goto10 wrote:It's interesting though. As far as we know, the latest internal Resolume build is currently the only software that can actually do an A/B test of both codecs in the same software, and we found CPU usage was identical.
How have you tested the CPU usage? Maybe it's just that VDMX has a lower CPU overhead?
For everything except DXV it was measured in a custom app which programatically measured usage over 40 seconds for a range of different styles of 24 FPS 1080p clips. I then subtracted the "idle" usage of the app. For DXV I used Activity Monitor on its longest interval (5s) and played back the same range of clips in Avenue, averaging a series of readings for each clip. I then measured and subtracted the "idle" usage of Avenue (ie discounted as much of the overhead as possible).
We did a lot of testing of various secondary compressors during development and went with Snappy because it offers substantially lower CPU usage than anything else. An interesting test would be the number of simultaneous 1080p streams of DXV vs Hap you can achieve from the fastest SSD you can get your hands on.