Jon gave me some great critique regarding the 'stamping' in my walk cycle, suggesting that I drop the leg on the passing positions in order to give a more subtle step. First are a couple of attempts I made to correct the issue beforehand, then are the ones following his feedback where I tried to integrate his advice - all with varying success, but things that don't work are just as important as those that do. So here we go.
Let's blast through this:
Version 27
First, here's the default version, just so you can see what I'm trying to correct. I've made some slight alterations since V25 - couple of things regarding the arms, just to loosen them up a tad. The legs do look pretty off in comparison to the otherwise fairly casual arms. His legs have a definite 'kick' to them.
Version 28
Here, all I'd done is very slightly adjust the foot roll on the up position, trying to bring it a little lower and closer to the contact pose. It's only a very subtle change and it does appear to have smoothed things out by a fraction - but not really enough to be noticeable to anybody that hasn't been staring at it for 4 million hours.
Version 29
First attempt following Jon's feedback where I tried dropping both the legs on the contact pose and also bringing the leg on the contact a little closer to the up position. It didn't exactly work; the issue I'm having is that dropping the leg of the passing position, the foot scrapes or goes straight through the floor. I can adjust the foot roll to try and bring it back up but rotating it back too far (in favour of the previous pose) causes it to twist uncomfortably, and also raises the knee again - which puts me straight back to square one. Rotating it forward lessens the height of the knee but this is problematic in that the foot, rather than delaying and dragging, then looks very robotic as we can see above. It needs to drag a little for flexibility. It also means that the space between the trailing leg on the down and passing positions is very small, causing the leg to delay at the crossover.
I also had to adjust the bend in the trailing leg on the down position to compensate for the lower passing position - otherwise the knee and leg were bent more on the down position, causing a bit of a jump/'pop' as it tried to interpolate the two keys. Unfortunately lessening the bend in the knee now means there's too little movement to stretch equally across the frames - as a result the knees on both legs start to bend at exactly the same time and move almost in perfect synch, which looks a little peculiar and quite robotic to me?
It may also be me but I kind of think that dropping the leg looks a bit like he doesn't bring it up far enough - the walk just seems to lose a lot of the feeling of weight?
I'm actually wondering if the stamping is actually more to do with the large gap between the legs on the up and contact keys - closely observing it, I can see it almost looks as if he's kicking his leg out. I think perhaps I just need to alter those positions so that the legs are just a little closer together in terms of posing so that it's a bit smoother?
Version 31
All of these captures look exactly the same... it must look like I haven't changed anything! I guess it's one of those things; when you've been staring at something frame-by-frame for so long you can pinpoint every tiny little thing in each frame. Things that nobody else in their right mind would ever notice.
Anyway, here I just tried adjusting the contact pose very slightly so that the step was a little less wide, allowing the leg enough time to travel smoothly from one pose to the next. It seems to have gone some way to removing that 'kick,' but it also plops me back to square one where my step wasn't quite wide enough. I think that, really, it's the up position that needs to change more than anything. If I can just tinker with the foot roll and leg position I might be able to figure out a way to smooth it out a little.
Showing posts with label softimage. Show all posts
Showing posts with label softimage. Show all posts
Friday, 20 April 2012
Tuesday, 17 April 2012
Sneak cycle arm breakdown and reference
I'm still not totally sure on the whole arm movement thing so I've been examining some other reference a little more closely. I found this lovely example on Youtube:
The flexibility is amazing and puts me to shame. My cycle looks like arse in comparison.
Ignoring my bruised ego, the arm movement is very nice! I promptly shoved it into Photoshop and brutally ripped it apart for the purposes of analysis.
The contact poses are interesting because of the forearm placement. They're poised up and ready rather than lowered, as I'd been toying with, which I think was giving me problems as it was making things a bit 'busy.' It's also a nice pose in general — having the arms almost tensed like that gives a much more cautious look to the character. The wrists are very nice too!
As the character drops into the down position the torso moves forward and the elbow follows. The forearm relaxes very slightly and comes down.
As the body starts to rise again the forearm is brought closer to the chest and the elbow, again, starts to come back. The character rises into the 'up' position and the forearm again lowers very slightly.
The second contact is slightly further forward than the first one, to counter the leg.
As previously observed most of the shoulder movement seems to come from the twisting of the chest.
That diagram is a bit cluttered, so it's probably easier to see with the body:
Serves to give a better idea of the shoulder and elbow in relation to the body.
It seems to have been helpful, though now I'm worried I've mucked my cycle up completely and that this method won't work. It's certainly given me something to toy around with, though, which can only be a good thing.
The flexibility is amazing and puts me to shame. My cycle looks like arse in comparison.
Ignoring my bruised ego, the arm movement is very nice! I promptly shoved it into Photoshop and brutally ripped it apart for the purposes of analysis.
I've done my best to identify where the key poses in the cycle are — not sure if I hit them 100% accurately but they seem to be roughly in the right place. Obviously there are differences in my cycle and this one — the torso here is much more flexible but in a way it's a good thing. Viewing a more exaggerated animation means it's much easier to examine how arms are following the body.
The contact poses are interesting because of the forearm placement. They're poised up and ready rather than lowered, as I'd been toying with, which I think was giving me problems as it was making things a bit 'busy.' It's also a nice pose in general — having the arms almost tensed like that gives a much more cautious look to the character. The wrists are very nice too!
As the character drops into the down position the torso moves forward and the elbow follows. The forearm relaxes very slightly and comes down.
As the body starts to rise again the forearm is brought closer to the chest and the elbow, again, starts to come back. The character rises into the 'up' position and the forearm again lowers very slightly.
The second contact is slightly further forward than the first one, to counter the leg.
As previously observed most of the shoulder movement seems to come from the twisting of the chest.
That diagram is a bit cluttered, so it's probably easier to see with the body:
Serves to give a better idea of the shoulder and elbow in relation to the body.
It seems to have been helpful, though now I'm worried I've mucked my cycle up completely and that this method won't work. It's certainly given me something to toy around with, though, which can only be a good thing.
Friday, 6 April 2012
Moom sneak cycle V7 — V15
Version 7
First attempt at some head and arm movement. The head was sort of cobbled together and it's nowhere near done, so please disregard how choppy that looks. It was just annoying me floating there doing nothing... probably should have left it alone. It's distracting.
The arm is... okay, I guess, and I think if there was a little wrist overlap it would be nice. I find it very static, though, very lifeless, it just sort of bobbles around and doesn't do much. It also goes kind of against what I've been seeing on other sneak cycles — the forearm, I think, should tuck against the body as he drops into the down pose and then come down as he leans back.
Posing the arm is tricky, though — I'm having a little difficulty figuring out which control objects to keyframe for maximum effect. I end up adjusting all sorts of pieces — up vectors, shoulders, arm effector — and get the pose right, but then when I keyframe the next one the pieces don't necessarily flow smoothly so it looks alright but then when you view it from another perspective the shoulder's going crazy and the elbow is flapping like some sort of demented chicken-man.
Version 12
Jumping ahead a bit (previous versions consisted of me getting immeasurably frustrated with things and not making much progress) I've tried to introduce a little more flexibility into the arm by using shoulder rotation and getting it to swing back and forth a bit.
I think the back swing is okay, but the arm coming forward makes absolutely no sense. It just looks... weird. I can't even figure out what needs to change, really, all I know is that it doesn't really make sense. Maybe it's because the arm doesn't really counter his weight — it should be used for balance, more than anything, I think, and if he was really leaning all his weight back like that, with his leg and arm outstretched, he'd probably topple over. It's just not a natural thing to do.
Version 13
Oh goodness this is even worse. It makes a tiny bit more sense... sort of... Other than the fact that the arm appears to counter the "wrong" leg, but I don't know if that makes too much difference. In some sneak cycles it seems like the arms move together in synch, almost.
I don't think this would be too bad but there's a bit of a jerk forward. Looks really peculiar. I think that the elbow should probably come backwards, rather than forwards, then the forearm should sort of come up towards the chest?... which answers my earlier question about the legs opposing the arms...!
Version 14
Side
Perspective
And here we have an attempt at the arm opposing the leg. There ain't much overlap right now; quite stiff and robotic. I was mostly just trying to really peg the shoulder movement more than anything. I think it's looking... slightly better, on reflection, but there's just something bothering me about it.
The arm seems to make a little more sense now, at least! I'm still not totally happy with it - looks to me a bit like it either goes back too far or doesn't come forward enough. I might also have got it backwards, again! I think maybe it needs to oppose the legs, but I don't know if that will look weird?! My brain is so tired I don't even know what I'm doing. Logical thought escapes me…!
The more I look at the body movement the less I'm happy with it. It seems to bump around all over the place and I really want to get more flexibility in his spine, but I'm not quite sure how to go about doing that, exactly. I think if I mess around with the curvature of the chest and hips I might be able to come up with something!
Version 15
Side
Perspective
I switched the arm so it opposes the leg and I don't really know what to make of it?! It kinda looks weird to me? I'm not sure why, but I kind of prefer the other version. I don't know if it makes sense. I'm very tired.
The arms are just completely baffling me at this stage. I'm really struggling to understand how they're supposed to work — I think my brain is fried, I keep trying to break it down and analyse it and act it out myself but it just doesn't seem to be going in...
What's confusing me most, I guess? The arms kind-of-sort-of oppose the legs, but they don't go very far forward at all. Contrary to a regular walk cycle they appear to be at their furthest extension on the contact pose when the body is leaning back. As the body comes forward on the down position, they move forward as well, and are drawn close to the chest. Then the body rises and the forearms come down and start to move back as the body leans back to balance for the contact.
... Wow, that actually makes a lot more sense now that I've typed it down. I must be making a mountain out of a molehill.
To the Batmobile!
First attempt at some head and arm movement. The head was sort of cobbled together and it's nowhere near done, so please disregard how choppy that looks. It was just annoying me floating there doing nothing... probably should have left it alone. It's distracting.
The arm is... okay, I guess, and I think if there was a little wrist overlap it would be nice. I find it very static, though, very lifeless, it just sort of bobbles around and doesn't do much. It also goes kind of against what I've been seeing on other sneak cycles — the forearm, I think, should tuck against the body as he drops into the down pose and then come down as he leans back.
Posing the arm is tricky, though — I'm having a little difficulty figuring out which control objects to keyframe for maximum effect. I end up adjusting all sorts of pieces — up vectors, shoulders, arm effector — and get the pose right, but then when I keyframe the next one the pieces don't necessarily flow smoothly so it looks alright but then when you view it from another perspective the shoulder's going crazy and the elbow is flapping like some sort of demented chicken-man.
Version 12
Jumping ahead a bit (previous versions consisted of me getting immeasurably frustrated with things and not making much progress) I've tried to introduce a little more flexibility into the arm by using shoulder rotation and getting it to swing back and forth a bit.
I think the back swing is okay, but the arm coming forward makes absolutely no sense. It just looks... weird. I can't even figure out what needs to change, really, all I know is that it doesn't really make sense. Maybe it's because the arm doesn't really counter his weight — it should be used for balance, more than anything, I think, and if he was really leaning all his weight back like that, with his leg and arm outstretched, he'd probably topple over. It's just not a natural thing to do.
Version 13
Oh goodness this is even worse. It makes a tiny bit more sense... sort of... Other than the fact that the arm appears to counter the "wrong" leg, but I don't know if that makes too much difference. In some sneak cycles it seems like the arms move together in synch, almost.
I don't think this would be too bad but there's a bit of a jerk forward. Looks really peculiar. I think that the elbow should probably come backwards, rather than forwards, then the forearm should sort of come up towards the chest?... which answers my earlier question about the legs opposing the arms...!
Version 14
Side
Perspective
And here we have an attempt at the arm opposing the leg. There ain't much overlap right now; quite stiff and robotic. I was mostly just trying to really peg the shoulder movement more than anything. I think it's looking... slightly better, on reflection, but there's just something bothering me about it.
The arm seems to make a little more sense now, at least! I'm still not totally happy with it - looks to me a bit like it either goes back too far or doesn't come forward enough. I might also have got it backwards, again! I think maybe it needs to oppose the legs, but I don't know if that will look weird?! My brain is so tired I don't even know what I'm doing. Logical thought escapes me…!
The more I look at the body movement the less I'm happy with it. It seems to bump around all over the place and I really want to get more flexibility in his spine, but I'm not quite sure how to go about doing that, exactly. I think if I mess around with the curvature of the chest and hips I might be able to come up with something!
Version 15
Side
Perspective
I switched the arm so it opposes the leg and I don't really know what to make of it?! It kinda looks weird to me? I'm not sure why, but I kind of prefer the other version. I don't know if it makes sense. I'm very tired.
The arms are just completely baffling me at this stage. I'm really struggling to understand how they're supposed to work — I think my brain is fried, I keep trying to break it down and analyse it and act it out myself but it just doesn't seem to be going in...
What's confusing me most, I guess? The arms kind-of-sort-of oppose the legs, but they don't go very far forward at all. Contrary to a regular walk cycle they appear to be at their furthest extension on the contact pose when the body is leaning back. As the body comes forward on the down position, they move forward as well, and are drawn close to the chest. Then the body rises and the forearms come down and start to move back as the body leans back to balance for the contact.
... Wow, that actually makes a lot more sense now that I've typed it down. I must be making a mountain out of a molehill.
To the Batmobile!
Characterised walk cycle reference and planning
Taking a bit of a break from the standard walk cycle for a bit I decided to make a start with a more "characterised" walk cycle. I opted to do a "sneaky" style walk, because... well, I just did. There's plenty of reference already available too - The Animator's Survival Kit has quite a nice section that goes into the mechanics of sneaks fairly well:
It's not brilliant but it does serve to give you more of an idea about how the weight shifts in relation to the movement of the legs and so forth.
I also did a sketch of the key frames from the diagram - it's pretty much just a carbon copy but I found that having to draw the poses myself gave me a much better understanding of how the body was poised, how the hips were tilted, where the weight was, etc. Even if you're just copying it's always good to draw things yourself - it gets you actually thinking about the mechanics of what's going on.
Following these examples, I started blocking out the walk cycle using Softimage's "stepped" interpolation method. Being able to set up and view the key poses without Softimage's rather distracting automatic keyframes. It allows you to much more easily check the flow between each pose and quickly see where things don't align.
Getting the poses right was actually a lot trickier than I originally anticipated. Due to the way the rig is set up it's really tricky to try and get the curve in the spine just right — I tried all sorts of methods including some ridiculously over-complicated method involving translating his chest object to bring it in line with the hips, but it just looked weird. Ther doesn't seem to be any way to adjust how much he leans forward with this rig — tilting the hips and chest does not re-position his upper body at all, and translating his upper body control moves his entire body. As a result he doesn't lean forward quite as much as I'd like, but I think it's looking okay.
There are also some slight issues with the foot slipping around — his back leg goes back too far on what I think is the passing position, when he leans forward. It needs to stay stationary to give the impression that he's pushing himself forward.
Switching to splined interpolation allows us to see that his foot definitely slips back, and there's also a bit of a weird 'bump' in his torso around one of his contact positions. I think there's a stray translation keyframe in there somewhere that got left behind.
Similar to our conventional walk cycle, Richard Williams also provides us with a solid formula to work from:
I thought this would be an interesting kind of walk to attempt to emulate with the Moom character, especially with all the bending and twisting going on in the torso. I find the arms quite interesting too, I love the way they're posed almost timidly. They remind me of a little hamster! I think it would be possible to get some really nice head and wrist movement with this walk.
To save myself a great deal of stress, I decided to adapt this formula for my characterised walk cycle. Just copying the frames is no good by itself though, it doesn't really give you any indication of the timing or any real sense of the weight. Observing and performing the action yourself is always the best way to go.
There's a disappointing lack of reference footage online for this particular sneak, but I did find a short clip from the DVD set that kind of demonstrates it:
It's not brilliant but it does serve to give you more of an idea about how the weight shifts in relation to the movement of the legs and so forth.
I also did a sketch of the key frames from the diagram - it's pretty much just a carbon copy but I found that having to draw the poses myself gave me a much better understanding of how the body was poised, how the hips were tilted, where the weight was, etc. Even if you're just copying it's always good to draw things yourself - it gets you actually thinking about the mechanics of what's going on.
Following these examples, I started blocking out the walk cycle using Softimage's "stepped" interpolation method. Being able to set up and view the key poses without Softimage's rather distracting automatic keyframes. It allows you to much more easily check the flow between each pose and quickly see where things don't align.
Getting the poses right was actually a lot trickier than I originally anticipated. Due to the way the rig is set up it's really tricky to try and get the curve in the spine just right — I tried all sorts of methods including some ridiculously over-complicated method involving translating his chest object to bring it in line with the hips, but it just looked weird. Ther doesn't seem to be any way to adjust how much he leans forward with this rig — tilting the hips and chest does not re-position his upper body at all, and translating his upper body control moves his entire body. As a result he doesn't lean forward quite as much as I'd like, but I think it's looking okay.
There are also some slight issues with the foot slipping around — his back leg goes back too far on what I think is the passing position, when he leans forward. It needs to stay stationary to give the impression that he's pushing himself forward.
Switching to splined interpolation allows us to see that his foot definitely slips back, and there's also a bit of a weird 'bump' in his torso around one of his contact positions. I think there's a stray translation keyframe in there somewhere that got left behind.
Wednesday, 4 April 2012
Moom walk cycle [upper block] V17 & V22
I was really scratching my head over my previously posted capture - no matter what I did I simply could not nail it and the arm just looked utterly wrong. I am ashamed to admit how long it took me to spot the mistake, and rest assured my face glowed like the setting sun once I'd clocked it.
I had the arm swinging with the wrong leg. I got it backwards. Amateur mistake.
As soon as I'd finished bashing my head against the wall for not spotting it sooner, I flipped it to the other side and it instantly looked a billion times better:
Version 17
More to the point, it was much easier to "read" for mistakes. It's still very much a work in progress, but simply fixing that arm has made one hell of a difference.
What followed was a series of tedious and very samey captures that I can't really talk much about aside from saying "the arm is too far forward," or "his shoulder pops here." I don't know, maybe that's valuable, but I talk too much as it is. I find it much easier to reflect/critique this stuff when there're more changes, so I'm just going to jump ahead a few versions.
Version 22
Lots of minor tweaks to the position of the arm here. The biggest change in this version is the shoulder - I wiped all previous keyframes and approached it in a much more straightforward manner. Before, I keyframed almost every aspect of it by hand which led to a slew of inconsistencies in speed and spacing that I believe may have contributed to some of the random limb popping and slight jerkiness in the previous versions. All I did this time is slap a couple of keyframes at the furthest points of its swing, then added a one to the first contact pose to have it move in a slightly smaller increment before it swung forward, giving it a little delay. I think it looks much, much better now, and the arc of the arm is clearly visible.
I think it's safe to say that the blocking of the arm is done at this stage. It's still not perfect, of course, but I think I'm ready to add in the opposing arm swing and also start looking at fine-tuning the timings and getting some wrist movement in there. The head is too static, too, so I'd like to get a bit of follow through going on that. I may also return to the legs a bit later as well and see if I can do anything to improve those.
I had the arm swinging with the wrong leg. I got it backwards. Amateur mistake.
As soon as I'd finished bashing my head against the wall for not spotting it sooner, I flipped it to the other side and it instantly looked a billion times better:
Version 17
More to the point, it was much easier to "read" for mistakes. It's still very much a work in progress, but simply fixing that arm has made one hell of a difference.
What followed was a series of tedious and very samey captures that I can't really talk much about aside from saying "the arm is too far forward," or "his shoulder pops here." I don't know, maybe that's valuable, but I talk too much as it is. I find it much easier to reflect/critique this stuff when there're more changes, so I'm just going to jump ahead a few versions.
Version 22
Lots of minor tweaks to the position of the arm here. The biggest change in this version is the shoulder - I wiped all previous keyframes and approached it in a much more straightforward manner. Before, I keyframed almost every aspect of it by hand which led to a slew of inconsistencies in speed and spacing that I believe may have contributed to some of the random limb popping and slight jerkiness in the previous versions. All I did this time is slap a couple of keyframes at the furthest points of its swing, then added a one to the first contact pose to have it move in a slightly smaller increment before it swung forward, giving it a little delay. I think it looks much, much better now, and the arc of the arm is clearly visible.
I think it's safe to say that the blocking of the arm is done at this stage. It's still not perfect, of course, but I think I'm ready to add in the opposing arm swing and also start looking at fine-tuning the timings and getting some wrist movement in there. The head is too static, too, so I'd like to get a bit of follow through going on that. I may also return to the legs a bit later as well and see if I can do anything to improve those.
Friday, 23 March 2012
Shuttlecock V7
I have no idea what happened to version 6; either it got overwritten or (more likely) my naming conventions suck and I accidentally skipped a number. Whoops. I'll keep it labelled thus for now until I figure out what happened. Preferably when I'm not so tired.
I was getting very flustered by this point so I scurried off and found yet another reference video. This one, I must admit, is pretty brilliant.
I cut it from what I think was a student film (11 minutes long!), and unfortunately they put some sort of peculiar stagger effect on it, so it kind of jars the image. If you spend long enough scrubbing through it you can sort of get a feel for the overall motion, though.
Immediately it's apparent where I was going wrong last time - you can see that the shuttlecock hits the ground and flips forward before landing on the corner, which then knocks the cork towards the ground, where it knocks the shaft back again and so on. Very, very, VERY similar to the motion in the previous video, so I was able to put two and two together and kind of understand the overall mechanics of the thing. Subsequently:
Version 7
Side
Medium closeup
Now that's better
The new reference video clearly helped, I think. It's not perfect by any stretch of the imagination but I think it's looking a lot better.
I think it possibly bounces a little too far forward (though I suppose it depends on the force of the launch) and perhaps it could stand to bounce a little higher?
The other thing I notice is that maybe it rotates a little too much. Currently, as it flies through the air, one of the corners brushes the ground but it keeps rotating until the other corner hits the floor, then it starts bouncing - I could have it bounce off the first corner and it might look a little better. If that makes sense.
After it bounces off the corner and lands on the cork, it kind of skids along the ground before rotating back onto the feathers - I think I probably need to have it bounce a little more instead of skid.
But, other than that, I'm feeling a little happier. I'm concerned with how simple it is... it just sort of bounces in a straight line. I feel like I should fancy it up a bit and have it bounce or rotate in a slightly different direction but I don't know. Maybe that's not important. I'll try to worry about that later.
I was getting very flustered by this point so I scurried off and found yet another reference video. This one, I must admit, is pretty brilliant.
I cut it from what I think was a student film (11 minutes long!), and unfortunately they put some sort of peculiar stagger effect on it, so it kind of jars the image. If you spend long enough scrubbing through it you can sort of get a feel for the overall motion, though.
Immediately it's apparent where I was going wrong last time - you can see that the shuttlecock hits the ground and flips forward before landing on the corner, which then knocks the cork towards the ground, where it knocks the shaft back again and so on. Very, very, VERY similar to the motion in the previous video, so I was able to put two and two together and kind of understand the overall mechanics of the thing. Subsequently:
Version 7
Side
Medium closeup
Now that's better
The new reference video clearly helped, I think. It's not perfect by any stretch of the imagination but I think it's looking a lot better.
I think it possibly bounces a little too far forward (though I suppose it depends on the force of the launch) and perhaps it could stand to bounce a little higher?
The other thing I notice is that maybe it rotates a little too much. Currently, as it flies through the air, one of the corners brushes the ground but it keeps rotating until the other corner hits the floor, then it starts bouncing - I could have it bounce off the first corner and it might look a little better. If that makes sense.
After it bounces off the corner and lands on the cork, it kind of skids along the ground before rotating back onto the feathers - I think I probably need to have it bounce a little more instead of skid.
But, other than that, I'm feeling a little happier. I'm concerned with how simple it is... it just sort of bounces in a straight line. I feel like I should fancy it up a bit and have it bounce or rotate in a slightly different direction but I don't know. Maybe that's not important. I'll try to worry about that later.
Shuttlecock V5
Version 5
Oh dear...
Here I started trying to add the rotation to the shuttlecock... you can tell I wasn't thinking clearly. It makes absolutely no sense whatsoever. Though, to be fair, I'm typing this after watching a much clearer reference video (which I'll post shortly) so I can now understand what was happening in that blurry video - when I originally animated this one all I had to go on was the super tiny video and had to guesstimate most of it!
First big problem; why does it rotate to the side after the first bounce? It should, logically, have rotated forward, landed on the corner of the shaft and then flipped over onto its side. It should then probably have bounced along on its side, or hit the corner again and then bounced over to the right.
It's floaty and has no physics. I opted to just give up on this one; it was doing my head in and I needed more footage. Back to Youtube...
Oh dear...
Here I started trying to add the rotation to the shuttlecock... you can tell I wasn't thinking clearly. It makes absolutely no sense whatsoever. Though, to be fair, I'm typing this after watching a much clearer reference video (which I'll post shortly) so I can now understand what was happening in that blurry video - when I originally animated this one all I had to go on was the super tiny video and had to guesstimate most of it!
First big problem; why does it rotate to the side after the first bounce? It should, logically, have rotated forward, landed on the corner of the shaft and then flipped over onto its side. It should then probably have bounced along on its side, or hit the corner again and then bounced over to the right.
It's floaty and has no physics. I opted to just give up on this one; it was doing my head in and I needed more footage. Back to Youtube...
Shuttlecock V4
I took a step back this time. I'm still feeling very anxious and overwhelmed by this whole thing - seems Softimage really isn't my forte - and I'm just beginning to panic a little about the amount of work I still have to do... logically I know I'm not behind but I'm so desperate to get this part of the brief over and done with so I can just concentrate on the next bit. Unfortunately that clashes with my desire to be absolutely perfect.
Agghh.
So, anyway, I was poking around on Youtube and I managed to find a couple of reference videos that I thought might help give me more of an indication of the shuttlecock's movement as it strikes the ground.
I took this one from Will's blog (sorry Will!!) and trimmed it down so I was able to loop the point of interest. It's a little difficult to see, but you can kind of get a feel for the motion - looks to me as if it lands cork-down, then bounces to the side and lands on the feathers. When the feathers strike the ground, it bounces back forward and is knocked in an arc, landing back on its other side where it bounces very slightly and rolls around a little until it comes to a stop.
Rather similar to the observations I made from the video uploaded onto myUCA. The main difference here is the direction it bounces and lands caused by the trajectory/angle it lands at.
Version 4
I tried to mimic the overall path of the bounce, opting not to tinker with any rotation at this stage. The speed and distance travelled in each bounce are a little off, but I did try to match the overall timing with the video as best as I could. I wasn't sure how I felt about it at the time - I was a little hazy and still getting myself very worked up over it, so I had trouble seeing whether it had potential to work or if I should just have scrapped it entirely and started fresh.
Looking at it now I think that perhaps it bounces forward just a little too far, and not far enough when it bounces from left to right. I think that it could have stood to bounce a little higher, too, bearing in mind that it was going to be rotating forward to land on the tip.
See? This is why I need to calm down and take breaks. Looking at it with fresh eyes I can much more easily see what could have been fixed, rather than rushing straight ahead and making a billion different half-finished versions. Oh well...
Agghh.
So, anyway, I was poking around on Youtube and I managed to find a couple of reference videos that I thought might help give me more of an indication of the shuttlecock's movement as it strikes the ground.
I took this one from Will's blog (sorry Will!!) and trimmed it down so I was able to loop the point of interest. It's a little difficult to see, but you can kind of get a feel for the motion - looks to me as if it lands cork-down, then bounces to the side and lands on the feathers. When the feathers strike the ground, it bounces back forward and is knocked in an arc, landing back on its other side where it bounces very slightly and rolls around a little until it comes to a stop.
Rather similar to the observations I made from the video uploaded onto myUCA. The main difference here is the direction it bounces and lands caused by the trajectory/angle it lands at.
Version 4
I tried to mimic the overall path of the bounce, opting not to tinker with any rotation at this stage. The speed and distance travelled in each bounce are a little off, but I did try to match the overall timing with the video as best as I could. I wasn't sure how I felt about it at the time - I was a little hazy and still getting myself very worked up over it, so I had trouble seeing whether it had potential to work or if I should just have scrapped it entirely and started fresh.
Looking at it now I think that perhaps it bounces forward just a little too far, and not far enough when it bounces from left to right. I think that it could have stood to bounce a little higher, too, bearing in mind that it was going to be rotating forward to land on the tip.
See? This is why I need to calm down and take breaks. Looking at it with fresh eyes I can much more easily see what could have been fixed, rather than rushing straight ahead and making a billion different half-finished versions. Oh well...
Shuttlecock V3
Version 3
I'm really sorry, I know it's stupid and I'm an idiot, but I cannot for the life of me remember what I changed in this version. It looks exactly the same but I did it a couple of days ago and was getting really flustered with getting this right so I just carried on revising them one after another without pausing to type down my observations. I know I must have changed it, or I wouldn't have saved it as a new version... I'll leave it here for the time being anyway...
I'm really sorry, I know it's stupid and I'm an idiot, but I cannot for the life of me remember what I changed in this version. It looks exactly the same but I did it a couple of days ago and was getting really flustered with getting this right so I just carried on revising them one after another without pausing to type down my observations. I know I must have changed it, or I wouldn't have saved it as a new version... I'll leave it here for the time being anyway...
Bowling ball V9
Medium closeup
Closeup
I think this is the version I'm most satisfied with thus far. I lengthened the roll of the ball and tweaked the curves so that the speed was relatively consistent when coming out of the last bounce, and had it slow to a stop. The rotation was a little tricky to get right, and I'm still not sure it looks correct. I opted to have it roll for about one frame after it came to a stop, just very slightly, in the hopes that it would be a little more realistic. In reality the ball wouldn't stop moving at the exact point it stopped rotating... I think?
However, looking at the closeup shot, you can see a huge problem - it goes off-screen! It shouldn't be too difficult to fix (hopefully) - I should be able to use the fcurve editor. If I select all the x position keyframes and just drag them up/down to slightly alter the position, I should be able to re-frame it and bring it back into the centre.
Closeup
I think this is the version I'm most satisfied with thus far. I lengthened the roll of the ball and tweaked the curves so that the speed was relatively consistent when coming out of the last bounce, and had it slow to a stop. The rotation was a little tricky to get right, and I'm still not sure it looks correct. I opted to have it roll for about one frame after it came to a stop, just very slightly, in the hopes that it would be a little more realistic. In reality the ball wouldn't stop moving at the exact point it stopped rotating... I think?
However, looking at the closeup shot, you can see a huge problem - it goes off-screen! It shouldn't be too difficult to fix (hopefully) - I should be able to use the fcurve editor. If I select all the x position keyframes and just drag them up/down to slightly alter the position, I should be able to re-frame it and bring it back into the centre.
Wednesday, 21 March 2012
Bowling ball V5
Version 5
Side
Medium close-up
I know, I said I was going to try spacing out the bounce... but I just ended up getting confused. The speed of the bounce was bothering me - my keyframes were all over the place, and after spending the better part of 45 minutes trying to tweak what I already had, I just ended up saying 'screw it' and doing the entire thing from scratch.
I still don't know. I think the speed is looking a little better - but again, the bounce - I think it's just a little too choppy. I think it needs to ease very slightly into the peak of the bounce and then hold it for just a fraction of a second. The only danger with that is then it might look too slow?!
I seem to be having a lot of difficulty expressing any form of weight with this. Or is it just me?
Side
Medium close-up
I know, I said I was going to try spacing out the bounce... but I just ended up getting confused. The speed of the bounce was bothering me - my keyframes were all over the place, and after spending the better part of 45 minutes trying to tweak what I already had, I just ended up saying 'screw it' and doing the entire thing from scratch.
I still don't know. I think the speed is looking a little better - but again, the bounce - I think it's just a little too choppy. I think it needs to ease very slightly into the peak of the bounce and then hold it for just a fraction of a second. The only danger with that is then it might look too slow?!
I seem to be having a lot of difficulty expressing any form of weight with this. Or is it just me?
Sunday, 18 March 2012
Bowling ball drop V3 & V4
Hooray for Vimeo processing!
Version 3
Here, I've adjusted the timing very slightly so that the ball doesn't take so long to ease into the drop. As a result, the ball enters the frame much earlier. I think this speed's better, but I'm not sure about the bounces. They seem too... "floaty" to me. I can't put my finger on what it is, exactly. I think that the ball drops from the peak of its first bounce a bit too slowly and doesn't seem to have much impact with the ground. It's a heavy object - the drop should probably be quite snappy or choppy to really give it the feeling of weight.
Version 4
Version 4
I wondered if spacing the ball out would help me to figure out why the bounces look a bit weird. I think they look a little better now; still slightly floaty? I'm not sure if trying to add the roll at this stage was a bad idea - it's easy to get distracted by too much going on. In this case though I think it might help - makes anything that looks off stick out a lot more.
Anyway. I don't think the ball really rolls or moves far enough. If it's going to roll, I think it probably needs to roll properly - if it was dropped straight down it probably wouldn't roll at all, like in the reference video - so it's probably only going to look right if I animate it rolling as if it's been dropped with a bit of a push. If that makes sense.
I think I could make this version work if I were to add just a little more rotation to the ball, but I might experiment with making the bounces and rolling a little more prominent, as if the ball has been dropped with some forward force.
Tennis ball bounce V13
All work and no play makes Alex a dull tennis ball
All balls and no tennis makes Alex a tennis ball
Ball tennis and no balls make tennis a ball
balllllsss
Medium close up
Oh goodness. I've fixed up the ball's rotation slightly, so now it begins to rotate after the first bounce. I think it's looking alright? I'm not sure. I still don't know if I'm terribly happy with it, and the more I look at it the worse it looks.
I think, in all honesty, I'm going to have to put a lid on this one for the time being - at least until I can get some outside feedback. It's still not brilliant, but there's only so much I can do with it using my own tired eyes. I'm feeling quite aware of the looming bowling ball and shuttlecock tasks, and if it's taken me this long to get a tennis ball going, I think I'm going to need to start getting on with those as soon as possible.
I'll more than likely end up coming back to it after I can get some feedback from others, but for now I think I can safely tuck it away.
Saturday, 17 March 2012
Tennis ball bounce V11 & V12
Going to have to make this the last one for tonight. I should do a bit more and try to finish it off but I'm incredibly tired, and as previously evidenced, I make stupid mistakes when I'm tired.
I actually think the rotation on this one looks alright! It needs to roll just a teensy bit more towards the end, I think, but it's not too bad (as long as you don't think too hard about realistic physics). The problem comes when I try to apply squash and stretch...
It's the same rotation, just with additional squash and stretch on most of the bounces. I don't know why - maybe it's just me - but it looks really odd to me now. The way that it doesn't spin at all on the first bounce somehow seems very out of place when I add in the squash. Like the force of the contact with the ground should start it rotating? I don't even know! I've been staring at this too long. I've started to go blind.
I also noticed one horrible problem, I've managed to break the rig somehow - probably during my cock-up with the rotation keyframes. If you look closely:
The position control object has magically floated away from the ball and now sits somewhere in front of it. I have absolutely no idea what's happened - I don't recall the point it happened. In fact, I only just noticed when I went to adjust the position to maintain the ball's contact with the ground after applying the squash. It doesn't seem to have broken the animation at all but I'm starting to worry that moving it back into the proper place (when/if I can figure out how to do that) will mess things up. And that will be a whole day wasted. Note to self to ask Jon how to correct it.
Doh... I'm stupid.
Anyhow. I'm going to try one more time to correct the rotation tomorrow and polish things off a bit, then I'm going to need to move on for a while. I can only stare at this silly green ball for so long...!
Tennis ball bounce V10
Using the video I shot as a guide, I thought I might try and be really clever and use a quick diagram I drew as a Rotoscope image within Softimage.
I think it's actually looking a lot better. The biggest issue that I can see now is the rotation - it looks completely wrong, and nothing I'm doing seems to be fixing it. I've tried speeding it up, slowing it down, changing the points it starts rotating - nothing. It just looks weird. I think that the spin just seems to change speed quite randomly - for example, towards the end it doesn't spin much at all but then suddenly rotates quite a lot. I don't know. I'm really tired and I've been staring at this far too long. I'm starting to go blind.
I also noticed that I made a huge cock up and set the rotation keyframes on the translation control - I have absolutely NO idea how I managed that as I was 110% certain that I was selecting the rotation control each and every time. There was one point where, thinking I was being clever, I had two dope sheets open - one locked to display the translation keyframes, so that I would be able to see both sets simultaneously. It's probable that as I was keyframing the rotation control it was getting added to the translation dope sheet as well.
I'm erasing all the rotation keyframes from the translation control now; I just hope I've not somehow completely broken the rig...
I also noticed that I made a huge cock up and set the rotation keyframes on the translation control - I have absolutely NO idea how I managed that as I was 110% certain that I was selecting the rotation control each and every time. There was one point where, thinking I was being clever, I had two dope sheets open - one locked to display the translation keyframes, so that I would be able to see both sets simultaneously. It's probable that as I was keyframing the rotation control it was getting added to the translation dope sheet as well.
I'm erasing all the rotation keyframes from the translation control now; I just hope I've not somehow completely broken the rig...
Moom character pose #3
For the third and final pose, I chose an actual superhero. Michelangelo from Teenage Mutant Ninja Turtles.
Though I'm not entirely keen on some of the modern depictions of the characters, they do have some more dynamic poses. I foolishly thought that this one, compared to Captain Planet, would be pretty straightforward. How wrong I was!
Trying to get him to touch the ground was interesting. I was initially too afraid to drop his hips any lower in case it messed up the position of the legs and feet, but eventually I bit the bullet and did it anyway. Happily, it was fine!
The fingers were quite good fun, as weird as that sounds. I simply bent the first segment of each digit downwards in its 'natural' direction, then bent the last two segments upwards slightly to give the impression that they were pressing into the floor.
I'm actually pretty pleased with how it turned out. The thumb is bothering me, it's just sort of floating there. I probably should have tucked it out of sight or something. Oh well!
Moom pose 2
Despite all shady allegations regarding his five young Planeteers, I like Captain Planet, so I opted to use him in all his star-studded spangliness as a reference for my next Moom pose.
Of course, in keeping with the very essence of Captain Planet, I chose the campest image Google could possibly offer:
Setting all sparkly kawaii anime-ness aside, it is actually quite an interesting pose. The angle of the head and arms is quite extreme and dynamic and I thought that, again, it would be interesting to see how I might be able to replicate this pose giving the limitations of the rig.
I must admit that I worried, at first, whether or not it would actually be possible, but fortunately my perseverance paid off.
It's not an exact replica and it certainly lacks some of the dynamism (is that a word?) of the original. The biggest issue I found was actually getting the body to lie at that angle. The chest control isn't designed to be translated. You can do it, certainly, but it usually results in... well...
Mass noodleage.
However, I could see no other option to get the body to lie flat in the air. Moving his hips back still resulted in the chest sticking up and no amount of rotation would fix it. He just looked like he was kissing his crotch, and that makes nobody happy.
So, I did break one of the first rules: I translated a control not intended to be translated. Only a very, very tiny amount, just to bring the chest in line with the hips. I didn't move it beyond any constraints that would otherwise have been set so it looks alright, I think.
The other nice thing about this pose was that it gave me an opportunity to use the synoptic viewer, which is a nice little interface that makes selecting the numerous rig controls much more intuitive. By selecting any rig objects and hitting F3 the panel opens. It's a much nicer way of working, I think - I frequently find that after moving joints around some of the smaller objects (especially the Up Vectors, which control the angle of the knees and elbows) get lost somewhere in the 3D space and it becomes tricky to figure out which one corresponds to which elbow or whatever, so the synoptic viewer keeps everything in place and you can quickly see which one is which.
An additional point of interest within the viewer is the 'hand' tab. This allows much greater control over each individual digit of the hand - essentially, you can control the bend and flex of each joint in the fingers. How cool is that?
I also got the chance to use the clavicle controls for this one! The clavicle controls are two small eliptical objects floating behind Moom. You can use them to relax or hunch the shoulders, leading to much more natural looking shoulders. His arm in this pose is stretched behind him so I dropped the shoulder a little to give it more reach.
Again there are a few problems with the angles - the front elbow should point forward a little more so that the forearm and hand are in front of his face, but for a relatively complex pose I don't think I did too horribly!
Moom character pose #1
In order to familiarise ourselves with the Moom character rig in preparation for the walk cycles and character action sequences, we were asked to replicate a series of dynamic superhero poses using the rig's control objects.
I, of course, chose the greatest and most highly respected superhero of all time.
I, of course, chose the greatest and most highly respected superhero of all time.
(Image from http://fuckyeahflashgordon.tumblr.com)
GORDON'S ALIVE!!!
Erm. Yes. Ahem.
It's an interesting pose that I thought would be quite fun (or possibly an absolute nightmare) to replicate. It's quite a basic pose but there are some quite interesting angles going on in the hips and legs.
Breaking down the pose into a basic skeleton demonstrates the angles and lines a little more clearly, making it a little easier to understand what each limb is doing. His hips are slanted in favour of the left (our right) leg - where most of his weight is. The shoulders slightly counter that angle, tilting on the opposite direction to maintain balance, with his arm high and outstretched. His front knee is raised very high and pointing slightly outwards - you can't even see much of the thigh behind it.
The torso is quite interesting - he's leaning back but looking straight at the camera. His neck is probably craned forward to maintain eye contact. He should be careful - he'll seriously strain his neck if he ain't careful.
It's certainly not brilliant, but I had a lot of fun with it! It's interesting trying to replicate such ridiculous muscular characters with a rig as lanky and weedy as this fellow - he has a lot more torso and less leggage than Flash Gordon. There are, unfortunately, some limitations with the rig - you can achieve most things if you play around with it enough, but there aren't individual controls for each section of limb which can make really fine-tuning the angles of things like the shins or forearms a bit difficult.
There are some issues with the angle of the front arm - I couldn't find a way to sort of foreshorten it without the elbow bending, so it's not quite as dynamic as the original. His knee isn't quite as high as I'd have liked, and the angle of his shin is a bit wrong - it should be pointing inwards, but as said previously there is a limit to the amount of control you're given. I'm sure that if I'd played around with it enough I could have figured something out, but at this stage I'm just trying to familiarise myself with the rig at a fairly basic level.
That's not to say I'm complaining though - for all its minor limitations the rig really is quite amazing. What you can do with it is absolutely incredible - the foot controls and hips, especially.
Sunday, 11 March 2012
Tennis ball bounce V7
Another entirely unsuccessful attempt to procure a tennis ball today. You wouldn't have thought they'd be so damn difficult to get hold of. Looks like I'm going to have to exercise my imagination a little bit longer...
I've sorted out weird jumping-forward problem from the previous post, so now the ball rolls forward at the end fairly nicely. I made a few more timing adjustments as well, speeding up the drop on the second bounce. I also slightly adjusted the height that the ball reaches on its first and second bounce - it seemed to go a little too high before, so I dropped it down just a tad and it seems a little better. I think the bounces are slowly getting there - though I think the third bounce (and perhaps the entire set of small bounces) are a bit too slow, but I'm not entirely sure? Might look at speeding them up. This is hard without reference.
The other big change is the addition of a little squashing and rotation. The ball squashes a lot more than I'd originally intended - at first, I only squashed it a very tiny amount and because the impact with the ground is so quick and instantaneous it didn't really 'read.' I slightly adjusted the curve so that the ball would remain close to the ground for a fraction of a second longer and also increased the amount of squash - looking at the capture it stands out far more than I thought it would, so I may reduce it slightly. Tennis balls are pretty firm!
I'm not sure if the rotation is at all accurate - probably not! I think it looks alright though (so far at least) so I'm probably not going to do too much tweaking to it, unless necessary.
The other issue that probably isn't apparent from the angle of the shot is that the ball isn't actually touching the ground when it bounces. It's an issue I encountered whilst adding the squash to the ball - I could have fixed it straight away, but I decided to talk about it on the blog a bit first. I don't really know why. Maybe because it gives me an excuse to post some screen captures (which I've been neglecting so far)...
Basically, the rig seems to be set up in such a way that locks all its transformations to a local axis. The ball's centre point is positioned right on the bottom of the ball, so that when scaled along the y axis it squashes quite nicely. However, when a rotation is applied to the ball, the centre point follows - i.e. it's no longer on the bottom of the ball. This means that it scales from a slightly different point, meaning that the bottom of the ball is no longer touching the ground (as seen above).
It's not a big problem at all and I don't have any wonderfully clever solution - I'm just going to move the ball down so it's back in contact with the ground. I just thought it was worth talking about.
Maybe I'm just boring.
Sunday, 26 February 2012
Post-tutorial wall bounce amendments
Following Jon's (very helpful) critique on Friday I amended my first wall jump exercise!
We both agreed that my ball hit the peak of its stretch on the drop a little too early, and he also pointed out that my ball remained in the squashed position once it began the jump, which looked a bit weird — it needed to stretch as soon as it left the ground. There was a similar issue with the landing — the ball squashed before it even hit the ground.
My first amendment was better, but the ball was still stretched on the drop a little too early, and Jon commented that he felt the drop itself was perhaps a bit slow. I do think that the speed of the jump and fall is a little too uniform, so on my next amendment I made an attempt to speed it up slightly by reducing the number of frames between the highest and lowest points and slightly altering the fcurve.
The faster speed seems better to me, but I think the stretch is still too early! It also seems to me like the ball comes down a little too quickly — I may try increasing the ball's delay at the peak of its jump, or just slow down the beginning of the drop a little.
I also went back and tried to correct the same squashing/stretching issues in my other bounce, the one with multiple jumps:
Something still seems at odds to me about this one, but I'm not quite sure what it is exactly. I think that maybe the ball hangs in the air a little too long on the first (and smallest) bounce, or maybe that it doesn't hold its initial downwards/preparatory squash long enough. The ball's squashes on each impact with the ground are also quite quick/sudden, so maybe I need to look at holding the squashes on-screen for just a little bit longer?
We both agreed that my ball hit the peak of its stretch on the drop a little too early, and he also pointed out that my ball remained in the squashed position once it began the jump, which looked a bit weird — it needed to stretch as soon as it left the ground. There was a similar issue with the landing — the ball squashed before it even hit the ground.
The faster speed seems better to me, but I think the stretch is still too early! It also seems to me like the ball comes down a little too quickly — I may try increasing the ball's delay at the peak of its jump, or just slow down the beginning of the drop a little.
I also went back and tried to correct the same squashing/stretching issues in my other bounce, the one with multiple jumps:
Something still seems at odds to me about this one, but I'm not quite sure what it is exactly. I think that maybe the ball hangs in the air a little too long on the first (and smallest) bounce, or maybe that it doesn't hold its initial downwards/preparatory squash long enough. The ball's squashes on each impact with the ground are also quite quick/sudden, so maybe I need to look at holding the squashes on-screen for just a little bit longer?
Subscribe to:
Posts (Atom)