Hi Topaz Labs,
Sometimes, we need to select different models to the same video content, for solving specified issues, scene by scene.
One solution is almost there.
You have a scene detection that splits a video content in different scenes for interpolation and avoiding to interpolate scene cuts!
Video Flex could have a scene detection button, so this feature would create regions in the timeline for showing each detected scene.
When this done, we could move the in and out point of scene if we want to adjust the length of each scenes by grabbing the region edge to extend or reduce it. The previous or next region will change appropriately.
Then in the timeline region ruler, we could have a pull-down button menu which show which Model the Video Flex function suggests! The model suggestion would be done while detecting the scenes.
If the scene is a landscape or urban type, for instance Proteus.
If the scene is a close-up on one of more characters, for instance Iris.
If the scene is dark with compression artefacts or continuous tones banding, for instance Dithering.
And so on, and so worth.
Then if there is a cross-fade between two regions, we could set manually a cross-fade duration to blend the rendering of one model to another one.
While rendering, the engine will cut and send each scene to the right engine, then concatenate the output, sharp cut or smooth cross-fade.
It could save a lot of time.
Best regards, Vincent.
