HAP Performance

It is no secret that video playback can be very performance intensive. HAP codecs however, perform decompression using the machines' graphics hardware system, consequently reducing the CPU usage required to play video. This leads to higher resolution playback and higher framerate projects.

Encoding

Videos encoded with the HAP codecs are typically exported into standard .mov or .avi containers. As a media server, LightAct only decodes videos. The encoding part of videos is usually done in another software, previous to importing the video asset into LightAct. However, the encoding settings are still significant to video playback performance in LightAct.

Chunks

To achieve smoother video playback with HAP codec, we have to reach what is called multi-threaded CPU decoding. In the HAP community, this feature is called 'chunked' encoding. In general, the number of chunks should be equal to the CPU cores of the machine used for video playback. That is because each 'chunk' takes advantage of a different CPU core to be decoded instead of using only a single CPU core.

Good to know: Depending on the encoder, the 'Chunk' feature may or may not be included in the encoding settings. If your encoder, with the HAP codec chosen, does not include a 'chunk' section, you can try using FFmpeg. For encoding to HAP from Adobe Media Encoder, you can try the AfterCodecs plugin.

Server Performance

In our testing environment, we used a 30sec section (00:00:10:000-00:00:40:000) from the famous Sam Swordsmith videos, encoded with Hap, .mov container.

reActor rt

Resolution	Chunks	FPS	MB/s	Max Videos
HD	32	30	21.4	54
UHD	32	30	74.5	16

reActor xr

Resolution	Chunks	FPS	MB/s	Max Videos
HD	36	30	21.4	58
UHD	36	30	74.7	18