Apologies for not updating on my progress in a while. Last week was a particularly busy and unusually stressful week. As we near Christmas our teaching is leading us into 3D and allowing us more time for our assignments. Last week I was working on some shots for a short film that is to be released in 2018 called Turning Tide. Naturally the film is still in production and so I have not been able to share any of the work with you, but I can tell you it has been a good experience so far. If you are curious about the film you can read about it on it’s kick starter page.
We have amassed a significant amount of knowledge in these short weeks, and though patchy I feel like the breadth of the software is becoming easier to envisage with each passing day. I deem a lot of essential work in this stage of the course to be in exterior learning. The supplementary knowledge and clarity I gain from tutorials is most helpful and can assist my current project.
This weekend I refreshed my memory on basic 3D geometry within Nuke. Nuke is not inherently designed to manage and produce complex 3D geometry, but the tutorials I have followed allowed me to understand the capability of what Nuke can provide. Even with basic geometry and shaders you can create interesting compositions and basic motion graphics. I am aware that the next part of the learning is taking this a step further and using 3D geometry to help ‘rebuild’ scenes for the purposes of clean-up and projection.
We’d all like to not be doing too much over the christmas break, so I imagine there will be a lot of heads down this week.
We are moving into deeper and more complex compositing techniques now. Nuke has a remarkable ability to work within 3D space and manipulate 2d elements in the quest for seamless composites. Whilst we would never do any bulk 3D work or modelling in Nuke, there are a plethora of reasons to use it’s capability. This week we have been learning about 3d camera projection, allowing us to create virtual cameras that mimic the motion of the original cameras. From there, we can composite 3D geometry or cards into the scene in real time and with great accuracy. This leads towards projecting new elements or backgrounds which is particularly important for clean-up.
My clean-up and CG shot is under-way and I’m employing much of what we are learning. We need to be aware of all the 2d techniques that we learnt as early as week one also. I have had to refresh my memory on certain basic techniques!
An example of the 3D camera track and point cloud generator. This data will for the base calculations for a lot of projections. Below is some traditional 2D cleanup using match-move and patch roto’s. I’m currently working on painting out the lamp-post which is a tricky and time-consuming task as I have to guess the background. Might have to project a new one instead.
Well, the Roto is finished. It was very difficult but I learnt a lot. I feel in a much better position going forwards, which is ultimately a good thing. You can read about the assignment here: https://wp.me/P9fsKB-88
Class hasn’t stopped and now we have moved into 3D within Nuke. We started off with basic geometry and how to get 3D elements into the 2d space.
We then moved onto displacement mapping, allowing you to manipulate and distort 2d images to fake a sense of depth, perspective or scale. This is id done through luminance values and is quite interesting.
It is particularly useful in making static 2d backgrounds look 3d.
I also got myself set up today for the next assignment which is clean-up and CG composition. This assignment asks us to demonstrate our ability to clean-up elements and seamlessly blend CG elements into a shot. I found this great bit of footage that gives me a lot of opportunities for clean-up. The CG elements will either be sky traffic or a CG drone. Haven’t got long to get going on this so will have to get started straight away.
Only one more to go.
Unfortunately it continues. A journey of discovery and fear as I attempt to put together two moving characters and a racecar. I had a panic earlier but upon review it’s not going bad at all. You need lots of motivation to see it through that’s for sure, but I’m positive the end result will make it worthwhile.
It’s revision week to conclude the half-way mark for short-course/vocational students. This is a great opportunity to recap on what we’ve learnt, but I have not had a moment to digest it. I have been working flat-out on my Roto. I received good feedback from my tutor Allar during the presentations on Friday. I’m quite happy with the racecar but I still have two organic humans to Roto, a presentation to prepare and a write-up to present by Friday!
It’s coming along. There are very different obstacles with organic elements. Luckily I have saved a lot of time by doing my retrospective daily so the essay should be fine. I think a great deal of my attention is focused towards getting this Roto shot squared away.
Not much to report today. We went over our Roto shots to prep for the mock presentation on Friday. As a result of the feedback I got I tried something new in the afternoon. It turns out it is probably a tracking technique I should have employed right from the start. The matchmove tracker has made the area I tackled in the afternoon a lot easier, and I think I will have to go back and revisit other areas that could do with a similar treatment.
There is a lot of chopping, cutting, re-arranging and restarting. This is all being documented in my retrospective. I think these obstacles and revelations are good and will stand me in good stead when I tackle the human figures of the shot. I have also switched to Silhouette FX 6.2, the latest version of the software. This seems to be slightly more powerful and intuitive. Importantly, it didn’t break my project! I have to get a shift on so throwing in the late hours to accomplish this Roto.
This week we start learning about Colour Grading and Theory. This is a fortunate stroke of luck for me because for a while I thought I was going to be a Colour Grader! I even did all the prep, like reading Alexis Van Hurkman’s critically acclaimed Colour Theory book ‘Color Correction Handbook: Professional Techniques for Video and Cinema‘, and learning DaVinci Resolve. I have even done one or two projects with independent film-makers. That doesn’t necessarily mean there isn’t more I won’t learn this week, but from a pressure standpoint it really helps me relax with regards to being able to put more time towards my Rotoscope shot.
Incidentally, I am aware I haven’t shared any progress of the Rotoscope on here. As part of my assignment I am keeping a full retrospective log of events and decisions which is due in as a write up alongside my work. In this retrospective I detail the motivation behind certain decisions, challenges faced and crucially where I failed. I am also taking the time to read around the subject of Rotoscoping, hopefully to validate or throw into contention reasoning behind my work. Crucially, doing this contextual reading around the topic lends a hand to the real outcome of this all, becoming a Master of Arts. I won’t consider it a good retrospective piece unless I can draw some parallels or insights from whats happening in the industry now.
For my first Roto shot I eventually (after a great struggle!) narrowed it down to a topic I have a soft spot for – Motorsports. Yes, I somehow managed to pick a subject that will not leave me bored after staring at it for 80+ hours! The shot in question is of a Formula car (I think Formula 4 or an Indy Lights Car) rolling away down the pits, with a mechanic or friend hopping away just before. This combines two useful elements. Firstly, the inorganic nature of the race-car will allow me to demonstrate my ability to bifurcate accurately and will test how uniform and harmonious my shapes can be. Whilst on paper it is just a car moving from point a to point b, there are a lot of nuances to pick up and edges never keep as straight as you think. Secondly the organic nature and movements of the mechanic will test the opposite technical skills. Crucially, as a result of the mechanics motion, the shutter speed of the camera and the quality of the shot, there is a lot of motion blur on the character which is an important skill to conquer. According to my tutors, being able to accurately motion blur is half the battle and impressive to potential employers.
I’ve managed to get the first half of the car done. I have restarted from scratch a total of two times to get to this point. This is largely due to going in blind to begin with and realising that, actually, reading around the topic and watching a few tutorials before diving in wouldn’t do any harm. I took the trade-off of losing a day or two to this preparation over the estimated time saved not mucking about or constantly reinventing the workflow. I’m really glad I did. Benjamin Bratt’s book ‘Rotoscoping Techniques and Tools for the Aspiring Artist‘ was a phenomenally sound purchase. Concise and to the point, it really gave me a good grounding with regards to how to tackle my shot. It has reflected in the speed at which I am now starting to work on the shot. I feel confident with my judgements knowing that I have a mental checklist to tick off as I go along. Perfection of Rotoscoping will be a product of time, but getting the mechanics down will get me there in half of it.
As you can see, coming along nicely! I’ve enjoyed working on the shot thus far. I’m no mechanic but I am interested in race-cars. I think my careful study of the shapes of the race-car is reflected in the roto. In all honesty I think the fill area between the focus edges is overly complicated, but I simply got carried away dissecting the race-car into it’s shapes and forms. It’s not a bad thing when you get a bit lost in your work, getting into ‘the zone’ is one of the great feelings of being an Artist.
Another foray into Keying today. This time we looked at the IBK keyer. This tool is very useful for generating greenscreen clean-plates if the production crew neglected to shoot one in the first place. Not only can it do this, it can retain a lot of luminance detail that other keyers such as Primatte struggle with. This is particularly true for motion blur and areas of high transparency.
There are so many varied ways to apply the IBK, and indeed they can even be used in conjunction with other keyer nodes. Of course that is one thing, but making them all play nicely is a different story. Knowing the strengths and weaknesses of each key node is important, but I also think perhaps personal preference comes into play also.
For my own part, when I practised putting my own face into a TV screen, I drew great satisfaction from the luma despill generated by the IBK. It effectively made ‘reflections’ on the rim of the TV screen, with minimal effort on my part. What’s better was I did this by mistake, so it was a good surprise.