Add transcription for: week06 14 building a low res cage pt2.wav

transcriptions/week06 14 building a low res cage pt2_transcription.json

@@ -0,0 +1,8 @@
+{
+  "text": " Okay, so we are back for the second part of creating a lower-risk cage. I have created this cage here now. And you can see it also includes the fingers. And as we talked about before, I've tried to make everywhere where we'll get some bending. I've tried to add three edges, three edge loops, so here in the elbow area, the shoulder area. I've added a few more here actually. And then the spine for each of these spine joints, I've added one loop, probably not matching 100%. So perhaps I might have to move them around a little bit more. Also added some more in here. But what's important here to see or to notice that it's pretty much all quads. There are no triangles and no five-sided polygons. It's all made out of quads. Actually, here might be one that is five-sided. If we look at just that, we could probably try to resolve that too. Maybe here we have two five-sided ones. So I think we actually more. There's also one. So we can perhaps try to resolve that as well. Make this a little bit different. Actually, there are a couple of five-sided ones, so forget what I said. But in the areas where we will get a lot of deformations, we have all quads, and that's really what matters. And you will also notice that we have one line that goes like from the bottom of the arm to the side of the body all the way down. Same thing with the front here. So we have one in the middle for one side of the body and it goes all the way down to the foot here. Same thing on the other side. And here we have one on the top which basically goes all the way across over the head and then comes down on the other side, also on the top here. same thing for the back, if one continuous line going through. So that will make for good deformations. And now we can try to use that to skin to the joints. And then we'll use that as an influence object. The other thing that I wanted to point out here is that if we show everything again, I did not only snap it as I did before. So before we said, OK, we're going to make the main geo here be our be live. So we had it set to as a live surface. And then we said, okay, now we're using quadraw. Come back here, Polygons, mesh, quadraw tool. And then I can move these points around and it will snap to the surface. But what I did instead is I moved those points afterwards once I was done with everything. I moved those points out a little bit so that they very closely, you know, the subdivided surface area, the smooth preview should match pretty closely to our skin. The closer the matches, the less the offset we have, the better the wrap deformable work and the better the formations we will get the end out of it. So we have two surfaces that are of different topology, different resolution, but in terms of the shapes they're matching quite closely. And now we can try to skin that. Before we do, I'll make two small adjustments inside the GUI for a sec. The thing that I like to do is I like to color that a little bit differently. We don't have to do that, but I typically like to create a shader. You can think of this as like a flesh layer. So we can actually make this red. Again, we don't have to. I did it in the past. Let's create a blend here maybe, and assign it to the Geo. And then let's make this red. What we can also do is I might also like to make the wireframe a little bit different, So that when we view it in BioFrame, we can see a difference between the regular skin, if we turn the live object off, between the regular skin and where our main or low-risk cages, so we'll select a mesh low here and go into the attribute editor. We could potentially also set it on a different layer, but I'll just go in here and I will enable the overrides. You can probably do it in the shape too, enable the overrides, drawing overrides, enable overrides and we'll set the wireframe color to red that will make it clear that we have two different mesh-ish here, models. And this is our flesh model. Sometimes what I also like to do is make it a little bit transparent. We don't necessarily have to do that, but maybe later on. Okay, so now here we have our flesh and a little risk-age. And now we want to skin that. So we select our skinning joins, and we can use the same script that we used before. I think we put it on a shelf here. So we're selecting all the joins. Actually let's hide this here, run it again. So we're selecting all the joins, all the drain T's, deselecting all the end joints, deselecting anything that is IK in it, anything that is FK in it, anything that has refK in it, and also DI joins. We're selecting our low-risk age now and we'll skin that under Animation menu, Skin, Smooth Bind, still to Select, Draw and still Interactive. And then we'll have to skin that a little bit differently. To come in here for arm. It's kind of like the same workflow that we had before. I'm surprised it doesn't work better out of the box. But, we can maybe try to do is we detach this one more time. try to run our script here again. Select the skin, smooth bind and see if we can change the maximum influences to 2. Maintain max influences off and we'll let it apply here and see if we get a better starting point from that if it only can have two joints, not really. Then we want to come in here and paint all that to this blend joint here, blend mid. 100%. The great thing about this is now that all of these edges will be very clear what they are going to be weighted towards that one here too. And then we have some partial joints here. Here we have the one in the middle. So we can actually go to select the whole loop and weighted 100% to the partial joint. Not this one, the main one here in the middle. And we can even use flood here, flood, 100%. OK. And then those guys, they should probably be weighted to this partial joint. I think that is partial joint as B probably. We'll also weight it 100%. And then those ones here will select these, the whole loop. I'll wait until partial join A. And a percent. Here we go. And then this one will be weighted to the upper arm or root. Okay. Just like that we continue stepping through and doing all of our waiting here on this low-risk cage. I won't do it all here in the video otherwise it will take too long, but I guess you get the idea. So here I think we were missing a partial joint. We haven't added one yet. So let's do that now. That's why it's usually a good idea to color code. Let's create an for our low res layer. Add that on here. That's why it's a good idea to color code your partial joints, and maybe also scale them up a little bit. We have quite small joint size. Because then it will make it very easy to see where we have partial joints partial joints and where we are missing them still. So let's maybe take this one, duplicate it and parent it to this joint here, parent and set it all to zero. And that's the joint four right there. And we can connect this at through a multiply, divide node and add half of the rotation of the wrist to this partial trine. I appreciate, bring those in here, multiply and divide. I will connect the rotation, 10 put one, and I will multiply with 0.5 and that gives us the rotation that we will need to apply to our partial 50% joint. Okay. Seems to be working fine. Then skin that guy here. So that whole roll loop should be weighted to the arm joint, the mid, and that whole roll. It should be weighted to the hand, to the wrist 100%. Wrists are, I think we call that hand joined actually. Or just this joint right here, which follows 100% with the hand, because it's just parented under there. So we'll weight that 100%. And then this one here we'll go to the partial joint. Actually, we should probably rename this to... I'm having some issues here selecting this properly. I'm not sure why. Here we go. I think this was set wrong. OK, so here we have our partial joint, which at the moment is still called arm. Something, so we have to call this hand partial, or wrist partial joint. And that has to go, so now we can paint. Change the weight in here for the middle area. So this here should all be hand. And this one here should also be hand. And then the 50% one or the one in between should be on the partial. So hand partial. Oh, we didn't add it as an influence yet. So let's do that. Send it to zero. Get this joined. And the skin. Go to Add Influence. Not use Dramatry because it's not, it's a joint and unlock weights, apply. Now it's in there. Now we should be able to paint to it. Here we go. We can unlock it and we can paint the middle here 100% to the partial joint, the wrist. Okay, let's see how it looks. We bent the arm down or the hand down. Let's hide the joints here for a sec. So we can see we're getting something that looks quite nice. Keeps the volume nicely here of our wrist. We bend it all the way down, and we bend it all the way up. It looks quite nice. And plus on top of that, as I told you before, if we come in here, we can also scale the joined up if we wanted to. To that partial join, we could translate now. Or we could scale it up, down, if we want to take saturated or scale and translate, for example, and move it around that way. I think even without that, it doesn't look too bad. We should be able to also use side motion. Actually, here I think something is wrong in terms of the connections. Side motion y, that should probably connect to a different rotation on the partial joint. not the y itself, because that will be twisting for this partial joint. It should rather be x movement, so let's fix that. So if this goes left and right, so y, I want to connect y to the rotation of the joint. Make this unselectable here. The joint will be rotation x because it's rotating around the red one. So y, we connect to x. output y goes to x. And then the x should probably go to y. That is correct. So y is working now. We're getting this 50% rotation from the partial joint. That looks good. And for twisting, we have to probably work on it twisting anyways, but even here you can already see that the partial join is rotating in the wrong direction. So if we rotate it this way, the partial join goes the opposite way. So we should probably multiply that with minus. So the x channel will set it to minus 0.5, and then it will rotate in the correct direction with 50%. But what we'll probably have to do is we'll probably have to add another join so that we can twist the whole flash here on the lower arm around. So let's add that. So here what I'll do is I'll probably also take just this one, duplicate it, and parent it to the arm mid-joint. Let's see where the other one is or where it's at a moment. Actually, these are under FK. They probably shouldn't be under FK. They probably should under the blended arm. So blended arm and also not the end, but rather the mid. So blend arm mid, apparent them there. Because then they will go along for the right in both modes, in IK in an FK, which is what we want. And now that additional partial joint under the mid. The first one is for the hand. And the second one, what we will do is you will notice that we have a translation in Y now, now that is a child of the blend mid joint. This is how much it had to translate to go to this position. So if we set this back or move it back, you can see as we decrease it, it comes closer and closer to zero here. So if we set it all the way to zero, then it will be exactly in the same space as the parent, the blend mid. So what we can do with this is, we know that the distance from here, from the parent to the child is 20.754 units. So now if we want this drawing to be exactly in the middle, we can do is we can just divide that by two. So we just say divide equals two and then it will be exactly 50% between those two. And now this could become our twist join. So let's call this maybe low arm or arm low partial. It's more twisting. We also added So it has an influence to our low model here, add influence once again. Then we can come in here and we can change this to be partial. Here is our low partial joint and we'll skin all that to there. also this area down here because at the moment it will be weighted to the wrong thing. So if we set this to blend, or arm blend mid, arm blend mid, without a partial now. We can see that this obviously doesn't make sense, the arm mid is kind of here. So instead we want to weight all that also to the twist partial joint. Little arm partial. So weight that here as well. Now what we can do is, if this is rotating or twisting, we can connect half of that twist to this joint. But only for the twisting, because if this goes up and down, obviously we don't want this to change. Only for twisting up and down and left and right, nothing should happen. But for if we twist, then we want a half of the twist be applied here as well. So let's hook that up. So we'll select that twist partial joint, bring it in here. And then we'll select the hand, control, and bring it in here as well. And we'll add another multiplier because we only got half of that. Multi. I was, I think, rotate x. And then the output goes into, I believe, it's x2. for the twisting here you can apply so we go output x into rotate x and we have to divide it by 2 or multiply it with 0.5 looks like here is also something not correct just a zero. Now this seems to be the wrong axis. We actually have to probably be in this axis. It might be that it has some weird rotations on here. Here we can see it's not the same orientation as it's parent, actually it does. So it's twisting here. OK, so we have to connect the x to the twist, output x to the twist, which would be rotated y. And then we scale this with 0.5, multiply it. And now let's see what we get from there. OK, it goes into the opposite direction, wrong direction. so we multiply this with minus. Almost. For the twisting now we can probably apply a little bit more on this partial joint. We want it to twist a bit more. So we set the twisting here to 0.75. That I think looks a little bit better. And we can play now with these values. We can also apply a little bit less twisting on that partial joint. Select both of them and show the incoming connections. Probably too much here to see. Actually, let's just select the two joints, show the incoming connections here. So one is for our hand partial joint, and the other one is for our lower arm partial joint, arm low. So now we can play with this and set the multiplier to whatever we feel looks kind of correct. And we could potentially also add another partial joint in here, also for this one, so that we can control that separately. up down, in out, and the twisting, which takes the arm with it. Now this, for this case, might work or might not work. I think a better idea would probably be instead of relying on these rotations and these rotation values that we connected directly. It might not work so well. Actually here in this case it does work. It depends also a little bit on what your rotation, rotate order is set to on this control. So if we had to set to something else instead of x, y, z, then it might not work so well to rotate in both ways like rotate it or bend it down and then apply by the twisting here on the arm. But then what we should probably do is add a post reader in here, which is a lot more reliable than using the connections directly. It's kind of like the same thing that we already talked about if we set this to like 180 or something. This might be a bad example. Well, even here you can see it right now. I'm rotating this 360 degrees. And with this partial drawing, we're getting a lot of messed up stuff, because this is now thinking, oh, I'm rotated 180 degrees when, in fact, it's kind of like the same as if it was zero. So with a post reader, we would kind of get rid of this issue. Of course, you might say, well, but the animation really be rotating this 360 degrees anyway, which is correct. But just showing you once again that a post reader can be really, really helpful in a lot of these cases here. And what are you using? the one that I showed you, which you can do or build yourself manually in Maya, or whether you're using some sort of a plugin or what other people wrote. As long as you have a good way of reading the exact post that the body part is in, will be a lot more reliable and robust than connecting it directly to the rotation. It will take a little longer to set up though. And then we just go through and kind of like skin everything to you know those partial joints here for example for the fingers and so on. And we have to do a lot of cleanup here. But once we're done with all of that, and once we have everything kind and weight painted, which again will be a little bit easier than doing it. Everything should be working mostly okay. It will be easier to skin that low-res mesh here than if we were to skin our high-res mesh. And as I said before, our high-res mesh here that I'm using is not that bad, so we can still skin it fairly easily, but if you have a higher rest then obviously it will make much more sense. And then once you're done with everything then we'll just take the final geo and glue it on. So I'll just take this here and I'll delete all the history, so I'll set the skin cluster to zero and I'll set the corrective shapes to zero and I will delete all the history here on it so that there are no deformers anymore on our skin and then we will now only use our low-res mesh as an influence for the high-res mesh. And with this you could do two things. You could either kind of like select your skinning joints and select your skin, the final one, and skin it one more time, smooth bind to select the bones. And then what you could do is you could transfer the weighting that we did here. This already looks quite good by default, but you could take the low-res mesh that you've weights painted on, which is more easily to weights paint on, then high-res, and then you select the high-res and you just transfer the weights. Let's try that. Select this one. Select that one first. This one is kind of hard to select because they're really on top of each other. Here we go. Now I got it. So select the low rest first, then the high rest, and then you go to copy skin weights. And you can also Play with these closest components might give good results or closest point on surface. Probably better here, maybe raycast. And we hit that. And then it applies the skinning from the lower rest that your weights painted onto your high rest. Actually here it doesn't seem to be doing such a great job. Let's try the other one, or one other method. Let's try copy skin weights. point on surface. Let's see what we get here. Halls are not too great. Close this component. Probably also might not work too well. So I think it seems that transferring weights doesn't work too well, but I wanted to show you another method anyway. And I was using that as a wrap. So let's select the high res, detach the skin cluster. And instead, we will use the low res as a raptor former for the high res. If you've done raptor formers before in Maya, they're not very, very good. The raptor form in Maya is not very good. But what I found what improves the quality is if we smooth the driver. So if we smooth our mesh low here, and let's call this actually M mesh low GU. If we smooth that once under polygons, mesh smooth, either we subdivide it once or maybe even twice. Just keep in mind that the more we smooth it, the more points it will be and the slower it will probably get. But if something like that, which might be too much already, let's go one step back. This one here might be fine. I will be a little bit faster now. And then we use that version as a raptor former. Now if you smooth it, then you're losing the ability to kind of still change your So for that, what you can do is if we remove that smoothing effect and we go to delete non-deform history that we get read just off the poly smooth, but we get to keep our skin cluster. Let's try this again. Delete but type non-deform history. So that should keep the skin cluster on. So if you want to keep the ability to still change your weighting later on, what we could do is we could duplicate that out, and then we can connect the two meshes together. So the moment you can see one is deforming, the other one is not. So if we now select that and select this one and bring them into the node editor, for example or hyper shade and then we connect we can call the duplicated one mmesh-low-g01 we can call this mid maybe and then we connect actually we needed to shape nodes here so I'll just use the down arrow key that should Actually, it doesn't work here, it seems. Down arrow key while you're in the viewport here, or can obviously also go to the Outliner, reveal selected, and then just show shapes and get it that way. But we want to have the two shape nodes, here they are. And we go from the low mesh or from the original one that we had that has the skin cluster on it. the one that doesn't, the duplicated one, which we called mid. And then what we connect is we connect the out mesh from the skinned one into the in mesh of the other one. So we have to expand it all the way. Then we have out mesh up here. And then we have in mesh down there. So connect out mesh to in mesh. Now we should see both of them rotating and they do both rotate. Also the mid one, although the mid one doesn't have a skin cluster, it just has this incoming connection from one mesh to the other mesh. And now we can take the mid one and apply our smoothing on this one. Mesh. Smooth. Here we go. Okay, make this a set this to one or two or whatever we feel works for us. But now you can see we still have our old one that we can change the weighting on if we have to. And then we have the smoother version on our mid where we apply the smooth now. Let's hide the low res one and just work with this mid level. You can already see it's kind of like getting slower now because it has to evaluate a lot. So let's set the smoothing back to one, which will make it a little bit faster. might be sufficient, we have to check, we have to test. And I will select our skin and glue that to the smooth version. I'll wrap the former. I think we have to select actually the skin first, and then the influence. I believe it's kind of like the other way around. It's a little bit weird. Let's try this animation. and then create a former wrap. It will take a little while. Then we should have the skin following whatever the low-res is doing. You can also see inside that one. So now we don't have to do any skinning on the final skin. We did everything on the low-res version, but you can see, or maybe it's not so clear here when I'm demoing it, but it's getting pretty slow. So for animating that, it will be pretty painful, I guess, but what animators now can do is they can animate the mid or the low and only show the final skin when they absolutely need to see it. And even here on the mid we could potentially apply a skin-like shader instead of the red here that we did earlier. We try to find our skin, body skin. And then it's actually not that bad here in terms of speed. We haven't even done anything with those joints yet in terms of scaling. Actually, I think we did scale it a little. I think there is something a little bit weird going on. See if we can improve that a little bit. It seems to be still losing volume. I think I know why. Let's go to the low. Wait it a little bit differently. Instead of what I did here, let's try painting those, not to those outer partial joints. Paint skin. So I'll go to the blend mid joint and wait for this one. That's always, always helpful to create a little test NM. Perhaps 100% might be too much. Perhaps we have to wait at 50-50 between the joint and the partial joint. Although then I think we'll still lose volume. So let's try this. And then the root one can weight all that. So now we should be a little bit better here in terms of keeping that volume preserved. And you will see that whatever we're changing here or doing here, it will go through to the mid. And that will then go through to the final one, which we have here. You can see that it doesn't look too bad without doing any weighting on our final geo. Let's also try this one here too. See what we get there. Also doesn't look too bad. going up. And then as I said before, we can use that partial join to kind of scale it up a little bit or move it around based on that rotation that we're getting here. Move this up or down. So this pretty much is the workflow that I would be using to rig this character with a low-risk cage. The great thing about that is also that we can now do anything that we want in terms of deformation on this lower sketch so we can apply corrective plane shapes, for example, for our biceps bulging or additional joints, do any kind of deformations. And we're pretty simple, pretty easy. We can also do maybe even things like, for example, do things like muscles and stuff like that on this layer and will kind of all come through to the final to you. Again, one of the downsides is that the internal Maya raptiform is not very fast and you will also probably run into trouble when you have two things that are very, very close to each other when the geometry cannot figure out, okay, where do I need to bind to, to this finger, do I need to be raptiform by that finger, do I need to be raptiform by this finger. But if your geometry is pretty close like what I tried to do here between the low res and the final geo, then it usually does an okay job. Here obviously we have the low res already moving with, or I guess this is the low res, so we haven't done any weights painting on that one. So once you have that all done and set up, then it will actually work quite nicely. The other great thing here about this that I found is you're getting a lot of skin-like behavior for free. you you",
+  "segments": [
+    {
+      "text": " Okay, so we are back for the second part of creating a lower-risk cage. I have created this cage here now. And you can see it also includes the fingers. And as we talked about before, I've tried to make everywhere where we'll get some bending. I've tried to add three edges, three edge loops, so here in the elbow area, the shoulder area. I've added a few more here actually. And then the spine for each of these spine joints, I've added one loop, probably not matching 100%. So perhaps I might have to move them around a little bit more. Also added some more in here. But what's important here to see or to notice that it's pretty much all quads. There are no triangles and no five-sided polygons. It's all made out of quads. Actually, here might be one that is five-sided. If we look at just that, we could probably try to resolve that too. Maybe here we have two five-sided ones. So I think we actually more. There's also one. So we can perhaps try to resolve that as well. Make this a little bit different. Actually, there are a couple of five-sided ones, so forget what I said. But in the areas where we will get a lot of deformations, we have all quads, and that's really what matters. And you will also notice that we have one line that goes like from the bottom of the arm to the side of the body all the way down. Same thing with the front here. So we have one in the middle for one side of the body and it goes all the way down to the foot here. Same thing on the other side. And here we have one on the top which basically goes all the way across over the head and then comes down on the other side, also on the top here. same thing for the back, if one continuous line going through. So that will make for good deformations. And now we can try to use that to skin to the joints. And then we'll use that as an influence object. The other thing that I wanted to point out here is that if we show everything again, I did not only snap it as I did before. So before we said, OK, we're going to make the main geo here be our be live. So we had it set to as a live surface. And then we said, okay, now we're using quadraw. Come back here, Polygons, mesh, quadraw tool. And then I can move these points around and it will snap to the surface. But what I did instead is I moved those points afterwards once I was done with everything. I moved those points out a little bit so that they very closely, you know, the subdivided surface area, the smooth preview should match pretty closely to our skin. The closer the matches, the less the offset we have, the better the wrap deformable work and the better the formations we will get the end out of it. So we have two surfaces that are of different topology, different resolution, but in terms of the shapes they're matching quite closely. And now we can try to skin that. Before we do, I'll make two small adjustments inside the GUI for a sec. The thing that I like to do is I like to color that a little bit differently. We don't have to do that, but I typically like to create a shader. You can think of this as like a flesh layer. So we can actually make this red. Again, we don't have to. I did it in the past. Let's create a blend here maybe, and assign it to the Geo. And then let's make this red. What we can also do is I might also like to make the wireframe a little bit different, So that when we view it in BioFrame, we can see a difference between the regular skin, if we turn the live object off, between the regular skin and where our main or low-risk cages, so we'll select a mesh low here and go into the attribute editor. We could potentially also set it on a different layer, but I'll just go in here and I will enable the overrides. You can probably do it in the shape too, enable the overrides, drawing overrides, enable overrides and we'll set the wireframe color to red that will make it clear that we have two different mesh-ish here, models. And this is our flesh model. Sometimes what I also like to do is make it a little bit transparent. We don't necessarily have to do that, but maybe later on. Okay, so now here we have our flesh and a little risk-age. And now we want to skin that. So we select our skinning joins, and we can use the same script that we used before. I think we put it on a shelf here. So we're selecting all the joins. Actually let's hide this here, run it again. So we're selecting all the joins, all the drain T's, deselecting all the end joints, deselecting anything that is IK in it, anything that is FK in it, anything that has refK in it, and also DI joins. We're selecting our low-risk age now and we'll skin that under Animation menu, Skin, Smooth Bind, still to Select, Draw and still Interactive. And then we'll have to skin that a little bit differently. To come in here for arm. It's kind of like the same workflow that we had before. I'm surprised it doesn't work better out of the box. But, we can maybe try to do is we detach this one more time. try to run our script here again. Select the skin, smooth bind and see if we can change the maximum influences to 2. Maintain max influences off and we'll let it apply here and see if we get a better starting point from that if it only can have two joints, not really. Then we want to come in here and paint all that to this blend joint here, blend mid. 100%. The great thing about this is now that all of these edges will be very clear what they are going to be weighted towards that one here too. And then we have some partial joints here. Here we have the one in the middle. So we can actually go to select the whole loop and weighted 100% to the partial joint. Not this one, the main one here in the middle. And we can even use flood here, flood, 100%. OK. And then those guys, they should probably be weighted to this partial joint. I think that is partial joint as B probably. We'll also weight it 100%. And then those ones here will select these, the whole loop. I'll wait until partial join A. And a percent. Here we go. And then this one will be weighted to the upper arm or root. Okay. Just like that we continue stepping through and doing all of our waiting here on this low-risk cage. I won't do it all here in the video otherwise it will take too long, but I guess you get the idea. So here I think we were missing a partial joint. We haven't added one yet. So let's do that now. That's why it's usually a good idea to color code. Let's create an for our low res layer. Add that on here. That's why it's a good idea to color code your partial joints, and maybe also scale them up a little bit. We have quite small joint size. Because then it will make it very easy to see where we have partial joints partial joints and where we are missing them still. So let's maybe take this one, duplicate it and parent it to this joint here, parent and set it all to zero. And that's the joint four right there. And we can connect this at through a multiply, divide node and add half of the rotation of the wrist to this partial trine. I appreciate, bring those in here, multiply and divide. I will connect the rotation, 10 put one, and I will multiply with 0.5 and that gives us the rotation that we will need to apply to our partial 50% joint. Okay. Seems to be working fine. Then skin that guy here. So that whole roll loop should be weighted to the arm joint, the mid, and that whole roll. It should be weighted to the hand, to the wrist 100%. Wrists are, I think we call that hand joined actually. Or just this joint right here, which follows 100% with the hand, because it's just parented under there. So we'll weight that 100%. And then this one here we'll go to the partial joint. Actually, we should probably rename this to... I'm having some issues here selecting this properly. I'm not sure why. Here we go. I think this was set wrong. OK, so here we have our partial joint, which at the moment is still called arm. Something, so we have to call this hand partial, or wrist partial joint. And that has to go, so now we can paint. Change the weight in here for the middle area. So this here should all be hand. And this one here should also be hand. And then the 50% one or the one in between should be on the partial. So hand partial. Oh, we didn't add it as an influence yet. So let's do that. Send it to zero. Get this joined. And the skin. Go to Add Influence. Not use Dramatry because it's not, it's a joint and unlock weights, apply. Now it's in there. Now we should be able to paint to it. Here we go. We can unlock it and we can paint the middle here 100% to the partial joint, the wrist. Okay, let's see how it looks. We bent the arm down or the hand down. Let's hide the joints here for a sec. So we can see we're getting something that looks quite nice. Keeps the volume nicely here of our wrist. We bend it all the way down, and we bend it all the way up. It looks quite nice. And plus on top of that, as I told you before, if we come in here, we can also scale the joined up if we wanted to. To that partial join, we could translate now. Or we could scale it up, down, if we want to take saturated or scale and translate, for example, and move it around that way. I think even without that, it doesn't look too bad. We should be able to also use side motion. Actually, here I think something is wrong in terms of the connections. Side motion y, that should probably connect to a different rotation on the partial joint. not the y itself, because that will be twisting for this partial joint. It should rather be x movement, so let's fix that. So if this goes left and right, so y, I want to connect y to the rotation of the joint. Make this unselectable here. The joint will be rotation x because it's rotating around the red one. So y, we connect to x. output y goes to x. And then the x should probably go to y. That is correct. So y is working now. We're getting this 50% rotation from the partial joint. That looks good. And for twisting, we have to probably work on it twisting anyways, but even here you can already see that the partial join is rotating in the wrong direction. So if we rotate it this way, the partial join goes the opposite way. So we should probably multiply that with minus. So the x channel will set it to minus 0.5, and then it will rotate in the correct direction with 50%. But what we'll probably have to do is we'll probably have to add another join so that we can twist the whole flash here on the lower arm around. So let's add that. So here what I'll do is I'll probably also take just this one, duplicate it, and parent it to the arm mid-joint. Let's see where the other one is or where it's at a moment. Actually, these are under FK. They probably shouldn't be under FK. They probably should under the blended arm. So blended arm and also not the end, but rather the mid. So blend arm mid, apparent them there. Because then they will go along for the right in both modes, in IK in an FK, which is what we want. And now that additional partial joint under the mid. The first one is for the hand. And the second one, what we will do is you will notice that we have a translation in Y now, now that is a child of the blend mid joint. This is how much it had to translate to go to this position. So if we set this back or move it back, you can see as we decrease it, it comes closer and closer to zero here. So if we set it all the way to zero, then it will be exactly in the same space as the parent, the blend mid. So what we can do with this is, we know that the distance from here, from the parent to the child is 20.754 units. So now if we want this drawing to be exactly in the middle, we can do is we can just divide that by two. So we just say divide equals two and then it will be exactly 50% between those two. And now this could become our twist join. So let's call this maybe low arm or arm low partial. It's more twisting. We also added So it has an influence to our low model here, add influence once again. Then we can come in here and we can change this to be partial. Here is our low partial joint and we'll skin all that to there. also this area down here because at the moment it will be weighted to the wrong thing. So if we set this to blend, or arm blend mid, arm blend mid, without a partial now. We can see that this obviously doesn't make sense, the arm mid is kind of here. So instead we want to weight all that also to the twist partial joint. Little arm partial. So weight that here as well. Now what we can do is, if this is rotating or twisting, we can connect half of that twist to this joint. But only for the twisting, because if this goes up and down, obviously we don't want this to change. Only for twisting up and down and left and right, nothing should happen. But for if we twist, then we want a half of the twist be applied here as well. So let's hook that up. So we'll select that twist partial joint, bring it in here. And then we'll select the hand, control, and bring it in here as well. And we'll add another multiplier because we only got half of that. Multi. I was, I think, rotate x. And then the output goes into, I believe, it's x2. for the twisting here you can apply so we go output x into rotate x and we have to divide it by 2 or multiply it with 0.5 looks like here is also something not correct just a zero. Now this seems to be the wrong axis. We actually have to probably be in this axis. It might be that it has some weird rotations on here. Here we can see it's not the same orientation as it's parent, actually it does. So it's twisting here. OK, so we have to connect the x to the twist, output x to the twist, which would be rotated y. And then we scale this with 0.5, multiply it. And now let's see what we get from there. OK, it goes into the opposite direction, wrong direction. so we multiply this with minus. Almost. For the twisting now we can probably apply a little bit more on this partial joint. We want it to twist a bit more. So we set the twisting here to 0.75. That I think looks a little bit better. And we can play now with these values. We can also apply a little bit less twisting on that partial joint. Select both of them and show the incoming connections. Probably too much here to see. Actually, let's just select the two joints, show the incoming connections here. So one is for our hand partial joint, and the other one is for our lower arm partial joint, arm low. So now we can play with this and set the multiplier to whatever we feel looks kind of correct. And we could potentially also add another partial joint in here, also for this one, so that we can control that separately. up down, in out, and the twisting, which takes the arm with it. Now this, for this case, might work or might not work. I think a better idea would probably be instead of relying on these rotations and these rotation values that we connected directly. It might not work so well. Actually here in this case it does work. It depends also a little bit on what your rotation, rotate order is set to on this control. So if we had to set to something else instead of x, y, z, then it might not work so well to rotate in both ways like rotate it or bend it down and then apply by the twisting here on the arm. But then what we should probably do is add a post reader in here, which is a lot more reliable than using the connections directly. It's kind of like the same thing that we already talked about if we set this to like 180 or something. This might be a bad example. Well, even here you can see it right now. I'm rotating this 360 degrees. And with this partial drawing, we're getting a lot of messed up stuff, because this is now thinking, oh, I'm rotated 180 degrees when, in fact, it's kind of like the same as if it was zero. So with a post reader, we would kind of get rid of this issue. Of course, you might say, well, but the animation really be rotating this 360 degrees anyway, which is correct. But just showing you once again that a post reader can be really, really helpful in a lot of these cases here. And what are you using? the one that I showed you, which you can do or build yourself manually in Maya, or whether you're using some sort of a plugin or what other people wrote. As long as you have a good way of reading the exact post that the body part is in, will be a lot more reliable and robust than connecting it directly to the rotation. It will take a little longer to set up though. And then we just go through and kind of like skin everything to you know those partial joints here for example for the fingers and so on. And we have to do a lot of cleanup here. But once we're done with all of that, and once we have everything kind and weight painted, which again will be a little bit easier than doing it. Everything should be working mostly okay. It will be easier to skin that low-res mesh here than if we were to skin our high-res mesh. And as I said before, our high-res mesh here that I'm using is not that bad, so we can still skin it fairly easily, but if you have a higher rest then obviously it will make much more sense. And then once you're done with everything then we'll just take the final geo and glue it on. So I'll just take this here and I'll delete all the history, so I'll set the skin cluster to zero and I'll set the corrective shapes to zero and I will delete all the history here on it so that there are no deformers anymore on our skin and then we will now only use our low-res mesh as an influence for the high-res mesh. And with this you could do two things. You could either kind of like select your skinning joints and select your skin, the final one, and skin it one more time, smooth bind to select the bones. And then what you could do is you could transfer the weighting that we did here. This already looks quite good by default, but you could take the low-res mesh that you've weights painted on, which is more easily to weights paint on, then high-res, and then you select the high-res and you just transfer the weights. Let's try that. Select this one. Select that one first. This one is kind of hard to select because they're really on top of each other. Here we go. Now I got it. So select the low rest first, then the high rest, and then you go to copy skin weights. And you can also Play with these closest components might give good results or closest point on surface. Probably better here, maybe raycast. And we hit that. And then it applies the skinning from the lower rest that your weights painted onto your high rest. Actually here it doesn't seem to be doing such a great job. Let's try the other one, or one other method. Let's try copy skin weights. point on surface. Let's see what we get here. Halls are not too great. Close this component. Probably also might not work too well. So I think it seems that transferring weights doesn't work too well, but I wanted to show you another method anyway. And I was using that as a wrap. So let's select the high res, detach the skin cluster. And instead, we will use the low res as a raptor former for the high res. If you've done raptor formers before in Maya, they're not very, very good. The raptor form in Maya is not very good. But what I found what improves the quality is if we smooth the driver. So if we smooth our mesh low here, and let's call this actually M mesh low GU. If we smooth that once under polygons, mesh smooth, either we subdivide it once or maybe even twice. Just keep in mind that the more we smooth it, the more points it will be and the slower it will probably get. But if something like that, which might be too much already, let's go one step back. This one here might be fine. I will be a little bit faster now. And then we use that version as a raptor former. Now if you smooth it, then you're losing the ability to kind of still change your So for that, what you can do is if we remove that smoothing effect and we go to delete non-deform history that we get read just off the poly smooth, but we get to keep our skin cluster. Let's try this again. Delete but type non-deform history. So that should keep the skin cluster on. So if you want to keep the ability to still change your weighting later on, what we could do is we could duplicate that out, and then we can connect the two meshes together. So the moment you can see one is deforming, the other one is not. So if we now select that and select this one and bring them into the node editor, for example or hyper shade and then we connect we can call the duplicated one mmesh-low-g01 we can call this mid maybe and then we connect actually we needed to shape nodes here so I'll just use the down arrow key that should Actually, it doesn't work here, it seems. Down arrow key while you're in the viewport here, or can obviously also go to the Outliner, reveal selected, and then just show shapes and get it that way. But we want to have the two shape nodes, here they are. And we go from the low mesh or from the original one that we had that has the skin cluster on it. the one that doesn't, the duplicated one, which we called mid. And then what we connect is we connect the out mesh from the skinned one into the in mesh of the other one. So we have to expand it all the way. Then we have out mesh up here. And then we have in mesh down there. So connect out mesh to in mesh. Now we should see both of them rotating and they do both rotate. Also the mid one, although the mid one doesn't have a skin cluster, it just has this incoming connection from one mesh to the other mesh. And now we can take the mid one and apply our smoothing on this one. Mesh. Smooth. Here we go. Okay, make this a set this to one or two or whatever we feel works for us. But now you can see we still have our old one that we can change the weighting on if we have to. And then we have the smoother version on our mid where we apply the smooth now. Let's hide the low res one and just work with this mid level. You can already see it's kind of like getting slower now because it has to evaluate a lot. So let's set the smoothing back to one, which will make it a little bit faster. might be sufficient, we have to check, we have to test. And I will select our skin and glue that to the smooth version. I'll wrap the former. I think we have to select actually the skin first, and then the influence. I believe it's kind of like the other way around. It's a little bit weird. Let's try this animation. and then create a former wrap. It will take a little while. Then we should have the skin following whatever the low-res is doing. You can also see inside that one. So now we don't have to do any skinning on the final skin. We did everything on the low-res version, but you can see, or maybe it's not so clear here when I'm demoing it, but it's getting pretty slow. So for animating that, it will be pretty painful, I guess, but what animators now can do is they can animate the mid or the low and only show the final skin when they absolutely need to see it. And even here on the mid we could potentially apply a skin-like shader instead of the red here that we did earlier. We try to find our skin, body skin. And then it's actually not that bad here in terms of speed. We haven't even done anything with those joints yet in terms of scaling. Actually, I think we did scale it a little. I think there is something a little bit weird going on. See if we can improve that a little bit. It seems to be still losing volume. I think I know why. Let's go to the low. Wait it a little bit differently. Instead of what I did here, let's try painting those, not to those outer partial joints. Paint skin. So I'll go to the blend mid joint and wait for this one. That's always, always helpful to create a little test NM. Perhaps 100% might be too much. Perhaps we have to wait at 50-50 between the joint and the partial joint. Although then I think we'll still lose volume. So let's try this. And then the root one can weight all that. So now we should be a little bit better here in terms of keeping that volume preserved. And you will see that whatever we're changing here or doing here, it will go through to the mid. And that will then go through to the final one, which we have here. You can see that it doesn't look too bad without doing any weighting on our final geo. Let's also try this one here too. See what we get there. Also doesn't look too bad. going up. And then as I said before, we can use that partial join to kind of scale it up a little bit or move it around based on that rotation that we're getting here. Move this up or down. So this pretty much is the workflow that I would be using to rig this character with a low-risk cage. The great thing about that is also that we can now do anything that we want in terms of deformation on this lower sketch so we can apply corrective plane shapes, for example, for our biceps bulging or additional joints, do any kind of deformations. And we're pretty simple, pretty easy. We can also do maybe even things like, for example, do things like muscles and stuff like that on this layer and will kind of all come through to the final to you. Again, one of the downsides is that the internal Maya raptiform is not very fast and you will also probably run into trouble when you have two things that are very, very close to each other when the geometry cannot figure out, okay, where do I need to bind to, to this finger, do I need to be raptiform by that finger, do I need to be raptiform by this finger. But if your geometry is pretty close like what I tried to do here between the low res and the final geo, then it usually does an okay job. Here obviously we have the low res already moving with, or I guess this is the low res, so we haven't done any weights painting on that one. So once you have that all done and set up, then it will actually work quite nicely. The other great thing here about this that I found is you're getting a lot of skin-like behavior for free. you you"
+    }
+  ]
+}