hytham sameer's Blogs

hytham sameer
The original Magic Bullet Colorista brought the 3-way color corrector to After Effects and improved upon the one built into Final Cut Pro. With the recent release of Colorista II and the addition of secondary color correction, more flexible masking, and UI refinements the plugin has solidified its place as our go-to favorite for grading within those desktop applications.
Working in commercial post, I used flame as my primary tool with a seat of After Effects at its side to do tasks I felt were easier to accomplish in the desktop app. On almost every shot I finished, easy access to 3-way color correction in the Flame Colour Corrector was critical for getting the job done. Autodesk's Colour Warper took this a step further with familiar track-ball like grading controls and built-in secondary color correction. It's this secondary color correction that really helps takes grading to the next level. It's the type of control that is familiar to professional colorists and is quite integral to the grading process.

With the premiere of fxguidetv several years ago, as well as the creation of our online training site fxphd.com, we've been finishing a ton of material for web distribution in Final Cut Pro and After Effects. For grading, plugins have been incredibly useful extensions to the built in functionality and we've utilized Magic Bullet's Colorista and Looks as well as The Grading Sweet.

The original Colorista was a single three-way color corrector with built-in simple masking features which helped make grading in After Effects and Final Cut much easier. Enter Colorista II. This release builds upon the original offering in a big way by bringing secondary color correction to the plug-in as well as additional sets of 3-way adjustments. In the new plug-in there are now effectively three layers of color correction: Primary, Secondary, and Master. Each of these layers contains a 3-way color corrector.


The benefit to having effectively three old-school Colorista adjustments within a single effect is that you can set up a look and then adjust each individual scene to balance scene to scene. This type of workflow can be especially useful when grading a spot or section of a film, where you have a look you wish to end up with but the individual shots within the sequence vary in brightness, saturation, or in other ways.

Stu Maschwitz, Red Giant Software Creative Director, recommends the following approach when grading:

* Primary: Balance and compensate for differences between individual shots
* Master: Set up your overall look for a scene
* Secondary: Adjust a specific color, range of colors, or region in the scene
The benefits of Secondary

It's this last area, Secondary, where Colorista II has seen major additions. For years, profesional colorists using high end machines have relied on secondary color corrections to bring scenes to life. At the most basic level, this is a second layer of color correction applied to a scene, either using masking or isolating a certain color or range of colors and then adjusting only this area. Whether or not the colorist knew it, they were effectively doing a key to isolate a range of colors and then adjusting this isolated part of the scene accordingly.

In the past, if you wanted to do some type of secondary color correction using Colorista you would have to add footage as another layer, create a key using a different plugin, and then grade the layer appropriately to get the color you want. Red Giant has developed a new keyer as part of Colorista II that makes this selection process quite straightforward. Clicking on the Key Edit button takes you into a floating window that shows the result, source, matte, and interactive UI elements for creating the key.

The first step is to select a base key color in the image. You can then easily modify the key by pressing the < > or <-> buttons and then clicking on pixels in the image to either add or subtract to the key. Working directly on the image is a great way to quickly get a decent key in progress.

There is also a graphical cube, vector scope, and luminance/saturation indicators which allow one to gesturally drag to refine the key. A major reason I've found to use these UI elements is for adding softness to the key. I should say what I like to call softness, not what Colorista calls "Softness" (see later). By clicking and moving the dashed lines, you can expand the transparent areas (varying levels of gray in the matte) to the key, helping blend from one color to another and get smoother edges.

There is also a key lift and gain adjustment (a levels adjustment of sorts for the key) as well as what Colorista calls "Softness". Personally, I feel this Softness should more accurately be called "Blur", since it applies a blur to the mask and is not a softness based upon key color range and tolerance. This adjustment happens post-lift/gain.

All in all, this keyer is quite effective when working on grading. It's not the be-all or end all keyers of keyers, but it does make for fast and efficient grading tweaks. It's a tool built for its job and to that end it performs admirably. When doing most grading, secondary selections can be quite forgiving when adjusting color correction. Consider the case of a red box in a scene. If you wish to have the red box a bit more saturated, the key can be forgiving. However, if you're wishing to swing the color from red to blue, you'll probably need more robust keying and masking tools that are available in After Effects.

A well-thought out and easy to use workflow is something that Red Giant tried to pay attention to in the development of Colorista II. "One thing that I've come to realize is that color correction is repetitive work. Any little snag that slows you down gets multiplied over hundreds of shots and becomes a big bummer," says Maschwitz. "So when we designed the keyer for Colorista II, we set out to make it so easy that you'd never think twice about using it. Many keying tasks can be performed in a handful of mouse clicks. That's been the design philosophy of Colorista II from day one: It's one thing to have the color correction power, but its another thing entirely for it to be so easy to use that you don't have time *not* to use it."

The secondary section also has masking controls that Colorista users will be familiar with. This provides an adjustable rectangular or elliptical mask with size and feathering controls. The Power Mask on on the secondary layer has several modes, the default being Key And Power Mask which uses the key, but restricts it using the mask shape. There is also Key Only (similar to having a Key with no Power Mask active), Power Mask Only (similar to having a Power Mask without a Key), and Key Minus Power Mask, Power Mask Minus Key, and Power Mask Plus Key. The mask positions can also be tracked using the AE tracker.

The good news with this release is that there is also a Master Power mask which can, if desired, be combined with the Secondary one. In fact, you have the option of using the Master Power Mask to restrict the entire effect, restrict the Master effect, restrict the Secondary Effect (similar to intersecting two standard masks), Add to Secondary (adds the mask to the secondary selection), and Join Secondary (which acts as an additional mask on the secondary layer). Between the two masks and the various modes you can apply them, there are a multitude of ways in which you can mask your imagery.

For most purposes, these shapes provide a flexible way of isolating the secondary mask for grading. While obviously it would be ideal to have full roto-spline support within Colorista II, the decision to concentrate on the grading tools is a good one. Building a complete masking/roto system seems overkill, especially with the complexities involved when doing so. Simple shapes can been used effectively to make great looking images. And if you're working in After Effects, you still have the capability to add additional layers in if you wish to use the more advanced masking techniques built into the app.

Other Improvements
Within the Primary area of Colorista II, there is a new Highlight Recovery feature. What this feature does is find flat (blown-out) highlights and does iterative image processing on them to build "pyramids" on the plateaus. This processing helps ease transitions into the blown out areas, giving a bit more range when grading. This new highlight info is true HDR, so it only shows up if subsequent color grading tints or darkens the image. It also explains the features placement in the Master area of the plugin, because it is done as a first step so it is available later in the processing pipeline.

In order to more easily show how this works, we created a black and white image in Photoshop where the white is at 1.0. This can be seen in the bottom half of the image to the right. We then applied a highlight recovery value of 50 to the image using Colorista II. This actually creates the plateaus with values greater than 1.0. In order to see the effect, we then reduced Secondary Exposure to -.66 to bring back the highlight details. Once this is done, you can see the steps which ease transition to the highlights. With actual footage, this transition helps areas which are blown out. This is no magic bullet (no pun intended), and as you can see from this testing will work best in smaller areas of blown out data -- it won't magically bring back a fully blown out sky.

The Secondary area also has a feature called "Pop", which is effectively a local contrast adjustment that is similar to Lightroom's Clarity slider. "Where a standard contrast control makes bright things brighter and dark things darker, Pop works with a soft radius -- so a color near a bright area will get darker, and a color near a dark area will get brighter," says Maschwitz. "This is basically the inverse of a diffusion filter effect, so the result is that any kind of haze or diffusion can be mildly corrected away."

As far as other minor improvements, the color wheels have added functionality for adjusting hue, saturation, and brightness for each of the three way controls. In addition, numerical access is easily available by clicking on a calculator icon.

The new Master HSL controls in each section allow adjustment of individual colors. This enables you to tweak the saturation, brightness, or hue (limited) of Red, Orange, Yellow, Green, Cyan, Blue, Purple, and Magenta. This might be useful if you wanted to bring up the saturation of grass in a scene or pop the blue in a sky. Think of it as localized color color correction based upon a color's hue.

There is also a new option to utilize GPU processing for the effect, which helps tremendously with interactivity when using Colorista.

All in all, this is a very solid update of the After Effects/FCP workhorse. The software is available for $299 (upgrade for $99) at the Red Giant Software web site. I've spoken with a few artists and they've somewhat balked at the price. But in the grand scheme of things, it's well worth the money. While it's true that apps such as Apple's Color and the upcoming Resolve from BlackMagic can be had for around $1000, there is a huge benefit of integrated grading within a single application. I find this workflow much more efficient than exporting a sequence to an external application and baking in a grade at the finish. Of course, there are times when using an external app makes sense, but lot of the work can be easily handled within Final Cut or After Effects and there is a tremendous benefit for working within the same UI as opposed to switching between apps.

Current fxphd members can get 30% off the plugin and fxphd members taking the RGT201 course can purchase it at 50% off list.
hytham sameer
In our continuing and popular "Art of..." series, we tackle wire and rig removal. Far from being the most glamourous of visual effects, in reality wire removal is extremely challenging. We audit the current approaches and techniques of the craft. In this week's podcast, we speak with Aaron Rhode from The Orphanage about the complex 2D and 3D solutions for scenes with wires.

Techniques for General Wire Removal

1. Wire Removal By Painting Frame By Frame

The primary comment made about wire removal is "we'll just paint it out in post." But painting out on a per frame basis is extremely difficult when done over a series of frames, potentially causing the images to seem as if they are boiling. While a clone tool works well on a single frame, the lack of frame to frame cohesion means the 'fix' boils and becomes very visible once the clip is played. Of course there are times that wires are only catching the light for a moment,  and thus an isolated frame is still very effectively fixed by a manual painting, either by cloning or delicate painting. Cloning is much more effective, as it clones not just colour but grain or noise that is also almost always  present.


2. Rig Removal By Patching Over The Top
One solution using this method is based on finding a clean frame from some other point in the clip and  pasting it over the top of the offensive rig or wire. This works best on static cameras, and does not work on people in movement. Since the fill for the patch is static, considerations need to be made for grain and noise. Various tricks can used to solve this, including compounding or averaging several clean frames, reducing the grain/noise so it can be reintroduced at a later stage.

Another approach is to patch the wire and then use this new clip as a "reveal back to" clip in a manual painting environment, or roto the patch back in with finer attention to detail than the main patch. This has the advantage of replacing the minimum amount of the overall image but can suffer if there are other light changes that the 'patch' is not matching and tracking in terms of colour.

At its most extreme implementation, the patch approach becomes a 2D environment replacement which we discuss later.


3. Stabilise and Paint Back



One approach to solving a moving shot is to stabilize the image, fix it, and then invert the stabilization. If a wire is moving relative to the background, this approach can be very effective. Once the background is stable, the artist can paint back to a previous frame where the line or wire isn't seen. So unlike painting back to a still as in the patch version discussed previously, the artist paints back to moving clip offset in space to the foreground. This 'moving reveal' can be slightly noticeable when done at this stage but when the original camera motion is reintroduced it can seem flawless.

Problems arise with motion blur. A stablised image still has the blur contained inside that frame from any original camera movement. If this is offset in time - the foreground may have a different camera motion blur than the offset clip below at any point in time. Reveling the 'offset background' will then revel the wrong amount of blur, which can be a problem.

4. Roto Clone Tools/Source Nodes


To solve the problems in approach one (manual painting), there are faster computer-assisted approaches to cloning and painting. A common approach is to create an accurate matte of the wire or rig via rotoscoping. This matte is used to cut a hole in the iamage. To fill the hole, the same frame is slid or shifted to reaveal a part of the frame that is wire-free. In flame this is done by using source nodes, which move the matte or foreground independently of each other. There are many similar approaches available in other applications.

The reason this approach works so well is that it matches motion blur or light changes since each frame draws from the same frame for the fix. It can be thought of as a dynamically updating patch approach. The disadvantage is the issue with sharp edge transitions. Since the entire wire matte region is replaced with the same scale patch, what works on camera right may not align camera left or vice versa. This means that some shots work amazingly well, but a very similar shot with different structural solid lines at a differents angle may produce very poor results.
5. Automated Tools: Matador to Furnace

Matador

This early paint product had specialized roto, paint and wire removal. Matador was originally developed by British developer Parralax, and was acquired by Avid along with Parralax’s compositing application Illusion. Available only on the SGI platform and priced around $15,000, Matador was one of the first digital rotoscoping tools which gained a wide acceptance in the film post production pipeline. Matador started as a tool made for editing still images, so many of the tools used for motion work were not well thought out. Matador provides excellent matte creation tools including b-splines, motion tracking, and a full set of painting and cloning tools, with full 16bit/channel support. Avid stopped development of Matador in the late 90’s.

The original developers tried to spin it off into a new company called “Blue”, but that never took off. Matador had strong wire tools that allowed for median painting that would remove a wire from a frame via a brush that performed median filtering.


Commotion

Developed by Scott Squires,  at the time an Industrial Light and Magic visual effects supervisor, Commotion was used for years at ILM before Squires formed Puffin Designs and released it to the public. Commotion, then called Flipbook, was sighted often at ILM and mistakenly referred to as the “secret ILM motion version of Photoshop”. Though Commotion looked very similar to Photoshop in some respects, Commotion’s interface and tools were designed for moving images, and was the first tool on the desktop to offer realtime ram based playback. This realtime core functionality was the foundation for all of the wire removal tools added as the product developed.

Advanced wire removal tools include raster based paint, spatial and temporal cloning, clone to center wire removal tools (which painted from the outside of the brush stoke to the center of the same brush stroke), auto-paint, unlimited bezier and natural cubic b-splines, motion blur on rotosplines, and a very fast and accurate motion tracker. Commotion quickly became the de-facto roto tool in the industry, replacing Matador in most post facilities. Commotion curves could even be exported and imported into AfterEffects. Puffin Designs was acquired by Pinnacle Systems in 2000, but sadly development has stopped on the product. All the original developers left and no new work ever been done on the product.

Most paint work done in the wire removal process is used for touching up film or video footage. This includes removing wires and rigs, removing logos, dust busting, scratch removal, etc. Clearly, roto plays a key role in solutions beyond hand painting. Roto-based approaches have the huge advantage in that they are repeatable, or able to be rendered, so a shot can be reworked without requiring starting from scratch. Roto is therefore often  central to any wire removal process. In these circumstances, the roto tool must provide a procedural temporal and/or spatial cloning. Spatial cloning is a type of cloning which takes pixels from one position of the frame, and paints the source onto another position on the frame. Photoshop’s rubber stamp tool is an example of spatial cloning, but it is not easily automated. Temporal cloning allows one to paint pixels from one frame in a sequence to another frame. Commotion’s Super Clone tool is an example of temporal cloning.

A good wire/roto tool should provide both of these options so users can offset position and frame number together. Other cloning tools include wire removal tools which allow you to draw a line to zip out a wire. Typically, wire removal tools clone pixels from a specified value on either side of the line, then smear the outside pixels together to cover up the wire or scratch. More advanced wire removal tools will add advanced cloning techniques to the wire removal process. For example, Commotion looks at a specified number of pixels on either side of the line, flips those pixel values then cross dissolves to cover up the wire.

Furnace
One of the few companies to aggressively try and produce specific wire removal tools is The Foundry in London. Released initially as plugins for Flame, The Foundry now produces a large variety of solutions for a range of applications such as Shake and Nuke.

The Foundry's Furnace tools are based on producing a rotoscoped or animated spline path that can then have a variety of techniques deployed. The Foundry's spline defines the shape of the wire (straight or curved) and the width. It will then use one of four techniques to clone, average, clone to center or temporally remove the wire. In so doing it attempts to not remove film grain and instead just replace the wire. The techniques increase in computational complexity, and are generally agreed to be currently some of the best implementations anywhere, "best in class".

It is openly agreed - even by The Foundry, - that the tools are not one button press fixes that always work perfectly. Rather, the logic is that the wire removal tools can successfully remove a large amount of unwanted wire, but the last 20 percent will need manual or human intervention. As you can hear in this week's fxpodcast, nearly all professional wire removers use Furnace for this reason, and universally agree that getting the job 80% done is invaluable. However, there is a significant gap between any computer solution and an acceptable feature film final shot.


Mokey
As the task of wire removal extends to rig removal, The Foundry and companies such as Imagineer Systems also produce planar trackers. These trackers are widely used to solve wire and rig removal but are not solely aimed at these tasks. Imagineering produces a set of very impressive planar tracker solutions such as Mocha which will track a region of a shot. These regional tracking approaches allow much better "patch' style solutions since the 'patch' can now move in three dimensions and much more seamlessly blend with the original plate.

Imagineering also produces automated roto tools such as Mokey which can aid in separating an object from the background. The combination of this and rotoscoping allow the 2D and 3D environment techniques discussed next to work well. Products such as Mocha and Mokey are not wire removal tools in and of themselves but impressive and valuable tools that aid more advanced solutions which would otherwise be impractical.

7. 2D Background Replacement

Some of the most advanced solutions do not try and minimize the region of the frame which is being painted or retouched. Rather, they remove the entire background and replace it with a clean background tracked in 2D to look real. An example would be wire work done for a martial arts film. In these films extensive wire work is often required. Chinese senior VFX artist August Zhuang gave the example of a fight on wires staged in a small clearing, in front of bamboo. The complex nature of the vertical bamboo made many techniques fails and the amount of complex wires and rigs needed to be removed meant the shot was always going to be very complex. Rather than tackle the wires individually, Zhuang and his team roto-ed the actors off the background and then replaced it.

Using multiple frames from the sequence which revealed clean sections of the background, the team was able to piece together a single long panorama of the bamboo clearing. They would then track this matte painting back in to every shot, replacing the original background with a nearly identical flat 2D cyclorama without any wires. They were careful to create accurate 2D tracks and apply the correct amount of motion blur on each shot. This approach proved more effective and less time consuming than wire removing 20 individual wires one at a time.

8. 3D Environment Replacement
The logical extension of 2D set replacement is full 3D set replacement. Some productions choose to start this way and film the original plate on greenscreen, while others seek more realistic lighting by filming without greenscreen and manually rotoscoping all the key elements such as actors, and then dropping in 3D behind them. In cases where primary plate photography is not greenscreen, 2D background replacement techniques such as the ones listed above can be used and the matte painting is projected or mapped over 3D geometry to allow more complex 3d spatial camera moves. 3D environment replacement has grown in popularity in large part to the advances in 3D tracking and 3D automated camera software from companies such as Pixel Farm, 2d3, Syntheyes, Realviz and 3d Equaliser.

For Die Hard 4, the elevator shaft behind Bruce Willis was extended and a virtual 3D set added to build out from the real set. The set extension had to be three dimensional to work with the complex camera moves that the team at the Orphanage was presented with. The shot works due to the quality of the wire removal on the foreground, the accuracy of the roto, and the skill of the Orphanage's 3D team in matching the lighting and textures so accurately.

In this week's podcast we speak to Aaron Rhodes, the Roto/Paint Supervisor at the Orphanage. We discuss the approaches below in terms of the real world of feature films and the Orphanage's work on Die Hard 4 in particular.

hytham sameer
CGSociety :: Game Production Focus
22 July 2010, by Peter Rizkalla

A good role-playing video game is probably one of the hardest genres to develop. Extensive time and effort needs to be afforded to every aspect of the game. The gameplay needs to be solid, the UI needs to be intuitive, the experience system and the AI needs to be balanced. The graphics and sound must both be brilliant cos players will most likely put in a minimum of 40 to 50 hours into the game to beat it. So you can see why it’s going to have to look and sound pretty. Nothing can be overlooked in developing an RPG!

That being said, many gamers and developers alike have to tip their hats to a single development house for never overlooking anything in their RPG titles and for also putting together some of the game industry’s most renowned RPGs. BioWare is known for some of the most sophisticated RPGs anyone can ever play; games such as Jade Empire, Star Wars: Knights and the Mass Effect series are just to name a few. One of BioWare’s crowning achievements is Dragon Age: Origins; an RPG which won multiple awards for 2009 RPG of the Year and was also nominated for multiple awards at the Game Developer’s Choice Awards. With the recent announcement that BioWare is currently in development of Dragon Age 2, we take a look back at the development of Dragon Age: Origins. To do that, I got to have a little chat with Art Director Matthew Goldman on the development of the Dragon Age series so far.

Art and Design
Going back to the BioWare roots of creating a fantasy/party based RPG, the first step was establishing the feel of Dragon Age: Origins. “Sung Kim and Fran Gaulin worked with then Art Director Dean Anderson referencing various world and fantasy cultures,” says Goldman. Dragon Age: Origins features three playable races; humans, elves and dwarves. Each race has their own art style and feel to them. For example, the human race feels very 'old world' with their surroundings being rich in big timber and rough stone while the dwarf race feels more mystic with lots of angular architecture and even angular design tattoos on their faces. Goldman goes on to tell us more about the intent of Dragon Age: Origins. “We wanted this to be a ‘spiritual successor’ to Baldur’s Gate 2. However that was a high fantasy setting and the goal early on was to bring an edgier more mature feel to the universe. Whimsical was out and bloody was in.”




The Dragon Age: Origins world is filled with thousands of different races and creatures. Controlling your party of warriors through DAO’s UI is also a breeze considering how deep the game can get. “Characters and creatures are certainly memorable and there are a lot of them too. The UI turned out really well too. Both systems are incredibly dense and complex,” says Goldman. But for some reason, horses are not a main part of the Dragon Age experience. You would think that horses would be everywhere in a medieval style game. “Well we keep trying to do horses and it never seems to take off. One day we’ll figure out how to make this fun with multiple party members. We almost got it this time,” says Goldman, smiling. “Our games are about evolving worlds and compelling character interaction which are extremely complex in their own right.” True enough!

Basically, Matthew Goldman sums up the direction for Dragon Age: Origins, and any BioWare developed game for that matter, in just a few short words. “For gameplay? Emergent complexity within an elegant set of rules. For art? Silhouette and motion coupled with interesting behavior.”



Development
Typical tools of the trade apply at the BioWare studios. “Max, Photoshop and ZBrush did the heavy lifting for the authoring of individual assets,” says Goldman. “We have a propriety toolset we use to construct the scenes, manage lighting and pathfinding. We had our own animation system as well because at the time, Character Studio couldn’t do a lot of the things we wanted.” In a previously released developer diary, the BioWare team reveals some of the tricks they used to create gorgeous looking models without sacrificing resources. In one instance while modeling the Cry Demon, the modelers would create a low polygon model based on the 2D concept art. They would then take the low poly model and flesh it out into a very high poly model and sculpt out all the little details of the demon. They would then flatten out the high poly model into a 2D projection of the sculpture from all sides creating a texture map. They would then take that map and apply it to the original, low poly model and have the texture map create the geometry by bump-mapping the highlights and low lights of the texture.

Creating the environments of Dragon Age: Origins was another story. “We create libraries of tile and objects that speak to a theme,” says Goldman. “There are 220 levels in Origins which means flexibility is paramount. We also released the tools to the community for free so that they can design new adventures.” What Matthew Goldman is referring to is that BioWare released a toolset which allowed modification and customization of the game. This toolset is much like the kind of toolsets you will find in a copy of Unreal Tournament 3 where many aspects of the game’s visuals and playability can be modified. Obviously, this feature is only available for PC users.



Last March Bioware released an expansion pack for Dragon Age called Awakening. When asked what was the main difference in development this time around, Goldman replied with this. “The main difference is that the build environment is mature and a lot more stabile. That let the art team spend more time on doing art instead of worrying about pacing, framing opportunities and visual story telling.”

BioWare developed a proprietary character generation tool specifically to handle the task of creating a living society of many characters in the Dragon Age world. “The character generation tool is quite an accomplishment. We like to bandy around the number 1.52x10(52) head variations. In the end we only made 1,700 different characters.” At this point I replied with “ONLY!?” “so maybe that was a bit overkill. 1,700 is still a lot of talking characters in a game!” I’ll say! Speaking of character creation, BioWare released the Dragon Age Character Creator on October 13, 2009 which is about three weeks before the game’s launch date. This tool allowed players to create their own character before the game was officially released and then import that character into their game once it was released. Much like the mod toolset referred to earlier, the character creator proved to be a very robust piece of character creation software. In fact, many players loved creating multiple characters and uploading pictures of their creations online, which was kind of what happened with the SPORE creature creator, just on a much smaller scale.



A lot of the environments in Dragon Age: Origins are epic; in fact, the artists intended that many instances when the player walks into a significant environment for the first time that they are flooded with color and size. All of these grand entrances require appropriate lighting which brings us to the rendering. “I’m an art guy and my engineering background is limited to one especially unimpressive year of university,” You're not the only one Matt! He goes on to say, “But I’m told that the engine is forward rendering. In our case we are generally managing an enormous set of assets in each scene so this style of engine makes it a lot easier to parse memory usage.” In the development of the lighting of the game, BioWare did all the baking and rendering in Illuminate Labs’ Maya plug-in, Turtle.
A big problem with developing Dragon Age is that the game engine they started with was not the same game engine that they finished with. “Swapping out the engine mid-project was pretty hard going. It made the game look and feel better but imagine, if you will, doing that with an airplane and you can imagine the response; a tremendous amount of screaming. We responded by putting our noses to the grindstone. BioWare ships awesome games because the people band together in the face of adversity and git ‘er done.” I love the Larry ‘The Cable Guy’ Reference. “I suspect this behavior may be a product of our grueling winters.” Possibly. BioWare is based in Canada. Early development of Dragon Age was done in the same game engine as the Neverwinter Nights series, the Aurora engine. Later on, the engine was switched out to the newly developed Eclipse engine.



The Grand Scheme Of Things
I asked if any other studios helped in putting together the ‘Dragon Age pieces’, so to speak. Goldman’s response was “Liquid was a big help to us. They operate out of Portland Oregon.” Keep in mind, BioWare’s standards are very high; any studio that attempts to work with them needs to be prepared to meet that same standard. In fact, Goldman left me with a little comment on the BioWare standards. “If you stand still you can’t grow your fan base so we always striving to create unique and interesting experiences for our fans,” he says.

Now with Dragon Age 2 on the horizon, gamers can very well expect another deep RPG experience if BioWare intends on staying true to their tradition. A while back a BioWare rep commented on how Dragon Age: Origins runs by saying that it runs 'very well' on Windows 7. Sure enough, it does! Let’s hope the same thing can be said for the sequel.

Dragon Age: Origins is available on PlayStation 3, Xbox 360 and PC as well as the Dragon Age: Origins expansion pack, Awakening. The latest downloadable content to be released for Dragon Age: Origins is the Leliana's Song pack.
hytham sameer

The SIGGRAPH panels on game development started off with Stylized Rendering in Video Games.

Three of my absolute favorite titles were used as examples in this particular panel and for very good reason; the 2008 Prince of Persia title, Mirror’s Edge and Borderlands.

Prince of Persia
Ubisoft Lead Graphics Programmer, Jean-Francois St-Amour, breaks down Prince of Persia. This specific Prince of Persia title was lauded for it’s gorgeous palette and beautiful cel-shaded presentation which is a dynamic twist from previous PoP titles. Ubisoft devs intended on creating a very stylized world with inspiration taken not from harsher elements such as “lightning” but from softer elements such as “calligraphy”. For example, the secondary character Elika’s magic attacks emit self lit calligraphy-esque trails. As for the environments, "The world had to feel integrated into nature and not built on top of it." Says Jean-Francois. Monsters on the other hand had needed to look like they were made from the black goo called “Corruption”. Funny story; the idea for Corruption actually came from an episode of Big Bang Theory where a gooey mixture of cornstarch and water seemed to have a mind of it’s own as it was bounced around from the bass of a subwoofer.

Original concept ideas for Elika actually had her as a monkey that sat on the Prince's shoulder. When the first pitch trailer was produced, Elika was then settled on being a magical female as we all know. However, the original art style had Elika looking much different with long white hair and fair skin. The Prince also looked quite different. The first pitch was turned down by Ubisoft heads critiquing that the characters and environments looked too anime. "Miyazaki-ish" was their exact description. To me it looks a lot like something from the creator of Ico.

A year later many changes were made. Black outlines around characters and objects were developed by rendering the models twice while backfacing the models. Also bias helped eliminate most of the uglier edges. For lighting, each character received specific sunlight and ambient light which is probably why there are so few human characters in the game. Finally a very aggressive diffuse ramp was also added to get the style that we see in the final render.

Shadows needed to be sharp so the devs created a white outline around the shadows. To do this, they took a composite image and added an edge filter result which created an outline, not only the shadows, but around everything. A little tweeking of the edge filter and the shadows came out great
hytham sameer
Matthew Baldwin moved to New York City after getting a liberal arts degree, to get an MFA in painting at Pratt Institute. All he could see was huge paintings and even bigger school debts. This was a big turning point for him, graduating 20 years earlier into an economic recession and staring down the barrel of poverty and no practical skills.

"Pants-wetting fear gave me the incentive for bluffing my way into Pratt’s nascent computer graphics lab and learning digital apps, hoping to build some marketable job skills," explains Baldwin. "Once I discovered the digital paint programs I was hooked. Here were these seductive tools that allowed me to create art, luminous and wonderfully flexible."
The conundrum for Baldwin was that it was stuck in a monitor. He's been trying to figure that out for the last two decades. One answer is create things for moving graphics and animation, and let film and video be your medium. "But I’m a still image guy at heart: rather than tell narratives over a length of minutes, I love filling a single frame up with a ponderous sense of time," he says.

"That was a year and a half ago and we’d just moved to Austin, Texas. I was back in the middle of the country, because I grew up in the woods, hills and farmland of rural Minnesota. 20 years of dense NYC urbanism really makes the contrasts pop. I was thinking about living in horizontal landscapes again, the frontier prairie, layers of cultural history here and where they’ve gone, the spread of suburban development."
At the same time Baldwin was playing around with 3D objects. Someone online had done an instance test with a cow mesh, sticking replicated cows all over a bigger cow. It was just a five minute experiment, but was kinda cool. It reminded him of the cosmological myth of the giant turtle with the world on it’s back.


Putting both these inspirations together, it felt natural for Matthew Baldwin to construct this bison-as-world-to-native-Americans thing. "Someone on the forums had mentioned the final piece was an eulogy," explains Baldwin. "I like that notion, the feeling of something passing away. It’s not accidental that the bison/cart is slipping off the backdrop, pointed east instead of west."
Matthew Baldwin was also paying homage to all the Victorian photo portraits of Native Americans. But by working within this photographic style, He found a way of finishing 3D renders that he really liked. "Less pristine digital and more fuzzy analog."

BISON
The technique was running start to finish without getting distracted, sort of skimming the surface of modo! "Seriously, I must have a thing for glowing lights and technology–I’ve been perpetually seduced by 3D modeling for the last ten years. I’ve tinkered and puttered and fiddled with thousands of dailog buttons and tools. Like Rip Van Winkle, I woke up and looked back dismayed by the lack of finished pieces. I made a promise to myself: more art, less winkle."

The trick with this piece was to simplify. Baldwin kept the geometry low and rough so he could play with it quickly in modo. Move cameras and lights around quickly. He also cut the chroma down so he didn't have to assign appropriate colors to everything. This also had the happy side-effect of amplifying the emotion of the piece. He could ditch render-intensive materials and camera effects so high-res renders could be rendered quickly. "I could do a lot with a final color render, and an ambient occlusion render, even a surface ID render to grab masks from, describes Baldwin. "Then throw it all into Photoshop and work out a comp of the final that felt balanced and had the emotion I wanted. Once I had that target to hold on to, I went back to modo and added the modeling detail. But it was great having locked down the camera position and lighting early."

Plausible surrealism
For Matthew Baldwin, 3D is about how much you want the viewer to suspend their disbelief. "I love how utterly believable you can make objects in 3D," he says. "I like the realness but also want to tell stories, so I’m aiming somewhere between product render and fish-riding-bicycle. Too close to the fantastical and it starts becoming illustration. On the other hand, unless you painstakingly sweat every surface parameter on a straight product render, it winds up looking just wrong. Ultimately, I hope it makes a viewer buy into the believability, the realness of the piece."

The Long Voyage
As a Minnesota youth, Baldwin had a picture of a 50’s cruise ship hanging on his wall. This image of the Santa Rosa from Grace Line would beckon the viewer to escape, like all good travel art does.

Matthew also remembers stumbling into an antique store in his home town when he was 14. Duluth, Minnesota is a harbor city on the Great Lakes. "Some old sailor had welded a six-foot steel replica of a freighter in his spare time," he recalls. "It was crude and sloppy, detailed and awesome all at the same time. He was near-fanatical modeling every hatch, cabin, ballast tank, light and rudder. I understand the impulse, and thought of the dude as I was assembling my own sea-faring vessel. Hopefully you get a sense of longing looking at the piece, but also get that itchy adolescent vibe of building a cool Revell model and then setting it on fire with an M-80."

Long Voyage process
The process of recreating the Long Voyage, is similar to that of the Bison piece, where finalizing lighting and composition FIRST (using rough geometry) is crucial. Baldwin's work starts in modo, but much of the compositing is done in Photoshop using multi-render pass techniques.

"Since I’m purposely simplifying color schemes down to something more monochromatic, the shapes need to read very quickly. I’ll add a layer in the shader tree for a diffuse amount trick using incidence angle. I’ll apply this globally to everything in order to get lights and darks to show up across the smokestacks or define where the corners of ships are. It reminds me of the chrome Mario in the N64 days. I’m trying to deploy it a little more subtly, of course!"

"Since I’m purposely simplifying color schemes down to something more monochromatic, the shapes need to read very quickly. I’ll add a layer in the shader tree for a diffuse amount trick using incidence angle. I’ll apply this globally to everything in order to get lights and darks to show up across the smokestacks or define where the corners of ships are. It reminds me of the chrome Mario in the N64 days. I’m trying to deploy it a little more subtly, of course!"


There was a time effect Baldwin was hoping for, "to have the cruise ship chase the setting sun westward, and find a way to have both day and night depicted in the piece." Rather than totally solve this with lighting in modo, he created a ‘day’ version of the ship and a ‘night’ one, then composited them together in Photoshop. Flicking lights on was achieved simply by grabbing the surface ID for the windows and using a mask to illuminate the lights towards the stern.

There’s been an effort to create atmosphere in all the work from this series. One drawback of a straight 3D render is that it lacks any hints of the alchemy going on in traditional photography, or the meaty materiality of traditional painting. Lens blooms and artifacts, exposure, grain, ghosting, all remind a viewer of the liquid quality of light. "One way I like to add a little viscera into the final look is to dupe a layer of the final image in Photoshop, blur it, and set the above image to ‘Overlay’. I usually then mask off the highlights and darks to control the effect. The result is an image whose lights and darks pool. It’s also a great way to emphasize broader tonal areas, downplaying the chatter of finer detail."

Displaying 1 to 5 of 33