Pretty much done with the reworked floor panel. I'll probably end up making another design to alternate with this one, since I'd wager that the repetition of the circular element will look a bit much in the actual scene.
Looking good. A LOT of people try to pull this stuff off and just water down the idea of steampunk until he becomes code for slopy mismatched pieces when it can still be a great style.
The Art Deco flourishes bring a lot to this, push more of that, the clearest example is that new Nemo grill this image.
Anyway, good luck with this one, the building style seems pretty authentic so far.
Looking good. A LOT of people try to pull this stuff off and just water down the idea of steampunk until he becomes code for slopy mismatched pieces when it can still be a great style.
The Art Deco flourishes bring a lot to this, push more of that, the clearest example is that new Nemo grill this image.
Anyway, good luck with this one, the building style seems pretty authentic so far
Thanks, Kevin. One of the main reasons I've decided to revisit certain assets is to make sure that everything 'belongs', and to avoid that mismatched feel that you mention. It's been a fantastic project to research, and I've stayed as true to authentic art deco as possible.
It's certainly been a challenge, integrating art deco elements with more industrial features. One luxury that Bioshock's environments had in this regard was the fact that Rapture is a city, whereas obviously the Nautilus is a vessel, and therefore has to reflect that both in terms of design and materials.
I think I'm getting there, though! Thanks again for everyone's feedback and encouragement.
The thing I'm struggling with in terms of these pieces is achieving a balance of visual interest versus the reality of the surface properties. I didn't want to overdo the scratches or surface damage, but I wanted it to be a little more interesting than simply idiot-smooth. In terms of general color palette, I'm trying to match this reference from the original Disney movie:
Are you using the left wall as reference? I think its catching some light from around the corner. Either that or they tried to give the illusion of distance by painting things farther off, lighter colors, kind of a fake fog or "baked in lighting". I think you should use the metal closer to the camera on the right as a color guide. That's how I remember most of the ship.
Copper, bronze, brass, a little rusty brown but not corroded. In that pic it even looks like they tried to make it look like iron. http://landandseacollection.com/id52.html
A lot of metal won't come alive in just the diffuse until you add spec. In the spec highlight edges of things especially if there isn't any geometry to define the detail.
I was using the walls on the right for a general color guide, as that seems to be the dominant color of the ship's interior, but I found the disparity between what I saw in Photoshop and what it looked like in UDK made adjusting the diffuse pretty tricky.
I got some funky light errors when I tried using a simple point light to get a better idea of the overall hues, so I figured I'd try and add some surface details first before moving on to the dominant overall color. Probably not the best idea, but I wanted to at least make a tentative start on the texturing.
This is what I mean. No parameters have been altered on the point light.
I'm guessing this is a shader issue (just diffuse and normal right now), but from the little part you can see of the surface as it is right now, color-wise it's not too far away from the reference. I didn't want the scene to be overly saturated, and I was going to make a separate thread in Technical Talk for the lighting issue before making any major adjustments to the diffuse color.
Thanks for the feedback, though, and any suggestions.
Humm weird. I'm not a UDK expert (you might want to post this problem in the UDK thread) but to me it looks like your normal map is having some trouble. What happens when you unplug it and use only the diffuse?
Yeah, I've got a thread going over there about this - definitive solutions are a little thin on the ground right now, though.
I used auto-group for the smoothing groups on the mesh in Max, then re-exported / imported it into UDK, which helped a little. I tried unplugging the normal map, and this is how it looks right now.
I know you wouldn't light a scene just using point lights, but even though it's an improvement over the last attempt, there's still some very evident banding between the wall meshes, even though they and their material setups are identical. Yes, it won't be as evident with a proper light setup and other meshes to break up the contiguous wall surfaces, but it shouldn't be happening, and I'm not sure what's causing it. :poly127:
Anyways, still got plenty to do aside from wrestle with UDK. If I focus on these little problems too much, it'll end up killing my enthusiasm and I've still got some models to rework and a crap-ton of texturing to get on with. Think I'll use a viewport shader until I can gather my patience to really get to the bottom of this.
Just had a look through and think it's a great concept and looking very impressive, I'm a fan of Bioshock so I think I know what you're looking for. I look forward to see how it develops!...
I'm guessing the polycount is high for this to be a gaming environment, is the idea to make normals for all the assets?.. I'm currently working on my very first CG project and I wasn't quite sure whether to model everything without regard for the polys, so interested to know what your workflow is.
Just one more query, can you tell me what is the process you use for presenting your assets?
Hey Gooner, sorry for the delayed reply - missed your post.
Right now, I've only presented the high-poly models. The idea is to model the high-resolution models first, then create the low-poly, in-game meshes based on the high-res models and use normal maps to capture a lot of the intricate details. There's some great info in a huge thread on optimization in Technical Talk that you should read.
This technique is just one of several workflows that people use - some choose to model the in-game mesh first, to see how the models work in terms of silhouette and general efficiency with regard to poly counts and create the high-resolution mesh from that, whereas others choose a method somewhere in the middle, creating the low and high-res models from a medium-res mesh.
There's no definitive, right or wrong way to approach it, it varies on how you want to work and what you find most productive. It definitely helps to have an idea of how you're going to approach the scene before you start working on it, though, but this will come with practice.
Personally, I prefer to get the high-res modeling done first, then optimize - and yeah, I use as much geometry as needed to get it looking good, but I try to bear in mind how certain features will work best for the low-poly model and plan accordingly. As EarthQuake says, you're not going to win any prizes for making an 'optimized' high-resolution mesh.
As for my presentation, I use the material from this tutorial with the specified settings and a simple gradient map.
Just a quick update today - my machine has been complaining, so I'm guessing I'm probably due to blow another PSU sometime soon. Reworked the support struts for the corridor to suit the scene a little better:
So aside from finishing off a couple of models here and there, I'm going to start texturing full-time in the next day or so, as well as revisiting some of the issues I was experiencing with lighting the scene in UDK.
As always, thanks in advance for any feedback or suggestions.
Hi Danshewan, thanks for your response and the links!..
At which stage are you unwrapping UV's?, presumably after you make low polys of your high polys?... and then from that baking normal mapping onto the low polys, okay I'm just thinking it through, so many stages but I'm getting my head around it. I'll have a good read of that first link, should clear things up.
I can't see the update that you have, did you post an image?
Yeah, you unwrap and layout your UV shells once the low-poly mesh has been created. To do this, I typically begin with the high-poly mesh and start removing the support edges, removing unseen or unnecessary faces and edges and simplifying the mesh as much as possible while retaining the overall forms and silhouette.
Again, it's important to be aware of optimization issues at this stage, as these days texture resolution is considered as important, if not more so, than polygonal optimization when it comes to memory efficiency. This becomes really evident on curved surfaces - sometimes you can definitely go too far with optimizing to the point that the low-poly suffers, and can also create problems when baking out your normal maps.
Once the UV's are laid out, I use xNormal to bake the normal maps. xNormal is a free and very powerful little application that handles baking out normal and ambient occlusion maps (amongst other things) that I find much less hassle than using Max's render to texture system. You can grab it here:
Another small update before moving on to some serious unwrapping and texturing.
I'd originally modeled this some time ago, but it looked too much like a pull-down projector screen, even with some reasonably ornate metalwork. It's a wall-mounted map chart showing the location of Nemo's base in Vulcania that'll be situated on the wall behind Nemo's desk, to the left of the ship's power core-type thing.
The majority of the texture will be a large, old map showing the seas around south-east Asia, where Vulcania is supposed to be located. The end of the arm will be centered on the exact location of the island base.
There's (literally) a couple of other bits I have to tidy up, but it's mainly smaller pieces like a ship radio-type communications unit, a depth gauge and a couple other bits and pieces, but this was the last piece I wanted to completely re-do.
I promise, the next update will involve textures in some way.
Hey danshewan, this is looking really awesome, and i see that you are using UDK. Take a look at this thread. It may help you out with your workflow when you set up all your materials.
Thanks for the kind words, Ryan. I'll be sure to check out the master material tutorial as soon as I've got some texturing done. Very generous of you to provide that free of charge - bookmarked!
Thanks for the advice danshewan, I'm actually working on my first personal project now which is an environment. I've been meaning to put it here on this forum for feedback, I can then ask all my noob queries on there and not hijack your thread!:\
I know of xnormal, haven't got so far as to need it yet, I've also heard of crazybump. Also for unwrapping I've used in the past Headus UV layout, are you using something to help with unwrapping or just the Max tools?.. I've found Max a bit crap for unwrapping compared to Maya but maybe I just haven't had enough experience of it yet.
For your images, I can't actually see any from anyone on this computer, I'll wait till I get home and have a look at what you've done.
I'm using Max's UV layout tools, with renderhjs' TexTools plugin. I've used UV Layout Pro before, and I do like the keyboard shortcut method of cutting and stitiching, but constantly exporting and importing meshes just got to be too much of a hassle.
You should definitely start posting your work. I checked out your 2D work, and it's really impressive. I'll look forward to hearing your thoughts on my work when you can actually see it!
thanks for the link, I'm doing unwrapping now so I need it. I'm doing a course in Max at the moment and unable to get any plugins working for it here, but this one installed!..cheers!
Thanks for the compliment, my 2D stuff is vastly different to what I want to do in CG, somehow environments intrigue me more in CG whereas wildlife/portraits interest me more in traditional art
I have seen your work before, just not seen the recent updates till now. I think it's looking like a cracking piece of work, if your texturing can match the concept and your modelling it'll be very impressive. I think you mentioned texturing was what you weren't so hot on, will be great practice though to get working on it, I presume you have a wacom?
Just out of interest, are you working in CG, or a student?
Yeah, I do use a tablet when texturing, and part of the reason I chose such an intricate, prop-heavy environment was to really get some solid practice in texturing. As I've said before, it's pretty obvious I'm much more comfortable modeling than texturing, and the scary (and potentially frustrating) thing is the weight of my own expectations in terms of nailing the texturing and doing the models justice.
As for me, I'm still very much a student, entirely self-taught.
Thanks again for everyone's feedback and encouragement.
Long time, no updates, long story, won't bother. Suffice to say that using a desktop with the computing power of a low-end laptop and getting ready to move countries isn't helping matters.
A simple lighting test.
Main reason I posted this was to motivate myself. Felt like I was getting more done in the earlier stages of the project. I'll hold off posting images of the actual office area until some large-scale changes have been completed, namely the majority of the furniture and wall decorations. As interesting as it was to explore an entirely art deco theme, it just wasn't looking like the Nautilus I'd imagined and was straying a little too close to Bioshock's aesthetics.
Hopefully updates will be a little more frequent from now until the completion. I've learned a fucking ton so far from this project, but it's dragged on far too long. A lot of assets had to (and still have to) be remodeled - some things seemed like they'd be a good fit initially, but needed significant reworking to fit the feel of the scene. For example, the support struts in the corridor are actually the fifth iteration of the design.
Still a lot of work to do, but I feel like the end is almost in sight.
Looking good, I like the ideas and pools of light you have going so far. I think it needs more ambient light. It doesn't look like the porthole light is doing much?
Hey man havent seen this project in a while but Im glad your still on it. I love the atmosphere and the style of the framework you have in the recent shot. You could check on use emissive for static lighting for those portholes to get that green light to bounce in that corridor. Good stuff.
Yeah, the scene does need a little more ambient light, and some meshes to motivate them which I'll be adding shortly, along with a whole mess of pipes to break up some of the wall surfaces. Then I'll be moving on to the actual office area, which needs some serious work.
I'm planning on adding some subtle volumetric lights to simulate the greenish light from the ocean coming through the portholes and the larger observation windows in the office itself, which should offer some more illumination and compliment the overall color scheme.
I like your last shot, it looks good. Though there are some strange normalsmap errors alongside the edges of your arch mesh. and those circles on your floor i think can be pretty repetitive when put after each other like that. skip one in between or make it less obvious. but that’s just my opinion
Keep up the nice work. Looking forward to see it finished.
Not sure if it's the normal maps need some extra lovin' or whether I need to look at the lightmap UV layout, but I'll get to it either way. I suspect it's the lightmap.
Moving on to the staircase's handrail today - had to abandon the previous design as it just wasn't practical to optimize it. After that, moving on to remodeling the furniture, primarily the desk, then texturing Nemo's diving helmet and underwater rifle. Almost all textures are placeholder - just trying to get the overall scene to more of a finished state before tidying everything up.
I am really liking the stylization of everything. I would push the normals a little bit more as you mentioned, maybe with a Bump offset? Any way keep up the good work!
Added the slightly-redesigned handrail. Again, all textures (and some geometry) are placeholder.
Still have to fix the uniformity of the wall panel textures and the metal part of the staircase, and replace the floor panel adjacent to the first step. Will also be adding bolts to the curving corridor support struts.
Still needs a lot of work - not sure about the paintings right now, or at least how they're displayed. I don't want it to look too much like a museum, but I do like the pools of light. Also still debating the position and design of the desk. Working on a suitably-ornate trim to sit beneath the light over the stairs, bridging the transition between surfaces of the archway.
The large 'rib' supports will likely be reworked - right now, they're just not interesting enough. I'll be adding a dumbwaiter and a standard lamp to the left of the stairs, and am debating whether a rug might help break up the empty floor space (which also needs to be less obviously tiled, and the hues shifted towards a darker, more cherry wood). Also deciding between wall-mounted shelving or actual bookcases.
Some small updates. Lighting and textures are still very much placeholder.
I'll likely be reworking the 'rib' supports to be a little more interesting (and the size of the bolts), as well as adding a lot more clutter to the desk. The chair needs some more work, too, as does the desk. I'm also not happy with the base metal texture - it's just not reading well, and needs to be cleaner.
One thing I'm learning a lot about as I go is scale, and how to model objects to seem of a realistic scale but large enough to read well, which some objects (including the drinks cabinet and pressure gauge to the right of the map) clearly don't at the moment.
It's coming together slowly, but it's starting to feel like a fucking marathon. I've still got a lot to do in terms of modeling, and I haven't even started texturing, something I fear my inexperience will delay completion even longer.
Even though it's my first environment, I have to admit to struggling with motivation at times. Considering how long I've been working on this, I thought I'd be happier with where it's at. I keep telling myself that I'll get faster with time and practice, and that I won't get better by rushing, but right now it feels like the whole scene is kinda crappy.
And I thought this scene would be a 'smaller' project. :poly127:
Anyways, better get some sleep. I'm sure I'll feel less emo in the morning. As always, thanks in advance for any crits, suggestions or insights.
I think that there are some people who do small projects and learn a lot. There are also people who do bigger projects and learn a lot. The motivation factor is key (in my opinion)
I personally think that going with larger projects is a little more commendable, and this one looks great. I just read the whole thread. One thing you might try doing is just blocking out the level first. You seem to model everything with a lot of detail, and I'm not sure if half the models are still in the scene. Even if it's just boxes, at least you can get a sense of "hey is this possibly going to work here?" Plus it might save you some hours modeling. (which by judging your models you've already had plenty of practice)
Also, go back to the original image. I see you've deviated from it a little bit, but you may be able to find some motivation by just staring at that pic for a while.
I'm looking forward to seeing the outcome of this. Don't give up man, the last 10% is always the hardest, but worth its weight.
I recently stumbled upon some Disney blueprints of their 20K Leagues Nautilus and was going to start modeling it. After seeing this thread, I think I better stick to my day job...
Incredible detail work! Cant wait to see it finished.
Nice going Dan, missed this one. Looking really promising. Whenever you run out of steam, focus on making the fun stuff first - don't force yourself to plow through the more boring parts. Something exciting usually comes out well and having it completed serves as a good boost to move forward. Then if you're really stuck, take a good break and come back to it with a fresh pair of eyes and compare it to your original concept and focus on what's most important from there.
Thanks, guys - I knew I'd feel better after a good night's sleep and a cup of coffee.
I think my biggest problem is that I'm judging the overall quality of the scene based on placeholder art. There's quite a lot that still needs to be added before it will even start to look finished, and I guess I'm pretty anxious to start improving the quality of individual assets to raise the quality of the overall scene before the scene is properly populated.
I'll keep plugging away at getting all the assets into the scene, then focus on polishing. I'll post Marmoset renders of completed assets as and when they're done so I can hopefully get some feedback on my texturing.
Thanks again for everyone's encouragement, and apologies for being such a whiny bitch last night.
Just found this thread today and can't believe I haven't seen it before. Lots of really inspiring work in here, especially your dedication to making sure all your assets look like they belong together. Can't wait to see more progress.
Thanks, guys - I knew I'd feel better after a good night's sleep and a cup of coffee.
I think my biggest problem is that I'm judging the overall quality of the scene based on placeholder art. There's quite a lot that still needs to be added before it will even start to look finished, and I guess I'm pretty anxious to start improving the quality of individual assets to raise the quality of the overall scene before the scene is properly populated.
I'll keep plugging away at getting all the assets into the scene, then focus on polishing. I'll post Marmoset renders of completed assets as and when they're done so I can hopefully get some feedback on my texturing.
Thanks again for everyone's encouragement, and apologies for being such a whiny bitch last night.
Best way to approach this, really. The quality of the scene has definitely gone done, BUT you'll kick back and make this a beauty.
Glad to see this project is still in the works, been a while since Ive seen anything. The thing I like most about this project is the style and tone of it. Some notes on what I see(I know textures are just placeholder).
-I'm digging the design of the back wall with the map.
-Wood floor definitely needs some breakup, I feel like its somewhat out of place with the style, perhaps the color adjustment and breakup might change it.
-The ribbed supports on the lower level corridor are well designed, you could try to incorporate some of that design in the office
I think I'm going to ditch the trim around the main map, and instead focus on making the rib supports more interesting, and simply transition from a rib support directly into the map wall. It seemed like a good idea, but it just doesn't look right.
I also agree with the comment about the floor - right now, it's literally just a cropped stock texture applied to a plane to give me an idea of scale. I'm thinking much darker wood, and maybe slightly larger boards, but we'll see.
I think I've got three more main focal props to add (the helmet and stand, the mounted underwater rifle and a large lever to sit adjacent to the desk), a few general props (books, bottles and some papers for the desk), and a couple of structural pieces to add / cleanup, then I should be ready to start polishing.
As I said earlier, I think the reason I'm so gung-ho to start polishing assets before the scene is properly populated is that I know my texturing work will probably require the most critique in order to improve, and I guess I'm a little concerned because the textures are going to make or break this scene. I'm determined not to hide sub-par work behind masses of post processing, even though I do intend to do quite a bit of post.
Anyways, thanks again for everyone's feedback and encouragement. I'm aiming to get this scene finished before I leave the country in a few weeks, so I'm definitely on a serious deadline.
Small update. Thought I'd switch gears and focus on an individual asset for a while to give me a break from blocking in and unwrapping.
Still got a lot to do. Been tweaking the maps endlessly, trying to achieve the right look for the brass. Still far too much like gold for my liking right now, though I'm trying to hit the sweet spot between pristine reconditioned antique and recovered-from-the-bottom-of-the-ocean look that my reference images span.
Once I've got the metals looking right, I'll be moving on to the glass and rubber piping elements, before adding some wear.
Still playing with the spec, and I'll probably be going back and reworking the size of the grain on the large front panel, as well as playing with the normal intensity of the sculpted elements. Oh, and adding some books.
Looking good great progress! The modeling is great, I love the lamp
The texture on the helmet needs some work. I think it needs some green and white oxidation from being submerged in salty sea water, and maybe some dents and dings to show some use. Probably not as extreme as I took it when I created an antique helmet a few years ago, but probably something between the two extremes?
You might want to flip the grain on the doors of your cabinet, it lines up with the grain that's running on the rest of the cabinet which counter to the strength of the wood and odd that it would be cut from the same piece. looks like you flipped the shelves and the drawer but missed this?
Cheers Vig, exactly the kind of feedback I was hoping for.
The helmet's wear is proving the most tricky, but you're right, definitely needs a bit of beating up. I'm thinking the mallet brush at a low intensity in ZBrush might be a good fit for that. I wasn't sure about oxidizing the surface, as I read that this can take a long time of constant exposure, but I think it needs some. I really like the blend of colors you've got happening on yours, very nice.
The only asset that's really 'finished' is the lamp, so definitely going to be revisiting them, as well as chipping away at some of the other assets along the way. I'm finding I'm getting more done and sustaining my interest by tinkering and refining a variety of assets right now. Still got a buttload to to, though....
Thanks for the tip about the wood grain - I guess I thought it would be more logical to create the main body that way, but I'll switch the flow of the grain and tweak the intensity of the normals before posting again.
Finally got back to working on this stuff after a lengthy hiatus. Pretty much done with this one, I think, barring any punishing crits that need sorting. Been working on this asset for a bit too long overall, so moving on for now.
Thought it'd be best to present the various assets individually to begin with, before working on the larger scene overall. The underwater rifle is next on the list.
Gosh! I am a really big fan of your bronze and overall metal texture, ie: your walls. Anyway you can grace us with some texture sheets? Really looking foreward to you finishing this.
Well well your back..I almost forgot about this thread happy to see it popped back up. I know you been working on it for a while but it has alot of potential the environment overall. The texture on the helmet doesn't look that bad, it does look a bit low res for 1024 and the seam at the back is a bit noticeable. You should post some wires of the helmet and how many sides are those bolts?
This is awesome. Helmet is looking good too, I like the oxidation on the copper especially. I would expect a hell of a lot more water damage on the other parts though. Especially considering this is something made of metal and goes underwater. Maybe even some seaweed hanging, or a starfish stuck on some area.
The helmet looks a lot better I'm really digging lighter colours and oxidation.
The bake looks slightly messy though; there is some nasty colour bleeding (or is that supposed to be dirt?) at the edges of the helmet. It makes it look like you don't have enough padding.
Regarding Zbrush; I wouldn't bother, I believe it might be quicker to just paint a heightmap in PS and bring it into Crazybump, or use nDo.
Keep on cracking mate This scene keeps getting better and better :thumbup:
Replies
Pretty much done with the reworked floor panel. I'll probably end up making another design to alternate with this one, since I'd wager that the repetition of the circular element will look a bit much in the actual scene.
Back to work....
Can tell a lot of thought is being put into the modular pieces, and overall layout.
keep up man, good job.
The Art Deco flourishes bring a lot to this, push more of that, the clearest example is that new Nemo grill this image.
Anyway, good luck with this one, the building style seems pretty authentic so far.
Thanks, Kevin. One of the main reasons I've decided to revisit certain assets is to make sure that everything 'belongs', and to avoid that mismatched feel that you mention. It's been a fantastic project to research, and I've stayed as true to authentic art deco as possible.
It's certainly been a challenge, integrating art deco elements with more industrial features. One luxury that Bioshock's environments had in this regard was the fact that Rapture is a city, whereas obviously the Nautilus is a vessel, and therefore has to reflect that both in terms of design and materials.
I think I'm getting there, though! Thanks again for everyone's feedback and encouragement.
Copper, bronze, brass, a little rusty brown but not corroded. In that pic it even looks like they tried to make it look like iron.
http://landandseacollection.com/id52.html
Anywhere you have copper or brass and hints of corrosion you can add tinges of green.
http://www.copper.org/applications/rodbar/images/ct-fig7a-large.jpg
A lot of metal won't come alive in just the diffuse until you add spec. In the spec highlight edges of things especially if there isn't any geometry to define the detail.
I was using the walls on the right for a general color guide, as that seems to be the dominant color of the ship's interior, but I found the disparity between what I saw in Photoshop and what it looked like in UDK made adjusting the diffuse pretty tricky.
I got some funky light errors when I tried using a simple point light to get a better idea of the overall hues, so I figured I'd try and add some surface details first before moving on to the dominant overall color. Probably not the best idea, but I wanted to at least make a tentative start on the texturing.
This is what I mean. No parameters have been altered on the point light.
I'm guessing this is a shader issue (just diffuse and normal right now), but from the little part you can see of the surface as it is right now, color-wise it's not too far away from the reference. I didn't want the scene to be overly saturated, and I was going to make a separate thread in Technical Talk for the lighting issue before making any major adjustments to the diffuse color.
Thanks for the feedback, though, and any suggestions.
I used auto-group for the smoothing groups on the mesh in Max, then re-exported / imported it into UDK, which helped a little. I tried unplugging the normal map, and this is how it looks right now.
I know you wouldn't light a scene just using point lights, but even though it's an improvement over the last attempt, there's still some very evident banding between the wall meshes, even though they and their material setups are identical. Yes, it won't be as evident with a proper light setup and other meshes to break up the contiguous wall surfaces, but it shouldn't be happening, and I'm not sure what's causing it. :poly127:
Anyways, still got plenty to do aside from wrestle with UDK. If I focus on these little problems too much, it'll end up killing my enthusiasm and I've still got some models to rework and a crap-ton of texturing to get on with. Think I'll use a viewport shader until I can gather my patience to really get to the bottom of this.
Thanks for the suggestion, though.
I'm guessing the polycount is high for this to be a gaming environment, is the idea to make normals for all the assets?.. I'm currently working on my very first CG project and I wasn't quite sure whether to model everything without regard for the polys, so interested to know what your workflow is.
Just one more query, can you tell me what is the process you use for presenting your assets?
Apologies for the 'noob' questions, but I am one!
Right now, I've only presented the high-poly models. The idea is to model the high-resolution models first, then create the low-poly, in-game meshes based on the high-res models and use normal maps to capture a lot of the intricate details. There's some great info in a huge thread on optimization in Technical Talk that you should read.
http://www.polycount.com/forum/showthread.php?t=50588&highlight=optimization+triangle+count
This technique is just one of several workflows that people use - some choose to model the in-game mesh first, to see how the models work in terms of silhouette and general efficiency with regard to poly counts and create the high-resolution mesh from that, whereas others choose a method somewhere in the middle, creating the low and high-res models from a medium-res mesh.
There's no definitive, right or wrong way to approach it, it varies on how you want to work and what you find most productive. It definitely helps to have an idea of how you're going to approach the scene before you start working on it, though, but this will come with practice.
Personally, I prefer to get the high-res modeling done first, then optimize - and yeah, I use as much geometry as needed to get it looking good, but I try to bear in mind how certain features will work best for the low-poly model and plan accordingly. As EarthQuake says, you're not going to win any prizes for making an 'optimized' high-resolution mesh.
As for my presentation, I use the material from this tutorial with the specified settings and a simple gradient map.
http://www.artemstudios.com/08Portfolio/Tutorial/MaterialRenderingTut.htm
Just a quick update today - my machine has been complaining, so I'm guessing I'm probably due to blow another PSU sometime soon. Reworked the support struts for the corridor to suit the scene a little better:
So aside from finishing off a couple of models here and there, I'm going to start texturing full-time in the next day or so, as well as revisiting some of the issues I was experiencing with lighting the scene in UDK.
As always, thanks in advance for any feedback or suggestions.
At which stage are you unwrapping UV's?, presumably after you make low polys of your high polys?... and then from that baking normal mapping onto the low polys, okay I'm just thinking it through, so many stages but I'm getting my head around it. I'll have a good read of that first link, should clear things up.
I can't see the update that you have, did you post an image?
Again, it's important to be aware of optimization issues at this stage, as these days texture resolution is considered as important, if not more so, than polygonal optimization when it comes to memory efficiency. This becomes really evident on curved surfaces - sometimes you can definitely go too far with optimizing to the point that the low-poly suffers, and can also create problems when baking out your normal maps.
Once the UV's are laid out, I use xNormal to bake the normal maps. xNormal is a free and very powerful little application that handles baking out normal and ambient occlusion maps (amongst other things) that I find much less hassle than using Max's render to texture system. You can grab it here:
http://www.xnormal.net/2.aspx
There's also a mega-thread on how to use it and overcoming certain problems in Technical Talk:
http://www.polycount.com/forum/showthread.php?t=41776
And yeah, I did post an image in the last post. I use Image Shack to host my images, and it seems to be working fine for me.
Should have some textures to post later today!
I'd originally modeled this some time ago, but it looked too much like a pull-down projector screen, even with some reasonably ornate metalwork. It's a wall-mounted map chart showing the location of Nemo's base in Vulcania that'll be situated on the wall behind Nemo's desk, to the left of the ship's power core-type thing.
The majority of the texture will be a large, old map showing the seas around south-east Asia, where Vulcania is supposed to be located. The end of the arm will be centered on the exact location of the island base.
There's (literally) a couple of other bits I have to tidy up, but it's mainly smaller pieces like a ship radio-type communications unit, a depth gauge and a couple other bits and pieces, but this was the last piece I wanted to completely re-do.
I promise, the next update will involve textures in some way.
As always, thanks in advance.
cheers!
I know of xnormal, haven't got so far as to need it yet, I've also heard of crazybump. Also for unwrapping I've used in the past Headus UV layout, are you using something to help with unwrapping or just the Max tools?.. I've found Max a bit crap for unwrapping compared to Maya but maybe I just haven't had enough experience of it yet.
For your images, I can't actually see any from anyone on this computer, I'll wait till I get home and have a look at what you've done.
Cheers!
I'm using Max's UV layout tools, with renderhjs' TexTools plugin. I've used UV Layout Pro before, and I do like the keyboard shortcut method of cutting and stitiching, but constantly exporting and importing meshes just got to be too much of a hassle.
http://www.renderhjs.net/textools/
You should definitely start posting your work. I checked out your 2D work, and it's really impressive. I'll look forward to hearing your thoughts on my work when you can actually see it!
Thanks for the compliment, my 2D stuff is vastly different to what I want to do in CG, somehow environments intrigue me more in CG whereas wildlife/portraits interest me more in traditional art
Just out of interest, are you working in CG, or a student?
Yeah, I do use a tablet when texturing, and part of the reason I chose such an intricate, prop-heavy environment was to really get some solid practice in texturing. As I've said before, it's pretty obvious I'm much more comfortable modeling than texturing, and the scary (and potentially frustrating) thing is the weight of my own expectations in terms of nailing the texturing and doing the models justice.
As for me, I'm still very much a student, entirely self-taught.
Thanks again for everyone's feedback and encouragement.
A simple lighting test.
Main reason I posted this was to motivate myself. Felt like I was getting more done in the earlier stages of the project. I'll hold off posting images of the actual office area until some large-scale changes have been completed, namely the majority of the furniture and wall decorations. As interesting as it was to explore an entirely art deco theme, it just wasn't looking like the Nautilus I'd imagined and was straying a little too close to Bioshock's aesthetics.
Hopefully updates will be a little more frequent from now until the completion. I've learned a fucking ton so far from this project, but it's dragged on far too long. A lot of assets had to (and still have to) be remodeled - some things seemed like they'd be a good fit initially, but needed significant reworking to fit the feel of the scene. For example, the support struts in the corridor are actually the fifth iteration of the design.
Still a lot of work to do, but I feel like the end is almost in sight.
Yeah, the scene does need a little more ambient light, and some meshes to motivate them which I'll be adding shortly, along with a whole mess of pipes to break up some of the wall surfaces. Then I'll be moving on to the actual office area, which needs some serious work.
I'm planning on adding some subtle volumetric lights to simulate the greenish light from the ocean coming through the portholes and the larger observation windows in the office itself, which should offer some more illumination and compliment the overall color scheme.
More updates be comin'.
Keep up the nice work. Looking forward to see it finished.
Not sure if it's the normal maps need some extra lovin' or whether I need to look at the lightmap UV layout, but I'll get to it either way. I suspect it's the lightmap.
Moving on to the staircase's handrail today - had to abandon the previous design as it just wasn't practical to optimize it. After that, moving on to remodeling the furniture, primarily the desk, then texturing Nemo's diving helmet and underwater rifle. Almost all textures are placeholder - just trying to get the overall scene to more of a finished state before tidying everything up.
Thanks for everyone's feedback.
Added the slightly-redesigned handrail. Again, all textures (and some geometry) are placeholder.
Still have to fix the uniformity of the wall panel textures and the metal part of the staircase, and replace the floor panel adjacent to the first step. Will also be adding bolts to the curving corridor support struts.
Still needs a lot of work - not sure about the paintings right now, or at least how they're displayed. I don't want it to look too much like a museum, but I do like the pools of light. Also still debating the position and design of the desk. Working on a suitably-ornate trim to sit beneath the light over the stairs, bridging the transition between surfaces of the archway.
The large 'rib' supports will likely be reworked - right now, they're just not interesting enough. I'll be adding a dumbwaiter and a standard lamp to the left of the stairs, and am debating whether a rug might help break up the empty floor space (which also needs to be less obviously tiled, and the hues shifted towards a darker, more cherry wood). Also deciding between wall-mounted shelving or actual bookcases.
Back to work.
I'll likely be reworking the 'rib' supports to be a little more interesting (and the size of the bolts), as well as adding a lot more clutter to the desk. The chair needs some more work, too, as does the desk. I'm also not happy with the base metal texture - it's just not reading well, and needs to be cleaner.
One thing I'm learning a lot about as I go is scale, and how to model objects to seem of a realistic scale but large enough to read well, which some objects (including the drinks cabinet and pressure gauge to the right of the map) clearly don't at the moment.
It's coming together slowly, but it's starting to feel like a fucking marathon. I've still got a lot to do in terms of modeling, and I haven't even started texturing, something I fear my inexperience will delay completion even longer.
Even though it's my first environment, I have to admit to struggling with motivation at times. Considering how long I've been working on this, I thought I'd be happier with where it's at. I keep telling myself that I'll get faster with time and practice, and that I won't get better by rushing, but right now it feels like the whole scene is kinda crappy.
And I thought this scene would be a 'smaller' project. :poly127:
Anyways, better get some sleep. I'm sure I'll feel less emo in the morning. As always, thanks in advance for any crits, suggestions or insights.
I personally think that going with larger projects is a little more commendable, and this one looks great. I just read the whole thread. One thing you might try doing is just blocking out the level first. You seem to model everything with a lot of detail, and I'm not sure if half the models are still in the scene. Even if it's just boxes, at least you can get a sense of "hey is this possibly going to work here?" Plus it might save you some hours modeling. (which by judging your models you've already had plenty of practice)
Also, go back to the original image. I see you've deviated from it a little bit, but you may be able to find some motivation by just staring at that pic for a while.
I'm looking forward to seeing the outcome of this. Don't give up man, the last 10% is always the hardest, but worth its weight.
Incredible detail work! Cant wait to see it finished.
I think my biggest problem is that I'm judging the overall quality of the scene based on placeholder art. There's quite a lot that still needs to be added before it will even start to look finished, and I guess I'm pretty anxious to start improving the quality of individual assets to raise the quality of the overall scene before the scene is properly populated.
I'll keep plugging away at getting all the assets into the scene, then focus on polishing. I'll post Marmoset renders of completed assets as and when they're done so I can hopefully get some feedback on my texturing.
Thanks again for everyone's encouragement, and apologies for being such a whiny bitch last night.
Best way to approach this, really. The quality of the scene has definitely gone done, BUT you'll kick back and make this a beauty.
-I'm digging the design of the back wall with the map.
-Wood floor definitely needs some breakup, I feel like its somewhat out of place with the style, perhaps the color adjustment and breakup might change it.
-The ribbed supports on the lower level corridor are well designed, you could try to incorporate some of that design in the office
Keep going, you have some interesting work here.
I think I'm going to ditch the trim around the main map, and instead focus on making the rib supports more interesting, and simply transition from a rib support directly into the map wall. It seemed like a good idea, but it just doesn't look right.
I also agree with the comment about the floor - right now, it's literally just a cropped stock texture applied to a plane to give me an idea of scale. I'm thinking much darker wood, and maybe slightly larger boards, but we'll see.
I think I've got three more main focal props to add (the helmet and stand, the mounted underwater rifle and a large lever to sit adjacent to the desk), a few general props (books, bottles and some papers for the desk), and a couple of structural pieces to add / cleanup, then I should be ready to start polishing.
As I said earlier, I think the reason I'm so gung-ho to start polishing assets before the scene is properly populated is that I know my texturing work will probably require the most critique in order to improve, and I guess I'm a little concerned because the textures are going to make or break this scene. I'm determined not to hide sub-par work behind masses of post processing, even though I do intend to do quite a bit of post.
Anyways, thanks again for everyone's feedback and encouragement. I'm aiming to get this scene finished before I leave the country in a few weeks, so I'm definitely on a serious deadline.
Back to work.
Still got a lot to do. Been tweaking the maps endlessly, trying to achieve the right look for the brass. Still far too much like gold for my liking right now, though I'm trying to hit the sweet spot between pristine reconditioned antique and recovered-from-the-bottom-of-the-ocean look that my reference images span.
Once I've got the metals looking right, I'll be moving on to the glass and rubber piping elements, before adding some wear.
Still playing with the spec, and I'll probably be going back and reworking the size of the grain on the large front panel, as well as playing with the normal intensity of the sculpted elements. Oh, and adding some books.
Thanks for looking, and any suggestions.
The texture on the helmet needs some work. I think it needs some green and white oxidation from being submerged in salty sea water, and maybe some dents and dings to show some use. Probably not as extreme as I took it when I created an antique helmet a few years ago, but probably something between the two extremes?
You might want to flip the grain on the doors of your cabinet, it lines up with the grain that's running on the rest of the cabinet which counter to the strength of the wood and odd that it would be cut from the same piece. looks like you flipped the shelves and the drawer but missed this?
Making good progress
The helmet's wear is proving the most tricky, but you're right, definitely needs a bit of beating up. I'm thinking the mallet brush at a low intensity in ZBrush might be a good fit for that. I wasn't sure about oxidizing the surface, as I read that this can take a long time of constant exposure, but I think it needs some. I really like the blend of colors you've got happening on yours, very nice.
The only asset that's really 'finished' is the lamp, so definitely going to be revisiting them, as well as chipping away at some of the other assets along the way. I'm finding I'm getting more done and sustaining my interest by tinkering and refining a variety of assets right now. Still got a buttload to to, though....
Thanks for the tip about the wood grain - I guess I thought it would be more logical to create the main body that way, but I'll switch the flow of the grain and tweak the intensity of the normals before posting again.
Thanks again.
Thought it'd be best to present the various assets individually to begin with, before working on the larger scene overall. The underwater rifle is next on the list.
Thanks in advance for any input.
Good stuff man.
The helmet looks a lot better
The bake looks slightly messy though; there is some nasty colour bleeding (or is that supposed to be dirt?) at the edges of the helmet. It makes it look like you don't have enough padding.
Regarding Zbrush; I wouldn't bother, I believe it might be quicker to just paint a heightmap in PS and bring it into Crazybump, or use nDo.
Keep on cracking mate