With users barely familiar with the new features in Max 5, Discreet has released yet another upgrade to its popular 3D animation system. There are a few important new features, lots of minor ones, and an overall refinement of the interface and workflow. Interface-wise 3DS Max 6 looks much the same, except a new Reactor tool bar that’s docked to the left edge by default. A simplified Layers system and toolbar has been introduced and a Render Presets menu in the Extras toolbar makes choosing a rendering setup quicker and easier than before.
The Layer manager is a major improvement over the old. It’s now a non-modal dialog so you can have it open as you work, though like other Max panels it can’t be docked to the interface. Any kind of modal panel is a serious bind in a 3D workflow, so it’s a very good move by Discreet.
Max 6 improves and refines existing modifiers and includes the new Shell modifier, which lets you extrude selected polygons to give objects thickness. In the Shell modifier, there’s an option to define the edge contour using a spline, which offers many creative bevelling possibilities. EditablePoly and MeshSmooth both have a new isoline option that reduces the density of the mesh display. This provides a simpler depiction of the contours of the object and its topology without the need to see every single polygon edge.
The Schematic view has undergone a bit of a makeover too. It’s becoming an increasingly important part of scene management, and now boasts a grid background and snapping, plus you can load background images. This is great for character rigs, where you can display an image of your character in the schematic view with nodes positioned on respective parts of the image. Not exactly as elegant as Softimage|XSI’s synoptic view, but it lets you have a greatly zoomed-out view of the hierarchy while still being able to see which nodes are which. Node text is easier to read though, so zoom levels are less restrictive in this respect.
Layouts can be saved and restored. This leads to the option of creating multiple layouts for specific parts of a character’s rig. For example, you can have an overall view of the model with details like hands enfolded for clarity, then separate layouts saved with their own background images for the face and hands. The ability to shrink unneeded nodes to small dots, leaving them accessible but not cluttering the view, is a useful feature.
It would be great if you could add hyperlinks that can switch to a different layout (a la Softimage|XSI). You could then have had a link on the hand to open its layout, and a corresponding one to link back to the main layout. It’s small but important details like this that Max tends to miss out on.
For some users, the greatest new feature is the inclusion of the mental ray renderer. It used to be a costly add-on but Discreet is now bundling it with Max 6 as a way to combat the likes of Maya and XSI, which have mental ray as standard. Mental ray 2.2 adds a huge boost to rendering in Max, though it’s not for
the faint of heart.
Mental ray is a serious production tool and, as such, it has a bewildering number of controls for setting render quality and optimization. To accommodate this Discreet has redesigned the Render Panel, which now features tabs to hold all those extra features. The old design was visually cluttered and confused, and this one is not much better. Discreet has adopted the awful Windows tab behaviour, where tabs change position depending on whether they were on the top or bottom row when selected. It’s bad enough that you have new buttons and new features to learn, but with tabs conspiring to change position when you click on them makes it even more difficult.
Despite this, mental ray offers impressive features such as advanced anti-aliasing control, depth-of-field, 3D motion blur and sub-pixel displacement maps for hyper-detailed surface bumps. Mental ray adds its acclaimed global illumination, Final Gathering and caustics features to Max. It adds two new light types, MR Area Omni and MR Area Spot. Mental ray integration is tight, and marginally better than Maya’s, but both pale in comparison with XSI.
That said, scenes rendered with mental ray should look exactly the same as the default scanline renderer if you change the assigned renderer in the Rendering palette. 3DS Max lights and (most) materials are translated and exported to mental ray and the renderer displays its progress as normal in a render window.
Mental ray shaders
A lot of mental ray’s power comes from its shaders, and you can access these through special mental ray materials, or Phenomena. These are coloured yellow in the Material Browser making it easy to locate and select them. Most Max materials and maps are compatible – only a few are not, and are coloured grey to let you know. Other shaders include those for camera lenses and volumetrics.
Mental ray is fast, and its bucket rendering means it’s very efficient when rendering across a network. Because a frame is divided into buckets (square patches of pixels) multiple render servers can work together on a single frame, easing bottlenecks for render-intensive frames and reducing the possibility of servers standing idle.
Reactor, the third-party dynamics plug-in bundled with 3DS Max, has been improved with the addition of Rag Doll dynamics. This constraint provides incredible dynamics simulation for digital stunts – animating the behaviour of passive characters as they fall and collide with scene objects. It’s an impressive system, and very fast, allowing you to view the action in real-time. Vehicle dynamics is improved with the new Car-wheel constraint giving you all the dynamics of a typical car suspension.
Vextex painting is important for games production and with the new Vertex Paint modifier in 3DS Max 6, it’s a lot more powerful. This now behaves as a floating panel, and has a greatly improved interface featuring layers and blending modes. There’s full tablet support for opacity and size, and brushes can be customized in the Painter Options dialog to suit.
3DS Max 6 gains a host of features specific to architecture, particularly for compatibility with scenes created in Autodesk Viz. A host of new procedural architectural models are available which will be useful for all kinds of 3D work involving buildings, not just in the architecture market. The AEC objects include doors and windows, banisters and railing, stairs, walls – even foliage. When rendering, a camera correction modifier provides vertical perspective correction to images so that vertical edges are always parallel. The new Architectural material grants you access to a cornucopia of material combinations via a greatly simplified interface.
3DS Max 6, its fair to say, has an improved workflow. But, and this is a but, Max still suffers from a lack of conceptual clarity in its operation. In most 3D programs there’s a very simple metaphor at work – that of the object and the tool. You have an object and you work on it with a tool. Tools are not objects, and objects are not tools. It’s all clear and simple and conceptually well-defined.
With 3DS Max, you have geometry and tools, known as modifiers, both of which are presented to you by the interface as objects. For example, in order to do some polygon editing you need to add a poly-editing modifier to your model. The tools are part of the modifier (which
is essentially an object, or data node) and not independent from it. With some modifiers this makes sense, with others it doesn’t. It adds quite an overhead to workflow, since you’ll often have to wade through the modifier stack for an object every time you need to access a tool.
Objects can contain modifiers other than those for modelling, such as animation modifiers, radiosity subdivision modifiers, and texture modifiers. These are all jumbled in the same part of the interface, and frankly it’s a mess. Eons ago programs such as Softimage|3D divided up the interface by task so that tools relating
to animation, rigging, modelling and texturing were in specific locations. XSI and Maya still do the same and it makes sense for workflow. Even low-end apps such as Bryce and Carrara have the concept of rooms, and LightWave even goes as far as having two totally independent applications for modelling and animation. Why Discreet does not address this confusion of tools and features is a mystery.
From a professional’s perspective, 3DS Max 6 is extremely capable, and there are indeed segments of the program that work extremely well. Once you’re within a modifier, for example, modelling can be very good. Dynamics are getting better all the time, and the rendering is world-class. It’s when you look at the overall picture and how the program fits its features together that things are less compelling. Next to its competitors, its workflow is not quite as slick it should be. Basic application design seems opaque and inconsistent, and ergonomics seem like an afterthought.
Scene management and object selection are an obstacle. Though the Schematic view helps, a simple non-modal object list would solve so many problems much more efficiently. It’s the small details, the glue that hold an interface and workflow together, that Max seems deficient in, and Discreet really need to spend some time thinking more about this in forthcoming versions. When you spend most of your working day sat in front of a 3D program week in week out, the frustration that poor design and ergonomics can induce is not trivial.
However, if you’re used to Max’s idiosyncratic workflow, you’ll be pleased by many of the new features in this latest release, and it’s a compelling 3D system
if you take its features on face value, and are less critical about their implementation.