Concept Art of Khumba

Its been quite sometime since I have actually sat down and updated my site. Its been an interesting few months of late. We are roughly 15 months into the production of Khumba, Triggerfish Animation Studios second fully animated feature film. Zambezia being set for release Q4 2012! Never sure of what footage or production details are allowed to mentioned online makes updates difficult. Khumba is nicely into Animation with a good few scenes already having had final animation approved, so we should be seeing renders coming out in the next few weeks.

Ric Capecchi, Storyboard Artist Extraordinaire!

The next 5 months are dedicated to animation and so my time is spent nearly entirely at my desk working on the setup and polish of stereoscopic cameras within every shot of the film. Stereo having such a large impact on cinematography however, I have been involved since storyboarding in a DOP role, right through previz up until this point. I must admit that I miss storyboarding, for an animated film its really the most flexible time during production. Exploring the enviroments for interesting shots whilst resolving story issues that you did not expect during script writing. It also has the quickest turn around, shots can be reworked and redrawn during a daily review session, making it a busy and exciting time during production!

Concept Art of the Karoo

Working on Zambezia left me with so much knowledge, and a desire to make the 3D just so much better on Khumba. Having said that the stereo on Zambezia was started just after Avatar had come out, and the trends in stereography have changed drastically in the last two years. Khumba will be up to date with all the current trends, thats for sure!

Its feels like before I know it, I will be sitting in stereo dailies again looking at the final rendered shots for stereo disparities and render issues… its exciting and daunting at the same time, seeing how much work still needs to be done. We’ve done it once before and of course we’ll do it again!

C

Related Posts:

1. Zambezia in the Media Again
2. Week with Bernard Mendiburu
3. Stereo Pairs and 3D Video Storage

I have a confession. I am a stereographer who doesn’t watch 3D movies….well Live Action ones that is. Not because the content, visual effects, or story are particularly bad, but because I am bound to walk away feeling a little despondent about the field in which I work. Stereoscopic imaging may not be new to the film world having been around since the ’60s, but unfortunately the Live Action guys have so many physical constraints that directly influence their final stereo footage. So let me preface this by saying, of course I am biased… chances are everyone reading this on AnimationSA.org will join in behind me and agree, simply because, well, we’re all in Animation!

For those new to stereoscopic cinema, at its simplest we film the same movie twice once for your left eye, and once for your right. Obviously there is a lot more to it than just that… but that’s all you’ll need to know to understand most of the issues that Live Action suffer from. In a CG environment we have a better than ‘perfect world’ no dust, minute accuracy, and best of all no physical constraints (our cameras can sit inside each other). The average humans eyes are separated by roughly 65mm, no two Cinema Cameras are small enough to have the centre of their lenses 65mm apart thus most commonly Live Action will shoot through a mirror rig (or beam splitter) where one camera films straight like normal, but through a mirror and another films down reflecting off the mirror. This is just the start of their issues, these rigs do their best to make camera alignment as accurate as possible but often small geometric distortions occur, keystoning, rotational, zoom difference between cameras can instantly create stereoscopic problems. You say what’s a pixel or two disparity (differences between left & right eye) between friends, I say, ‘Work out the mathematics of 4pixels magnified to an average 40ft (12m) screen!’ Basically its like having a something the size of a match box thrown at your face and only hitting you in the one eye! Cinema projection is all about magnification, ‘always better on the big screen’?!?

Lets dig deeper, disparities in all their incarnations I believe are the gradual causes of headaches and eye fatigue during a film. Geometric distortions are just the beginning. Transformers 3 was riddled with reflections in cars often being brighter in one eye than in the other, specular hot spots on actors faces (also reflections) being visible only to one camera and not to the other! Subtle colour variations between eyes can also become a problem along with lens artifacts like lens flares, etc are all potential stereoscopic problem areas.

I would hate to see a sunset shot over the ocean filmed in stereo, all those little reflections would be a nightmare. Technically it is these disparities between eyes that actually makes the ocean look like the twinkling diamonds you expect to see in real life. The human brain has learnt to accept these disparities as normal, but when they are captured and transferred to film our brain is less accepting of them.

The benefits to animation are obvious we get so much for free, never bothered by colour variations, geometric distortions, rotational or zoom differences, that’s well over 75% of our way to a perfect stereo pair. Dreamworks committed to a fully stereoscopic workflow years ago and their movies are real evidence of excellent use of 3D. We sadly don’t escape scott free! Gnomeo and Juliet had a few shots in the long grass where the thin blades didn’t quite match up in left and right frame. Pixel size disparities are a constant curse (ask any one of the compositing artists that work at Triggerfish Animation Studios) often a small bright spot between two leaves (on the tree of Zambezia) will only be visible in a single eye and so we send in a compositor to clean up the disparity. Like any good criminal knows, you have to clean up the scene of the crime

There is a whole other and further discussion to be had from the fact that in Animation we can preview our entire environment in 3D before going to render. We can position objects and adjust our stereo depth bracket as we see fit based on every unique shot. At the end of the day Animated films are bound to have better stereo than live action. The odds are just generally stacked against live action, and that leaves the door wide open for us lucky few placed in the prime location of 3D Animation.

Related Posts:

1. Guest Lecturer at The Animation School
2. Mechanics & Mathematics of Stereoscopy
3. Stereo Pairs and 3D Video Storage

The whole process is an oddly emotional roller coaster ride, you slowly edge yourself closer and closer to completion with minor deadlines along the way, each deadline is a mini accomplishment but from each phase to the next you are tested and tried! Sounds very much like a classic storyline or script where you guide the viewer through the many Ups and Downs of your character ultimately building to the grand finale, and such is the film making process! Only the final film, sound and all is that grand finale! Oh and the process takes maybe 2 years instead of 2 hours! Its a journey totally worthwhile though, and although frustrating, fighting against machinery and software and having to work some oddly strange hours the results speak for themselves.

I thoroughly enjoyed the two days (although fraught with issues) spent at the refinery in joburg whilst we made our final Stereoscopic Grade adjustments to Zambezia. Finally seeing the film in Full HD on a large colour accurate projector was amazing and I was so excited to see how well the stereo performed when we viewed 100% of the rendered pixels. You see throughout production I have viewed all stereo Over-Under which is effectively only showing me half the usable pixels… I specifically chose Over-Under to attempt to maximize the horizontal pixel resolution which should have maximized the stereo viewing. (See my post regarding Over-Under vs Side-by-Side) Seeing it in full HD was amazing though, and I think once you finally see the movie you will know what I am talking about!

Zambezia Premiered to the American Film Market on the 5th and 6th of November, and we are expecting a 2012 Release

Related Posts:

1. Stereoscopy in Live Action vs Animation
2. The Karoo goes 3D in Khumba
3. Stereo Pairs and 3D Video Storage

The Scipt

this version saves out 32bit TGA files with their alpha channels

//set up
preferences.rulerUnits = Units.PERCENT
displayDialogs = DialogModes.NO

//Doc Info
var docRef = app.documents[0]
var typeLayer = app.activeDocument.artLayers.add();
typeLayer.name = app.activeDocument.name.slice(0,-4);

// Save as a new TGA file with these options.
var Options = new TargaSaveOptions();
Options.alphaChannels = true;
Options.resolution = TargaBitsPerPixels.THIRTYTWO;
Options.rleCompression = false;

//make HalfRes
var resize = docRef.resizeImage( 50,50 )
var saveFile = new File(decodeURI(activeDocument.fullName.fsName).slice(0,-4) + "_med.tga");
app.activeDocument.saveAs (saveFile, Options, false, Extension.LOWERCASE);

//make QuaterRes
var resize = docRef.resizeImage( 50,50 )
var saveFile = new File(decodeURI(activeDocument.fullName.fsName).slice(0,-4) + "_low.tga");
app.activeDocument.saveAs (saveFile, Options, false, Extension.LOWERCASE);

//close without saving
app.activeDocument.close(SaveOptions.DONOTSAVECHANGES);

a modified version by simply changing the save options & file extensions allows you to just about any format, we also use the PNG format so therefore

// Save as a new PNG file with these options.
var options = new PNGSaveOptions();

is switched out to save PNGs.

// Save as a new TIF file with these options.
var Options = new TiffSaveOptions();
Options.alphaChannels = true;

is switched out to save TIFs.

Download Scripts

I have put them both into a Zip File with instructions on how to install and use, and that is Downloadable Here!!

Enjoy!!

~C

Related Posts:

1. DCP Image Converter
2. XRef in 3D Max
3. 5 Tips for XSI Pass Setup & Management

This first phase automates the entire process of converting an image sequence to the cinema jpeg 2000 format. It can handle 2K stereoscopic and 2K monoscopic flat formats (not scope yet) and exports the image sequence ready for DCP Packaging. Currently the conversion from sRGB .tif takes roughly 1min 30secs for 24 stereo frames (effectively 1sec of film). When stereoscopic is enabled it converts each image stream independently considerably speeding up the conversion time. Originally these conversions took nearly 6mins per 24frames when run as a single process.

As it stands the converter takes care of a few XYZ colour space conversions, outputting a 12bit XYZ Image following the Cinema 2K specifications. I am quite interested in colour space so hope to expand this area in the future.

Roadmap

The roadmap for the next phase includes, Implementation of 4K Cinema standards under monoscopic conversion, MXF Wrapping for both Video & Audio, and the creation of the DCP Package xml files. Ultimately I hope it to be a full DCP conversion and packaging tool!

I am interested to know what other options may be needed, what colour space’s do most people work in, would the loading of a custom LUT be of value, should I investigate input formats such as DPX and EXR?

~C

Related Posts:

1. Auto Image Resizing via Photoshop Scripting

To watch the 3D version you will need to visit Youtube by clicking here -> S3D Promo Trailer make sure to check it out or download it and watch it in Stereoscopic Player!

2D Promo Trailer

Related Posts:

1. Zambezia in the Media Again
2. The Callsheet Stereoscopic Article
3. Stereo Pairs and 3D Video Storage

‘To be honest we spoke to a few customers about their stereo requirements and you are the first (that I am aware of) to mention over-under for stereo previews. Could you maybe explain the advantages of over-under for you guys and why you settled on that standard?’

Now I truly feel that the over-under method for storing stereo paired video footage makes far more sense than side-by-side… and here are my reasons.

Often with stereography you require HIT (Horizontal Image Translation) in post, effectively sliding your images left or right to move the depth bracket back and forth in Z Space. Now with images stored in a side-by-side format at 1920×1080, each eye is halved (960×1080) on the horizontal resolution this to start with is a bad as our eyes are separated horizontally so any loss in pixel resolution on the horizontal has the most direct effect to viewing quality when re sampling these images back to 1920×1080.

Secondly and referring to the HIT if you slide your left image right & right image left your pixels are crushed down the center and lost because the stereo re-sampling method always cuts at the halfway mark of your image resolution. Over-Under avoids both of these issues by maintaining full Horizontal Pixel Resolution at 1920×540, and also avoids destructive behavior to the resolution of the image because the HIT happens down the Horizontal not the Vertical. (sounding very outer limits there, lol)

The introduction of black pixels down the edges of the frame occur in both side-by-side and over-under formats, however they are overcome by the over rendering of frames. We render a buffer of 20px on either side allowing for an overall HIT of 40px left & right. The most important fact to note here is the destructive tendencies of side-by-side footage when images are HIT inwards, and the fact that most editing software assumes that anything beyond one side of the halfway mark on the horizontal is the opposite eye therefore even with the over rendering of pixels, black bars centrally are unavoidable.

Thirdly, many of the new passive monitoring systems such as the Zalman Trimon 3D Display and the JVC GD-463D10 use a horizontal interleaving polarization method. This effectively makes every alternate row of pixels polarized for a specific eye. When converting a side-by-side stereo pair to be viewed on a interleaved monitor you loose a quarter of your total video resolution. The side-by-side video is halved on the horizontal (1 half lost) and then when the monitors interlace the footage each alternate row of pixels is lost due to it being used to show the opposite eye (1 half lost) this effectively means you only see a quarter of your total resolution.

The Over-Under format stores the images with full horizontal resolution and has the pre-compensated for the loss of pixel resolution that occurs during the interleaving process. Therefore there is no loss of pixels. We use Stereoscopic Player to preview the captures in stereo weather it be on our Polarised Horizontally Interleaved Monitors or Active Projector (Quad Buffered) as it deals with converting the input formats to the required output methods. The same applies in our edit suit where we monitor stereoscopic on a Panasonic TH-P50VT20P

Just on a final note, the HDMI 1.4 stereoscopic standard is two full resolution frames Over-Under…

~C

Related Posts:

1. Making the Grade
2. Mechanics & Mathematics of Stereoscopy
3. Zambezia in the Media Again

A recent article in Screen Africa Magazine regarding Zambezia and Stereoscopic

Recently animationmagazine.net also wrote a great article on Zambezia and Triggerfish Animation

“We are opening the window into a world not many children have seen – the mighty Zambezi River Valley, the awesome Victoria Falls and the varied and rich cultures (and birdlife) of Africa,” says director Wayne Thornley.

We also received some exciting news regarding Khumba, Triggerfish Animation Studios second animated feature film. The Industrial Development Corporation (IDC) approved the much needed funding. This means that as from 1st October we are officially in production and it’s all systems go!

Related Posts:

1. The Callsheet Stereoscopic Article
2. Zambezia Movie Promo/Trailer 3D
3. The Karoo goes 3D in Khumba

Simple intro to the Mechanics & Mathematics of Stereoscopy

Stereoscopy at its simplest form is straight forward, we create two image sequences one catering specifically to your Left Eye & one catering specifically to your Right Eye. Those images are created in such a way that they attempt to recreate the best possible S3D effect for all who watch the film. Basically our work mostly concerns the z-axis (objects in front or behind the screen), behind screen is known as the Positive Parallax and in front of screen is known as the Negative Parallax.

But why does the area coming towards you get called negative and the area moving away from you get called positive? Well that’s where the above scene geometry comes into play. We all work on screens, weather its your computer screen or the cinema screen, and that screen we call the Zero Parallax, its the one point where what both your Left & Right eyes are seeing is perfectly matched. Now if we attached red lasers to each of your eye balls (very terminator style) and asked you to look behind your screen and then measured the distance between the two red dots (at screen position) we would find a positive number, but if we did the same with you looking at an object in front of the screen the Left dot would cross over the Right dot and thus, remaining in the same mathematical space, the new measured distance turns up negative.

Now speaking of the positive parallax, where your eyes are looking parallel to one another, and you are looking at a distant object. Although comfortable and natural, most of the time your eyes spend their days converging on objects, but if you attempt to diverge (make your eyes move apart, as seen above) the eyes to much, this is what I consider the primary physical danger zone of S3D, no one likes having their eyes forced apart.

This is where the mathematics come into play, we need to know the maximum positive parallax the human eyes can handle at the screen size of your final deployment screen. Its a simple ratio, if the human eyes average distance apart is 65mm (Interocular) then you need to know what percentage 65mm would be on your specific screen. Lets take two examples, firstly a 500mm computer monitor, and secondly a 9meter cinema screen.

Percentage of Screen = (Interocular / Screen Width)

0.13% = (65mm / 500mm) or 13%
0.0072% = (65mm / 9000mm) or 0.72%

Now if your Screen Resolution is say HD (1920×1080) you can work out the how many pixels would make your eyes sit in parallel in the resolution you are working.

Pixel Parallax = Resolution x Percentage of Screen

249px = 1920 x 0.13
14px = 1920 x 0.0072

So effectively what this says is that at a screen size of 9m wide, the maximum pixel separation you should go to is +14px when showing HD Footage. The next area of concern is distance towards the viewer in the negative parallax, but in my opinion this area is a little more flexible. Obviously extremes are just as dangerous as extremes in the positive, but the pixel values in the negative can be quite a bit larger without putting to make strain on the viewer. I plan to write a post specifically about the negative parallax because it alone is a large topic of discussion in its own right.

This being the simple intro I will stop right here, but it does get quite a bit more complicated when you begin to take into account the physical location of the viewer relative to the screen on the z-axis, etc

C

Related Posts:

1. Stereoscopy in Live Action vs Animation
2. Stereo Pairs and 3D Video Storage
3. Zambezia in the Media Again

Click for Larger Version

Its been a while since my last proper Gaming, Animation or 3D related post… its been tough working under pretty strict NDAs. I have however been very busy developing, growing and writing tools for our in-house pipelines and furthering the stereoscopic pipeline.

C

Related Posts:

1. Zambezia in the Media Again
2. Stereo Pairs and 3D Video Storage
3. Zambezia Movie Promo/Trailer 3D