Jake Korosi (jakeblended)

87 replies · posted

Polybook - Jake

Heeey there guys.  Remember when I used to actually make CG stuff and post it on this site?  Man, those were the days!

Well, after Stuff Happened that got in the way for a while, I'm finally back.  Here is my Polybook.  Can't wait to make some stuff and share it with you guys!

  • Okay so, today I started Kent's Realistic Character Modeling course.  I haven't sculpted in a few months, so it took some re-acclimating, but in little time I regained the hang of it.

    Here's where we're at, end of day 1.  I've got my initial (very) rough blocking pass most of the way done.

    Tomorrow I'll get the hands and feet blocked at least, and hopefully the head if possible.  Then it'll be time to refine.  

  • For day 2 I didn't have as much time to work this evening as yesterday; so I managed to finish rough-blocking the hands and the feet, but I only just barely got started on the head.  That's okay though; I'm off all weekend, so I'll have two whole days to work on this guy.  Tomorrow it's the head, finishing up the blocking and moving to refining the whole thing.  

    Funny thing about the feet; working on the toes I decided to skip the Boolean method Kent recommends in the course video and decided to see if I could brute-force chisel the gaps between the toes into the mesh with the crease brush.  Well, it can be done, but....I really don't recommend it, it takes forever to carve a crease that deep.  And the weirdest part was, I ended up leaving some floating geometry behind somehow....I've never seen that happen before.  I've seen sculpting problems leave missing faces - holes in the mesh that are a pain to patch up - but I've never seen just random faces left floating in the air, completely detached from the rest of the mesh.  The mesh cleanup tools handled it easily enough though.

    My character is going to be an ancient Egyptian soldier, by the way.

    • jjakeblended How did you approach learning anatomy so far? I know the big muscles, but especially around the elbow / knees my results always get really sketchy.

    • jjakeblended You're off to a good start. I'm not that into Ancient Egypt but it would be cool to see a cool 3d  soldier nonetheless :D

    • numbernine It remains to be seen whether I'm good enough to be giving anyone else advice; but I will let you and everyone else know about the anatomy resource I found and like most.

      It's a series of videos on a YouTube channel called Proko, which teach the forms of human skeletal and muscular anatomy for artists.  The channel is dedicated to charcoal drawing; HOWEVER it takes no imagination to realize the information is just as useful for sculptors - including the digital kind; anatomy is anatomy.  The lessons are also quite watchable too, with lots of humor that trends on the campy/dad-joke side.  Proko has a paid "premium" service, run from his website (he plugs it at the end of EVERY video)  which has some longer videos and goes over some extra information, for instance explaining some of the deeper and hidden body muscles, that aren't covered in the YouTube playlist - but the YouTube channel DOES cover all of the muscles you can see on the surface of the body, do not worry.

      The videos are each only a few minutes long and they start with some anatomical basics before getting into first the bones, going through each part of the body, and then shifting to muscles.  The series is still in-progress, but the only parts that have not been made yet as far as I can tell are the lower-leg muscles and the foot muscles.  

      You can find the anatomy playlist here.  I really personally recommend taking the time and watching the whole playlist, including all the bones; but I suppose if you want to save time, you can watch just the bone episodes which cover the bony surface joints (elbows, knees, hands, feet), but make sure you watch all the muscle anatomy and form videos of course.  You can feel free to skip the assignment example and critique videos, since again those more specifically focus on drawing.

      And for sculpting, even after watching the anatomy lessons, nothing beats a good reference.  In the course Kent recommends the website 3d.sk, and I've found nothing better for model photos.  The super high-res photos and sets cost money of course; but I've found that you can open and save the lower-res photos on the product pages  for free, and while they're not huge by any means, they're often decent enough for using as sculpting references.   As I'm becoming more serious about my art though, I think eventually I will fork over the money for a few high-res sets; those paid photos are large enough that you can see like fingerprint-level detail in them.

    • jjakeblended 

      Thanks, I'll definitely check out that anatomy playlist. The introductory video seems promising :)

  • Day 3; finished blocking and moved to refining.  Got some real progress done on the head and face:

    Still a fair bit of work to do, but we're getting there.  Now that I've gotten my ear and jawline placed at least, I think tomorrow I'll just fix one or two shape issues on the face and then go ahead and give the whole body a pass, and then come back for a final polish pass to tie up the little stuff at the end.  

    I'm finding that when I start working on the small scale for intricate details like the ears and eyelids and that kind of thing, it's really easy for me to get distracted and obsess over all the little sculpting artifacts - I call them "nicks and dents" - that sometimes make these shapes look rough and janky when you're zoomed up close.  I waste loads of time trying to smooth those things out; I keep forgetting that the retopology stage fixes all of that for free.  

  • Okie dokie, I'm satisfied with this mesh as a "75% complete sculpture" base for my course project.  After finishing the body, I revisited the face; I decided it looked too generic, so I went and gave it some more definition and character.  He's a warrior after all, he needs a Game Face.

    So tomorrow, I start giving this guy something to wear.  I'm a little apprehensive though; as a Bronze Age soldier, this guy wouldn't be nearly as kitted out as say the example character Kent makes during the course, and I wonder if that's going to mean there won't be much work for me this week.  We know what Egyptian soldiers wore, and I don't want to stray too far from reality and into fantasy.  Not that there's anything wrong with doing so; I just already decided that I wanted my guy to look reasonably historically accurate.  So, as far as clothing goes, he's going to have a kilt of the proper style, a light headwrap, and a kind of peculiar longish padded waist-guard that these soldiers wore.  I will probably give him some sandals (they did or did not wear them, it could go either way), and I'm also going to give him a distinctive Egyptian sword and shield.  Holding these items means he'll have to be posed, though; so I'll model them this week, but they'll probably have to wait until retopo week or later before I can actually screenshot the character carrying them.  

  • Working on the clothes.  First, the simulated cloth.  After a....couple of hours (I think?) trying to get my cloth physics actually working correctly at all, I got the kilt and the headdress done.  I think these are the only pieces I'll need to simulate; the rest will be modeled.  The way the simulator laid that headdress is just about perfect!  

    • jjakeblended I was waiting for an update! I'm doing Kent's course as well right now, you're inspiring me to try my best at the whole anatomy business. I'm still working on my blockout, my aim is to get it on the free critique stream on the 28th.

      The face looks good too by the way. The expression's already there.

  • Progress update!  Continuing to add stuff.  Here's how we're looking so far:

    Believe it or not, historical-accuracy-speaking, he's wearing all he should be.  But I think more detail would make the character more compelling, so now I'm starting to invent things - for instance, those cloth streamers, which I imagine will be colored as a way of showing what unit this guy belongs to from a distance.  After taking the time to model them though, I think I'm actually going to relocate them to the other side, because his sword needs to be belted on this side (he's a righty) and I need some detail on the right for balance.

    Speaking of which, I also need to invent a plausible way to attach his sword to his belt, because how exactly they did that isn't really known.  All of the Egyptians' art depicting soldiers always shows them with weapons in-hand.  And the Egyptian sword is...well frankly it's weird; it's a very odd shape that it seems to me doesn't lend itself well to a simple scabbard (there are modern scabbards for replicas of these swords, but they don't look plausible for the time period, or for actually wearing generally).  Oh well, I'll think of something.

    Tomorrow I'll be doing the above work, plus modeling his shield and weapons.

  • So the character project is kind of on hold, you may have noticed - in fact all of my sculpting is on hold, due to a broken tablet that I need to replace.  It wasn't a QC issue - the breaking was definitely my fault, we will leave it at that. :)

    But I want to do something while I'm saving up to get another.  So I'm going to geek out a little and make the spaceship from Star Trek.  This is partially inspired by Chris Kuhn's recent model of a different ship from the same movies, but I'm making the Enterprise.  

    Meh news: This ship has been done before, lots of times, so I'm breaking absolutely no new ground here CG wise and this is for personal challenge and development only.

    Better news:  It's a spaceship, which is a vehicle, which means it qualifies for this month's challenge!  I honestly have some serious reservations as to whether I'll be able to fully flesh this thing out AND get it materialed/textured by the end of the month (and I'm already a day behind!), but that's what I'm aiming for, so we'll see. 

    Good news:  I found some of the best possible reference schematics you could ever wish for, for this thing.  If there's one thing Star Trek fans like to make it's blueprints of stuff in the shows/movies, so there was never any doubt I found one.  But I did even better than usual, and instead of simple orthographic projects, I found some actual CAD sheets, designed for making a physical model - out of wood or plastic or whatever.  And the model the instructions are intended for is a replica of the 100-inch(!) filming model used for the movies, so it's quite detailed.  

    There are only two catches - the first is that all of the measurements given in Imperial.  Which, yeah Blender can "technically" use Imperial units, but it can't parse fractions, so I have to convert these manually outside of Blender on the fly.  Not a show-stopper  but it's an annoying extra step that slows the workflow somewhat.  The second catch is that the instructions are for shaping solid objects, which isn't a 1-to-1 match for edge-modeling, process-wise.  I COULD replicate some of the process by using Boolean subtractions and such, but the problem is that clean mesh is a thing I have to worry about.  So it's a struggle sometimes, especially when the shapes get weird.  

    But anyway.  Let's get to it!

  • Here was yesterday's work so far.  I'm working from the top down, and then back.

    You may notice from the grid guides that I'm quite a ways above the world origin.  Interesting thing about these CAD plans; the measurements for each individual component are all relative to a single common origin point.  That means every part is built "in place" and I shouldn't have to adjust or move any of these parts around to put the larger ship together once they're finished, they'll just magically be together.  Neato.

    Also, getting the smooth curves on that piece, especially those ramp-looking dealies on the sides of the dome where they meet the base and the sides, and getting nothing pinching with the subsurf, was a NIGHTMARE of epic proportions.  I had to restart a couple of times to try different approaches in order to get a nice smooth shape.  I hope the rest of the ship, or at least most of it, isn't that hard...

  • Okay, so the first component is completed.  The rest of this part was easy to model, no more nightmares.  There was just a lot of details to model.  

    Everything is pretty straight-forward; the only unusual thing would be those little vertical vent slots near the tops of the ramps.  Those are an unapplied Boolean.  It can't be applied, because that particular piece of mesh isn't solid - it's not enclosed on the bottom.  Just out of curiosity I tried it, and it didn't work right.  In this case it's okay, because I don't need any special materials for the insides of those slots, so I won't have any UV issues or anything like that. 

    Almost 170,000 faces in this project so far, just from that component.  My computer might be crying before this project is done...

    • jjakeblended Looking good so far! 🔥

      Just thinking out loud, couldn’t you just close the mesh with a big n-gon, apply the Boolean and then delete the n-gon again..?

      Wow that’s a lot of geometry...

      Keep up the good work Jake! 😎👍🏻

  • Thanks!  I might try your suggestion of temporarily solidifying it just long enough to try to apply the Boolean; but I think I'll wait until the end when everything's modeled and I'm cleaning up.  If push comes to shove, it doesn't absolutely NEED to be applied - we get to cheat like that when we don't need it to be game-ready. :)

    To solve the poly count I'm considering turning down my subsurf levels on some parts.  I need it cranked up for the nice and close-in WIP shots naturally, but for the final composition some of the smaller pieces especially won't need it to be nearly that high.  I bet I could trim it down considerably that way.

    Okay, today I finished the second component - which also is a funky curvy piece but was really nice and simple to do, it only took me an hour and a half or so.  The window holes are cut completely through, I've got a piece in there that I intend to add an emission shader to so I can have nice lit windows for the end composition.

    And I've started working on the third piece.  This one IS another weird combination of odd flat angles and curves.  But hopefully I learned enough struggling with that first piece that it won't be too hard of a battle.

  • Hoo boy.

    This piece was less of a pain than the first one, but all that detail took a while.  A great while.

    Next up:

  • Okay, so I expected this part to go quickly - and it was really easy to block out, it only took minutes.  But just modeling those grid lines into the surface has taken me all day.  

    Like with the first component in the very beginning, I had to keep starting over and trying different things.  The subsurf modifier just does not like to make things easy when chiseling small, sharp-cornered details like this into curved surfaces.  The bevel modifier really helps, generally, but it also has issues when used in combination with a shrinkwrap.  Surrounding topology is super important if you want all your nice smooth outside curves to stay nice and smooth and curvy.  You can't just add as many holding edges as you need to fix subsurf artifacts because even when using a shrinkwrap, too many edges too close together WILL create unwanted effects on non-planar surfaces.

    The way I got this working is, there are no faces in the bottom of those grooves - only sidewalls.  The bottom is filled with a second mesh piece that's just a copy of the shape of the flat, unaltered surface shape that's been separated and moved downwards slightly (or upwards, in the case of the grooves on the bottom surface), and intersecting the sidewalls.   That completely eliminates subsurf artifacting in the bottoms of the grooves, and left me only needing to deal with the hull surface topology.

    Seems like a whole day wasted for such a simple solution in the end, but I suppose it wasn't wasted if I learned something.

    Tomorrow I need to detail the rim and cut some windows there and on the bottom surface, and make the gizmo that caps the bottom of this part of the hull, and the whole dish-section will be complete.  Still plenty of work to go....

    • jjakeblended What matters is you found a solid solution, and you will never ever forget it because you worked so hard to get there! Will always be on the back of your mind if you run into similar issues in the future, so I say your time is well spend 😊👍🏻

      Can’t wait to see the pieces together! 

  • Sorry about no screenshots last night; had some internet issues.  Nevertheless the work got done, with the finishing of all the windows, details, and the gizmo on the bottom, the entire saucer part of the ship is finished, modeling-wise.

    I'm starting to love Booleans the way Oscar loves trash.  

    Anyway, I *believe* this marks about the halfway point in all the modeling.  I could be wrong though; we'll see.  I'm confident this project could be finished by the contest due date.  In time for the live critique, though?  Not so sure...I'll have to really hustle, but I MIGHT be able to squeeze it in...

    So now, moving on down the model.  Next part up is the neck.

  • Thought I'd finished the upper neck section between yesterday and tonight, but the more I look at it, I'm seeing some unevenness on the surface that I think I want to fix.  I'll do that tomorrow, and then start on the lower neck section.

  • I smoothed out the surface of the upper portion and completed almost all of the lower portion, the torpedo bay:

    The windows are Booleans as usual, but the airlock has interior detailing so I couldn't cheat on it, I had to actually cut it into the mesh.  Luckily I only had to cut the hole itself; all the airlock geometry I just copied and merged in from the one I already modeled on the back of the bridge, early on in the project.  I'll have to do this one more time, as there's another airlock on each side of the lower hull section as well.

    Of the lower neck section, I just have to make the front of the torpedo launcher.  It's not exactly greeble-y, but it's got a lot of small details and will take a little bit of time to model, but I think I should be able to finish it tomorrow.  

    After that, it's on to the secondary hull.  I'm on schedule to finish by the contest due date, but very apprehensive about making it in time for the critique.  It'll be close if I make it.  Expecting some rough weather over the weekend too, that might cause problems.  But, I'm thinking happy thoughts!

  • All right, launcher detailing done, and with that the neck section is complete.  

    LookDev mode makes that look very satisfying!  I love LookDev mode.  I love so darn much about 2.8.

    To be honest, I expected to be able to do more on the next section than just lay out the guides and a few frame circles, but that's all I was able to do today.  This part might take me a couple of days - but I'm still hustling as best I can to make the critique, keeping in mind I still have to shade this mess AND build an actual scene by the 21st.  I think I can I think I can....

    • jjakeblended How are you maintaining correct topology?  It seems like good edge flow would be impossible with all those shapes stuck into other shapes.  Could I please see the wireframe?

    • williamatics Sure:

      The "secret" is, it's not one contiguous mesh.  It's all the same mesh object, as far as Blender is concerned (well, two objects - the small tubular parts together make their own second object); but the object is composed of separate mesh pieces whose surfaces intersect when they're put together but aren't actually connected to each other.  

      I can get away with this in this case because this model is not intended to be a game asset. As you can see from the earlier screenshot, Blender's lighting interprets this kind of arrangement just fine, with no weird artifacts.  And since the separate pieces are all "joined" to the same object, your modifiers still work on all of them as intended, with the only real exception being that your bevel modifier won't create bevels where separate parts intersect, so you'd have to actually model a bevel into one of the pieces if you wanted that.  

      I use this same general method for most of the ship components.

      Although usually each "piece" is actually a separate object, rather than separate meshes joined into a single object.   The only reason I did that for the torpedo launcher is because it's made up of so many different little pieces, it would clutter up the Outliner, plus I'd have to think of names for all those individual little bits.  Joining them all into only one or two objects (without actually merging the meshes together) was more of an organizational choice on my part.  The other thing you have to consider is that all separate mesh pieces that are joined into a single object have to share the same modifiers.  That's why I didn't join the little tubes into the rest, but made them their own object - I didn't want them using a Bevel modifier like the rest of the pieces.  You could also join them anyway but use a vertex group to exclude them; either way works.

      For windows, I use objects modeled into the shapes of the windows I want, combined with a Boolean modifier.  See here:

      Here, the "Saucer Rim" object has the Boolean modifier, with the windows object as the affecting object.  The "trick" is, the Boolean modifier is never applied; like the Subsurf modifier, it just lives permanently in the modifier stack.  The windows object is simply turned invisible.  So the mesh of the Saucer Rim object is never actually cut - the principle of non-destructive modeling.  The downside to this method is, when it's time to UV-unwrap the mesh for texturing, those windows don't actually "exist" and therefore they don't appear in your UV map.  You can't texture the inside of the Boolean cut differently from the rest of the object.  That's okay for me in this application because only a sliver of the "window frame" is actually visible and it would be the same material as the rest of the spaceship hull around it anyway.  But if you need details inside your cuts, then you need to actually apply the Boolean and clean up your mesh issues the "hard way".

      For the window glass, I use basically a copy of the surface the Booleans are cut into, separated and scaled back just slightly so it intersects the cut, like so:

      This both reduces the profile of the window cuts and solves the problem of the fact that the window glass does need to be a different material from the rest of the hull.  When it's time to shade, I can give this "glass" object some emission and presto - lighted windows.

      The Booleans plus the glass:

  • I got an okay amount of work done on the lower section today.

    Tomorrow I need to finish this part.  I have to at least get the front and back end details done, and the windows.  If I'm lucky, I might be able to start on the engine struts.  I'm trying my best to work as fast as possible without leaving anything out.  I've hit something like a stride workflow-wise, but the calendar is my mortal enemy at the moment and it's a fairly even match.

  • Well I did not get as much done as I wanted today.  Been dealing with electricity issues all day from the tropical storm - I'm not in an area where I'm in any danger, but my local power grid is just kind of dodgy and susceptible to problems when it rains all day long.  Still, I got SOME work done.  I finished the front details:

    And I got the tail all blocked in at the very least.

    But there's a lot of greeble-y stuff on either side of those doors that I have to model; and I didn't even get started on the windows and other small body details.  Have to get all that done tomorrow.

  • Okay, tail end detailing done, along with all the windows, weapon turrets, and other bits and bobs.  Secondary hull is complete.

    Next up, the engine towers.  The final parts!  Let's see if I can get them done in only a couple of days.  Still a long-shot to have a fully completed scene by the 21st, but not giving up yet.

  • All right.  Finished the struts and working on the engines now.  

    If I'm really lucky, I might have this last part finished tomorrow!   If I do, then I'll go back over the whole thing to make sure I didn't forget anything, or miss any stray bad geometry.   Can't wait to grab some screenshots of the finished model!

    Then, it'll be shading time.   I've got my work cut out for me, if I want my shading to be as accurate as I've tried to make the mesh...