Depth Data Enhancement
Note: There's a little bit of dithering introduced by compressing as a GIF.
Our depth enhancement service allows you to improve the quality and resolution of your depth map so that any holes are filled, edges are smoothed and the resolution matches your camera (1080p–4k+). We currently offer enhancement as a service – you provide us with your data, a matching white on black key and your data will be processed in five business days after confirming the receipt of your data.
- Processed footage must be shot against a green screen and properly keyed. The quality of the depth maps relies heavily on the quality of the key, so make sure the provided key is properly prepared.
- To be delivered to us is the color footage, the color key, and the depth frames, as well as the calibration you generated before capturing. Essentially, it's a normal DepthKit project folder, plus the keyed video.
Keyed Video Example:
You can see above an example of the what the keyed video should look like. What you want to be shown should be white, and everything else should be black. NOTE: The keyed video should also be the same resolution as your source video. For formats, H264/MP4 is fine. If you have any questions, reach out to us.
The cost is calculated based on how many clips you provide and how long the clips are (total). We highly encourage people to limit the amount of clips they send in to better cut down on processing costs and processing time. The best way to do this is to first cut your project using non-enhanced footage, then getting only your clip selects enhanced once you have picture lock. You can find a calculator below.
If you plan to deliver footage in single clips, there are a few constraints.
- The color videos must be stitched to be a single video.
- The keys must be stitched to be a single video.
- The depth frames must be in a folder with frame numbers that sequentially match the stitched color/key sequences.
- The calibrations must be the same for the individual takes inside of a stitched sequence.
- The bounding boxes of the individual takes in each stitched sequence need to be the same. The enhancement process works best when sequences have the same relative motion across the viewing plane, meaning that takes of stationary people should not be paired with those that exhibit more motion across the frame. Doing so will reduce the output resolution of the final depth map.
To better help you with your production, we'll process one five second clip of your choosing for free so you'll get a better sense of what is possible with your data.
Calculate your total data enhancement cost
Enhancement Price Calculator