Jetpack framerate dependency tests

atman
878
Posts
57
Location
Luleå, Denmark
Joined
7 March 2006
11 April 2018 - 17:32 CEST
#1
Date: 2018-APR-11
NS version: 3.2
HL version: Latest patch 2017-AUG-01
Listen server, sv_cheats 1

Method:
100 and 500 fps tested
Standing in marine start on ns_tanith. Holding +jump until jetpack fuel bar fully discharged then letting go.
Measured time for complete discharge and recharge.
Also measured time from floor to ceiling to verify vertical acceleration dependence on fps.

Recorded video @ 60 fps, timestamps examined in media player classic.

Note: Custom, accurate, jetpack fuel sprite was used (at least back when i made this one, original jp fuel sprite had a "hidden" part and wasn't 100% accurate).

Sample size: 1 run per test case

Results:
100 fps
Discharge time: 6.021 s
Recharge time: 9.277 s
Floor-to-ceiling time: 0.654 s

500 fps
Discharge time: 6.03 s
Recharge time: 9.216 s
Floor-to-ceiling time: 0.646 s

Conclusion: No dependency on frame rate for any of the tested cases.

Any differences in time measured likely attributed to human error and accuracy of method (16.66 ms per frame of video).

Video: https://www.youtube.com/watch?v=mA9lK-LVNG0
atman
878
Posts
57
Location
Luleå, Denmark
Joined
7 March 2006
11 April 2018 - 17:37 CEST
#2
[3:59 PM] relakks: regarding what @king_yo said earlier about high fps jps flying longer distance, i think that given a straight line flight, 100 and 500 fps jps are almost equal. however a 500 fps jetpack accelerates so much faster that it spends more time at top speed than a 100 fps jetpack
[4:01 PM] relakks: did some quick and dirty testing flying back and forth in a vent, 100 fps jetpack made 4 times the distance of the vent, while 500 fps jetpack made 5 times the distance
[4:01 PM] relakks: in one full fuel bar
[4:04 PM] relakks: so while they are more or less equal for a long distance without stops, short distances with many stops/accelerations heavily favour 500 fps jetpack. i.e. real game situation
atman
878
Posts
57
Location
Luleå, Denmark
Joined
7 March 2006
11 April 2018 - 21:28 CEST
#3
Date: 2018-APR-11
NS version: 3.2
HL version: Latest patch 2017-AUG-01
Listen server, sv_cheats 1
cl_showspeed 1

Method:
Flying forward and backward in a vent on ns_tanith.
Measured time from full speed -> 0 speed -> back to full speed.
Tested fps: 500 (discarded bc i wasnt keeping stable frame rate), 400, 300, 200, 100, 50

Definitions:
Full speed = 576 units/s
Full speed to 0 speed = deceleration time
0 speed to full speed = acceleration time
deceleration + acceleration = total time for a complete direction change

Recorded video @ 60 fps, timestamps examined in media player classic.

Sample size: 4 runs per test case, values averaged in results section.

Results:
fps 400 300 200 100 50
decel time 0.072 0.090 0.111 0.142 0.175
accel time 0.124 0.155 0.237 0.462 0.922
total time 0.196 0.245 0.348 0.605 1.096

Graph:


Conclusion: Jetpack horizontal acceleration is clearly dependent on FPS. Deceleration is less affected but still measurable.
A 400 fps jetpacker will switch direction and get back to full speed 3 times faster than a 100 fps jetpacker.

Video: https://www.youtube.com/watch?v=Y0HEP492KvE
Note: 500 fps test data discarded as i couldn't keep a stable frame rate.

Minor errors in data attributed to human error (manually switching direction, not using scripts) and accuracy of method (16.66 ms per frame of video).
New Reply