Kadano's SSBM Audio+Video input lag test measurements
 Share
The version of the browser you are using is no longer supported. Please upgrade to a supported browser.Dismiss

 
Comment only
 
 
ABCDEFGHIJKLMNOPQRSTUVWXYZAAABACADAEAFAGAHAIAJAKALAMANAOAPAQARASATAUAVAWAXAY
1
↓Sample number→Test series, vsync
Console VGA CRT monitor 60hzDirect3D11 3xIR HLE XAudio2Direct3D11 3xIR HLE DSoundDirect3D11 3xIR HLE OpenALDirect3D11 3xIR LLE XAudio2Direct3D11 3xIR HLE XAudio2Direct3D11 3xIR HLE XAudio2Direct3D11 3xIR HLE XAudio2Direct3D11 3xIR HLE XAudio2Direct3D12 3xIR LLE XAudio2Direct3D11 3xIR LLE XAudio2Direct3D12 3xIR HLE XAudio2Direct3D11 TV40 HLE XAudio2Direct3D11 TV1 HLE XAudio2Direct3D11 TV4 HLE XAudio2Direct3D11 2xIR HLE XAudio2Direct3D11 3xIR HLE XAudio2Direct3D11 3xIR HLE XAudio2Direct3D11 3xIR HLE XAudio2
2.5x,int.Audio,W10,FM5.66
2.5x,int.Audio,W10,FM5.66
2x,int.Audio,W10,FM5.66
2x,int.Audio,W10,FM5.66
2x,int.Audio,W10,FM5.66
2x,int.Audio,W10,FM5.66
2
Monitor →N/A (direct video to oscilloscope)120hz CRT (video to oscilloscope)120hz CRT (video to oscilloscope)120hz CRT (video to oscilloscope)120hz CRT (video to oscilloscope)120hz CRT (video to oscilloscope)120hz CRT (video to oscilloscope)120hz CRT (video to oscilloscope)120hz CRT (video to oscilloscope)120hz CRT (video to oscilloscope)120hz CRT (video to oscilloscope)120hz CRT (video to oscilloscope)180hz (dual photodiode)180hz (dual photodiode)180hz (dual photodiode)120hz CRT (video to oscilloscope)120hz CRT (video to oscilloscope)120hz CRT (video to oscilloscope)120hz CRT (video to oscilloscope)120hz CRT, Dolphin-vsync120hz CRT, Dolphin-no,NV=adaptive120hz CRT, Dolphin-no,NV=adaptive
120hz CRT, Dolphin-no,NV=adaptive
120hz CRT, Nvidia-vsync-on120hz CRT, vsync disabled
3
GPU, game, Dolphin version, buffer
Wii-PAL, SSBM PALGTX960,FM4.9,0Via integratedGTX960,FM4.9,0
Via integrated
GTX960,FM4.9,0Via integratedGTX960,FM4.9,0Via integratedGTX960,FM4.9,0Via integrated
GTX960,926 (2b2cf0369)
Via integratedGTX960,FM4.9,0Xonar DGGTX960,D5.0,0Xonar DGGTX960,D5.0,0Xonar DG
GTX960,D5-3566,0
Xonar DG
GTX960,D5-3566,0
Xonar DG
GTX960,FM4.9,0
Via integrated
GTX960,FM4.9,0
Via integrated
GTX960,FM4.9,0
Via integratedG4560,750Ti,FM4.4,0Via integratedGTX960,FM5.5,0Xonar DGGTX960,FM5.5,8Xonar DGGTX960,FM5.5,9Xonar DGC2Q-Q8200, 9600GT
C2Q-Q8200, 9600GT
C2Q-Q8200, 9600GT
C2Q-Q8200, 9600GT
C2Q-Q8200, 9600GT
C2Q-Q8200, 9600GT
4
VideoAudioVideoAudioVideoAudioVideoAudioVideoAudioVideoAudioVideoAudioVideoAudioVideoAudioVideoAudioVideoAudioVideoAudioVideoAudioVideoAudioVideoAudioVideoAudioVideoAudioVideoAudioVideoAudioVideoAudioVideoAudioVideoAudioVideoAudioVideoAudioVideoAudio
5
171.638.226.411024.2159.227.5221.529.511622.510363105.519.51464615567.516650.5157.553.5149.535115601035910837111281146715070161418240832979297931753279
6
358.42930.6115.827.2163.633240421183211071118.521143.5471675515448.515476.5160.52610136754186321165715367158337627743380338035762466
7
465.632.632.611325.8161.224.523240.5117.5361165710127.5150.552.51724914755.516960.5154.53010829753981241126215577165377733782371237145853075
8
579.246.626.6105.426.6162.83624426.5105.527105.553.510017.5138.551.5173.563.5158.547.516959.5153.53310732783274341186615567157459131773173317345833581
9
665.632.635112.627.2165.430.523835.5116.523.5104.550.593.514131.551.5168.574.5157.530.5152.583165.52910641853676281186315275166317827743174317445862468
10
766.43628.2107.825.4160.831.5240.535.511425.5102.577119.530.5151.552.517166.5155.549166.565.5157.53110726723879311166515271169418035753881388134722566
11
86736.427.8108.226163.822.52333311530.510453.596.52914652169.579.517045161.553.51522910741903071361276114669153377138844080408042824284
12
97544.234.6112.828168.630.524139.51223410764.510632.514848166.551.514849.516757.5143.54012045903476311246815261153418338843372337236764081
13
1057.8€_&_(27.833.8114.637.2177.829237.537.5118.521.510367.5112.53415352.517175.5164.54616761.5155.54012039813477291226315862154468937823282328247873384
14
1170.437.833114.235.4171.822.52304212121.5103.564.5106.525.514430.5147.559.5161.539.516155.5143.53511447894784311245414558147448638783985398540803282
15
1255.823.230.6112.219.5144.545.5163.563.51593684368443813583
16
13724334.211828.51565817277.5163.53578357840873482
17
14582725.4108.227.5152.538.515066159.53581358133782975
18
1571.84023.610828.5155.545.516352.51534082408237822271
19
16643326.2108.222.51494916577.5164.54183418341794288
20
1765.636.229.4112.630.5157.539.515450.51453073307337802977
21
1866.837.416.2101.82014359.516880.51662872287233772671
22
1961.428.426.4111.626.5151.538.515656.51524386438638832875
23
207543.639115.433.515940.515473.51582067206743843882
24
2176.646.227.4107.236161.556.5166.563.5152.53070307042822873
25
226733.833110.617142.542.5153.552147.52975297537814185
26
2368.837.825.4100.820146.549.5162.5731573584358441804086
27
246026.629.2105.822.514947.515855.5149.53583358341844080
28
2577.849.228.6107.234.5161.555.5163651653278327838802473
29
2674.844.238.8117.230.5158.552.5164.576.5162
30
2759.228.826.210726.515160.5168.565159
31
2882.25234.4114.218.514247.515855154
32
2956.826.629.8108.2161405316980.5168
33
307546.629108.232.516039.515149146.5
34
3159.230.828.2109.622144.538151.554.5140
35
3255.223.835.8119.625.515440.515063160
36
Maximum lag in ms86.857.639119.637.2177.836244421223611677119.536161.560.5173.580.517055.516983165.540120601035910837111361276815877169469140844386438647874288
37
Minimum lag in ms55.223.216.2100.824.2159.222.5221.526.5105.521.5102.550.593.514131.530.5147.54914030.5152.553.5143.5261012672307137111241125414558147317127742067206731722266
38
Average lag in ms67.3153846236.4582278529.8516129110.51612928.3165.528.75235.7536.15116.427.4105.962.2105.9525.46774194149.419354847.79032258162.032258164.27419355156.903225846.15162.562.65153.5532.8110.539.683.83981.23711130.4119.162.6151.867.7158.339.681.334.478.933.2578.0416666733.2578.0416666739.3333333380.8333333332.2083333377.79166667
39
40
NTSCNTSCNTSCNTSC 1.02NTSC 1.02NTSC 1.02NTSC 1.02NTSC 1.02NTSC 1.02
41
TearingTearing not accounted forTearing not accounted forTearing not accounted forTearing not accounted forTearing not accounted forTearing not accounted forTearing not accounted forTearing accounted forTearing not accounted for
Tearing accounted for (double photodiode)
Tearing accounted for (double photodiode)
Tearing accounted for (double photodiode)
Vsync enabled, no tearingVsync=adaptive often teared, tearing accounted forVsync=adaptive often teared, tearing accounted forVsync=adaptive often teared, tearing accounted forVsync enabled (set to ON)Vsync disabled
42
(Console has no tearing)
43
44
Date2017-06-262017-08-182017-08-182017-08-262017-08-262017-08-262017-08-26
45
NotesNvidia Control Panel settings: Scaling=display, vsync=off, Win10, i5-750, single monitorXAudio2: Latency=0, Via integrated audio (Asus P55D-Evo) with Windows 10 integrated driversNvidia Control Panel settings: Scaling=display, vsync=off, Win10, i5-750, single monitorDSound: Latency=0, Via integrated audio (Asus P55D-Evo) with Windows 10 integrated driversNvidia Control Panel settings: Scaling=display, vsync=off, Win10, i5-750, single monitorOpenAL: Latency=0, Via integrated audio (Asus P55D-Evo) with Windows 10 integrated driversNvidia Control Panel settings: Scaling=display, vsync=off, Win10, i5-750, single monitorXAudio2: Latency=0, Via integrated audio (Asus P55D-Evo) with Windows 10 integrated driversNvidia Control Panel settings: Scaling=display, vsync=off, Win10, i5-750, single monitorXAudio2: Latency=0, Via integrated audio (Asus P55D-Evo) with v11.11 official drivers. There is a small improvement from installing the drivers from Via over using built-in, however with only 10ms difference it could just be random variation from small sample size. (105ms audio lag is still way too much to be even close to acceptable, so I didn't bother taking more samples. It doesn't matter whether audio lag is 105 or 115 ms, so there is no point in taking more samples to find out. I'll keep doing 10 samples until audio lag on Dolphin / FM is < 50ms.)Nvidia Control Panel settings: Scaling=display, vsync=off, Win10, i5-750, single monitor; Faster Melee video speed up settings / codes / hacks disabled in this build, left as supplied by Tino (dev)XAudio2: Latency=0, Via integrated audio (Asus P55D-Evo) with v11.11 official drivers. No noticeable improvement from installing the drivers from Via. The new alpha Ishiiruka seems to improve audio lag slightly at best, with 9 ms less minimal lag (might just be from small sample size with this series, though). In any case, 106ms average audio lag is still way too much.Nvidia Control Panel settings: Scaling=display, vsync=off, Win10, i7-2600K, multiple monitors (exclusive fullscreen on non-primary); D3D 11 in FM4.9 has a bugged aspect ratio on CRT monitors used as secondary monitors to primary LCD monitorsXAudio2: Latency=0, Asus Xonar DG with UniXonar drivers.All driver audio features disabled, sound set to "Hi-Fi" (seems to force-disable audio "enhancements"). Much more audio lag than integrated Via audio.TV = Timing Variance setting in /user/config/dolphin.ini30ms less audio lag than FM4.98 buffer: 4.6ms less video input lag than console+CRT9 buffer: 0.5ms more video input lag than console+CRT. Compared to 8 buffer, we have 5.1ms input lag increase, which is slightly more than the theoretical 4.167 ms (due to small sample size probably). Still, it's accurate enough to see that even with FM5.5, the old recommendation still holds true: 9 buffers for 120hz CRT, 8 buffers for 120hz LCD, 7 buffers for 60hz CRT and 6 buffers for 60hz LCD.Audio input lag is still massive. The Xonar DG is one of the slower cards, surprisingly, so with integrated audio or custom minimal audio buffers you should get up to 30-40ms less. Even then, you'll have 60-70ms sync delay over console, so I'd say the lag is still big enough that playing with completely mute audio is appropriate.CPU load: ~30-78%, GPU 40-60%Enabling "triple buffer" in Nvidia CP globally did not increase the input lag, but also did not remove tearing.
46
Testing procedure: I compared the button input that initiated an in-game full black to full white scene change (can be accomplished within develop mode) with the green line of the RGBHV video output and two photodiodes, one mounted at 25% from the top of the screen (120 Hz CRT monitor), and one mounted at 25% from the bottom. When only using one photodiode and no green video signal, tearing leads to sometimes getting measurements that are lower than justified (if the frame tears just a few lines above the lower photodiode's position, the player does not get to perceive this frame's relevant in-game cues in time). So whenever tearing happened between the two photodiodes, I kept counting until the next 120 Hz frame's lower photodiode activation. If a CRT was used, it's always a lagless computer monitor (Sony GDM-F520 [137khz] usually, sometimes Sun GDM-5510 [130khz]).Idle power draw: 80W
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
↓Additional samples (shifted down
81
to keep the chart readable)
82
3386.857.6
83
347745.4
84
3566.838
85
3665.233.4
86
3766.433.8
87
386737.6
88
3969.437
89
4065.432.6
90
4167.637.8
91
4259.428.6
92
4359.629.4
93
4465.234.2
94
4564.435
95
4673.244.6
96
4767.839
97
4867.436.8
98
4966.837
99
5071.239
100
5161.632.2
Loading...
Main menu