Week 31 – Final Going Live, Cymatics Project Production – Nuke Tracking, Maya Proxy Collision Objects

On Wednesday of this week, we’ll be presenting to Axis in Glasgow, so for the sake of completion, here’s the final film, Concerning Dragons:

Cymatics Production

So, the time has now come around to focus soley (apart from the odd presentation) on the production of my project. I have begun going through my video clips and organising what works and doesn’t, filtering out the unusable shots before deciding which ones are best. I will look at the storyboards again, and match up the shots with that timing-wise, before logging which ones need to be 3D tracked and so on. Early days yet, but much to be done!

I have also gone through the HD cymatics video of the reference footage, and taken stills of the best individual frames for modelling. Here are a few examples:

There are so many, I’ll have to limit myself somewhat to bear in mind the amount of time I have to model them. There are, however, recurring themes or glyphs throughout the piece, as the melody and harmonies are based on the same melody throughout.

Aside from the glyphs, I have also begun on one of the few scenes where no cymaglyphs are seen. At the beginning of the film, there will be some musical notation where the notes pop out from the page 2D into 3D. Here is a low resolution test:

Project Shot List

I have created an initial linear shot list that matches the storyboards, with a couple of changes, but the idea is that I can see what needs to be done in terms of tracking and effects etc.

Shot	3D Track	Camera Work Completed	Post Production	Completed
1. Writing notation	No	Yes	CC, effects
2. Overhead piano	No	Yes	CC, effects
3. Walk in with bosch	No	Yes	CC, effects
4. Place down bosch	No	Yes	CC, effects
5. Left right pan roof higher	Yes	Yes	Cymatics, CC, effects
6. Zoom out diagrams	No	Yes	CC, effects
7. 3D pop out music	No? tbc	n/a	CC, effects
8. Left right jib pan better	Yes	Yes	3D Sine wave, Cymatics, CC, effects
9. Jo beach no hat headphones	Yes	Yes	Cymatics, CC, effects
10. Long shot back church static	Yes	Yes	Cymatics, CC, effects
11. Beach aahs	Yes	Yes	Cymatics, CC, effects
12. Overhead piano	No	Yes	CC, effects
13. Beach aahs	Yes	Yes	Cymatics, CC, effects
14. 2 handed piano	No	Yes	CC, effects
15. Behind bosch	Yes	Yes	Cymatics, CC, effects
16. Zoom out from stage	Yes	Yes	Cymatics, CC, effects
17. Bosch box? tbc	Yes	Yes	Cymatics, CC, effects
18. Tilt from stage to roof	Yes	Yes	Cymatics, CC, effects
19. Blair roof? tbc	Yes	Yes	Cymatics, CC, effects
20. Static sound hole	Yes	Yes	Cymatics, CC, effects
21. Inside car	Yes	Yes	Cymatics, CC, effects
22. Drive past	Yes	Yes	Cymatics, CC, effects
23. Fire ground static tbc	No	Yes	Cymatics, CC, effects
Quick cuts of fire and church? tbc		Yes
24. Fire pan left to right	Yes	Yes	Cymatics, CC, effects
25. Fire people 2	Yes	Yes	Cymatics, CC, effects
26. Fire people ensemble	Yes	Yes	Cymatics, CC, effects
27. Fire sparks 5	Yes	Yes	Cymatics, CC, effects
28. Queens farmhouse	Yes	Yes	Cymatics, CC, effects
Total 3D Tracked Shots	18

Tracking and Building Proxy Collision Objects

The first shot I tackled was the car driving past the camera. The idea in this shot is that the sound bubble will be coming out of the window and distorting in the wind as it drives past, so I assumed I needed a tracking shot of the car, something close to the open window. A 3D tracking shot wouldn’t work here, as the car is moving, so that would compromise the track. I could try masking out the car, but first I thought I’d try a regular 2D track. There wasn’t much to focus on within the shot, so I ended up manually keying most of it (147 frames). I will return to that shot later.

I then 3D tracked a shot within the church to see what the process of tracking and setting up collision proxy models for the particles would be. Following the same procedure from before for tracking, I opened the fbx file and tried to judge from the Nuke file, where thepoints in space would roughly be. However, I had forgotten about the camera where I can assign images from the environment tab – use image sequence – so thje basic boxy shape I had crated for the church, I added a transparent material to it to see the images behind from the render and then rotated and adjusted the church shape to match the render, as seen below (the first one shows the image sequence details, then the church shape modelled to match the images):

The next stage then (or one of them at least), is to bring in the particles and make sure that the collisions look as though they belong within that space.

Nuke Tracking File

For the sake of completion again, here is an image of the Nuke file which includes the merge geo and write geo nodes to export the camera data out to Maya:

For some of the image sequences, there are the odd frames missing or broken for some reason, so I shall need to re-import them. In addition, I am batch renaming any files that do not begin at frame 0, as this will make things easier once in Maya.

I am also masking out the fire for tracking purposes, by:

• adding in a roto paint node
• using the bezier tool to roughly draw around it
• move it every few frames as necessary to create keys (select the transform tab)
• then selecting ‘mask alpha’ within the mask selection on the camera tracker

Going Live Presentation

A mention has to go towards the prsentation this week, and how well it was received. The number of views on the video now stands over 20000, and after getting a congratulatory email from the Tolkien Society, it seem that the film is taking on a little life of its own. So, well done to everyone, and I don’t think we could have done a much better job.

http://www.dundee.ac.uk/djcad/news/2014/article/djcad-students-animate-dragons

We also visited Digimania in Glasgow, who work on games, but at the moment are producing a 52 week childrens animation. They may be hiring animators later in the year…
Finally this week, I had a meeting with Phil Vaughan concerning the focus of the last semester, which essentially starts now. So, I have already begun, as noted above, on tracking and setting up proxy collision objects, the next thing to do is map out a plan for the next 12 weeks or so so that I can keep a close eye on what I have to do.

Week 30 – Going Live Finale, Advanced Production

Going Live

So, this week was the initial final date for the going live production, and as such we have finished pretty much everything. Unfortunately, and much to my frustration, I haven’t been able to produce smoke of a sufficiently qualitative nature for scene 2. This has not been without effort – following the fluid container efforts, I then tried using particles as a source for the smoke. And as much as it works, I felt that the quality wasn’t good enough – fluid containers produce more detail and I felt that the results I was getting with the particle smoke was too ‘c.g.’ looking, and I didn’t want to compromise the production – not that I would have been allowed anyway!

Here are 2 versions of the fluid container smoke followed by the particle smoke for comparison:

Particle Smoke:

Advanced Production

The weekend saw me, and a much appreciated helping hand from John Harrison, film scenes inside Heartspace – the renovated church space. This was to be the main space, and more than likely, main scenes for my Cymatics project, so I wanted to make sure that it went as well as it could do. Thanks to the generosity of Finlay Wilson, the director of Heartspace, we had the place to ourselves and only a small issue with the camera jib was a problem. The end of the jib that holds the camera in place was screwed on upside down, meaning that I couldn’t tilt the camera to counter for the slightly less than horizontal flooring. However, the shots look as though they will more than suffice – I have yet to organise them!

In addition, on the Friday night, I drove up to Pitlochry, to Queens View, where I shot the scene which will be the final shot I the project – looking down a valley with a couple of dwellings where music will emanate from.

Finally, on Sunday night, I added in a couple of the initial shots which include myself writing the musical notes on paper before the Cymatics take over.

Wednesday 23rd April

Following a few comments concerning then ‘rushed’ nature of the narrative of ‘Concerning Dragons’, the decision was taken to re-record the voice over, which we did in the recording studio here in the Matthew building. It has Pro Tools 10, so I was familiar with the set up. The new take has a few removed words to clarify some of the language. I think that this will be the last tweak. On top of that, we have been compiling material to be used in the Going Live presentation next week, including animatd turnarounds of the main models, videos of tests and various stills of the various stages of production. Kieran and Lilly are going to do the main presentation as they are the ones most likely familiar with all the stages of the production as an overview.

Thursday 24th and Friday 25th April

We have the dates in for the semester 2 assessment presentations, 21st May, so I have begun compiling materials from the beginning of the year, including images and test video etc. There was also a requirement for the Going Live project to write a 600-800 word report covering the individual work done and roles played within the team etc, which I have begun. In addition, I have completed my application for a possible PhD in Bournemouth, so we’l see if there is any more positive feedback in due course.

Week 29 – Going Live, Advanced Production

Advanced Production

Last week ended with a weekend filming session in various locations outdoors – mainly focusing on the camp fire shots and a couple of shots down at the beach with a lighthouse in the background. It was the first time I had used the Sony EX1, but after a few shots I felt more comfortable with it. One of the things I will need to look out for next time is the difficulty in focusing accurately just using the viewfinder, rather than a large screen. Use the peaking indicator to check where the edge of the focus lies to make it easier.

The weekend coming sees another round of filming, this time inside Heartspace, the renovated church space. Most of the projects filming will be based in there, so I have booked out the dolly track and jib for some smooth shots, hopefully!

Going Live

The final week for Going Live has me still working on fire and smoke. After several attempts to produce a realistic looking fire breathing dragon within a static fluid container, I have opted to use the original render but complement it with a 2D container to have some ‘trailing’ flames behind it. The static container renders looked too ‘c.g.’ unfortunately, so ‘faking it’ was the last attempt to improve the look.

As for the smoke, it was very difficult again to get any quick results as the calculation times for the simulations are so long. So, I ran a whole series of frames through the render farm to discover that it looked horrible – far too blocky and fast.

So, I tweaked the shader settings to improve the look and discovered various options:

• By caching out frames every 0.1 for the ‘evaluate every frame’ setting, when I played it back at a scaled up 10 times faster version, it should look good – alas, it didn’t – there is too much noise and too many artefacts produced when over sampling. I tried this at 10 times, 20, 50 and even 100 to see if it made any difference, I also tried changing the solver quality and substeps (to see if by adding substeps it would stabilise the fluid), but to no avail.

• Finally, I tried changing the frame rate to 300fps, with the idea that I could speed up the camera move and slow it down in the comp. Again, however, the evaluation of the fluids happens on a frame by frame basis, not a time based one, so the end result would have looked just as turbulent as the other efforts.
• The last ditch attempt now, is to hold the last frame of the cache for 200 frames and fly the camera over that – single renders have been proven to look good, it’s the fast motion of the smoke that has proved very difficult to contain at scale.

In addition, I have been working on the sound effects and editing the music for the final film – I recorded some paper for the bird flaps and I will use a number of sounds I have gathered over the past few weeks including: wind, water, squeaking leather (for the tree sequence) and breaking wood. Tom has recorded the voice over and I will pull all of the elements together this week. I recorded the music using Pro Tools, but I will put the final music edit together using Adobe Premier which allows me to synch the audio up with the visuals over several audio tracks. Here is the last but one version (hopefully) of the sound and effects along with the animatic.

http://youtu.be/NDg1eAr1Ykw

 

Week 28 – Going Live, Advanced Production

Going Live

Once again, this week i’m still tweaking the fire for scene 3 in Going Live. Having created a number of different versions, one of the most successful was the version without smoke, as the smoke tended to linger within the 3D container. The problem being when the contaner was animated the smoke moved along with the fire in an unnatural way as I have mentioned before. Kieran preferred the version with smoke so once again I tried changing the shading attributes by setting the incandescense to temperature so that the version with purely fire would have a little smoke as the age of the particles died out. This was the version I cached out and rendered, and as good as it looks, the problem still arises with the animated container. The only way around this would be to have a large static container with the particles moving within it. I tried this before with little success on a large scale, but I will try again in the knowledge that if it doesn’t work, at least we have a version that we can work with. The image below shows what happens with using particles in a static container, where the problem arises when the speed of the animation is too quick for the fire to render as a solid stream. I am going to try adding in some volume emitters to see if I can ‘fill in the gaps’. To be continued…again…

This one was using a scaled out volume emitter in the shape of a cone:

I looked again at scene 2 and the burning paper, this time adding more detail and using the knowledge from the last 2 weeks of tinkering with the dragon fire. The net result is that I have a much improved version that will take some time to render due to the 35 gigabyte cache created for the 2 fluid containers! Only thing I forgot was to turn off the glow – hopefully it won’t be a major issue.

For scene 2 in Going Live, I have begun creating some palls of smoke coming out of the castle area and from under the bridge, emanating up toward the camera and slightly back to the left. It took me a bit of tweaking to get the right density and colour – first of all I tried emitting from copied parts of geometry but for some reasons as yet unknown it didn’t work well. I swapped that out for volume emitters and roughly matched where the smoke was coming from (in terms of the geometry) and made the geometry collide with the fluid containers. Kieran also noted that the dragons would be flying through the smoke, so I cached out the geometry for the dragons first of all so that I could scale the whole scene down to 1/10^th of the original size to accommodate the scale of the smoke. Then making the dragons also collide with the fluids, I cached out the smoke before sending it off to the render farm.

On Friday, I took the Sony EX1 out for the weekend, with the focus on getting some outdoor shots for my own project. The shots I hope to get are the camp fire ones as well as the outdoor shots of the car and some scenic long shots with a few dwellings in the distance. Next weekend I will be doing the same but with the focus on the indoor shots for the church at Heartspace.

Meeting with  Nigel Johnstone

This week I also had a meeting with Nigel Johnstone concerning the possibilty of doing a PhD at Dundee, or at least getting some advice from him. The overall feeling is that, regardless of Nigel’s interest in my proposal, funding is scare and extremely difficult to come by. Worsened by the fact that the main sources of funding for 2014 have now closed. it means that my options are limited to going through the channels with an application and hoping that a suitable supervisor could be found along with some unseen funding. So, highly unlikely. It means therefore, that the best I could hope for is applying for a PhD start next year, but clearly I have to survive, so I’ll be looking for a job most likely in the meantime. The only other possibility is that one of the institutions down south expresses enough of an interest (I have had a reply from Bournemouth) to follow through on that.

Degree Show Gig

On an unrelated matter, I will be performing at the Degree Show opening night at Liar Bar between 9pm and 11pm roughly…