- Which workflow change most directly addresses the mismatch?
Answer: Export scene-linear floating-point passes, combine them linearly, and apply the view transform afterward
Whenever you see a question about render passes and compositing in Blender, the central concept to keep in mind is color pipeline order: display transforms (like ACES or Filmic) are non-linear operations, and non-linear math breaks additive compositing.
Here's why that matters. A beauty render is constructed by adding scene-linear light contributions together — diffuse, glossy, transmission, emission — and then applying the view transform for display. If you apply the display transform to each pass individually before exporting, you've baked a non-linear curve into each image. Adding curved values together doesn't equal the curve applied to their sum: f(a)+f(b)=f(a+b) when f is non-linear. That's exactly the mismatch the artist is experiencing. Option C solves this directly — export passes as scene-linear floating-point files (EXR), add them linearly in the external editor, and apply the view transform once at the very end, matching how Blender internally produces the beauty render.
Option A compounds the error by applying the display transform again after adding already-transformed passes, creating double-tonemapping artifacts. Option B discarding color information and restoring it from Combined defeats the purpose of working with separate passes entirely, and luminance addition doesn't reconstruct color accurately. Option D's approach of multiplying by alpha and averaging is a compositing technique for premultiplied transparency, not a solution to color space mismatch.
The study tip here: linear light, then transform. Any time passes are meant to be recombined mathematically, they must live in a linear space until the final display step. Watch for questions that describe "adding passes" — always ask what color space they're in first. - Which compositing method is most appropriate?
Answer: Alpha-over the rendered character onto the plate while respecting its premultiplied alpha
Whenever you see a compositing question involving transparency and rendered characters, the core concept being tested is premultiplied alpha and how it interacts with compositing operations. Blender's renderer outputs premultiplied (also called "associated") alpha images: the RGB channels are already multiplied by the alpha, meaning edge pixels carry color weighted by their transparency.
The correct approach is C — using Alpha Over with premultiplied alpha respected. Blender's Alpha Over node is specifically designed to handle premultiplied imagery. The operation it performs is essentially:
Output=Foreground+Background×(1−αfg)
Because the foreground RGB is already premultiplied, this formula composites cleanly without re-multiplying or artificially darkening edge pixels — preventing the dark fringe artifact described in the passage.
A is wrong because simply adding the character RGB to the plate ignores how alpha-weighted edges should replace background pixels, not stack on top of them. You'd get washed-out, overexposed results at full-coverage pixels.
B describes un-premultiplying by dividing RGB by alpha (converting to straight alpha), which is a valid intermediate step in some pipelines — but then it incorrectly discards the plate alpha and doesn't reassemble the composite correctly. Dividing by near-zero alpha values on transparent edges also introduces noise and blown-out fringes.
D multiplies already-premultiplied data by alpha again, double-premultiplying it. This darkens edges severely — exactly the dark fringe problem the question warns against — and the Screen operation doesn't fix it.
Your study tip: always ask whether image data is premultiplied or straight alpha before choosing a compositing node. Alpha Over in Blender assumes premultiplied input by default; mismatching these produces the telltale dark or bright fringe artifacts. - Which compositing change best meets the requirement?
Answer: Replace Diffuse Color with the new blue color, reconstruct the diffuse branch, and add the untouched branches
Whenever you see a compositing question about isolating changes to one surface property, think in terms of Blender's render pass reconstruction formula. The full diffuse contribution is built as:
Diffuse=Diffuse Color×(Diffuse Direct+Diffuse Indirect)
This means color and lighting are separate, multiplied layers — which is exactly what lets you swap one without touching the other.
B is correct because it respects this architecture precisely. You replace only the Diffuse Color pass with blue, then multiply it by the untouched Direct and Indirect passes to reconstruct the diffuse branch. Finally, you add the unmodified Glossy, Transmission, and Emission branches back in. The lighting pattern stays identical because the Direct and Indirect passes are never altered. Nothing else changes.
A is flawed because tinting the Combined pass affects everything — reflections, emission, transmission — not just the surface color. Subtracting the old Diffuse Color afterward doesn't cleanly undo those contaminated contributions.
C gets the math backwards. You'd be multiplying the lighting passes by the color instead of the color by the lighting passes — but more importantly, tinting Direct and Indirect blue bakes the color into the lighting data, which corrupts those passes for any future use and incorrectly alters the light-bounce information itself.
D is a brute-force mistake. Replacing every color pass ignores that many passes (like Glossy and Transmission) have no "color" component in the same sense, and averaging reconstructed branches produces physically meaningless results.
Your study tip: memorize the reconstruction formula for each pass branch. Blender exam questions frequently test whether you know that color and lighting are multiplied, not added — so changes must be made at the Color pass level, not the lighting level. - Which conceptual node operation produces the intended depth fog?
Answer: Mix from the render to blue fog using the Mist pass as the blend factor
When composing depth fog in Blender's compositor, think of the Mist pass as a blending weight: it tells each pixel how much fog to apply. Since nearby surfaces output 0 and distant surfaces output 1, a Mix node using the Mist pass directly as the factor will interpolate smoothly from your render (at factor 0) toward the blue fog color (at factor 1). That's exactly what D describes — and it's the clean, correct approach.
A is flawed because adding the Mist pass to the render brightens the image rather than replacing it with fog. Multiplying that sum by the fog color compounds the error and bears no logical relationship to depth-based blending.
B inverts the Mist pass before using it as the blend factor. An inverted Mist pass makes nearby surfaces 1 and distant surfaces 0, which causes the fog color to dominate the foreground and the render to dominate the background — the exact opposite of the intended effect.
C multiplies the render by the Mist pass, which darkens near pixels toward black (since Mist ≈ 0 there) and fully preserves far pixels — again backwards. Placing the fog "behind its alpha channel" is also conceptually muddled; alpha compositing doesn't work this way to produce smooth depth fog.
A useful pattern to remember: whenever a pass encodes a gradient from 0 to 1, you can plug it straight into a Mix node's factor field, where 0 = first input and 1 = second input. Match your desired transition direction to the pass's natural direction, and you'll rarely need to invert. - Where should the Ambient Occlusion factor be applied before reconstructing the beauty image?
Answer: Multiply it into the reconstructed diffuse branch, then add glossy, transmission, and emission unchanged
When compositing a beauty image from render passes, you need to think carefully about which branches of light contribute to the effect you're modifying. Ambient Occlusion darkens areas where surfaces are close together — it affects how diffuse light reaches surfaces, not how objects emit light, reflect mirrors, or transmit through glass. This distinction is the heart of the question.
The correct approach, A, multiplies the AO factor only into the diffuse branch before reassembling the beauty. The classic render pass reconstruction looks like this: Beauty = Diffuse × AO + Glossy + Transmission + Emission. This keeps contact shading darker while leaving emissive signs, mirror reflections, and light through glass completely untouched — exactly what the artist needs.
Answer B is a common and tempting mistake. Multiplying AO into the fully combined image darkens everything — glossy highlights, emission, and transmission included — which would dim those neon signs and flatten the mirror reflections the artist specifically wants to preserve.
Answer C misunderstands what AO models physically. Emission is self-illumination; it has nothing to do with how ambient light reaches a surface. Multiplying AO into the emission branch would darken glowing signs while leaving diffuse contact shading unchanged — the exact opposite of the goal.
Answer D confuses a compositing operation with a shading one. Using AO as an alpha channel controls transparency, not surface darkness, and compositing over black would simply cut out parts of the image rather than strengthening contact shadows.
Study tip: Always ask yourself which light type each pass represents. AO targets diffuse reception, so it belongs in the diffuse branch — not the full composite, not emission, and never the alpha channel.
- Which pass combination is designed for this task?
Answer: Use the Combined image with the Vector and Z passes in a Vector Blur operation
When working with motion blur in Blender's compositor, the key is identifying which render passes carry the right kind of data for the specific blur operation being used. The question describes a post-render motion blur workflow that needs both motion information and depth separation — that combination points directly to the Vector Blur node.
The Vector Blur node was designed precisely for this task. It takes the Combined pass (the rendered image), the Vector pass (which stores per-pixel 2D motion vectors describing how objects moved during the frame), and the Z pass (depth information) to intelligently blur motion while preventing color from bleeding across depth boundaries. The Z data lets Blender determine which pixels belong to foreground versus background objects, keeping blur localized to the correct depth layer. That makes D the correct answer.
Choice A is wrong because Normal and Mist passes describe surface orientation and atmospheric depth, not motion direction — a directional blur driven by these would produce artistically incorrect results with no relationship to actual object movement. Choice B fails because UV and Alpha passes describe texture coordinates and transparency, not motion; a standard Gaussian blur is also uniform and has no concept of per-pixel motion vectors or depth separation. Choice C is wrong because Object Index and Shadow passes handle masking and shadow data — an Alpha Over node composites layers together rather than applying motion blur at all.
A useful study tip: whenever you see "motion blur in compositor," think Vector pass + Z pass + Vector Blur node. That trio is Blender's dedicated post-render motion blur solution, and each component has a specific, non-interchangeable role.
- How should these passes be connected for the intended denoising workflow?
Answer: Connect Combined to Image, Denoising Normal to Normal, and Denoising Albedo to Albedo
When working with Blender's compositor Denoise node, think of it like filling out a form — each input socket has a specific purpose, and the passes you enable in your render settings are designed to slot directly into those corresponding slots.
The Denoise node uses three distinct inputs: Image (the raw noisy render), Normal (surface direction data to preserve edges and geometry detail), and Albedo (base color information to help reconstruct clean textures). The Cycles render passes — Combined, Denoising Normal, and Denoising Albedo — map directly onto these three inputs in a one-to-one relationship. So the correct workflow, answer B, is simply connecting Combined → Image, Denoising Normal → Normal, and Denoising Albedo → Albedo. Each pass feeds the slot it was specifically generated to serve.
Answer A is wrong because multiplying Normal and Albedo together produces meaningless data — these passes represent fundamentally different kinds of information (geometry direction vs. surface color) and combining them mathematically before input destroys what the denoiser needs from each separately. Answer C misunderstands the node entirely; there is no "blend factor" input on the Denoise node, and adding Albedo to Combined corrupts the raw image data the denoiser relies on. Answer D swaps Combined and Denoising Albedo, putting noisy beauty data where base color belongs and vice versa — the denoiser would receive garbage guidance and produce poor results.
A useful tip: whenever you see Cycles denoising passes in a question, remember the rule — each pass has one matching socket, and none of them are interchangeable.
- Which revision most directly stabilizes the effect across the sequence?
Answer: Map a fixed world-depth interval to the desired factor range and clamp values outside it
When working with Z passes in Blender's compositor, the core issue is normalization scope. A Normalize node maps the minimum and maximum values within the current frame to 0–1. This means the output is always relative — if a distant object enters the frame and becomes the new maximum depth, every other value rescales accordingly. Stationary foreground objects appear to shift their depth grade even though nothing about them changed.
The fix is to replace frame-relative normalization with absolute mapping: define a fixed world-depth interval (say, 1–100 Blender units) and map that range manually to your 0–1 factor range, then clamp anything outside it. Because the mapping reference never changes between frames, a new object at any depth cannot disturb the grade of existing objects. This is exactly what B describes, making it the correct revision.
A is a trap. Inverting the normalized output still leaves you with a frame-relative normalization — the instability isn't caused by which end is "largest," it's caused by the range shifting each frame. Inversion doesn't solve that.
C is a misunderstanding of what causes the problem. The Alpha pass affects transparency compositing, not depth stability. Blending it into Z before normalization would corrupt depth values rather than stabilize them.
D confuses color pipeline concerns with compositing math. Converting Z to display-referred space compresses its numeric range but doesn't eliminate the frame-relative rescaling — the Normalize node still shifts with each frame's content.
The broader lesson: whenever an effect must be temporally consistent, avoid any operator whose output depends on per-frame data ranges. Prefer explicit, clamped mappings over automatic normalization.
- What is the most likely correction?
Answer: Add the graded glossy branch to the other branches instead of averaging the branch pairs
When reconstructing a beauty pass in Blender's compositor, the correct mathematical operation is addition — you sum the individual light branches (diffuse + glossy + transmission + emission) to reproduce the original render. This is because each branch represents a separate contribution of light energy, and energy contributions combine by addition, not averaging.
The problem here is that Mix nodes with a factor of 0.5 perform a weighted average: 0.5×A+0.5×B. When you chain multiple Mix nodes this way, each additional branch gets diluted further. If you have four branches and average them pairwise, you lose a significant portion of the total brightness — hence the dimmer result.
A is correct because replacing those Mix nodes with Add nodes restores the proper additive combination. Each branch contributes its full value to the final image, matching how Cycles actually computed the beauty pass internally.
B is wrong because multiplying branches together has no physical basis here. Multiplication is used for masking or modulating signals, not for recombining light contributions — it would crush values and produce an even darker, incorrect result.
C is wrong because subtracting the original glossy branch from other branches would corrupt those channels with irrelevant data, introducing artifacts rather than fixing brightness.
D is wrong because setting every Mix factor to 1 would just output the last connected branch in each node, discarding all earlier branches entirely — not combining them.
Study tip: Whenever you decompose and recompose render passes, always ask yourself: was this originally built by addition? If yes, use Add nodes — never Mix — to reconstruct it. - Which RGB value reconstructs the pixel's diffuse contribution?
Answer: (0.60,0.24,0.10), found by coloring the sum of both lighting passes
When reconstructing diffuse contribution in Cycles render passes, you need to understand what each pass actually represents. The Diffuse Color pass holds the material's base color, while the Diffuse Direct and Diffuse Indirect passes hold lighting intensities — raw light values that haven't yet been tinted by the surface color. To get the final diffuse result, you multiply the color by the combined lighting, not add everything together arbitrarily.
The correct approach (C) first sums both lighting passes: (0.50+0.25, 0.50+0.10, 0.50+0.00)=(0.75,0.60,0.50). Then you modulate that total light by the diffuse color via component-wise multiplication: (0.80×0.75, 0.40×0.60, 0.20×0.50)=(0.60,0.24,0.10). This correctly reconstructs the full diffuse contribution.
Choice A simply adds all three passes together — (0.80+0.50+0.25, …)=(1.55,1.00,0.70) — which treats the color pass as additive light rather than a multiplier, producing an oversaturated, physically meaningless result. Choice B only multiplies the color by the direct pass and ignores indirect lighting entirely, giving (0.40,0.20,0.10), which drops the bounce-light contribution. Choice D multiplies all three passes component-by-component — (0.80×0.50×0.25, …)=(0.10,0.02,0.00) — double-applying the color as if it were a third light modifier, which has no physical basis.
As a study tip, remember the formula: Color × (Direct + Indirect). The lighting passes combine first, then the color modulates them — never the other way around.
