Rendering and Editing The Film

 

A still from Scene Five featuring my Pokebot character

After a lot of issues along the way, it's finally time to render my final film. Over the past week or so I had attempted to do some test renders on each of the scenes in the final film to see how long it would take for my laptop to complete.

This is where I ran into an issue - after running for over 90 minutes, my laptop still hadn't completed rendering a single frame - in fact it was only on 8%. It was then that I messaged some of the people in my class to ask them if they had any suggestions for improving render times as there was no way I'd get anything dome otherwise. 

That's when someone suggested using a render farm and even recommended a few that I could try. After looking into the options I settled on Fox Render Farm and the speed that I was able to get things rendered was like night and day. I couldn't render a single frame in 90+ minutes and the render farm managed to render Scene 1 - which is 200 frames - in a little over 6 minutes.

After checking through each of the 200 images in Scene 1 and seeing that there were no issues, I pressed ahead and set up and submitted each scene to be rendered. Rendering with this render farm was really convenient in that I would point the software to where the Maya scene I wanted rendered was, and then it would analyse it and read all of the asset and texture files using the project settings related to that scene. As for defining the actual render settings - it worked the same way - I would set the render settings in Maya as I would normally do if I was rendering it on my machine and the software would read those settings and apply it to the render job,

A still from Scene Three

All in - I would estimate that the rendering process for all 8 scenes took about 2 hours using the render farm - which consisted of 748 frames, something I definitely wouldn't have been able to do using my own laptop.

Now, with all the scenes rendered it was time to take all the image sequences I had gathered and turn them into a video. In class it was recommended that we use DaVinci Resolve for this as it's free and professional grade software. However, given the extremely limited space on my laptop I had to find an alternative option. After a lot of research I chose Shutter Encoder. It has an extremely basic interface but it is relatively straightforward to use. To make the video using my images sequences I did the following;

Add all the images to the queue on the left

Set function to h.264 or AVI or any output you want

Under 'Bitrates adjustment:'

• Set 'Scale' to 1920x1080 (or any other resolution you want)

• Click 'VBR' until it says 'CQ'

• Set the CQ value to 21

• Enable 'Max quality'

Under 'Image sequence:'

• Enable 'activate the image sequence to 24fps'

Under 'Advanced features:'

• Set 'Force tune' to 'stillimage'

Start function

 Once I started the rendering process on the video it took about five minutes to complete. I rendered out the video in both MP4 and AVI. I chose to do that because AVI is the format required for assignment submission and MP4 because if I ever decide to add this to my portfolio, MP4 is a more widely used format across different operating systems and video playing software.

With the render now complete I brought it into YouCut on my phone to be able to edit it on the go. As part of the editing process I cut out some of the awkward scene cuts to try and make it look a bit more cohesive and add some music in the background.

Overall, I'm very happy with how the film turned out.

Working With Blend Shapes

 

In today's class we looked at how blend shapes could be used when modelling. To do this we were given some basic shapes that resembled a smiley face and after being walked through the process of how to do it, we were tasked with creating new blend shapes that conveyed different emotions.

This techique would be especially useful when animating faces and I am planning to incorporate it into how one of the characters in my film communicates if time allows.

Covering The Basics of Animation and Playblasting

A ball rig in Maya

This week we covered the basics of animation in Maya by learning how to animate the movement of a ball in a couple different ways. We looked at how to animate using only the graph editor and by doing it in the viewport by setting keyframes on the timeline at important points in the animation cycle. I found that using the graph editor was a little more difficult to get used to that the keyframing method in the viewport but I'm sure with more practice, using the graph editor will be much more intuitive to me.

We also looked at how to Playblast in Maya. A Playblast is a quick preview of what your animation should look like when the final render is complete. It works by taking a screengrab of the animation in the viewport at each frame during playback and then 'blasting' those images back to you - typically in quicktime format, but this can be changed to something else in the settings menu. This allows you to quickly evaluate if your animation is working out as planned without having to commit time to a full render.

By default, Playblasting will generate the quicktime file using the active viewer and the time range that is indicated on the timeline below the viewer - this also determines the animation range that the playblast will render for you. The default resolution is quarter of the size of the active viewer.

As someone who experiences frequent software crashes and doesn't have the best equipment to render from - the Playblasting feature will be invaluable to me when checking if my animations are working instead of wasting potentially hours on a full render that may not be usable in the final cut.

Bringing My New Character to Life!

 

It wasn't easy but I was able to model, rig and aminate a brand new, as yet unnamed, character. As with everything I've done so far, it wasn't without it's problems along the way but new problems just mean finding new solutions!

The problem this time wasn't the modelling part, the unexpected software crashes (that I've become very accustomed to) or even putting in the skeleton - it was actually painting in the skin weights. I encountered an issue where every time I would finish painting a skin weight and move onto the next, the one I had just finished would reset itself and disappear. Now, I didn't notice this issue straight away - I didn't actually notice it until after I was almost finished painting the model - or at least thought I was almost finished. Confused, I went to see if maybe I had hit undo by accident - because sometimes the fastest path to a solution is the path of least resistance. What I should have known is that noting is ever that simple in Maya, at least not in my experience with using the software up to this point.

I spent all hours trying to figure out the issue - painting and repainting the skin weights only to watch them reset themselves over and over and over again. If I wasn't so stubborn I would've given up for the night but I just couldn't stop until I knew what the problem was. This was the last major hurdle between doing the animation and starting a render to see if it worked. It took much longer than I would have liked but I eventually found my issue. I discovered that the max influence when binding the skin was set too high and as a result, was causing the reset issue I was experiencing. My max was initially set to five so I lowered it bit by bit until I got a number that didn't cause a reset - that number, was two.

Where did I have to go to make this change? It was as straightforward as you'd expect. I had to do the following;

Select the object > go to the Attribute Editor > switch to the "skinCluster" tab > set "Max Influences" to a number that worked for this model.

With that problem now solved and the headache now behind me - it was time to do the fun part - an animation! Now, given the issues I had up to this point I didn't want to go too crazy on the animation just in case something else went wrong. With that in mind I did a simple enough animation set out over 100 frames and started the render process.

This took another few hours to complete because I'm not using the best of computers to run Maya but when it finally finished I was able to take all the PNG files I had rendered, compile and convert them into a gif and see the end result. A result, I couldn't have been happier with given all the hurdles I had to deal with on the way. I managed to bring my new character to life!

 

Working Through Some Problems

Picture of Netflix's Love, Death & Robots

 

Ok so, I made Cozmo-E and now I need to rig him. The problem? I didn't model him in a T or A Pose so rigging him is turning out to be pretty difficult.

I have discovered that because I modelled him in a specific pose, adding a rig and controls to him has been next to impossible so far because it just doesn't look right. So, after many hours spent on just as many attempts to make a rig work, I have decided it's not going to happen for me and I need to come up with a plan B because I do not want miss the opportunity to use Cozmo-E.

So what's the new plan? I'm going to make a different, even smaller robot to use in the same scenes as Cozmo-E, sort of idolizing him - similar to how the little green aliens look up to Buzz in Toy Story.

This is obviously unplanned so I didn't do any drawings in advance of making something in Maya - I used the character sketch I made for Cozmo-E as the inspiration for this - as well as the animated series Love, Death & Robots, which I have been watching recently.

So, with no plan, I hopped into Maya and got to work. The building process went okay, there were a few crashes which at this point, I have come to expect on a fairly regular basis. What I didn't expect was for my save file to corrupt itself to a point where it was lagging making it almost unusable. To counteract this, I exported the model from maya and imported it into a new scene as one complete mesh and I can continue to work on it in the limited way doing this will allow. As it turns out, it was a good thing I did this as the original Maya file now crashes the software anytime I try to open it, so while I lost the original file, at least I didn't lose everything.

Now, time to go see if I can rig this model without any issues...

Bringing Cozmo-E to Life

Drawing done, character sheet created, brought into Maya and orthographic view set up - it was time to bring Cozmo-E to life!

This was a challenge. The toughest one so far. There were a lot of times where I was ready to give up on the character and do something that might have been easier to make - especially since I still consider myself to be very much a beginner in terms of 3D modelling.

I started of okay, but still finding it challenging enough - doing all the work in orthographic view didn't help either because this too was different to how I would normally model - using perspective view and rotating the camera as needed. I think I got caught up in the different camera views too much and I was flustered by it to a point where I was splitting my focus between that, and making what turned out to be a pretty difficult model.

When I finally let myself forget about the different viewing angles I was using - I still wasn't really getting anywhere with my model, I just couldn't get it to look how I wanted it to. Frustrated with my lack of progress I stepped away from it for a while and thought about how to 'fix' it so it at least looked something like my drawing. After running through a few different options in my head I came to the conclusion that I was making the process so much more difficult than it needed to be. I kept looking at my drawing and looking at it as 'one' thing - but it's not one thing, it's lots of smaller things that combine to make a bigger thing.

With that in mind I did what I was trying to avoid doing the entire time. I deleted what I had done already, which admittedly wasn't a lot, and started again. This time though, I went into it thinking about the process a whole lot differently. I broke down my character in a way that allowed me to build a very basic version of him just by laying out the primitive poly modelling shapes in a way that would look like him, Now I was getting somewhere,

Did the model look terrible at this point? Yeah. Did it look like I didn't have a clue what I was doing? Pretty much. Did it also lack any kind of detail that might allow you to figure out what I was actually doing? Absolutely. But, the details would come later when I was ready to add them. I had to remind myself that there's nowhere for details to go if you don't have a solid base to work with.

So, little by little I worked on my character. I selected each shape I had placed down and worked on them bit by bit until each piece was modelled into something that more closely resembled my drawing. This was an extremely slow process and I did get confused a lot while I was doing it - I would mix up which tool performed which function and would need to resort to some quick searches to help me out, but this is a learning process and I've definitely learnt a lot up to this point.

Now as things stood, I was pretty happy with where I was at - I had half the model done and there were a few things I was able to duplicate and just reorientate to make the process faster, and as it was late at night, I did exactly that, saved it and shut it down until the next day.

This is when I hit my next issue - when I reopened the file the next day several parts of my model were appearing as black on screen. In order to get a better look at the problem I assigned a temporary material to my model to see what was fine, and what needed to be fixed. Once I did this, I was able to see exactly what parts of the model need my attention. Working my way through the list of potential issues, I was able to determine that to fix the problem with these parts, I needed to go into the Mesh options and hit reverse to flip the 'normals' back around. Doing this one by one to all of the affected areas of the model, I rectified the issue and was able to get back to where I was before the issue began.

All that was left for me to do then was to fix up the positioning of some of the pieces and I would be done. After doing this for quite some time, getting everything just how I wanted it, I was done. I had created Cozmo-E!

The only thing I hadn't yet done was combine everything into one object - but as I still have to rig this character, something I'd never done before, I'm not sure if combining everything is something that needs to be done before or after the rigging is complete. This would be a job for next week when we begin to look at how to rig a character.

A Treasure Chest, Gears and a... Lightsaber?

All the preparation was complete - the research was done and the sketches were completed - it was time to build.

I approached the build process the same way I did with the preparation - start with the chest and gear because it was an easier task to complete compared to the extra prop I was hoping to make. I started with the treasure chest and gear.

This process was pretty quick and straightforward as I was just redoing what I had done a few days before in class - the longest part was fighting with the controls that would bug out sometimes and second guessing myself on some of the choices I was making. Other than that the process of actually creating the chest and gear was going well - add a shape, manipulate, duplicate and repeat until I got to where I wanted to.

With the first 2 props complete, I set about animating it. I chose to do a simple open and close animation as I had never animated in Maya before and didn't want to aim too high only to find myself stuck. Following some more trial and error with the placement of the chest lid that I hadn't noticed before I started working on the animation itself, I eventually figured it out and set something up that plays out over 30 frames.

Once that was done, I added a sky dome light to the scene and set up the render. This is when the auto save feature had saved me from disaster. During the render process, Maya crashed - a lot. Thanks to the auto save I was able to rescue my file from another potentially corrupted original version and in the end, get the render finished. I then used all the images I rendered, compiled them into a GIF and was able to use this to show the animation I had created.

Not wanting to tempt faith with yet another Maya crash, I decided to build my additional prop in its own Maya file, save it and then import it into the scene with the chest for some screenshots of everything together. It was time to build a lightsaber.

Preparing For The First Challenge (Again)

After the problems I experienced last week when I was saving my work and realising that auto save is not enabled by default, I immediately set out to enable that feature during the project set up phase in Maya. That was the easy part - now for something slightly more tricky - actually making something, or rather two somethings.

To recap, our challenge is to make two game assets - a treasure chest and a mechanical gear. Once we have that done and we're comfortable with the controls, the chest needed to be animated to show it opening and closing and if we really wanted to challenge ourselves, we could build one additional prop of our choosing to go with it.

As I was doing this outside of class time I decided to take a different approach to how I was going to design and build what I wanted to build. I've always been someone who has to put my ideas down on paper before I commit to starting anything because it helps me sort through the chaos, gives me more time to think about what I'm doing and sometimes, see potential issues before they happen in hopes it makes the process of actually creating them smoother. With that in mind - it was time to draw.

I started by drawing the 'easier' stuff first - some treasure chests and some gears that I could use as inspiration when making my model - I also added some notes to the side for things I might like to add to the model when building it if I thought it would look good. With those done reasonably quickly it was time to think about what I wanted to make to go with these assets.

I knew I wanted something completely different to the required items, something that would really challenge me to get the basic controls down and perhaps force me to use controls we hadn't looked at in detail yet. So I did exactly that - I chose sci-fi, I chose something futuristic, yet retro at the same time - I chose, a lightsaber.

Lightsabers are not something I draw everyday - or at all really so before I even thought about sketching out an idea, I looked up a ton of reference images because I wanted it to look similar to the ones used in the movies, but I also wanted to put my own unique spin on it.

There was a lot of trial and error involved in doing these sketches but I eventually got to a point where I was happy with what I had and set my sights on actually building in Maya.

Getting Started

Autodesk Maya Logo

This week was all about getting acquainted with Maya, its interface and the basic controls of the software. As someone who has a background in visual effects and compositing, I had some basic knowledge of modelling from college a few years ago going into this so the learning curve wasn't as steep as it could've been.

After relearning the controls a few times - the camera movement took a little getting used to - it was time to try and build something. We started off with a basic block castle using just the primitive shapes and the controls to move the shapes around. Now, while my castle won't win any prizes for creativity - I managed, with lots of trial and error, to do what was asked of me and in the process,  reinforce my knowledge of the interface and the controls.

From here, we moved onto something a little more complicated and that was using extrusions to add additional faces to our shapes with the hope of making some game assets - a simple treasure chest and a mechanical gear. After being walked through the process a few times because the controls for the extrusion technique weren't quite clicking in my head, I felt comfortable enough to give it a go myself and set about making something resembling a chest and gear.

By the end of class I had managed to make the assets required and was really happy with the results, however, just after class ended as I was preparing to save my work Maya crashed and I had lost what I had been working on all morning. It was also in this moment that I discovered that Maya does not save by default like I'm used to with other programs I use so as I prepared to redo everything all over again a few days later, the first thing I did, was turn on  autosave to prevent me losing so much work, again.

With class now over and disaster, basically confirmed, our task for the following week was to complete the chest and gear to get more comfortable with the controls and if possible, add in some sort of animation on the chest and then, make one additional prop if you have the time.

I took a day away from Maya so as not to overwhelm myself but in that time I had been working on some drawings of what I wanted to try and build for the tasks we were given. It was time to get to work.