So here is the result of the past months work, from idea to storyboard, previs to film shoot, modelling to compositing. Hope it isn’t entirely hopeless

This doesn’t mean I haven’t been working on the AVE ICA, merely that I’ve been busy to the point of not finding time to keep the blog as up to date as I probably should have.

Starting with the week leading up to the Easter holidays, Andreas Nor was kind enough to help me record foley sounds in the Phoenix building. The week after, home in Norway for a couple of weeks, I began building the 3D scene. At this stage I had ruled out Vue as it caused too many new problems that needed solving. Instead, I opted for 3ds max, using Mental Ray as the render engine.

During the easter holidays I created the first version of my 3D tree and tracked all five shots in PFTrack.

As I had only used PFTrack once before several months ago, I had to relearn a fair few things, but the tracks went well and I feel they ended up being pretty solid. Having created the tree and the tracks, I actually began rendering a few shots and performed some test comps.

I wasn’t exactly satisfied, but before I had time to make any changes, Easter was over and I returned to the uni merely five days before the final year project was to be handed in. This forced me to stop working on the AVE project for a while and as I later got an extension on the final year project, two weeks went by without any updates to the AVE ICA.

After delivering the final year project on 26 April I once again returned to the AVE project. During the for lack of a better word “break”, I had decided that the scene I made originally was too heavy, too boring, badly lit and lacking in detail. Starting up again, I therefore chucked nearly everything in the scene. The branches, the lighting and half the textures were some of the things I got rid of.  From there, I started with a tree from seek.autodesk.com, rotated it on its side, added moss and plants and tried to make it look more detail-rich.

Having now created one branch, I converted it to a Mental Ray render proxy and made ten instances in 3ds max. From there I started populating the big tree. Using instanced render proxies greatly sped up the viewport and reduced memory use, though it had the obvious drawback of looking repetitive. In hindsight, I probably should have made a few variations, but time didn’t allow.

As soon as I was reasonably happy with the finished tree, rendering begun. The shots were quite heavy to render due to light spilling through a lot of leaves. With rendering settings dialled lower that what I wished for, only a few days left until the deadline on 7 April and still without finished renders, I had to ask for an extension in AVE as well. By the time the renders actually finished, the deadlines in Studio Practice & Business were closing in fast and once again I had to take an unwanted break from AVE.

In the end, the last Studio Practice & Business deadline was late on a Tuesday and the AVE deadline was set for the following Friday. This effectively meant that I had less than three days to key and composite five shots, solve any unforeseen problems, cut the project together, write a report  and hand in the assignment before 4pm on Friday. I delivered 3:50pm.

Having finished I think the project turned out quite well. Sure, the pulled keys are imperfect in places, the leaves aren’t moving due to Mental Ray render proxy issues and the render and contact shadows/lighting could have been improved in places, but overall it could definitely have been worse. I also gained quite a few skills during the project, such as more advanced use of PFTrack, better understanding of roto and comp, and I used the 3D tools in Nuke for the first time.

Next and probably last post regarding the AVE ICA will contain the finished piece.

These are the slides I will use for my company pitch on Tuesday.

Download Company Pitch PDF (2.6 MB)

Download Company Pitch MOV (14.7 MB)

Download Business Plan PDF (690 KB)

Now I’ll go to bed happy

After some initial feedback to my first attempt at business cards, I decided to to go back to the drawing board and try again. This time, I waited until I had the logo ready and ordered the cards from moo.com, rather than using my £40 printer and flimsy paper. Like with the website, the cards might benefit from a tweak or two, but they should be OK for now.

Having come up with a name and logo, a company website was next on the list. The purpose for the site would be to make it easy for clients to find my company and contact details as well as well as give the company a place to show recent work.

The site was created using WordPress and a custom theme called Aperture and can be found at http://matteinverted.thimic.net/. The page will need more work in order to be used as a company website. I feel this is a good starting point however.

Another week has come and gone – way to quickly as usual. This week I’ve been looking at music for the scene, booked a foley audio recording session in the studio, experimented with keyers and possible looks and started matchmoving.

Here are the music sources I’ve investigated and pros and cons:

opuzz.com university catalogue

• wrong kind of music
• hard to browse through

incompetech.com

• free
• not the best music
• difficult to go browse

istockaudio.com

• some pearls
• a lot of junk
• costs money, but affordable

mobygratis.com

• limited
• some promising pieces
• need approval and haven’t gotten it yet, don’t know whether I’ll get it

www.unippm.se

• good music
• known artists
• £80 for 60 seconds of music for use at the university, another £80 for use on the web

Soundtrack Pro

• limited
• overused
• some good music
• not anything that fits too well

At this time I’ve found some pieces that might work, I just have to look through them and see what fits best. I found one track i really like on www.unippm.se, but there’s no way I’m coughing up £160 for a 50 second school project where the music is secondary anyway.

As eagle eyed readers might tell from the image in the beginning of the post and the video directly above, I’ve started keying the footage and trying different looks for the “apple” the actor is holding. As I didn’t feel quite up to scratch on the advanced keyers in Nuke, I asked Steven in my AVE class for some help and he showed me the basics of the IBK keyer and Primatte. Thanks Steven The key here isn’t finished yet. There is an edge to get rid of and one tracking marker left that needs some painting out. While doing this I ran into a problem where Nuke would crash every few frames for some unknown reason which made the work rather tiresome, but fortunately I found a solution for it in the end. The glow on the “apple” is a tad extreme, might reduce it by a bunch and remove the volume rays.

I’ve also started matchmoving this week and as it’s been a while, I’ve once again forgotten a thing or two about PFTrack. Might look at some of my fxphd classes to freshen my memory or be as radical as to actually try Boujou. Crazy, I know!

A problem with using this name however would be that I wish to work in the UK and that Thingnes doesn’t roll off the tongue that easily on this side of the North Sea. As a consequence, I decided on creating a business name. When a few weeks had gone by without any particularly good ideas, I started looking into words related to my line of work as a rotoscoping and compositing artist. It was then I came up with the name MATTE | inverted. Both compositors and roto artists work with mattes on a daily basis and going off the idea that a matte is black and white, this also gave me an idea for the logo; basically having one word on white and the other on black, just like two opposite mattes. A black and white logo also gave the extra bonus of working well on print.

Overall, I feel the name and logo work well enough, though the logo might be a bit boring or corporate looking. If I am to actually set up a company after Teesside, I might try create something a bit more creative-looking.

Source: businesslink.gov.uk

This week I’ve delved further into Vue, figured out how to undistort and redistort my plates properly in NukeX and chosen and cut together the takes I’ll be using for the final scene. The cut can be seen above.

To be able to get a good 3D track and also making the CG environment look correct, it is important to undistort the plate before tracking and redistort the CG plate before compositing. As the camera lens is curved and the image rectangular, all lenses produce distortion where what in reality are straight lines become curved to some degree. The amount of distortion varies with the field of view of the lens and also focus if the lens is breathing while focusing.

To figure out the amount of distortion I shot a lens grid for each and every one of the shots in the scene, making sure the lens properties were exactly the same as when filming the corresponding shot. This way, I was able to use the new LensDistortion node in NukeX to analyse the grid and produce a slightly larger, undistorted image; basically by pulling at the corners. I then plugged the actual plate into the node, undistorting the plate itself.

Now that this is done, I can render the undistorted plate and get a better 3D track in PFTrack. This undistorted plate is also the one I’ll be using as a background image in 3d studio max to match my CG environment to the plate. Then, when the CG is rendered, I’ll use the LensDistortion node in NukeX in reverse to distort the CG render; thus shrinking the CG down to the right size and making it look more believable due to slightly curved lines.

It must be noted that in this case, the distortion is relatively insignificant, meaning I could probably do the shot without bothering about it. However, as the distortion can be a lot worse, I feel it is good to get into the habit of doing the thing properly so I know how to do it in the future.

1 Business Summary

Lorem ipsum dolor…

1.1 Business Overview

• freelancer providing services as roto/compositing artist to vfx facilities
• industry: visual effects
• key benefit: flexible, both regarding working hours and the ability to relocate
• no plan on growth, will either continue freelancing or find a job

1.2 Service Features

• mainly do roto/compositing for vfx facilities
• might expand to matchmoving and some 3D work as well if requested
• flexibility and delivering consistently and within deadlines will give an edge

1.3 Market Analysis

• targeted market: vfx facilities and other companies in need of vfx artists
• business model proven to work by other artists in the same field and by Hans Rijpkema at Rhythm & Hues

1.4 Marketing Strategy

• showreel
• web site
• facebook/twitter/linkedin
• business cards

1.5 Key Objectives and Financial Overview

• Being able to scrape a living from year one

2 Detailed Plan

This is the detailed version of the Lorem ipsum dolor…

2.1 Market Analysis

• Clients: Vfx facilities, ArchVis companies etc.
• Geographical area: Europe, US and Australia. English-speaking countries
• Market is growing
• Main market – vfx facilities – somewhat segmented into specializations such as environments, creatures, photo realism and fantasy
• Market influences: economy, tax regulations, some times strikes
• Reach customers either by applying, like for a job, word of mouth or possibly agencies
• Seasonal trends: Summer blockbusters and Christmas movies makes the months leading up to these the most busy
• Pricing sensitivity: Depends on location, but will charge roughly the going rate, which at this time is approximately £3000 a month for a person in a beginners position
• Competitive edge: Can work “anywhere” & any time, quick delivery, high quality work
• Sources: fxphd.com, fxguide.com, Animex speakers
• Benefits: Can work “anywhere” & any time, quick delivery, high quality work

In the past two weeks I’ve been looking closer at Vue 8 xStream to see whether it can help me create the vegetation I need for the ICA. I’ve had a closer look at how good Vue is integrated with 3ds max, how easy it is to use and how well it copes when rendering out passes in 3ds max with Vue elements thrown in. Though I can’t say I know the application that well yet, I’m pleasantly surprised with what I’ve found so far. The 3ds max integration works really well and my biggest fear – that ambient occlusion passes and z-depth passes would fail because the tree leaves are just textured cards with opacity maps is put to rest:

This is a simple tree from Vue, rendered with Mental Ray in 3ds max. As you can see, the z-depth map works very well. The exaggerated zblur test done with Nuke underneath shows how this might help me apply depth of field during compositing.

All I really need to find out now is how I can customise trees to convert them into branches and that sort of thing. I’ll ask Chris about that next week.

I also had a quick look at FPHoe this past week as an easier and cheaper alternative to PFTrack, but found it to be too simple. PFHoe has no built-in tool for masking and the lens distortion tool feels very crude. I think I’ll stick with PFTrack.

Finally, because my Windows 7 RC installation was about to expire, I had to reformat a perfectly good system and install a fresh copy of Windows 7 this week. This turned out to be quite a hassle as the new Windows 7 drivers provided by the computer manufacturer caused the system to crash every 30 seconds or so. I’ve got everything running again now, but it took me almost 16 hours to get things back to how they were. Next up, now that the computer is all sorted is to create the final 3D scene. I have a fairly good idea of how to do this now, given my tests during the previs and poking about in Vue, so I hope to get that done fairly quickly.

I made this business card to have something to hand out on Animex if needed. Thought it might be used for the Studio Practice & Business module as well.

The press release can be downloaded as a PDF or viewed below.

Edit: The press release has seen a few updates. All are available here as PDF documents.

Download Press Release Version 3 PDF (45 KB)

Download Press Release Version 2 PDF (29 KB)

Download Press Release Version 1 PDF (27 KB)

MATTE | inverted
Michael Thingnes
+44 7717 454408 / +47 482 31 245
matteinverted@thimic.net

http://matteinverted.thimic.net

PRESS RELEASE

PROMISING GRADUATE STUDENT NOW TAKING ON CLIENT WORK
Middlesbrough, 15 June 2010 — Michael Thingnes is a fresh graduate student out of the Digital Visual Effects course at Teesside University. His primary skills are rotoscoping and compositing, but he also feels comfortable as a 3D modeller, render artist and matchmover. Michael has now set up a sole trader company called MATTE | inverted, which will aim to offer these services to visual effects facilities of all sizes, although he will happily consider working for other clients as well.

In addition to his Teesside University bachelor degree, Michael recently completed work as a compositor and 3D artist on the short film “Moving Day” by Jason Wingrove and has taken courses at fxphd.com involving compositing theory, Nuke, After Effects and Maya. He is planning on continuing taking courses to gain more experience and stay updated on software.

Experience wise, Michael considers himself fairly skilled in the use of Nuke, Photoshop, 3d studio max, Final Cut Pro and PFTrack. In addition, he has some experience using After Effects, Maya, Realflow and Massive.

Michael is currently based in London, UK although relocation nationwide or worldwide is definitely possible given that the client can guarantee six months of work or more. If you believe Michael’s services can be of use to your company, don’t hesitate to email him or give him a ring. He will be only too happy to talk to you.

FOR RELEASE 9AM 16 JUNE 2010

As I attended my first Animex ever this week and I wanted to see everything, this was a bad week as far as productivity is concerned. On Wednesday however, I attended a workshop with Pete Draper, the author of the Deconstructing the Elements books for 3ds max, and asked him his opinion on how to best implement realistic-looking foliage into my scene.

He recommended going to http://seek.autodesk.com/, find some foliage I like and then put it into my scene, either by hand or by using particles.

I’ve had a quick look at the Autodesk site and found some models I like so if they work well when I import them to 3ds max, this will probably be my approach for creating that part of the environment. If they don’t, I’ll look into Vue and some other foliage generators out there, or as a last resort, simply create the geometry myself.

Strengths

• Specialising in one particular field to become as good as possible, but have some secondary fields to fall back on if need be
• Location not an issue, I’ll relocate
• Relatively few expenses
• Saving the client money and paperwork involved in hiring someone

Weaknesses

• Business relies heavily on one persons ability to produce good work
• Might need to move/travel a lot
• Can be exhausting
• Irregular income

Opportunities

• Becoming a better roto artist/compositor
• Learn new software
• Producing a better show reel
• Finding clients
• Different clients (hopefully) means diversity in work

Threats

• Competition
• “Dry spells”
• Software I’m not trained on
• Lack of experience and clients when starting out

As mentioned in my last post, I shot the footage for my AVE ICA this past Wednesday. That was the plan anyway. Turned out we overshot on time with a lot and only got three out of five shots before the day was over. Might have been the workshop we had on lighting a couple of weeks ago. We seemed to spend an awful amount of time on it. Whether the lighting actually is any good I’ll leave for others to decide. Fortunately, Mark was very kind to me and freed up the Thursday so we could leave everything as it was and come back and finish the next day.

Apart from taking a very long time the shoot went smoothly enough. I think I got all the measurements I needed, I shot lens distortion grids for every shot and photographed the settings on the lens after every shot for good measure.

A big thanks to Thomas Utbjoa who agreed to be the actor and all the other guys who took a couple of days off to help me out.

After the shoot, I keyed and tracked the first shot and must say I am reasonably happy with the result. Next up is creating the scene in 3D which needs to be finished by  26 February if I wish to keep on schedule.

As I really want to find employment in a visual effects facility as a roto artist or compositor, the chance of this business idea being more than hypothetical is very slim. To get the most out of the module however, I’ll choose a business that will allow me to mimic being employed in a visual effects facility as closely as possible.

According to Hans Rijpkema at Rhythm & Hues in California, only a little under half of the Rhythm & Hues staff is currently employed on a permanent basis. The rest are brought in depending on the need for people and skills on each individual project. This, from what I’ve heard, is quite common in the visual effects industry today.

The natural choice for me would therefore be to set up as a freelancer and target client companies such as the ones I hope to be employed by in the future. By billing them instead of receiving salaries, I can work for them as a freelance artist while using their offices and equipment.

This is quite good the way I see it as it will reduce expenses at my end. I will have to get some basic equipment for use at home and other than thatI can leave most of what’s left to salary, taxes, pensions and whatever else. I don’t see the real need for an office space in this scenario. If I’m working at a visual effects facility, I’ll have my office there and if I were to go without studio work for a while and having to resort to smaller jobs, I think I’ll be just fine with working from home. Given that I don’t foresee very big expenses in such a freelance business, I will probably be fine as a sole trader.

As of yet I’m not too tied down to any location in particular so I am free to move all over the world for jobs should I wish to. At the moment I would prefer feature film work in an english speaking country which means that the UK, the US, Canada, Australia and New Zealand are highest on my list. Of course, I might have to aim a lot lower than feature film at first to get clients.

Regarding finding a name for the business I think I’ll make it as simple as just going with my actual name. With the kind of work I expect to be doing, all clients will know I am a single person anyway. Given that I also want to seek employment some day, it would be good (hopefully) to have people associating the business with my name.

Not much to report in AVE this week. After finishing the presentation I booked the green screen stage so I can shoot the footage I need for the ICA. The shoot will be on Wednesday this week. I’ve also found an actor.

To prepare for the shot I’ll have to print some A3 size lens distortion grids today or tomorrow. The reason for the A3 size is that the grid needs to fill the frame and that a smaller size will have to be too close to the camera to keep it in focus.

As I had some slight issues with sharpness on my test shoot, I want to spend half an hour shooting a test scene and doing a quick keying test to make sure the footage is nice and soft.

Given that everything goes according to plan on Wednesday I’ll be on schedule with the project. Next up will be creating the final environment, produce good keys for the shots and matchmove them.

Christmas is long gone and I’m back in Middlesbrough after a refreshing break. The first thing I had to focus on when I got back was to create a presentation for Advanced Visual Effects where we were to present a storyboard, previs, schedule and a couple of other things regarding the ICA. As I still didn’t have a previs because of work on my Practical Project, this became my main priority.

This is the first project I’m doing which combines CG and live action, green screen removal and 3D tracking. Because of this, I decided to make life a little more difficult for myself than was strictly required by doing all aspects of the work I’m supposed to do in the final shots right in the previs. I did this in the hope that I would gain experience which might help me when doing the real effects shots later.

Basically, I tracked all five shots in PFTrack, modelled and rendered a fairly simple scene in 3D Studio Max, quickly keyed the footage and composited it together using Nuke. As this is just a previs there is still a ton of errors and problems with almost everything. I have continuity errors, issues of scale, existence of contact lighting and shadows in odd places, lack of contact lighting and shadows in others, horrible keying/roto, no colour and grain consistency between shots and so on and so forth. There are also two different actors as I was shooting on two separate days and didn’t have the same people available.

The process of creating the previs took about three days and in hindsight I’m very glad I did it. As a result I will at least do these things differently when creating the final piece:

• Shooting in 1080p instead of 720p to get a better resolution for keying
• Make very sure all forms of sharpening in the camera are turned off
• Shoot a lens distortion grid for every shot to make sure I can accurately get rid of lens distortion (I actually went back and shot grids for the previs, but as the distortion changed quite dramatically with focus and I only had notes for focal length it didn’t help much)
• Be more relaxed about the subject being partly outside the green screen here and there or a cable/tripod being behind the subject as those things can be easily rotoscoped, but making all the more sure the green screen is evenly lit and that shadows are reduced to a minimum
• Adding less contact lighting on set and more in post as the light I got on set was generally too harsh
• Create and import a simple proxy scene of the CG environment to the matchmoving application to avoid scaling issues
• Make sure of continuity between shots – both in the studio and in the CG environment
• Read up on colour grading

Finally, after the previs was finished and the presentation was done, I opened Photoshop and have played around with ways of making the finished shots more interesting. Underneath is one version where I’ve added more glowing things and a slight spotlight on the apple, but I might dial things back a bit if this turns out to be too much.

This Thursday I headed over to the green screen studio in the Phoenix building to finish my test shooting for the Advanced Visual Effects ICA. The person I used as an actor on the first shoot couldn’t be there, so I’ll have two actors playing the same role in my previs.

This doesn’t matter much as the previs is meant for working out camera angles, timing, technical solutions, etc. and isn’t much about the performance. This time around I used Erlend which I might ask to be in the final shoot as well.

While on the shoot, I forgot to shoot test charts for lens distortion so I’ve booked an hour in the studio this Monday to get those. I’ve noted down focal lengths on all of the shots so it shouldn’t take long. Getting test charts for the test shoot might not be that important, but I’ve never tried undistorting footage in order to get a better track and better results in compositing with CG elements before so thought I’d try.

Here are some reference images I’ve gathered.

What I think I’ll do is to create a few trees in 3D, or rather just one tree and create some variations for it. Somewhere between six and nine trees should be enough. In the foreground I’ll use the 3D trees. Further back I’ll either find images of trees or render my 3D trees to 2D images and stick them on cards in 3D space. In the far background I’ll use a matte painting of some sort, mapped to the inside of a 3D cylinder or sphere. Hopefully, that should be enough to create an illusion of depth and keep the amount of work manageable.

What worries me most at the moment is the creation of the leaves. I think I’ll make some in 3D and use them sparingly in the foreground. Further back I’m hoping the depth of field will blur them quite a lot, so I’m thinking of either creating random particle clusters of them in 3D without actually sticking them on branches or alternatively extract them from images and stick those up on planes in 3D space.

I write this on a Sunday and I’ll be leaving Teesside early Wednesday to go home to Norway for the holidays. This, therefore, is likely be my last blog post for a while, but that doesn’t mean I won’t be working on the ICA. As I’ve been spending most of my time on Practical Project 60 and Studio Practice & Business lately, I’m planning on finishing most of the AVE previs and presentation over the holidays.

After the holidays I’m planning on approaching the Teesside Drama Society to see whether I can get an actor for my shot from there. If that falls through, I’ll probably use Erlend as mentioned earlier. As soon as I have an actor, I’ll book another shoot in the green screen studio to get that out of the way as soon as possible.