Mask Prompter 3

General workflow

You create mattes by specifying so-called prompts for the current frame. Prompts can either be points, rectangle masks around an object or textual descriptions of the objects you want to create a matte for.

Parameters

Points

Enable “Active” and specify the coordinates of the point. By keyframing the points position you can follow objects.

Boxes

Add rectangle masks to the layer and assign them to one of the five slots. This will create a matte for the object contained inside the rectangle.

Text Prompts

Describe the objects you want to mask by linking in masks and changing the masks name. To key e.g. a horse click “Add Text Input”, change the new masks name to “a horse” and link the mask layer in one of the slots. Sometimes the prompting requires some trial and error. You can detect multiple objects by separating words or phrases with a dot. E.g. “horse. red car.” will detect both horses and red cars.

Multiple

If active, will detect all object that match the phrases. Otherwise only one object with the highes similarity score is returned.

Detection Threshold

The lower this value the more objects will be detected, at the cost of more false positives.

Scale Rectangles

Increase the size of the detected rectangles around objects. This also has an influence of the matting results from the matting component.

Exclude Points

Mark areas in the image you want to exclude from the masks.

Models

Select the keying and object detection model.

Filter Width

The width of the filter around the corners. Use small values to apply minor corrections and larger values if the edges have great variance (e.g. hair in the wind).

High Resolution

Run the edge refinement on a higher resolution (consumes more VRAM).

Output

“Overlay” draws the matte as white overlay on top of the image (set opacity with “Overlay Strength”). “Matte” will output a simple black and white matte and “On Transparent” will give you the keyed object on alpha. “Detection Boxes” and “Detection Boxes On Transparent” will only render the detection boxes from text prompts, but no mattes.

Draw Rectangles

Draw rectangles / bounding boxes around objects detected from the text prompting model.

Color (only if rectangles are drawn)

The color of the rectangles.

Thickness (only if rectangles are drawn)

The thickness of the rectangles outline.

Tips & Tricks

If you use the “On Transparent” output the resulting matte can directly be postprocessed with After Effects internal effects like “Refine Soft Matte”.

Backend & Performance

Hardware Acceleration (on GPU version of the plugin only)

Run calculations on the GPU. This will give massive speedups compared to CPU mode.

Lower Precision

Compute with reduced precision if possible. This can save up to half of the memory and give you some speedups at the cost of sometimes slightly reduced quality.

Optimize for low VRAM

Enabling this will make sure only the ai model parts which are needed for computation are kept on the GPU. This might lower VRAM usage under some settings at the cost of moving ai models in and out of GPU memory.

Samples (not available for all settings)

The number of ai samples to calculate. This will improve the models accuracy.

Parallel (only available if Samples > 2)

This will render all samples at the same time (faster), if disabled computation might be slower but require less VRAM.

Computation Tiles (not available for all settings)

Split the computation into several tiles. This can help if you run out of memory.