Teaming up this time with the lovely vandalfish, the contest was too cool to pass up and we're looking forward to seeing everyone's entries. Stay Tuned!
Next we're gonna dig through these and pick out the direction we want to take, then do some palette ideas. I want to try some pretty far out stuff with the colors but it'll probably get toned down, we'll see
Some good ideas generally, and your sketches look nice. I would go with "A" and "H" because I like how there is less focus on his shoulders, kinda unexpected and sticking out.
Damn lots of great concepts! I'm doing PL too, and one thing to think about I'm finding hard is how low the poly limit is for him. I was looking at Lycan and he's like triple the count if not more. Anyway only saying that b/c some of those concepts might be hard with the poly constraints.
But nothing to say bad about any concept, they all look awesome! I kinda hate you haha kiiidding of course. The weapons are hard, b/c there are so many choices and its easy to riff on small ideas many many times.
Good luck you guys are one to watch!
Here's a few rough concept WIPs, combining some of the aspects we liked from our previous thumbnails. Also, some weapon silhouettes. I think the weapons are looking a bit too slender, so I'm going to need to fatten them up a bit on my next pass.
here's my contribution. Feel free to ignore the actual armor for the most part, I just wanted to try some bolder color options. I think we have a solid direction now and should have a design nailed down tomorrow. Hopefully.
Awesome work so far! Man, I wish I had a cool concept artist to team up with. From your latest post, I enjoy the color palette for choice A the most and the weapon silhouette for choice 1.
Your designs are really well done. I'm partial to design D. Really like the lighter color tones on his armor and the pink works well I think. The other concept with the skull antler shoulders is really interesting too. Can't go wrong with any of the concepts though really, they're all great! Keep up the good work.
Ah yes, thanks for correcting me. You two are working well as a team and can't wait to see further iterations. I have to get mine up as well. Trying out some silhouettes, but probably block something out in zbrush. Here's to your progress!
Another small update, this was drawn by vandalfish. We combined the elements we liked from previous ones and simplified some to make sure it would read well in game.
This isn't final, and there will be some more small tweaks to the color and some of the shapes, but it is indicative of what I'm going to model.
As for weapons, we think we're going with #1 from my latest post. It was a tough decision though as I'm happy with some of the other shapes as well, if there's time I might do a second one for fun. Should have a clearer concept for the weapon sometime this week.
Thanks for the comments everyone, we appreciate your feedback!
The most recent looks really nice, but I would lose the shoulder poof...not liking it. Try some of your earlier feather shapes, maybe straight up and back. Nice work otherwise
doh, I'm a little bummed that the poof didn't work out, but we agree and did another take, including back view for modeling ref. I've also started building some base meshes. My arm slot is going to be really tight, as is probably the rest, but I should be able to make it..
The main issue is that I didn't check if Phantom Lancer had alpha support before we started doing all the feather stuff. I think we can get by with just modeled feathers though.. we'll see!
and a quick feather test, simple geo and vert paint cobbled together to see if it would read right from player view. I think this could work.
love all the designs + colors and of the new set, though the new poof on the shoulder looks feminine to me.
I'd say try to play on the more cooler colors at least on those maybe a brown, orange, maybe yellow.
Just something about them throws me off otherwise the rest of It for me is good to go, either way I'll like the end result just based on the concept and colors.
Edit: The feathery test geo looks good, did you find out about them alphas?
I really like the direction you're going.
I know you're trying to make the poof of feathers work on the shoulder but I'd suggest avoiding that problem. DOTA2 doesn't support 2 sided materials, so that means each feather is going to need to be 2 sided (if not more), you can hide or simplify this issue on all the other poofs/feathers except the shoulder (big poof not under the armor which is ok) which will be an issue b/c its view from so many angles. It will take up more than double the poly's (of a single sided feather) for that shape and possibly give rigging issues.
@agentfx: I don't understand most of what you're saying. Why would the feathers need to be more than 2 sides? Just make the 2d feathers, flip normals. As long as the weight painting is the same on each feather and it's flipped counterpart, then it should be fine with rigging too. Unless there is some weird stuff with the Source engine that I don't know about, all of the concepted stuff should work fine if implemented correctly.
$!nz: No alpha support. . Makes this more challenging.
Agentfx: thanks for the feedback. I've been back and forth about it, and while I dig the idea the reality is it's not going to be done easily, and what's probably going to happen is either a gross simplification or an alteration altogether or what's happening. I'm just going to play with shapes and stuff on the model after I build the rest of the Shoulder piece so I know what my budget is. I think we were just having too much fun with the feather idea and it came and bit us in the ass haha .
@agentfx: I don't understand most of what you're saying. Why would the feathers need to be more than 2 sides? Just make the 2d feathers, flip normals. As long as the weight painting is the same on each feather and it's flipped counterpart, then it should be fine with rigging too. Unless there is some weird stuff with the Source engine that I don't know about, all of the concepted stuff should work fine if implemented correctly.
well what I was worried about when 2 poly's are in the same space is weird artifacting. It happens in other CG programs so I just try to avoid it. Rigging wise why its hard is if you want to paint the weights it may not paint both vertex points, unless you do rigid binding in maya and set each point by hand. I'm only saying it adds potential issues. If it's only bound to 1 joint might be ok. Shoulders in general are hard to rig, so I felt that it'd be saving unfun parts by removing it.
good news first. base meshes are done, started sculpting the things I want sculpts for, should be showing those soon. secondly, importing (at least the weapon) is way easier than I thought. I decided to just wing it before looking at tutorials and it really was as easy as parenting the weapon to the weapon bone and exporting an FBX. I didn't even change settings. This may be wrong and I'm definitely going to go through the guides soon but I was pretty stoked to see it work so easily.
no textures or anything, only a quick test:
looks cool from the above view but my screenshot button decided not to remember that one. Will have more soon I'm sure.
the bad news is, while I was going through the test weapon submission, I noticed that the LOD1 slots were for the GAME meshes. Shit. I saw LOD0 and 1 on the character workshop page and thought those were both for game, and that the cinematic one would be beyond those. Nope. Needless to say going from a planned 800 for the belt to 450, or worse for me - dropping the shoulder budget by 200 or so, is going to be difficult. The design will probably be simplified in some ways. For example, less bracelets or flat ones, less but bigger feathers (which will probably read better anyway), etc.
Thanks again guys, I'm excited to continue working on it. Hoping for big updates this weekend.
the good news is don't triangulate the mesh on export. The video tutorial shows doing that but its not necessary and you have have a much larger budget w/o triangles (altho that could be debated but I got my meshes in, the shoulder is 515 tris and 400 or even 300 something quads so there is a slight SLIGHT silver lining. But thats why I was worried about the feathers in your one design. There's only so many ways to reduce something like. Anyway yeah its an art to get the detail in there while staying at/under poly budget. I think UV space is easier than poly budget on this project.
Thanks guys, much appreciated. and wow, that's so cool to be on the blog! Thanks for the heads up.
@Managor: It's not too purple, I think. It's mostly red/orange with a few areas of mangenta-ish. Plus I asked in the Q&A and it sounds like we can do whatever palette we feel would work for our item set.
Now back to the model. I did a first pass on some of my shoulder stuff and wow I went way over the budget for shoulders by like 400 in terms of LOD1. After thinking about it I think I'm going to do LOD0 first and then cut that down to make the game mesh.
Here's what I've blocked out so far model wise, still very rough and wip:
I'm in the process of cutting stuff down like the necklaces, from 3 sides to one, unless I have extra polys in LOD0 later. I'm also going to be merging the skull and shoulder pad instead of keeping the model separate, It will be cheaper and probably read better after I bake it than them separately anyway. These feathers are killing me though haha, just going to keep them bigger and have less of them, they'll still fill their purpose of splashing color and cool shape. I think I'll be fine tri-count wise on the belt, weapon, and arms, but the shoulders are going to be a challenge since there's SO much going on there. Really excited to get to the baking and texturing part soon.
Thanks again everyone for the comments and feedback, you guys rule.
edit: apologies for the image quality, used the snipping tool this time and forgot how bad it's .jpg compression was.
It's not the rules I'm worried about. I'm just being cautious. It looks good so far and I like it but I'm trying to protect the color palettes so one hero won't be able to equip every color of the rainbow. It won't happen immediately but it could creep in.
I agree to an extent. If I were to wager a guess on how Valve will decide on the color changes (I hate when people talk like they know valve and what it will do) that said my GUESS is they want to make sure all characters are immediately recognisable as that character. Color palette will be part of that. If a character kit makes them look not like the character it wouldn't get approved. Thats just my guess. Adding one more color to the palette is a risk but I say go for it. Also make sure you've read how they work with color in the art guide http://media.steampowered.com/apps/dota2/workshop/Dota2CharacterArtGuide.pdf
You guys bring up a good point, so I submitted a question to be sure but as I mentioned in the Q&A it seems to be okay to do. Plus he actually has close to that shade in his palette in the dark areas of the armor, this is just a similar contrasting hue but brought to a much lighter value/saturation. I hope it's not an issue.
Getting the shoulders down is proving to be a hell of a challenge, barely making the LOD0 currently, I'm so greedy with these polys!. I have ideas for cutting down as previously mentioned.. but yikes! I hate to sacrifice the design, I'll see what I can do. I think I can get away with much fewer feathers and still retain the same feel.
Had a weird shading issue on one of the normals, not sure what caused it, but simply re-importing the low rez mesh fixed it:
the one on the right was the problem child:
here are some of the sculpts, couple more needed but nothing big:
trodding along. Can't wait to get to the texturing.
Thanks again everyone for your kind words, comments, and feedback.
awe crap! Looks amazing! I actually just learned something by looking at this, break up objects to sculpt! haha I guess people who sculpt all the time already know that. I've been struggling getting things to sculpt cleanly b/c i've been doing them all together.
Oh just noticed something. I found the belt to be the hardest to rig properly. I can already tell yours is going to be even harder. Please do this, test bind it and bring it in engine. You're going to see how bent the legs are at idle and how quickly the missing geo shows.
Thanks agentfx, I did what you suggested. I got most of it working, except I have no idea how to stop the center fabrics from drifting towards both of the legs. I've tried rigging each side of the mesh a little to each leg, and in this example they're rigged only to the root, plus some to the front_cloth nodes. Any idea what's causing this?
Yeah, I had the exact same issue. I "solved" it by a lot of trial and error and gently massaging the values until it worked. I'm at work atm but when I get home I can check my mesh and see what the weights are.
what I found is this, you have to put 100% center piece to center strap, legs to legs, then there's this gaping hole. If you don't do 100% center piece to center strap you get ugly stretching and distorting like you see. If you think about it, there's no way to avoid that if you don't give it to solely one bone.
That's where the filler piece comes in. Its sort of just cloth that can stretch and bridge the missing space. It will penetrate and be kinda nasty but it will do the job of hiding the missing geo, which I think is noticeable even in game camera, meaning far out and character is small. You have a bigger belt so hopefully the whole and filler cloth won't be significant.
Replies
here's my share:
thanks for looking!
But nothing to say bad about any concept, they all look awesome! I kinda hate you haha kiiidding of course. The weapons are hard, b/c there are so many choices and its easy to riff on small ideas many many times.
Good luck you guys are one to watch!
Small update for now, just showing the direction(s) we have decided to go, there should be another update later tonight, hopefully with some color.
My roommate who worked on Spellborn also picks 3!
I like no.4 weapon
This isn't final, and there will be some more small tweaks to the color and some of the shapes, but it is indicative of what I'm going to model.
As for weapons, we think we're going with #1 from my latest post. It was a tough decision though as I'm happy with some of the other shapes as well, if there's time I might do a second one for fun. Should have a clearer concept for the weapon sometime this week.
Thanks for the comments everyone, we appreciate your feedback!
- BoBo
The main issue is that I didn't check if Phantom Lancer had alpha support before we started doing all the feather stuff. I think we can get by with just modeled feathers though.. we'll see!
and a quick feather test, simple geo and vert paint cobbled together to see if it would read right from player view. I think this could work.
Thanks again guys for all your feedback!
I'd say try to play on the more cooler colors at least on those maybe a brown, orange, maybe yellow.
Just something about them throws me off otherwise the rest of It for me is good to go, either way I'll like the end result just based on the concept and colors.
Edit: The feathery test geo looks good, did you find out about them alphas?
I know you're trying to make the poof of feathers work on the shoulder but I'd suggest avoiding that problem. DOTA2 doesn't support 2 sided materials, so that means each feather is going to need to be 2 sided (if not more), you can hide or simplify this issue on all the other poofs/feathers except the shoulder (big poof not under the armor which is ok) which will be an issue b/c its view from so many angles. It will take up more than double the poly's (of a single sided feather) for that shape and possibly give rigging issues.
Agentfx: thanks for the feedback. I've been back and forth about it, and while I dig the idea the reality is it's not going to be done easily, and what's probably going to happen is either a gross simplification or an alteration altogether or what's happening. I'm just going to play with shapes and stuff on the model after I build the rest of the Shoulder piece so I know what my budget is. I think we were just having too much fun with the feather idea and it came and bit us in the ass haha .
well what I was worried about when 2 poly's are in the same space is weird artifacting. It happens in other CG programs so I just try to avoid it. Rigging wise why its hard is if you want to paint the weights it may not paint both vertex points, unless you do rigid binding in maya and set each point by hand. I'm only saying it adds potential issues. If it's only bound to 1 joint might be ok. Shoulders in general are hard to rig, so I felt that it'd be saving unfun parts by removing it.
- BoBo
so, some good news, some not so good news.
good news first. base meshes are done, started sculpting the things I want sculpts for, should be showing those soon. secondly, importing (at least the weapon) is way easier than I thought. I decided to just wing it before looking at tutorials and it really was as easy as parenting the weapon to the weapon bone and exporting an FBX. I didn't even change settings. This may be wrong and I'm definitely going to go through the guides soon but I was pretty stoked to see it work so easily.
no textures or anything, only a quick test:
looks cool from the above view but my screenshot button decided not to remember that one. Will have more soon I'm sure.
the bad news is, while I was going through the test weapon submission, I noticed that the LOD1 slots were for the GAME meshes. Shit. I saw LOD0 and 1 on the character workshop page and thought those were both for game, and that the cinematic one would be beyond those. Nope. Needless to say going from a planned 800 for the belt to 450, or worse for me - dropping the shoulder budget by 200 or so, is going to be difficult. The design will probably be simplified in some ways. For example, less bracelets or flat ones, less but bigger feathers (which will probably read better anyway), etc.
Thanks again guys, I'm excited to continue working on it. Hoping for big updates this weekend.
especially those feathers on his left shoulder!
great work!
@Managor: It's not too purple, I think. It's mostly red/orange with a few areas of mangenta-ish. Plus I asked in the Q&A and it sounds like we can do whatever palette we feel would work for our item set.
Now back to the model. I did a first pass on some of my shoulder stuff and wow I went way over the budget for shoulders by like 400 in terms of LOD1. After thinking about it I think I'm going to do LOD0 first and then cut that down to make the game mesh.
Here's what I've blocked out so far model wise, still very rough and wip:
I'm in the process of cutting stuff down like the necklaces, from 3 sides to one, unless I have extra polys in LOD0 later. I'm also going to be merging the skull and shoulder pad instead of keeping the model separate, It will be cheaper and probably read better after I bake it than them separately anyway. These feathers are killing me though haha, just going to keep them bigger and have less of them, they'll still fill their purpose of splashing color and cool shape. I think I'll be fine tri-count wise on the belt, weapon, and arms, but the shoulders are going to be a challenge since there's SO much going on there. Really excited to get to the baking and texturing part soon.
Thanks again everyone for the comments and feedback, you guys rule.
edit: apologies for the image quality, used the snipping tool this time and forgot how bad it's .jpg compression was.
Getting the shoulders down is proving to be a hell of a challenge, barely making the LOD0 currently, I'm so greedy with these polys!. I have ideas for cutting down as previously mentioned.. but yikes! I hate to sacrifice the design, I'll see what I can do. I think I can get away with much fewer feathers and still retain the same feel.
Had a weird shading issue on one of the normals, not sure what caused it, but simply re-importing the low rez mesh fixed it:
the one on the right was the problem child:
here are some of the sculpts, couple more needed but nothing big:
trodding along. Can't wait to get to the texturing.
Thanks again everyone for your kind words, comments, and feedback.
Looks great.
That's where the filler piece comes in. Its sort of just cloth that can stretch and bridge the missing space. It will penetrate and be kinda nasty but it will do the job of hiding the missing geo, which I think is noticeable even in game camera, meaning far out and character is small. You have a bigger belt so hopefully the whole and filler cloth won't be significant.