daniel-lim.co.uk

Just a quick word to say I’ve added a new plugin called Gravatar Signup, a nifty little feature which can automatically sign you up for a Gravatar icon/account when you leave a comment (if you want one). It just saves you having to fill out two lots of forms. Many thanks to plugin author, Mark Jaquith! If you don’t know what a Gravatar is, you’ve obviously not read my earlier blog posts as I talked about it back in August. Alternatively you can go to Gravatar.com and read about it there. So now you really don’t have an excuse not to get one. Oh and I’ve also made the site font a tiny bit smaller - I just prefer it. And now I’ve really run out of excuses to avoid working on my Innovations project! Oh well, catch ya later…

Hello there! Again it’s been too long since I’ve posted an update, so here we go! This week, instead of doing my Innovations project (which is now desperately in need of rescuing), I somehow managed to start rigging my character for my Major Project. I was only meant to be doing a skeleton and looking at how I could get a control rig to drive a fully skinned and weighted version with the high res mesh attached and BAM! Five days later and I’ve got a fully working, (almost) completed rig! I guess that’s good from the perspective of my major and bad from the perspective of my Innovations (which I’m currently taking a break from - I’ve got a tutorial tomorrow, so I’ve got to do something!).

Above you can see the entire rig with FK/IK controls, the skeleton (templated) and the character mesh (also templated), but I guess I should run you through everything in a bit more detail.

The Skeleton

Here we have the skeleton. It’s fairly basic: standard leg setup (hip, knee, ankle, ball and toe joints) and (working up) the pelvis, root, four spine joints, the arms, the base of the neck, the base of the head and the top of the head. Now let’s take a closer look at each section.

The Leg Rig

I wanted my leg rig to have FK and IK controls so the leg has 3 duplicates. They all share the same pelvis so there is a hip, knee, ankle, ball and toe joint for each of the “real”, FK and IK skeletons (shown above, templated). The FK is pretty simple, what you see is what you get: rotate the boxes and the FK joints follow. The IK leg control is a standard IK (RP solver) leg. There are set-driven keys for peeling the heel off the floor, standing on tip, toe-tapping, twist heel (rotating around the heel) and twist toes (rotating around the toes). Rather than using a traditional pole vector object to control the direction of the knee, I followed the Art of Rigging Vol.1 No-Flip IK Knee method. I created two custom attributes called Knee and Knee Offset. By running the Knee attribute through a plusMinusAverage node I can use it to control the twist attribute of the leg IK handle. With both the legs set up, I then orient constrained the real leg to both the FK and IK versions so that the real joints would sit between the two sets of controlling joints. I then parented the FK/IK Switch locator to the real heel joint and created a custom attribute that, allows me to blend between the two sets of bones. This was achieved by using set-driven keys to control the weighting of the orient constraints.

The Spine/Upper Body Rig

Moving up from the legs we have the pelvis control curve which affects the orientation of both the FK and IK leg joints. We then have the CoG (centre of gravity) control which moves the whole body (the FK controls follow it completely, the IK controls stay planted where they are). I decided fairly early on that I just wanted to have FK controls for the back as they’re all that are really necessary and I’m quite used to using them having spent quite a lot of time with Generi, IK Joe (using the FK controls) and LowMan. I really couldn’t be bothered trying to set up a decent IK solution for the back as I’ve never found IK Splines to be that useful on their own. Above that we have the neck and head controls which are pretty cool. Following a tutorial from Jason Schliefer, I can now rotate the head to a position (for example, to have the character look at something) and then rotate the neck and/or spline and have the head continue to face the same direction. It’s pretty hard to explain without a video, but it’s cool; trust me!

The Arm Rig

Again I’ve got 3 duplicates of the arm joints so that I can have a similar FK/IK switching method to the one that can be found in the legs. With all the controls needed for the arms and hands (more on that in a second), the arms are a little more crowded than I’d like, but I’ll manage I’m sure. All the arms are connected to the same clavicle joint which is controlled by FK (labelled in the Spine Rig image). I have no idea how to set up an automatic shoulder control, so I thought I’d avoid it. The FK clavicle controls both my FK and IK arm controls (as with the pelvis and the legs). It just fit in with my method for rigging the arms and I think it will give me more control in the long run. The FK arm control boxes which just work on rotation values and the IK arm control and the pole vector locator simply control where the wrist/arm is positioned in space. Unlike the legs however, the real arm has an extra joint in the forearm for twisting. The wrist control (which I’ll talk about in a minute) controls the rotation/direction of the hands and the twisting of the forarm for better mesh deformation. Unfortunately I couldn’t work out a way for them all to be on the same control curve so I had to split them up in this way.

The Hand Rig

As I’ve already mentioned, the wrist curve controls the rotation/direction of the hand (in Y and Z) and rotating the control in X provides the forearm twist. I’ve then got set-driven key attributes on the Finger Control curve for bending at each joint of each finger and FK controls the thumb as I couldn’t work out a decent way to set up driven keys for it (in fact, if someone knows a decent way to set up thumbs fullstop, let me know!). I’ve also got the FK/IK switch curve attached to the wrist controls which is in turn constrained to the real wrist joint.

Note: Not sure if I’ve got too many thumb joints or not, but I think I can achieve decent enough (believable) deformations, so I’m not too worried!

Conclusion

So there you have it. The only thing I don’t have yet is eyes and eyelid controls which will be driven with joints and not blendshapes like the rest of the face. I’ll be adding face controls later, probably to the final rig as opposed to this control rig. Basically I’ll be animating my project with this rig and then using it to drive the rotations for the master rig which will have the smoothed, skinned and weighted final model attached to it, hopefully eliminating the need for me to have my model completed and skinned and all the other stuff before beginning to animate. I’m now on Innovations duty so there may not be any major project updates for a while! Bear with me - who knows, my Innovations may even be presentable at some point!

3 Loops

Finally I have got the first draft of the whole block test done. It took a lot of pushing and forcing myself and even some pushing from Dan. I really don’t like doing block tests, but it is a “have to” to see what works and what doesn’t. So it’s just a matter of self discipline in the end I guess. Just too bad that I am in distinct lack of anything of the sort.

Anyway…here it is.

I’m not quite there yet though. As I have said before, I want the three loops to have different meanings. Where the first is supposed to be a clear chase, the second is there to confuse the viewer to what actually is happening. Then the third is supposed to make it all clear again. At the moment I think the first loop is working. It’s clear what is going on and it corresponds with what the viewer think is logical or plausible. The second loop on the other hand is not working as it is. I think there is more than one reason for this:

Short time memory

You as a viewer can only hold the information of very few shots in your head at one time. Therefore, in this case, when a loop ends what stays in your mind/your impression of the loop will mainly be based on the about 6 last shots of the loop. The subtle changes in the second loop, compared to the first, happen in the first half of the loop. Therefore when there is a slight pause, before the next loop starts, the shots the viewer holds in its head are the same as the ones for the first loop. The viewer might have a slight feeling that there were some differences, but won’t be able to tell what they were unless he/she watches the animation again.

To prevent these changes in the loop could be put more towards the end to make the ending of the loop different compared to the first. Another way to prevent this could be to have the excising changes more drastic to clearly show a change from one loop to the next.

Possible disadvantage of the medium

To be able to notice a subtle change in something you need time. Therefore to convey something through subtle differences suits still images better than moving images. This is simply because than the viewer gets more time to notice and understand the change. With moving images you therefore need to either be more obvious or to give the viewer more time. This could e.g. be done through a constellation that loops, where the viewer can stay and watch for as long as they want.

The solution to this could either be to simply make the message, which I want to convey with each loop, more obvious. Something to watch out for doing this is to avoid making it too obvious. I still want the viewer to walk away from watching the animation feeling a bit clever for understanding. In the same time the worst case scenario would obviously be the exact opposite. The context which something occurs in can change if it appears obvious or not though. For example when the viewer watch the second loop I assume that he/she has seen the first and therefore has an idea of what they think is happening. To make him/her change their mind it will therefore take more than if the second loop would have been watched on its own and therefore had its contents evaluated separately.

The third loop is not clear enough as it is at the moment, but it’s getting there. I have tried to link motions from the woman and the cheetah together through similar framing and motions and then cutting them together to one. This needs to be more obvious than it is at the moment. I will try to do this by matching the movements better and see if I can use dissolves instead of cuts at some points. As the third loop is at the moment it could possibly work as the second loop instead and then make another version of this loop that is clearer to use as the third.

Not content with just posting crap on the NCCAForum.co.uk, it would seem that the stupid Russian forum bots have started spamming WordPress sites too… mine to be more precise… Oh well, a little more moderation of comments to do. Perhaps I should be flattered… maybe my site is going up in the world?

I’ve been working flatout this week, so I’ll have some updates for you later today hopefully!

This is the first image I got when typing “Russian Robot” into Google Images:

In case anyone was getting a little confused, Marie is now posting updates on her major project as well! Normally this wouldn’t be too confusing as her posts aren’t shown on the main page by default but bear in mind if you navigate posts by going into one and then clicking the next or previous post links, the categories don’t have any prejudices as to what’s shown so make sure you check who the post is made by otherwise it could get confusing!

It’s been exactly one month since we released the well-received WordPress 2.0 release. In the past 4 weeks we’ve been listening closely to feedback, squashing bugs wherever we find them, and watching how 2.0 handled under different loads.

I’ve updated to the latest version of WordPress. Getting an error at the moment when it comes to posting sometimes, but it still seems to work - hopefully it’ll get fixed soon. Other than that, everything is still running smoothly and I’ve even done a bit of fixing elsewhere on the site (note the return of the nice Register/Admin and Login/Logout buttons). You can download the new zip here or you can check out the 114 bugs they’ve squashed.

Oh the shame! I had to postpone my Tuesday tutorial with Phill due to my general inaptitude and the inability of Maya to RENDER FASTER YOU BASTARD, which pissed me off quite a lot. Luckily he had some time for one today, but more about that momentarily.

Yesterday I had a tutorial with Phil “That’s Crackers” Spicer, as I’d heard he’s pretty helpful in tutorials. Not unsuprisingly, he was! This is where it’s going to get a little confusing: he gave me loads of ideas and brought my attention to a few interesting little things (such as Kylie’s Come into my World music video which I hadn’t seen before) and we also looked at the possibility of fundamentally changing the animation whilst keeping it… the same, if that makes any sense. Instead of having the animation start in the police station and then having Hyde go to the apartment to investigate and get killed, we thought it could be cool to have him start in the apartment, get a phone call rush out and arrive at the apartment only to find it’s the same one and then get killed. Ok, that wasn’t a great explanation, so here’s the blocktest that I was up ’til 5 finishing. It’s not exactly “polished” (even for a blocktest) and I guess I kinda forgot to include any kind of investigation - a bit of an oversight on my part and doesn’t help the impact of the piece. Oh well. It does however significantly cut down on the length of the piece - almost a minute. Which is quite a tempting prospect at this late stage in the project.

Either way, the tutorial proved useful. Firstly, along with giving me a much needed enthusiasm shot, Phil gave me some goals and ideas as to where to go next. The “foreseeable future” agenda now consists of possibly reworking my blocktest again, starting some rendering tests and actually working out exactly what the f*ck my project is going to look like as I’m still pretty aimless, although I love this image I found on ConceptArt.org. It’s been a major influence in terms of inspiration and what I want the mood and atmosphere to be like. I may try and replicate a similar style in 3D… if I can. It’s a daunting task for someone who’s not done texturing, lighting or rendering properly before :/ Oh and unfortunately I can’t remember where I found this image or who it’s by. If anyone knows, please get in touch and I can credit/link it properly! I’m also going to start looking into ways of rigging and skinning my character, whilst at the same time starting to model the clothes (f*ck Maya Cloth tbh) and working out how I can get it looking decent (Linux Karl has a nifty wire deformer solution that I’ll be investigating).

So, onto today and my belated tutorial with Phill Allen. He watched both the new blocktest (linked above) and the “old” one. I say old since I started it ages ago, but only finished it yesterday (yes, the new blocktest was completed in a day! :/). You can view the old blocktest here, it’s the one that most closes resembles my storyboard. We decided that this was in fact the better of the two (as you’ll hopefully agree) and talked about changes that could be made to it to tighten it up and improve it. So yes, I now (ideally) have to redo the blocktest again with the newly discussed changes. I will do it, but just not yet, I simply don’t think I could take any more blocktesting right now.

This week I’m starting render tests, maybe some lighting tests and getting started on modelling my characters clothes. Oh and I’ll be reworking my schedule as the old one ahs basically gone out the window. I’ll post them up for you so you can compare ;]

For my animation I need a model of an African woman and a cheetah. I started with the Woman and used reference images from 3d.sk.

I am not really sure how I will give her eye brows and eye lashes yet. At some point I will have to give her some hair as well. I am planning on keeping it to a short afro.

She also needs some clothes. I think I will make something being wrapped and tied around her.