ACU Maya common mistakes
ACU Maya · Understand the traps and how to avoid them.
When asked for the object's world-space position, the candidate reads the Translate X/Y/Z channel box values and treats them as world coordinates, ignoring that the object sits under a parent with non-zero transforms.
The channel box shows local translation relative to the parent, and candidates who learned the interface visually assume the numbers in the box are absolute. The distinction between local and world space is only visible in the attribute editor's world-space readouts or via the move tool's axis indicator.
What to do instead
Before answering, trace the parent chain in the Outliner and mentally compose the parent transforms, or switch the tool settings to world space and read the pivot's world position. Memorise that Translate channels are local and that world position requires accumulated parent transforms.
The candidate selects objects by clicking in the viewport and then reports the selection count or hierarchy, unaware that a parent was selected as a group or that a hidden child was included.
Marquee and click selection behave differently depending on select-by-component mode and whether 'select hierarchy' is active; candidates rarely check the selection mode indicator or the Outliner highlight before answering.
What to do instead
Confirm selection mode (object vs component) in the status line, verify the highlighted Outliner entries, and use the Outliner itself to count children rather than trusting a viewport marquee. State the mode you assumed when answering.
When asked to add edge loops to hold shape on a subdivided model, the candidate uses the Multi-Cut or Insert Edge Loop tool and places loops at arbitrary positions, then reports the resulting topology without checking that the loops are evenly spaced and parallel to existing edge flow.
Candidates know the tools exist but not the principle that support loops must sit close to the crease to control subdivision sharpness; they place loops wherever the cursor lands and assume any loop qualifies.
What to do instead
Snap loops to edge midpoints or use the tool's snap-to-percentage option, then verify spacing by orbiting and checking the smoothed preview. Explain that loop proximity to an edge controls the tightness of the resulting crease.
The candidate answers a question about polygon count or face count by counting what is visible in the viewport, ignoring faces hidden behind the model or inside a combined mesh.
Viewport display shows only front-facing geometry, and candidates forget that a closed mesh has back faces and interior geometry that still count toward the total. Display settings such as backface culling can also hide faces.
What to do instead
Read the polygon count from the Heads Up Display or the attribute editor's mesh statistics, which report the true total regardless of visibility. State that the count includes all faces, not just visible ones.
The candidate describes a material as 'just a Lambert' or 'just a Blinn' without distinguishing which shading attributes actually produce the described look, and attributes specular highlights to Lambert.
The material type names are familiar from the create menu, but candidates memorise them as labels rather than as models with different attribute sets. Lambert has no specular channel, so any highlight description must come from Blinn, Phong, or a shader network.
What to do instead
Map each material type to its channels: Lambert diffuse only, Phong and Blinn with specular, and note that Blinn's specular is controlled by eccentricity while Phong's is cosine power. Match the described visual feature to the channel that produces it.
When asked how a texture is placed on a model, the candidate says the file node is connected to the material and stops, omitting the place2dTexture node and its UV repeat, offset, and rotate values.
Candidates see the file node in the attribute editor and assume the connection is direct, not realising that Maya inserts a 2D placement node between the texture and the UV space, and that the placement node's values determine tiling.
What to do instead
Describe the full chain: file node to place2dTexture to the material's color channel, and cite repeat UV, offset, and rotate on the placement node as the controls for tiling and orientation. Check the node in the hypershade's input connections.
The candidate builds a skeleton and then reports that the skin deforms correctly, without verifying that joint orientation and bind pose were set before binding, so the mesh collapses or twists when posed.
Binding is a single menu action, and candidates assume it captures the current state correctly. If joints were rotated or scaled before binding, the bind pose is wrong and the deformation inherits that error.
What to do instead
Orient joints with the joint orient tool before binding, bind in the rest pose, and test by rotating a joint a few degrees to confirm smooth deformation. If the bind is wrong, unbind, reset the joint transforms, and rebind.
When asked to explain an animation curve, the candidate describes keyframes only and ignores the tangent handles that determine interpolation between keys, so a described motion that eases in and out is attributed to extra keys.
The graph editor's curve shape is produced by tangents, but candidates focus on the key dots they can see in the timeline and treat interpolation as automatic rather than authored.
What to do instead
Open the graph editor, select the keys, and inspect the tangent type (auto, spline, linear, stepped). Explain that ease-in and ease-out come from flat tangents at the keys, and that stepped tangents hold a value until the next key.
The candidate sets a light's intensity and reports the scene brightness without accounting for the renderer's exposure, the light's decay rate, or whether the light uses a physical falloff model.
Intensity alone does not determine brightness when decay and exposure are in play; candidates treat intensity as an absolute brightness value and ignore the camera's exposure controls and the light's decay setting.
What to do instead
State the decay rate and the exposure value alongside intensity, and note that physical lights fall off with distance squared by default. When comparing two setups, hold exposure and decay constant before attributing a brightness difference to intensity.
The candidate describes a render as 'using Arnold' or 'using the default renderer' without knowing which renderer is actually active, and answers questions about sampling or ray depth as if the settings were universal.
The render settings window shows the current renderer, but candidates rarely check it and assume the default is whatever they last used. Sampling and ray depth controls differ between renderers and are not interchangeable.
What to do instead
Check the renderer dropdown in the render settings before answering, and name it explicitly. When discussing sampling, specify the renderer's own sampling attributes (for example Arnold's camera AA samples and ray depth) rather than generic terms.