiling and seamlessness - that is perhaps the mostly useful thing about any texture. Hope that's more helpful than it is discouraging... CBR
  7. Thanks Sprial, it's getting there :) Yep, that was only first rough texture pass, and grass is a long way from finished - eventually I'll get those stripes into the front lawn by alternately brushing the mow lines... I think you'd be struggling unless you brought your own trebuchet. And the caretaker will be furious. CBR
  8. If you want best results at the cost of some manual work, then HB modelling bundle has an excellent retopo setup, but I am not aware of any plugin that can just fix this stuff automatically. CBR
  9. ***Gold Star*** :) Great work from both of you... CBR
  10. Refined my base mesh for some facial detail, and before I head in to sculpt, couldn't resist a little GI test render (IC/QMC, PR, Adaptive / Automatic / 12%). For a 2 minute render time I'm quite pleased with that...
  11. Oh I really like that. If your area of expertise is modelling then I reckon it's fair and reasonable to have modelled every damn thing in your showreel, but if you're more of a generalist, or specialise in some other area altogether I guess it's not so important. And it's not like you haven't a put a whole ton of work into the scene... And you had fun while you were doing it, and that's the important thing, hey ? CBR
  12. Yeah, you're spot-on there. I'm very conscious that I can't add too much detail without making it look newer, but stonemasons of the time would have done a better job on face and mane details, so will improve those, but not too much, as weathering will undoubtedly have damaged / worn away a lot of it. We're also dealing with what people in the 18th century thought lions looked like when they were mainly working from hand drawn pictures, and hadn't necessarily ever seen one; hence the slightly romanticised proportions and incorrect number of paw pads / toes, a surprisingly common mistake at the time ! :)
  13. Cerbera

    RFQ for a model

    PM sent... CBR
  14. ...and now with better textures... definitely needs some sculpt detail though, so that's what I'm doing next... CBR
  15. There are lot of lions dotted around Harlaxton, mostly astride various column heads, or reclining in alcoves around the building. Initially I thought I might leave them out, but it seemed a shame, and I had a spare weekend, so knocked up a quick base mesh to see how far I'd get. My first lion is one of the more content ones (others look considerably more cross, and some are actually attacking others) that will sit astride the right side pavilion column stacks. Once I'd built the base mesh I thought I'd just throw a quick cubic texture on to see if there was any chance I could get away with not UVing and sculpting it, but I can't, so I will :) But here's where I got to so far... CBR
  16. Yeah, it got me many a time before i discovered that the temporary fix is that Restrict to Selection does work, and completely eliminates the cut visible thing for the fairly minimal effort of selecting your polys before you cut... CBR
  17. I think you already can, my friend. I am fairly damn sure you can hold a key, and then temporarily select through the axis. Now if only I could remember what that key was, or find a single reference to it online, which I very much can't. But I am inexplicably sure there is one ! Come on cafe, put us out of our misery ;) I seem to remember I may have learned that from @3DKiwi a very long time ago, so maybe he's a good person to ask... CBR
  18. Well I just paid for it, without knowing what was in it, so discussions are fine with me... I REALLY hope it doesn't require windows 10, for example. And has a render pause button, like I've been asking for every year, for the last 7 years :) And most of all, I hope they made OpenSubdiv work with UVs, like R18 never managed, and that we at long last get a bulletproof knife toolset with zero problems (Cut Visible Only, I'm looking at you :) Oh, and some sexy new stuff as well, obviously ;) CBR
  19. The modelling looks OK (from what I can tell without wires), and the textures, especially the clothes are pretty much OK. However you could improve the clothing using the Irwan cloth reflectance models, at the expense of a lot of render time. Skin texture looks a bit weird, and makeup a bit overdone. Too shiny., so less specular / reflectance, and more roughness required. You could try adding some Subsurface scattering, and some additional texture to that, it's flat as well as shiny ! Hair isn't very well styled at the moment, some additional time doing that would also help. Then you could fix the lighting, which, whilst not bad exactly, is a little 'unremarkable' at the moment. Hope some / any of that helps CBR
  20. Yes they can, but annoyingly only in the most expensive version.
  21. The important thing to remember is that neither you, me, or the gardners mention ANYTHING about the top secret control bunker we've got under the main house.
  22. GI Vector, have you not noticed what the latin text on the balcony says ? We can't have slain Zombies on the front lawn of our IQP Terra Quadrant headquarters, can we ? What am I paying our gardening division for if they can't even keep our lawns corpse-free ? CBR
  23. That's a good point there, Lubo, and I did give it some careful thought before I started - of course I am aware the standard way to make arch viz buildings usually doesn't involve subdivision at all, but I felt that due to the age of the building, and the sheer amount of curvy bits and carved detail, and given the subtle point randomisation I was planning to put into it finally, that SDS would best give me segmentation everywhere I needed it. Got to go back to the client stuff for a while now, but I will be back to continue this, where I'll add that randomisation to both geometry and textures, hopefully losing something of the square look you mentioned, and then to finish it off with a healthy dose of flower borders, foliage, and possibly even some X-Particles ivy ! :) J
×
×
  • Create New...