Tweet
I'm just wrapping my head around motion blur subsamples, and I was wondering if anybody could help me confirm or refute that this is how this works:
Lets say each black notch here represents a frame, and each red notch represents a subframe that is being sampled for half a frame of motion blur. First black notch is f0.0, second black notch is f1.0 (where the playhead is placed) and the third is f3.0.
no motion blur, f1 is sampled:
|------------|------------|
Using MB with two geometry samples, frames f0.75 and f1.25 are sampled, like so:
|---------|---|---|---------|
For three geometry samples it's f0.75, f1.0, f1.25
|---------|---|---|---------|
For four geometry samples, f0.75, f0.9166..., f1.166..., f1.25
|---------|--|-|-|--|---------|
...and so on. Does this seem accurate?
Lets say each black notch here represents a frame, and each red notch represents a subframe that is being sampled for half a frame of motion blur. First black notch is f0.0, second black notch is f1.0 (where the playhead is placed) and the third is f3.0.
no motion blur, f1 is sampled:
|------------|------------|
Using MB with two geometry samples, frames f0.75 and f1.25 are sampled, like so:
|---------|---|---|---------|
For three geometry samples it's f0.75, f1.0, f1.25
|---------|---|---|---------|
For four geometry samples, f0.75, f0.9166..., f1.166..., f1.25
|---------|--|-|-|--|---------|
...and so on. Does this seem accurate?
Comment