this was originally graphed for the above linked app, but I know that katanga doesnt inform the user of gpu spiikes, scheduling, and dropped frames, and that makes the user think katanga is messin up. When its not katanga;s fault.
a conversation recently:
"The issue is on the game's side rather than steamvr/katanga/vrscreencap side, if for some reason a game has a spike in load on the gpu even in-between two frames it could lead to katanga/vrscreencap waking too late and dropping a frame. This happens a lot if the game doesn't use vsync but it can also happen in demanding games, reducing the resolution on katanga's side won't really solve the issue.
You could try and use an heuristic to detect if there's enough headroom on average, just to warn people running without vsync/running a game that is in general too demanding for their system. That could probably help in most cases (I think a big feature of katanga is that it also tries to set a frame limiter for the titles launched through helixvision?)
"
"average and peak load, it seems the user (understandably) blames the display software and doesnt look at gpu peaks or scale of resolution,
and the negative reviews on steam for heelixvision I bet are because people are dumb and if we inform the user how to soolve the problem, they wont blame the software and fix the issue themeselve (edited)
[10:49 AM]
i can loop through the log ffile. i think right now, you have the log file just pile up over time, I may store logs per session on my end, then loop through the file, detect time (divided by) number of dropped frames and inform the issue during or after game session
[10:50 AM]
even a month ago, I chose to use [vr_screen_cap] because i thought katanga had problems but I didnt have gpu overhead, there was no way for me to know untill I went into the forums
