- Home »
Fast Bokeh Documentation
- Fast Bokeh Documentation
- What is Fast Bokeh?
- What does "Pro" mean?
- System Requirements
- Blur Radius
- Depth Map Layer
- Invert Depth Map
- Focus Distance
- Use Focus Point
- Use Focus Point Every Frame
- Zone Control
- Intensity Boost
- Render Quality
- Linear Color Working Space
- Edge Behavior
- Use GPU
Please note: Make sure to download and install the latest version.
What is Fast Bokeh?
Fast Bokeh is a plug-in to simulate cinematic Depth of Field (also known as Bokeh) for 3D renders using a Depth Map.
Rendering Depth of Field directly in a Raytracer/Physical Rendering system is prohibitively slower for many projects. Fast Bokeh helps to simulate that easily in post-production using Depth Maps.
What does "Pro" mean?
If you purchase a "Pro" license and register it, you simply get more features like Custom Iris Shapes, Zone Control, Linear Color Working Space, GPU acceleration and more.
If Fast Bokeh Pro isn’t registered you’ll see a watermark rendered by the plugin if you utilize any of the Pro features. To remove these restrictions, you have to register the plugin with the license key you have purchased. To register, click on the “Register” button in the effect controls of Fast Bokeh.
For more detailed instructions checkout the registration page.
For Windows machines, Windows 7 and above. OpenGL 4.3 and above required for GPU Rendering. For Macintosh computers, macOS 10.14 (Mojave) and above is required. For GPU Rendering a GPU that supports Metal 2.
Blur Radius ∞
Amount to blur to applied. Internally, this is multiplied by 5 to calculate the blur radius in pixels.
Depth Map Layer ∞
The Depth Map Layer to use from which blur radius for each pixel is to be calculated. Depth values have to be in the range 0.0 to 1.0. A depth value of 0.0 indicates nearest to the camera and a value of 1.0 indicates furthest away from camera. If no depth map layer is applied, all pixels are blurred to the full radius.
Invert Depth Map ∞
Invert the values of the Depth Map. Useful to remap some depth maps produced by some applications.
Focus Distance ∞
The depth value that is currently in focus. You can either enter it manually or use the Focus Point Selection.
Use Focus Point ∞
If selected, select the Focus Point directly in the comp window using the Point picker. Focus distance values are automatically updated. Pro Registration Required.
Use Focus Point Every Frame ∞
This option enables to sample the Focus Distance actively from the location of the Focus Point, every frame. Useful if you need to track a certain object for focus.
This section enables you to choose the shape and other characteristics of the Iris used to simulate the blur.
Set the shape of the Iris to one of the following. You can visualize the shape using the Iris Preview parameter.
Gaussian: The blur looks like a Gaussian blur. This produces the fastest blur but visually not always the best.
Spherical: The blur looks like multi-blade full circle Iris shape. Requires Pro Registration.
Polygon: The blur looks like multi-blade Iris shape with fixed number of notches. Requires Pro Registration.
Polygon (HQ): Similar to Polygon option but renders at a much higher quality.
Custom Image: Use a custom image layer as an Iris. See further notes below
Iris Aspect Ratio
Set the Aspect Ratio of the Iris. Useful to create Anamorphic Bokeh.
Set the number of notches to the Polygon Iris shape.
Set the rotation of the notches in the Polygon Iris shape.
Set the curvature of polygon sides.
Custom Image Iris
Use a layer image as the shape of the Bokeh. Currently, only Luminance from the layer is supported.
We recommend using Image Layer of dimensions 128x128px. You can use bigger size images too, but might cause a slight drop in render performance. But if you do use lower than 128x128px, there will be a significant loss of quality especially in higher blur radii. We recommend setting the Image Sampling Detail between 25-100. You can enter much higher values but the results are not much different visually.
Iris Image Sampling Detail
Set the Image Detail co-efficient. The higher this number is, the more detailed the Iris shape is, but the render will become a lot slower. The level of detail follows a diminishing return i.e. for values higher than 100, the visual difference is imperceptible. We recommend setting this as low as possible.
Zone Control ∞
Zone Control or technically speaking Segmentation will give a lot of control over the parts of the image that have to be blurred. For example, you can have full blur in the far zone i.e. away from the focus and no blur in the near zone i.e. closer to the camera. You can also have a much wider focus zone so that all pixels that have a depth in that zone will not be blurred.
This visualization shows the amount of blur applied at the nearest zone to the camera, the focus zone and also the farthest zone. The blur is the highest at the near end and far end. It is completely zero in the focus zone.
Focus Zone Width
Set the width of the focus zone. The wider it is, the more depth range is in focus. If set to zero (default) only one depth value is in focus.
Focus Zone Offset
Set the offset of the Focus Zone. By default the focus zone is split evenly among near and far zones.
Near Blur Factor
Blur Factor to be applied to all the pixels in the near zone. If set to zero, near zone pixels will not be blurred.
Far Blur Factor
Blur Factor to be applied to all the pixels in the far zone. If set to zero, far zone pixels will not be blurred.
Render Each Zone Separately
If checked, each zone is Rendered Separately and the composited to a final image. This avoids a lot of pixel leakage from far and near zones into focus zones.
Visualize each zone or the final composite when rendering each zone separately.
Adjust the highlights in the image to get a more cinematic effect. For best results use 32bpc.
Minimum luminosity threshold required before brightness and saturation are modified for the pixel. Setting the threshold to 0 boosts all non-zero pixels. Setting it to 1.0, only boosts over bright pixels.
The amount of energy fed into pixels that are above the threshold. Depending on how bright a pixel is above the threshold, the pixel value is increased in brightness by this relative amount.
The amount of color retained in the boosted pixel. Setting it to 0 pushes the boosted colors toward white while setting to 100 attempts to retain as much color as possible. Useful to retain the color of highlights in the image.
Intensity Boost ∞
Boost the color intensity. Useful for large blur radii and transparent images.
Render Quality ∞
Multiple levels of quality. 1x is the fastest but low on details. 4x is the highest but slowest. 3x is the sweet spot for quality and render performance. Visually, 3x and 4x are almost identical. So 3x is recommended instead of 4x.
Linear Color Working Space ∞
Produces a result similar to a project set to a linearized working color space. Check this property (or set it from the project settings) to see realistic bokeh effects from your images.
Edge Behavior ∞
Change the way Image Boundary edges are handled by the plug-in.
None: The default setting. No special treatment for edge pixels and no change in dimensions of the image. Pixels beyond the image boundaries are considered as zero pixels. The rendered image is the same size as the input image.
Repeat Edge Pixels: Edge Pixels are repeated through out the length of the blur. The rendered image is still the same size as the input image.
Expand Image Boundaries: Pixels beyond the image boundaries are considered as zero pixels but the output image is expanded to include the entire length of the blur. The rendered image is significantly larger than the input image. Please Note: This can cause performance to slow down at high radii blurs.
Use GPU ∞
Use GPU to render. Please note that you need a compatible GPU that supports Compute. We don't recommend using this option if you're using Integrated GPU on a laptop. Results may vary depending on GPU.
You need a GPU that supports Metal 2 on macOS (10.14 and higher only). On Windows, you need a GPU that supports compute (GL 4.3 and above).
Due to differences in algorithms, GPU and CPU renders might have minor differences. For pixel accurate results across multiple machines and render farms CPU rendering is recommended.