ZEISS CinCraft LensCore

Group: Input_properties

Contains 2 elements: World to Meter, Extra Channels

Everything in this group tells the plugin, how to see the world. This currently are the world scale, as well as what additional channels to “see” and work on.

Knobs: World to Meter

world_to_meter, Slider, default: 0.000000

The value to enter is the multiplier needed to transform the image data into meters.

Example: The input data is in cm scale. To transform a centimeter scale to meters, we need to divide the actual number by 100, resulting in a “world to meter” factor of 0.01.

If the value is wrong, render time will increase drastically and the whole image will be in defocus.

Knobs: Extra Channels

input_channels, channelset

The extra channels option can be used to create matching masks for elements in the defocus chain.

Given the following example of the second chair in the example scene.

The alpha layer shall be used to create a mask to later recolor and finetune some elements of the scene. In this case the second chair. The mask needs to represent occlusion, distortion and lens breathing. All these transformations can be done to the mask alongside the actual lens effect application.

With this additional info of the second chair’s alpha channel, we can enable thee extra_channels feature in the plugin:

The resulting mask can be used by e.g. ColorCorrect node to manipulate the chair after the lens effects have been applied.

Group: Render_quality

Contains 5 elements: Quality Level (GUI), x, Quality Level (Farm), x, Disable rendering optimizations

The lens effect simulation is a complex algorithm. In order to find the best trade-off between quality and speed, the quality level can be tweaked here.

Knobs: Quality Level (GUI)

quality_level, Dropdown, default: Medium

Options: [PreVis, Low, Medium, High, Ultra]

This option only applies for Nuke instances with a graphical interface. It will be ignored for headless instances, e.g. on a rendering farm.

In order to get a quick preview or a spotless image, select one of the quality levels PreVis (pre-visualization), Low, Medium, High, Ultra. High and Ultra come with a multiplicator.

Knobs: x

quality_scale, Slider, min: 0.2, max: 5, default: 1.000000

This option only applies for Nuke instances with a graphical interface. It will be ignored for headless instances, e.g. on a rendering farm.

Only applying to quality levels “High” and “Ultra”, the quality scale increases the output quality over the given bounds.

Knobs: Quality Level (Farm)

quality_level_nogui, Dropdown, default: High

Options: [PreVis, Low, Medium, High, Ultra]

This option only applies for Nuke instances without a graphical interface. It will be ignored for local instances with a visible UI.

In order to get a quick preview or a spotless image, select one of the quality levels PreVis (pre-visualization), Low, Medium, High, Ultra. High and Ultra come with a multiplicator.

Knobs: x

quality_scale_nogui, Slider, min: 0.2, max: 5, default: 1.000000

This option only applies for Nuke instances without a graphical interface. It will be ignored for local instances with a visible UI.

Knobs: Disable rendering optimizations

disable_adaptive_ray_count, Checkbox, default: 0

Disables all optimizations. This will get rid of all artifacts that are due to optimizations to the lens effect core algorithm.

Important: Only check this box if visual artifacts are present, that cannot be fixed with other means e.g. inpainting. If the malfunction is not a rendering relict of the internal optimizations, uncheck this box again for better performance.

Group: Bounding_box

Contains 2 elements: Show overscan coordinates, Increase Bounding Box by pixels

Knobs: Show overscan coordinates

overscan_visualization, Checkbox, default: 0

Knobs: Increase Bounding Box by pixels

bounding_box_expansion, Number, default: 0

Group: Feedback

Contains 2 elements: Local telemetry log folder:, Report Issue

Knobs: Local telemetry log folder:

telemetry_folder, FilePicker

Our software collects telemetry data to aid support and further development. Telemetry is entirely opt-in; all collected data is stored locally on the user’s device and requires manual transmission by the customer, ensuring no data exchange occurs by default. This data is strictly unpersonalized, meaning it does not include any personal information or identifiers. Importantly, no sensitive information, such as image content, is collected during the process. We only gather metadata, such as usage patterns, system and feature configurations, to help us understand how our software is being used and identify areas for improvement. Additionally, the data is formatted in a human-readable manner, allowing customers or their IT departments to review and oversee the information before it is sent to ZEISS.

Knobs: Report Issue

provide_feedback_btn, Button

Use this button to get in contact with the ZEISS service. Please follow the instructions in the mail draft to provide as much information about either your feedback or issues. Your insights will help us enhance the plugins functionality and better meet your needs.

Group: Bokeh Preview

Contains 3 elements: Preview the bokeh, Use camera data for preview, Preview divider location

Knobs: Preview the bokeh

enable_bokeh_preview, Checkbox, default: 0

This checkbox enables the bokeh preview.

In order to see the behavior of the effect, two black walls with backlights are created: the left wall is placed in front of the focus, the right wall is placed behind the focus. The two walls will be separated by a white line. All together, this will look like a perforated plate over two depths:

The effect is applied to this just like any other scene. All lens effect settings are in place. It is not possible to render this preview in headless mode.

Knobs: Use camera data for preview

bokeh_preview_use_view_data, Checkbox, default: 0

While only enabling the bokeh preview will create perforated plates specifically to get a bokeh as large as possible, it’s using a fixed focus and f-stop. Using camera data will allow to preview the actual bokeh as it will be seen in the scene.

Knobs: Preview divider location

bokeh_preview_divider_location, Slider, min: 0, max: 1, default: 0.600000

This slider allows to move the white line (and with it the border between front and back defocus) across the image.

Group: Aperture

Contains 6 elements: Aperture Stop, Number of aperture blades, Aperture blade curvature, Aperture rotation offset, Simulated minimum fstop, Fstop rotation scaling

Knobs: Aperture Stop

stop_type, Dropdown, default: Fstop

Options: [Fstop, Tstop]

Allows the use of the Camera node’s f-stop as a t-stop (as it is typically used for real world footage). This option is not available for ZEISS DigitalLens lenses, i.e. ZEISS DigitalLens Lite and ZEISS DigitalLens Pro.

Knobs: Number of aperture blades

num_aperture_blades, Number, default: 0

Use 0 to disable and use no aperture blades. Invalid values will round to the next valid one, e.g. inputs of 1 or 2 will be treated as 3 blades.

Knobs: Aperture blade curvature

aperture_blade_curvature, Slider, default: 0.000000

This is treated as an interpolation factor of how close the blades are to a circle. The physical pendant to this feature is the curvature in the aperture fins. With a curvature of 1, the blade effect is fully deactivated independent of the setting in Number of aperture blades. The example images below are rendered with 3 blades.

Knobs: Aperture rotation offset

aperture_rotation_offset, Slider, min: 0, max: 360, default: 0.000000

Knobs: Simulated minimum fstop

override_min_fstop, Slider, default: 0.000000

This feature is only available for the ZEISS DigitalLens Pro. This allows to define the fStop that corresponds to full the designed cateye shape by using Cateye Bokeh Simulation. The same minimum fStop is used as a starting point for the blade rotation simulation. With increasing fStop, the cateye is less and less visible, as the resulting bokeh is limited by the aperture blades.

The following renders have been made all with the exact same lens settings with the ZEISS DigitalLens Pro with 5 aperture blades and cateye simulation active. The Camera fStop is set to 2.0 for all tests - only the Simulated minimum fStop option is modified in between renderings.

Knobs: Fstop rotation scaling

aperture_fstop_rotation_scaling, Slider, default: 1.000000

A value of 0 disables the fStop dependency on the rotation offset. Negative values will reverse the direction from clockwise to counter-clockwise.

Group: Focus & Defocus

Contains 9 elements: Anamorphic aspect ratio, Front defocus intensity, Back defocus intensity, Lens transmission, Focus sharpness, Color-code the focus in the image, Field curvature ^2, Field curvature ^4, Plot curvature

Knobs: Anamorphic aspect ratio

pseudoanamorphic_aspect, Slider, min: 0.25, max: 4, default: 1.000000

Anamorphic lenses have an aspect ratio and thus lead to ellipsoid bokeh throughout the image. The anamorphic behavior does not depend on the position in the image. This feature can be used to stretch the behavior of any given lens to aproximate the look of an anamorphic even if the lens itself isnt.

Knobs: Front defocus intensity / Back defocus intensity

front_defocus_intensity, Slider, min: 0, max: 5, default: 1.000000
back_defocus_intensity, Slider, min: 0, max: 10, default: 1.000000

These modifiers change the effect of the defocus to either emphatise or weaken blur elements in front or behind the focal plane. The effect works by scaling the physical world in the scene.

For front_defocus_intensity elements in the back of the focal plane stay untouched. The closer to the camera, the bigger the bokeh. Front defocus bokeh can scale into infinity so ensure to make little increments when using with this feature.

For back_defocus_intensity elements in front of the focal plane stay untouched. A lens always has a maximum bokeh size in the back defocus. This means higher values do not always lead to increasing bokeh. The physical limit of the lens cannot be overwritten with this feature.

Knobs: Focus sharpness

enforce_focus_threshold, Slider, min: 0, max: 15, default: 0.000000

Real, physical lenses do not have a 1:1 pixel match, even if the inspected area is in focus and in the center of the image. This is caused by non-perfect materials, build tolerances, wavelength specific behavior through the glass and other effects. To forcibly overrule those remaining (physical) aberrations this value is introduced. The higher the value, the bigger the forced sharpness around the focal plane.

Knobs: Color-code the focus in the image

focus_visualization, Checkbox, default: 0

As the effect is for previewing and editorial changes only, the quality-level will be automatically lowered significantly when using this overlay.

Knobs: Field curvature ^2

field_curvature_2, Slider, default: 0.000000

Adds a quadratic term offset to the focus behavior of the lens. Will vary with image position. No offset in center (0), full offset in diagonal (1). All offsets will be accumulated.

$$\text{offset} = \text{img\_pos}^2 * \text{field\_curvature\_2} \si{\m}$$

Group: Cateye Bokeh Simulation

Contains 49 elements: Enable bezier cateye bokeh simulation, Transition start, Transition end, Radius fall-off, Aspect ratio, Roundness x, Roundness y, Simulate cateye as circles, MOD focus pos, INF focus pos, INF pupil scale, INF pupil scale start focus [m], INF pupil scale end focus [m], Field interpolation pos fp0, Field interpolation pos fp1, Field interpolation pos fp2, Field interpolation pos fp3, mod Circle 1 radius field pos 0, mod Circle 1 offset field pos 0, mod Circle 2 radius field pos 0, mod Circle 2 offset field pos 0, inf Circle 1 radius field pos 0, inf Circle 1 offset field pos 0, inf Circle 2 radius field pos 0, inf Circle 2 offset field pos 0, mod Circle 1 radius field pos 1, mod Circle 1 offset field pos 1, mod Circle 2 radius field pos 1, mod Circle 2 offset field pos 1, inf Circle 1 radius field pos 1, inf Circle 1 offset field pos 1, inf Circle 2 radius field pos 1, inf Circle 2 offset field pos 1, mod Circle 1 radius field pos 2, mod Circle 1 offset field pos 2, mod Circle 2 radius field pos 2, mod Circle 2 offset field pos 2, inf Circle 1 radius field pos 2, inf Circle 1 offset field pos 2, inf Circle 2 radius field pos 2, inf Circle 2 offset field pos 2, mod Circle 1 radius field pos 3, mod Circle 1 offset field pos 3, mod Circle 2 radius field pos 3, mod Circle 2 offset field pos 3, inf Circle 1 radius field pos 3, inf Circle 1 offset field pos 3, inf Circle 2 radius field pos 3, inf Circle 2 offset field pos 3

The cateye simulation creates a physically correct (but not necessarily physically sane) cateye effect.

Several parameters are used to tweak the bokeh, aiding users in configuring their desired lens look.

Knobs: Enable bezier cateye bokeh simulation

cateye_simulation, Checkbox, default: 0

Knobs: Transition start / Transition end

cateye_simulation_morph_start, Slider, min: 0, max: 1, default: 0.000000
cateye_simulation_morph_end, Slider, min: 0, max: 5, default: 1.000000

“Transition” addresses the shape change from a perfectly round bokeh shapes to the characteristic cateye-shape:

The bokeh shape morphs in a linear fashion from the center of the image to the edges, increasing the cateye effect to its maximum on the outermost parts of the image. The controls for morphing start/end allow to change the begin and end of the area in which this transition takes place:

The values are a percentage of the image circle’s radius. Transition start can only be inside the image, i.e. a value between 0 % and 100 %. Transition end can also lie outside of the image, allowing to create the visual equivalent of a cropped shot.

Knobs: Radius fall-off

cateye_simulation_radius_falloff, Slider, min: 0, max: 1, default: 1.000000

The slider is a percentage to which size the bokeh radius will fall off.

Knobs: Aspect ratio

cateye_simulation_aspect_ratio, Slider, min: 0.25, max: 4, default: 1.250000

The aspect ratio controls the “cateye-ness” of the bokeh. This will always shrink the bokeh, never extend its size.

Knobs: Roundness x / Roundness y

cateye_simulation_scaling_x, Slider, min: 0, max: 1, default: 0.500000
cateye_simulation_scaling_y, Slider, min: 0, max: 1, default: 0.150000

The roundness controls allow you to adjust the horizontal and vertical scaling of the cateye simulation, creating different elliptical shapes for the bokeh highlights.

These controls work together to create the desired cateye effect:

• Roundness X: Controls horizontal scaling
• Roundness Y: Controls vertical scaling

Group: Bokeh Outline

Contains 2 elements: Gain, Thickness

Knobs: Gain

bokeh_simulation_outline_gain, Slider, min: 0, max: 20, default: 0.500000

Gain refers to the intensity or brightness of the bokeh outline. By increasing the gain, the edges of the bokeh become more pronounced and vivid, making them stand out more against the background. Conversely, reducing the gain results in softer, less distinct edges, which can be useful for creating a more subtle and natural look.

Knobs: Thickness

bokeh_simulation_outline_thickness, Slider, min: 0, max: 50, default: 3.000000

Thickness pertains to the width of the bokeh outline. Adjusting the thickness allows users to control how wide or narrow the edges of the bokeh appear. A thicker outline can give the bokeh a more defined and bold appearance, while a thinner outline can create a delicate and refined effect.

Group: Spherical aberration

Contains 4 elements: Mode, Intensity, Fall-off, Spherical aberration ray amplitude [mm].

Knobs: Mode

spherical_aberration_mode, Dropdown, default: Off

Options: [Off, Positive, Negative]

Knobs: Intensity

spherical_aberration_weight_intensity, Slider, min: 0, max: 1, default: 1.000000

Knobs: Fall-off

spherical_aberration_pow_exponent, Slider, min: 0, max: 10, default: 1.000000

Group: Vignetting

Contains 3 elements: Scale, Fall-off, Offset

The “vignette” effect in optics refers to a gradual darkening of the image towards the corners compared to the center. This phenomenon is primarily caused by the lens design and the way light is transmitted through the lens elements. As light rays enter the lens at oblique angles, particularly at the edges, they may be less effectively focused onto the image sensor or film, leading to reduced brightness in those areas. Vignetting can also be influenced by factors such as lens aperture, focal length, and the use of filters or lens hoods.

Both of the following parameters should be used in combination to reach the desired outcome.

Knobs: Scale

vignette_intensity_scale, Slider, min: 0, max: 10, default: 1.000000

Knobs: Fall-off

vignette_intensity_falloff, Slider, min: 0, max: 10, default: 1.000000

Knobs: Offset

cos4_vignette_weight, Slider, min: 0, max: 10, default: 0.000000

Group: Distortion & Breathing

Contains 6 elements: Distortion intensity, Distortion coefficients, Distortion clamp, Lens breathing, Lens breathing, Plot lens breathing

Knobs: Distortion intensity

distortion_intensity, Slider, default: 1.000000

Distortion Intensity is a parameter that allows to control the degree of optical distortion introduced by the lens. Optical distortion can manifest as barrel or pincushion effects, where straight lines appear curved or bent. By adjusting the distortion intensity, the unique characteristics of various lens types can be simulated.

Knobs: Lens breathing

lens_breathing, Checkbox, default: 1

This checkbox is only available for predefined lenses. It is used to enable/disable the breathing effect, i.e. field of view changes due to a slight variance in the physical focal length of a lens over various focus points.

Group: Chromatic Aberration

Contains 13 elements: Enable Chromatic Aberration, Draw color combinations, Front chroma color, Back chroma color, Chromatic aberration intensity, Lateral chromatic aberration intensity, Longitudinal chromatic aberration intensity, Transition minimum for inf CA, Transition maximum for inf CA, Chromatic aberration at inf back wl, Chromatic aberration at inf front wl, Back WL lateral CA scaling, Back WL longitudinal CA scaling

Chromatic aberration is an optical phenomenon where a lens fails to focus all colors to the same convergence point, resulting in color fringing around high-contrast edges in an image. This occurs because different wavelengths of light are refracted by varying degrees as they pass through the lens, causing them to focus at different points along the optical axis.

Knobs: Enable Chromatic Aberration

use_multiple_wavelengths, Checkbox, default: 1

Knobs: Back chroma color / Front chroma color

chroma_bg_color, color, default: 0, 0, 1
chroma_fg_color, color, default: 1, 0, 0

Only available for the ZEISS DigitalLens Pro.

The following visualization usese the Bokeh preview feature in combination with the Longitudinal chromatic aberration intensity.

Knobs: Chromatic aberration intensity

chromatic_aberration_intensity, Slider, min: 0, max: 10, default: 1.000000

Only available for physical lenses.

Knobs: Lateral chromatic aberration intensity

lateral_chromatic_aberration_intensity, Slider, default: 1.000000

Only available for the ZEISS DigitalLens Pro.

Knobs: Longitudinal chromatic aberration intensity

longitudinal_chromatic_aberration_intensity, Slider, default: 1.000000

Only available for the ZEISS DigitalLens Pro.

Knobs: Chromatic aberration at inf back wl

chromatic_aberration_inf_fstop_delta_back_wl, Slider, default: 1.000000

Knobs: Chromatic aberration at inf front wl

chromatic_aberration_inf_fstop_delta_front_wl, Slider, default: 0.000000

Group: Filter Kernel

Contains 1 elements: Kernel scale

Knobs: Kernel scale

bokeh_kernel_skew, Slider, default: 0.000000

For more detailed information about the kernel usage, refer to the Kernel section.

Group: Step 1: Choose edge detection algorithm and threshold based on the observed ‘depth shadows’

Contains 3 elements: visualize, Mode, Threshold

The algorithm works based on edges, and will only inpaint in regions around detected edges. Use the visualize feature to ensure the depth shadows observed are detected as edges.

Knobs: visualize

visualize_edges, Checkbox, default: 0

Knobs: Mode

edge_detection_mode, Dropdown, default: Relative

Options: [Relative, Absolute]

In relative mode, a pixel is classified as an edge pixel if:

$$\lvert\text{own\_depth} - \text{neighbour\_depth}\rvert > \text{own\_depth} \times \text{factor}$$

Example R1: Relative

• Pixel Depth: $2 \si{\m}$
• Threshold: $0.5$
• Neighbor Pixel Depth: $3 \si{\m}$
• Calculation: $\lvert 2 \si{\m} - 3 \si{\m} \rvert > 2 \si{\m} \times 0.5 \rightarrow \text{false}$
• Result: This will not detect an edge, as $1\si{\m} \not> 1\si{\m}$.

Example R2: Relative

• Same as Example R1 except for Threshold 0.5 $0.49$
• Calculation: $1 \si{\m} > 2 \si{\m} \times 0.49 \rightarrow \text{true}$
• Result: An edge pixel is detected, as $1\si{\m} > 0.98\si{\m}$.

In absolute mode, a pixel is classified as an edge pixel if:

$$\lvert\text{own\_depth} - \text{neighbour\_depth}\rvert > \text{abs\_value}$$

Example A1: Absolute

• Pixel Depth: $2 \si{\m}$
• Threshold: $0.5 \si{\m}$
• Neighbor Pixel Depth: $2.5 \si{\m}$
• Calculation: $\lvert2 \si{\m} - 2.5 \si{\m} \rvert > 0.5 \si{\m} \rightarrow \text{false}$
• Result: This will not detect an edge, as $0.5 \si{\m} \not> 0.5 \si{\m}$.

Example A2: Absolute

• Same as Example A1 except for Threshold 0.5 $0.49\si{\m}$
• Calculation: $0.5 \si{\m} > 0.49 \si{\m} \rightarrow \text{true}$
• Result: An edge pixel is detected, as $0.5 \si{\m} > 0.49 \si{\m}$.

Knobs: Threshold

edge_detection_threshold, Slider, default: 0.075000

Group: Step 2: Set inpaining area / affected pixels for inpainting impact

Contains 4 elements: visualize, Alpha threshold, Close threshold to original pixel, Enlarge inpaint area

This step allows for a more granular selection of how and specifically where the inpainting algorithm is applied.

Enlarge inpaint area is a common parameter to adjust to the specifics of the given scene.

Restricting the inpainted area is often not necessary and thus it’s not adviced to change the default settings of Alpha threshold and Close threshold to original pixel except there is good reason.

Knobs: visualize

visualize_heuristic_inpainting, Checkbox, default: 0

Knobs: Alpha threshold

inpaint_alpha_threshold, Slider, min: 0, max: 1, default: 0.985000

Use and change this parameter only if you have non seperable alpha layers in the image and observe inpainting artifacts behind those. It is adviced to overlay the alpha elements later onto the image.

The inpaint algorithm is designed to insert pixels in obscurated areas behind foreground objects. If the foreground object is transparent this behvaior will be visible and thus might cause trouble.

The example uses very low levels of alpha to have strong transparency and thus a more easy visualization for documentation purpose.

Knobs: Close threshold to original pixel

skip_if_too_close_perc, Slider, default: 0.010000

For an explanation on how the Relative calculation works, see Edge detection mode.

The following example shows when this parameter comes into effect. With a scene that already has original data in places where inpainting would happen, inpainting pixels are ignored if they would be inserted in the given range around the original depth pixel. Selecting a value of 0 will disable this optimization and inpainted samples will be inserted no matter what.

Knobs: Enlarge inpaint area

increase_inpaint_ray, Number, min: 0, max: 10, default: 2

The coverage of individual inpaint rays can be troublesome, as the edge detection algorithm loses information and precision by rasterization in the pixel grid. Especially curved edges and high frequent changes need a bigger coverage in the width of the ray to have all of the necessary area inpainted accordingly.

Group: Step 3: How many pixels get generated within previously defined area

Contains 1 elements: Inpainting intensity

Knobs: Inpainting intensity

inpaint_intensity, Slider, min: 0, max: 10, default: 1.000000

Group: Step 4: Repeat edge pixels in z-direction

Contains 4 elements: visualize, Mode, Edge repetitions, Step

Rendering objects front-faced only makes them two dimensional for the renderer, even if they realistically have a depth.

This algorithm allows to extend the detected edge pixels backwards in z-direction to simulate this depth of the foreground object. Opposed to the Heuristic inpaint approach in Step 2 and 3, this will extend the foreground object instead of the background.

Knobs: visualize

visualize_virtual_depth_on_edges, Checkbox, default: 0

Knobs: Mode

edge_repetition_mode, Dropdown, default: None

Options: [None, Relative, Absolute]

Knobs: Edge repetitions

edge_repetition_count, Number, default: 1

Relative

$$\text{new\_depth}_{\text{count}} = \text{repetition\_sample\_depth} \times (1 + \text{count} \times \text{edge\_repetition\_step})$$

Absolute

$$\text{new\_depth}_\text{count} = \text{repetition\_sample\_depth} + \text{count} \times \text{edge\_repetition\_step}$$

Knobs: Step

edge_repetition_step, Slider, default: 0.030000

Group: Clamp_to_edge

Contains 1 elements: Enable

This feature is designed to fix small black “shadows” at the edges of the image.

Missing scene information will be filled with the most recent data along the closer axis.

If this behavior is insufficient, Show overscan coordinates can be used to visualize the amount of missing scene information, allowing the content to be rendered accordingly.

Knobs: Enable

clamp_to_edge, Checkbox, default: 0

Group: Light_source_booster

Contains 3 elements: Enable light source modification, Detection threshold, Scaling

The contrast of any given bokeh depends on its size and the brightness of the source. The energy is distributed over the entire bokeh area, so the larger the bokeh, the less contrast it will have.

This feature allows for manual fine-tuning of bokeh contrast for highlights.

Knobs: Enable light source modification

light_source_modification, Checkbox, default: 0

Knobs: Detection threshold

light_source_detection_threshold, Slider, min: 0, max: 100, default: 1.000000

This parameter is the threshold value in grayscale that defines when a pixel is detected as a light and will be boosted. The light_source_detection_threshold shifts the scaling curve accordingly.

Knobs: Scaling

light_source_scaling, Slider, min: 0, max: 10, default: 0.100000

This parameter describes the strength of the nonlinear scaling applied to elements above the light_source_detection_threshold.

The curve below the detection threshold remains an identity to avoid manipulating the input data. At the threshold, both the identity and the nonlinear curves have the same slope to ensure continuous behavior.

The curve appears as follows:

Group: Troubleshooting

Contains 2 elements: Rescale alpha, Close alpha holes

Knobs: Rescale alpha

rescale_alpha, Slider, min: 0.9, max: 1, default: 1.000000

This algorithm is designed to fix very small deviations in the alpha channel, specifically those smaller than 5%. It remaps the alpha values of the entire image from the range [0; rescale_alpha] to [0; 1], converting values above the specified threshold to an alpha of 1 in the output image while linearly scaling all other areas.

Knobs: Close alpha holes

close_alpha_holes, Number, min: 0, max: 2.14748, default: 0

This enhancement algorithm is designed to close holes in the alpha channel and is recommended for use after inpainting. It specifies the number of pixels to be filled, with a value of 0 disabling the feature and values greater than 0 indicating the radius of pixels to close the holes.

This algorithm offers an alternative to Rescale alpha and is particularly effective for small holes or very fine edge artifacts. It operates on the entire image, so users should be cautious of potential artifacts at the edges.