Skip to content

Commit

Permalink
Update README.md
Browse files Browse the repository at this point in the history
v 1.3 update : new section about best practices
  • Loading branch information
moraell committed Jul 28, 2019
1 parent 9d58b5c commit 5922748
Showing 1 changed file with 12 additions and 1 deletion.
13 changes: 12 additions & 1 deletion README.md
Original file line number Diff line number Diff line change
Expand Up @@ -11,6 +11,17 @@ It has been developped for Blender 2.79 and 2.80 on Windows 10.
### A few words about the project genesis
A few monthes ago, I browsed the web in order to find a way to setup my own homemade capture studio with a Kinect sensor, for a videogame I'm working on. I am not a 3D artist and a real noob with Blender, so I wanted something to help me with the animation process. I found a few solutions but they were either very expensive (and not working very well anyway, at least the demos I tried) or not fitting my needs (not working on Windows 10 or requiring a specific armature that just wouldn't do with my target model). So I decided to learn Python, dig into Blender documentation (and I think I broke a few shovels in the process ;)), and here is the result.

## How to optimize the capture
It is best to have a large room with a flat floor, and a good lighting but not toot bright.
The sensor must be horizontal and it is best to place it around waist height.
The tilt angle doesn't really matter as long as the device can sense the floor.
Don't get closer than 1.5 meters to the sensor, especially if you track position.
The actor's clothing shouldn't be too loose, so that each limb is clearly visible.
Use Kinect Studio to check that your workspace is optimized :
- The floor grid (green and grey tiles) must appear and be stable
- The actor's skeleton must not be too shaky
- If not, try and modify the lighting or the sensor position and tilt angle.

## Demo and configuration video
[https://youtu.be/Zt8gJzSNSbw]

Expand All @@ -31,4 +42,4 @@ There are two parts in the project :
The library was designed to be built using Visual Studio 2017 (only the Release configuration has been properly configured at the moment).

## Current progress
The project is currently in version 1.2.
The project is currently in version 1.3.

0 comments on commit 5922748

Please sign in to comment.