Aim
To specify how video is processed during tracking. These settings are applied to all new experiments.
Background information
In EthoVision XT 17.5 and later versions you can set playback of video to be done through the Graphics Processing Unit (GPU or dedicated graphics card) instead of the CPU. This allows greater performance, especially when tracking in demanding conditions like with Behavior Recognition. Up to EthoVision XT 17.0 it was only possible to playback video with the CPU.
Procedure
1.Choose File > Preferences > Video settings.
2.Next to Play back and record video with, Select either GPU (default) or CPU.
▪Choose GPU if you want to have the GPU to process video.
▪Choose CPU if you want to have the CPU to process video.
3.Click OK to apply the change.
Notes
▪If you are in a middle of an experiment started with EthoVision XT 17.0 or an earlier version and some tracks were acquired before upgrading to EthoVision XT 19, you have two options:
▪Complete the experiment with the old playback method. Keep the option CPU selected (recommended). The reason is that if you choose GPU a video file could look slightly brighter than it would do if you chose CPU. In that case you would need to make new detection settings for the next trials. To keep consistent detection settings across trials, choose CPU.
▪Acquire all trials with the new playback method. Clear the existing trials (i.e., delete the tracks without removing the videos), and re-acquire all trials, with the option set to GPU for consistency. See Redo a trial
▪For all your new experiments, we recommend to keep the option GPU selected.
▪What option you choose next to Play back and record video with does not affect tracking based on Deep Learning, even though this relies on the GPU. So whatever option you choose, you can use the Deep Learning body-point detection technique, provided that your PC meets specific requirements. See Deep learning: Requirements
See also