The previous frames of the input clip are processed
with a pixel diffusion process and then combined with the current frame.
The output of each processed frame is stored and then combined with the
next frame. The trails are reinitialized whenever a non-consecutive frame
is processed, either the first frame, reprocessing a given frame, or
jumping to another frame. You must process multiple frames of a clip in a
row to observe the effect, and clearing your image cache before rendering
may sometimes be necessary.
In the Sapphire Time effects submenu.
Source: The current layer. The clip to be processed.
- Prev Brightness: Default: 0.8, Range: 0 or greater.
For each frame, the previous output is scaled by this amount
before it is combined with the new input frame. Normally this value
should be less than 1.0 which causes previous frames to fade out
over time. A value of 1.0 causes no fading, and values greater than
1.0 cause previous frames to become brighter over time.
- Prev Color: Default rgb: [1 1 1].
For each frame, the previous output is scaled by this
color before it is combined with the new input frame. This is
similar to Prev Brightness but affects the colors of the previous
frames instead of just the brightness.
- Prev Hue Shift: Default: 0, Range: -1 to 1.
Shifts the hue of the previous frames' colors, for each
- Combine New:
Popup menu, Default: Ave.
Selects the method for combining previous frames with the
- Ave: The current frame is averaged with the previous
output, smearing moving objects out over time. The output is scaled
by Fade and the input is scaled by 1.0-Fade for a weighted average,
so Fade must be less than 1.0 for this to work properly. Unlike the
other combine options, Ave should never affect the brightness of
stationary objects in the clip.
- Max: The colors of the current frame and previous frames
are combined with a maximum function. This makes the output frame
at least as bright as the current frame, and will make brighter
'trails' for example if you have bright objects moving on a dark
- Screen: The colors of the current frame and previous
frames are combined with a blend function. This can be used to
accumulate the colors of a moving clip. However, non-black regions
will become brighter with each frame.
- Add: The colors of the current frame and previous frames
are added. This can also be used to accumulate the colors of a
moving clip, with the non-black regions becoming brighter at each
- Over: The current frame is composited over the previous
frames using its Alpha channel. This uses pre-multiplied
compositing, so where the alpha is black the Source image should
normally also be black. If the input clip contains no Alpha
channel, the luminance is used instead.
- Under: The current frame is composited under the previous
- New Color: Default rgb: [1 1 1].
Scales the color of the current frame. Set this to the
complement of Old Color to offset overly colored trails.
- New Opacity: Default: 1, Range: 0 to 10.
Scales the opacity and brightness of the current frame.
- Diffuse Amount: Default: 10, Range: 0 or greater.
The previous frames are passed through a
pixel-diffusion process of this magnitude, for each new frame. This
has no effect unless it is positive.
- Diffuse Rel: X & Y, Default: [1 1], Range: 0 or greater.
The relative horizontal and vertical amounts
of the pixel diffusion process. This has no effect unless Diffuse
Amount is positive.
X & Y, Popup menu, Default: [Reflect Reflect].
Determines the method for accessing outside the borders of the
- No: gives black beyond the borders.
- Tile: repeats a copy of the image.
- Reflect: repeats a mirrored copy. Edges are often less
visible with this method.
Popup menu, Default: Normal.
Determines the method used for dealing with
- All Opaque: Use this option to render slightly faster when
the input image is fully opaque with no transparency (alpha=1).
- Normal: Process opacity normally.
- As Premult: Process as if the image is already in
premultiplied form (colors have been scaled by opacity). This option
also renders slightly faster than Normal mode, but the results will
also be in premultiplied form, which is sometimes less correct.
© 2007, GenArts, Inc. All rights reserved.