Combination

Now with all the playblasts completed it was tome to combine them all together with the title sequences over the top of the animatic to have a foundation of timings. It also was very nice to see the clips with a soundtrack! and this was how I was going to achieve the resolved part of my statement of intent. It is in scene order with all the timings correct to reach the resolved state, minus secondary sounds which I recorded such as: three different types of movement on gravel to increase the immersion and believability of the animation, because I think sounds are a vital part of an animation when it is story driven, it backs up the visuals to engage another sense into your experience of watching the animation. The clips were recorded after the playblasts were compiled. this helped get the timings right as I could use the animations for audio.

Initially I did want it all rendered for submission, but being as though it wasn't the ultimate goal or aspect of the project that I wanted to get graded on but rather the quality of direction, environment and the ability to present emotion through character animation in a 'resolved' animation, not necessarily a production quality render. But never the less to get the taste of the animations final aesthetic style I have rendered two scenes that will visually state the direction that the animation has taken. I rendered my favourite scene (scene 7) and the flashback scene. I chose the flash back scene to put another spin on the ways of provoking emotion. I am referring to various post effects that can also support the performance. So with the flash back I referred to various examples of flash backs in films:

As I descovered, there is a lot of blurring, especially of characters as well as some colour correction. I applied some effects to my scene using influence from certain references:

To slightly blow out the scene I added a light source, to blur I used the video effect blur. I am really happy with how it turned out. As the scene really needed to differ from the others to make it obvious that it was two weeks ago. This is something I had to think about at the very early stages of pre-production.

As the great Barry Purves said in his talk to my course, you have to consider everything at the start, literally every detail when directing an animation.

So he really helped my thought process when directing because it makes the production and post production stages a lot more solid due to a solid/pre-planned direction.

Scene 7, 8,9 and 10 playblast

(see submission folder for playblast)

Similarly to scene 2 and 4, I combined 7, 8, 9 and 10 together because they all fall in chronological order so it was prudent to use one scene file for all scenes. There were various things that I encountered with this playblast, not exclusively animation. But the animation was another challenge due to there being a very fast start and the rest quite slow again. The run into the apartment took a while to refine due to the limited amount of frames I had to complete the movement, so if I were to re-visit the scene I would try and get rid of the subtle jerk of Jon's back as he slams the door shut with his back. But when doing the referencing I did animate it to the movement but for some reason it didn't quite seem right. Obviously I was missing a small movement that would have completed the animation to a better standard. 

The animations that I am most happy about were when Jon is holding the gun towards the door and you can identify the emotion in his movements as he is shaking. Also when he shoots himself the death animation works with the addition of the cupboard door that is ajar that falls off due to the shock wave of the gun. 

The issue I had with the playblast though was that when I put it into lighted mode in the viewport to show the flash of the gun, it really darkened one shot almost to a point where you can't see anything but black. but I desperately wanted to show the gun flash.

If this scene has taught me anything is that I should find another way of quickly rendering my scenes rather than playblasting them due to the fact that some controllers have to stay on because an object is parented to it to make it look convincing that the character is holding the object. But also to either have a shorter animation to allow enough time to render the entire animation or to find alternative methods of animating such as doing a real time render in unreal for example which has amazing results such as their demo of the child with a kite that was all rendered in real time.

Scene 7 playblast

I was very excited to start animating in my favourite set, and it was a chance to improve on my walk cycle due to there being a walk and a run involved in the scene. Although I am not completely Happy with the outcome it was still an improvement from scene 5. I think it was because I wasn't over complicating the movements. It was also interesting animating quick movements, one with the turn around that Jon does to run away but also the thug that is keeping Sandra hostage, he demands his thug accomplice to go and get Jon so the movements had to be quick and intimidating to encompass the mood. This was quite a difference to all the previous animations being quite slow and emotional. But This is something I will have to do as an animator, adapting my animation style to accommodate different tasks. This scene helped me believe that there was still hop for my animation career as it was improving from that previous scenes!

(see submission folder for playblast)

Scene 6 playblast

As well as scene 2 and 4, scene 6 was meant to subtly capture the emotion of the character through slight movements of the body. I feel in this scene I might have over done it with the leg and arm movement, I added this in though because I wanted to add subtle audio effects of gravel moving under foot. There isn't much else to report on with this scene due to how small the scene is. But it was interesting to see how much emotion was possible to portray in such a small scene and environment.

(see submission folder for playblast)

Scene 5 playblast.

I HATE CREEPING!

I think I need a lot more practice with animating walk cycles or creep cycles in CGI as I just can't seem to get the motion fluid! Despite having reference, I think I try and over complicating things by littering the character with keys to capture each individual tiny movement that the body creates whilst walking which in turn conflicts with the other keys to create the jerky motion that I created in this scene. If I hadn't have set milestones I would have focused more time and effort into simplifying the movement to capture the principles of animation with a touch of reference. This is something I really need to persist with being as though I want to enter the video game industry, and according to a potential employer from a company that are working on a video game told me over a skype interview that technical animators are very sought after in the games industry. So I think it would be very prudent to crack the walk cycle before going onto bigger things ... Walk before I can run.

(see submission folder for playblast)

Scene 3 playblast

(see submission folder for playblast)

Now scene 3 was quite tricky to figure out despite having reference material. Due to the fact I was being cautious of camera angles I had to confine the movements for each character to keep them in frame and because my storyboard was just one frame with an explanation that the shots alternated between each characters over the shoulder I had to play the angles by ear. I also chose not to include the door actually opening in shot to be a bit more efficient and add the audio in post to give the effect. and to hold on to that intention of efficiency without detracting from the aesthetic or effect.

 My gripe with this scene is the way Jon reaches out after Sandra. He looks like he is waving goodbye rather than gesturing her not to go. I would definitely direct this scene differently to add slightly different movements to make it a bit less dull as opposed to just two characters waving their arms around as they speak. I could have had them acting out some survival activities such as trying to light a fire for example. but these are all lessons that I will take with me on my road to animator success.

Scene 2 and 4 playblast

I combined these two scenes as it was easier to just split the scene up in post being as though they are both in the same set and with the same character so I animated each scene back to back. Again this increased the efficiency in that respect as I could stay in my workflow and just edit the camera.

(See submission folder for playblast of scenes 2 and 4)

This is an odd one to document as there isn't much movement in the character, so it was quite interesting to see how much movement is needed to make a character look believable, it was mainly just head movements and the od sink of the shoulders and chest as he breathes.

Scene 1 playblast

The first scene was by far the most daunting scene due to the fact that it was the first one I was going to produce and it was equivalent to almost half the animation so it was by far the largest. The quality of animation would certainly increase from there after getting comfortable with the rigs and my workflow.

(see submission folder for playblast of scene 1)

First of all there was much demand for this scene due to the fact there are 4 characters in which there are 3 objects that are held and it all needed keying. I typically started by creating selection sets to make keying for each character neat and tidy instead of limbs going off on one randomly between action. I would start with the waist and animate up as the legs were Inverse Kinematics whereas the arms weren't, so if I wanted to slightly adjust the waist I would have to re-key the arms to avoid intersections. This shows my determination for efficiency to avoid any time spent re-keying limbs. I would then start on the wrists incorporating some follow through action. Then the fingers to animate their hands in a more believable way.

The last thing I would do is animate the face and eyes, this was because I already had the motion so I could solely focus on getting the facial expression right. 

To be perfectly honest it did throw me in at the deep end in terms of finding my rhythm so it benefited the rest of the project in that way.

One very strange thing that happened was that in viewport 2.0 (which is my preferred viewport renderer) when ever I scrubbed the timeline loads of geometry would disappear until I settled on a frame, so off course this would be the case, so changing the renderer caused the textures to not be accurate to what its like:

Lovely white cabinets... But in the viewport they are ugly panda coloured cabinets due to the renderer overriding the visible texture to be the specular map for some reason.

This made me realise that I should have really chopped the scene up into segments so I could do a clean export that might fix the problem.

Going back to my focus on efficiency, I was able to animate only the parts of the characters that were in frame, for example if someone's torso was in shot, their legs would be static, this saved me lots of time and kept things a little cleaner. The only downside was that sometimes it threw me off when a certain body part wasn't moving which made me conceive the animation as incomplete.

Playblasts

It was now time to playblast all my scenes after creating the animations. I will go through each scene to explain the encounters I faced with the animations. This gave me the chance to start working on secondary audio whilst I render a scene or two ready for submission. Being as though my intent was to create a believable and resolved animation that explores the directive abilities of myself and my ability to convey a story through appropriate techniques in the most efficient way, to have the whole thing rendered would not be feasible considering that it would take almost two months to render the animation if I were to render it solely on my own machine. (more about my own machine and how I shot myself in the foot later)  for submission due to time scale. But it will be rendered for the exhibition at the end, but obviously I won't be marked on that, but if I have two scenes that are rendered and the rest is a high quality playblast I can still convey the animations, it just won't look as pretty. But that wasn't my main aim, to create a prettily rendered animation. It was to convey a story through my directorship of a small team and my ability to portray emotion in my animations.

The reason I shot myself in the foot was how I did a good couple of weeks work on my own computer which was running Maya 2016, the latest instalment, whereas the computers at university were running Maya 2014. and because Maya binary files aren't backwards compatible all my work was useless on those machines. This also meant that render farming was useless. 

This has informed my practice in quite a subtle way but with up-most importance ... CHECK YOUR SOFTWARE AND SETTINGS ARE THE SAME IF YOU ARE USING MULTIPLE COMPUTERS!

In a professional context to give a bit of relevance to what I am saying and how render farming is the only way to spit out high quality renders in the most efficient way. the beautiful blockbuster film from James Cameron known as Avatar reportedly had:

Thirty four racks comprise the computing core, made of 32 machines each with 40,000 processors and 104 terabytes of memory. Weta systems administrator Paul Gunn said that heat exchange for their servers had to be enclosed

With each frame taking 2 days to render you can see why this is important.

Muzzle Flash

There was two techniques I could choose from to create the muzzle flash for when Jon shoots himself. 

Here is the reference image I used:

There was the method of creating some geometry that replicates the shape of a muzzle flash and apply an intense glow and material:

Where the shape is there the flash didn't effect the hole scene which isn't accurate, so the other way is to just create a point light and key its visibility and intenseness:

Although it look very blown out, I can imagine a muzzle flash going off right next to your face would blow out your face for a frame or two. This is the method I went with, I can add the image of a muzzle flash in post due to the fact its only present for 2 frames. 

This was a similar learning curve to everything else in this project, I was learning different ways to achieve a certain look and it was about choosing the most appropriate and efficient method in order to stick to my intentions of learning numerous techniques along the way that won't be forgotten, just not used.

Hole

One of the things I had to figure out how to do was to create a gaping hole in the first set and manipulate its contents to look more derelict. I achieved it by utilizing some tools and techniques I learnt through digital tutors such as split polygon and bridge for example, to re construct existing geometry. Here is a render of one of the scenes that uses the modified set:

This links with my statement of intent as I specified that I was going to manipulate an existing set to create this one. This is because I wanted to keep it identical in scale and texture other than the things I needed to manipulate to incorporate it into the story better. It was quite interesting manipulating an existing set. I'm not sure I liked it, I think I'd rather start from scratch due to the way you completely distort the UV's of the objects you manipulate which makes re-texturing them really hard without an extensive knowledge of UV mapping.

scene 5

I had big plans for the set of scene 5. I wanted to showcase the effects of the bombings in various different ways. But I had to cut various stuff out, mainly down to time restraints, but also the scene was so large with a lot of polygons it was very much stressing my machine out. Here is a screen grab of the development:

The contents of the scene included:

The entrance to the apartment building

A large number of bin bags

A body deposit sector 

a slight interior to the apartment building.

So because of the scale the scene started to lag, so I had to stop any additions. But I was able to still show some of the carnage from the explosions in the form of dead bodies with dirty bloody cloths draped over them:

But as you can see there isn't enough debris or carnage in the scene so its my least favourite. I could have applied various techniques that i've learnt across the module. I would definitely re-visit this scene to make it fit in the universe more appropriately.

Heart hugs

So after being on the phone to my mother about being stressed about work and going through a general depressed stage after burning myself out big time, I received a small parcel in the post a few days later. The contents of said parcel was a small heart with arms which have velcro on the hands to enable it to hug your arm. Recently it has become very useful to combat what I like to call 'animator wrist blister' where your wrist bone is touching the desk whilst you are using your mouse for 72 hours straight and starts to get very painful and sore. But if you wrap a small hugging heart around your wrist:

It provides the perfect amount of cushion to keep your wrist free of abrasions. How this has benefited my practice you ask? it has enabled me to be at the same level of animation focus without worrying about causing damage to myself. and how it benefits my career? if animation doesn't work out for me I can always produce items for people that are always at a computer that need that extra level of protection.

Scene 7

At first I was quite apprehensive about scene 7 due to the set that I had to create, but I honestly had the most fun creating this set over them all! I stayed up from 6pm until 3pm the next day working on this set (for two reasons, one because I was having so much fun and two because I didn't have long left to do it if I wanted to hit my milestone). I can't understand but I was in such a good zone, I had photoshop on my left monitor and Maya on my right and my workflow was impeccable, there were no complications and everything worked flawlessly and swiftly. This gave me a lot of confidence for my future career knowing that I can have an efficient workflow. But also in my statement of intent I mention that I wanted to create environments in an efficient way and this set couldn't have been more efficient.

Here is a couple of passes from the set:

Aswell as having a proffessional approach to this scene as well as keeping my scene and outliner clean and labelled I really went to town on creating my own textures and bump maps and specular maps which is what I wanted to refine through the project.

For example:

For one section of graffiti that says 'die like the rest' I started with the initial wall texture

Then I took this into photoshop to add some graffiti using different brushes and effects on the paint layer to make it look a little more believable:

But being as though I wanted this to be fresh blood, it would have to still be slightly wet. there is no better way than a specular map! So initially I set the whole material to a Blinn material which contains gloss:

As well as creating my own bump map by desaturating the image of the wall then adjusting the levels to create defined lines in which would become the cement grooves:

I then created a specular map by tracing over the graffiti with a white brush and applied a black background to really define what aspect of the image I wanted to be specular:

If I had a little extra time I definitely would have used different brushes to make it look that bit more sinister in the final result. The next step was to plug in the specular map to the texture node:

But because the original diffuse had a 90 degree rotation on it I would have to apply the same adjustment to the specular map to achieve the right result:

It looks blown out here because of the camera angle but its to clearly show where the secularity lies.

The way this procedure benefited my practice was the confirmation that I have got the method down and can utilise it to my advantage. The next steps would be to learn how to skin more complex meshes instead of just simple cubes or planes.

Photo re-image

Due to my refurbishment of my characters, of course I would have to re-skin the photo of Sophie that Jon looks at in scene 6. This gave me the opportunity to apply some skills that I picked up after creating other sets.

So here was the pose:

I had to open it photoshop to create a more 'photo' looking image:

I adjusted the levels and added a vignette to the frame to make it more believable. This would be the pre event version. But for the post version I needed some more wear and tear from Jon constantly folding and unfolding it to remind himself of her face. So here is the Diffuse for the photo:

Added scratches and creases to make it look more used. And I wanted to apply a specular map as the glossiness of photos tends to disappear on the crease lines when they become constantly handled:

Here is the final result:

The specular map can't really be seen having an effect in this single frame but it is having an effect.

This is just one example of how I was improving my skills as I was creating the environments and assets, which I can imagine is quite a good trait to have when approaching a job in the field that requires you to constantly keep learning new techniques and different ways of producing different effects and visuals.

Scene 6

Following on from my last blog about the renderers, it got me thinking about an opportunity to adapt one of my set designs. for two reasons: to simplify it to be able to allocate more time on the actual animation, and two, to see whether making a simple set would have a shorter render time due to the absence of copious amounts of polygons that the renderer has to accommodate for. Here is a render of one of the frames:

Its not how I wanted the scene to look in terms of lighting, but I was just testing whether one character plus two planes would render faster than four characters and about 600 thousand polygons. turns out that it only reduces the amount by a couple of minutes, so as opposed to a total render time of 55 days it would be 30. so in the long run it would be beneficial but its surprising to see that there wasn't a huge difference. This has benefited my project in quite a strange way, I was getting very upset and obsessed with focusing on the rendering of each scene, which doesn't conform to my statement of intent as I needed to focus on the animating and creating convincing environments, so it gave me that puch to start to relieve my focus on lighting and rendering and allocate my time more appropriately to directing the project to the performances completion.

Renderers

After struggling to reduce the render times. I thought it would be quite prudent to research into different renderers to see if there was one out there that could deliver photo realistic renders with a faster render time. I already knew of mental ray and Vray but I wanted to get a definitive answer as to which one would be more appropriate.

It seemed like there were particular renderers that were better for other processes. I learnt this through a forum on CGTalk:

sacslacker

07-01-2011, 05:56 PM

It really depends on what you're doing and what kind of budget you have really. Do you need GI/FG? You're going to want to look at VRay (though MR can do this well too) or even Maxwell. Are you doing lots of hair/fur or 3D motion blur, you're going to probably want to use Renderman (3Delight or what have ya). It almost all cases you're going to use specific shaders for the render engine so it's a good idea to learn the concepts and be prepared to apply them to different renderers. You're not wasting time learning Mental Ray since the concepts are very similar to other renderers really. That's just my opinion though, others might disagree.

Generally asking what is "best" isn't a solid question because render engines have pros and cons just like all software.

Another member of the forum posted this:

Kakkoii

07-01-2011, 09:48 PM

Since you already have Maya, go ahead with studying MentalRay. It's one of the most used renders in the CG industry. And most of the stuff you'll learn can be applied to most other rendering engines, especially shader wise. If you can learn to make great looking work with MR, VRay and others will be a breeze.

You'll get varying opinions on here about Renderman though. A lot of people are thinking it's slowly starting to die out and be replaced with pure raytracing renderers now that they are becoming more optimized and negating some of the benefits Reyes used to provide.

But it's still a very valid and useful renderer to learn, since it's still used quite a bit and isn't going to die any time soon. So if you can, learn both.

Renderman is the quickest renderer however requires more time to setup shaders/lights.
Mentalray is the slowest renderer that i have tried (i havent tried physically biased renderers) but easy for end users
Vray is in the middle, easy to use and quick to learn, with average rendertimes.

This was quite dis heartening knowing that I wasn't using the slowest renderer and still had pretty hefty render times. But to read that most CG companies resort to mental ray it made me think I should probably stick to exploring it due to the fact its like the mother of all renderers and if you know how to use mental ray you can transfer your knowledge to all the others. This will hopefully benefit my future because if I persist on the one main renderer then I wont need to constantly top up my knowledge due to less advanced renderers.

Rendering times

So with the first set close to completion, I wanted to start setting up the lighting for the scene to increase the believability of the textures and overall aesthetic of the scene. I started with a physical sun and sky which is my go to lighting system due to the accuracy of the render. Here is a frame:

Although this was a very appealing render with all the materials working the way I wanted them too such as the cupboards adopting the specularity I input into the texture node with a home-made specularity map but unfortunately the render took a whopping 39 minutes to complete. and although I have some fairly expensive materials I was hoping to get the render time down to about 5 minutes. This was going to be hard.

I decided to try different lighting.

I got rid of the physical sun and sky and created some point lights underneath the light shades.

With a much more tacky render it still didn't decrease the render time enough for me to settle.

I was then picking the brains of one of my tutors to see how to tackle this problem. He suggested that I try rendering it in passes, using an ambient occlusion pass:

Then to render a diffuse pass using Maya software:

Then combining the two to get the desired look.

This would help apply the same soft shadows as a mental ray render with hopefully shorter render times. Although the Maya software render didn't take as long, the combined render time for both passes was exactly the same as a mental ray render with nowhere near the accuracy that I wanted.

But this wasn't too important as I was intending to focus on the animation and story than the quality of the render ... but I'm a perfectionist. But also someone pointed out a very good point, if its daytime, why would they have the lights on? this was very true, the only problem was that the scene wasn't lighting completely so the characters were way too dark. So I decided that I should cut a couple of window holes in the side and stick with the mental ray physical sun and sky, this might also reduce render times or possible artefacts due to the lighting protons.

Pre and Post

As my statement of intent suggests, I want to stay true to storytelling. This is only possible with a conceivable timeline of events with evidence to back them up. An example of this is my use of the wide variety of features that adobe fuse has to offer when creating characters.

The event in the animation is the bombings of Leeds that leads to a destruction of society and basic civil services, gathered from factual information from the government that I researched into and the docudrama Blackout.

There is a time jump from the event to several weeks after where things are at the pickle of unrest. This needed to be evidenced in the presentation of my characters. 

So here is Jon before the event. a well presented adult, clean shaven with smart clothes. And here is 'post event' Jon:

As you can see the clothes are different, muddy and tattered. His beard has grown as well as his hair.

Here is Sandra before the event:

And here is 'post' Sandra

With both characters adopting a 'survivalist' aesthetic with dirty clothes due to electricity being unavailable. but also adopting more rugged clothing to survive longer and keep them warm.

I could have achieved a similar result by editing the diffuse files of the characters but the amazing features in adobe fuse allowed me to specify how much dirt, how much wear and what kind of clothing I wanted, taking half the time that it would have taken to do manually.

Like previously stated, because I want to enhance my skills in CGI animation and modelling and directing, I didn't want to spend time learning how to create characters like this without the tools of adobe fuse. Instead I can direct to aesthetic direction I want them to take to make the story more believable.

New characters

Despite the fact I was into the production stage of my project, I hadn't quite reached the animation stage so a complete over haul of my characters would add a little time to my project but would save me time in the long run due to the ease of use that the adobe fuse characters provide. Here are renders of my new characters.

 I am very very happy with the new aesthetics of my characters, and couldn't be happier with the way they are rigged. This not only benefited my ability to animate by having simpler tools at my disposal, but also the textures and meshes really fit in with the aesthetic I wanted to achieve. The autodesk characters looked very 'sim like' whereas these look more believable of being a group of 4 people living together.

A miracle

After battling between adobe fuse and audodesk character creation and feeling like i'd settled with something with less quality just for ease of use I discovered a miracle! There is a third party script for Maya which completely rigs adobe fuse characters in a very user friendly way. All one has to do is import the .fbx file of your character and run the script:

Then you are prompted with this window. You simply click on the 'rig character' button and it starts to calculate the algorithms that are needed with your specific character.

All the handles are created as well which gets rid of the need to go burrying into human IK. But what is even better is that it rigs the face as well

Which is a much more user friendly and efficient way of doing facial animations as apposed to blend shapes.

After animating the news reporter scene using an autodesk character and having quite a frustrating time due to the very meticulous way of setting keys and using human IK and the blend shapes window for the facial animations. I began a search for ways of using purely the adobe fuse characters in a more efficient way. this would also cut out the need to wrap different characters in clothes from the fuse characters. Miraculously I found this script which made my life so much more easier.

The only issue is that you need an adobe account to be able to use fuse and export fbx files as opposed to autodesk character generator which is free. It just meant I had to travel down to Birmingham to talk to Datomec to see if I could use their adobe account to create these characters.

I think there is definitely a big future for this kind of character creator. especially due to the quality of these characters. If I was to learn how to create these characters from scratch and rig them properly to the same level of professionalism I would need another 2 years ... this is not possible obviously. But this also applies to the CGI world for example if a studio had limited funds, an adobe fuse character would be a lot cheaper than to hire a character mesh creator, a character texture and UV skinner and rigger. It would also save tones of time. this is why I think this method has a big future. especially with its increasing professionalism and complexity.

Castle

Due to timing issues with the project, there were a few work arounds that we had to take resulting in slightly different results as to what was plans. for example: the main wall between the camera and the castle had to be made removable as only one side of it could be realistically completed and also the castle had to be simplified to allow the camera to hide the fact that it was half built.

The only reason this happened was the lack of communication between me and my colleague with the absence of setting regular milestone deadlines. But Due to the pressure that was placed on us both I think we will both be using this as a very valuable lesson when approaching future work. The plans might be comprehensive, but when regular deadlines aren't met then the development of the project can spiral into a stressful chore. This was beneficial to my own practice as it presented me with a first hand experience as to what it is like when regular milestones aren't hit, and with something like CGI taking into account render times then this was certainly something I wanted to avoid.

Walls

The universe in which was needed had to be consistent. This meant using a method of texturing walls the same along each set. Due to the fact that the scale was much larger than I am used to I had to experiment with acquiring a convincing brick texture. This came about by cutting hundreds of squares out of cardboard:

I then applied these squares onto a base of cardboard:

I experimented with different sizes and spacings between each brick. I then applied paper mache onto the whole thing to try and blend those spaces in-between each tile.

I realised that this made it look quite 'pillowy' which would suffice for outside walls that would have countless years of wear. but the inside walls I decided to just dry brush straight over the top of the cardboard tiles. 

The effect worked! The ground was simply painted grout to add that little touch of texture to make the set look a bit more believable.

Spinning Wheel

With the one set not having anything in view other than a spinning wheel, I wanted to put quite a bit of effort into the creation of the spinning when which also had to be functional. Here was the reference image I used:

Initially I was thinking that one sat on the wheel itself so it was beneficial that I looked up some reference images to keep it as accurate as possible.

The wheel was crafted using a curtain pole ring that was sanded to shape and a groove cut into the outer side of it to allow thread to be placed. The spokes were created using chicken skewers that were cut to shape and the rest was created using different scraps of mdf and dowling:

This process was really beneficial due to the confirmation that reference is key to keeping everything as accurate as possible. This also informed my own project as I built quite a heavy library of reference images for creating my assets.