Sound delay on Fairphone 5

Hello!

I tried several games and they all got a Sound delay around 500 milliseconds.

Anyone else noticed that Problem and or knows a fix?

Kind regards! :slightly_smiling_face:

2 Likes

Delay in what output device? The onboard speakers or a BT headset?
(I’m asking because I have a delay in my BT headset as wel on FP3)

Onboard speakers and wired headphones

You dont have this issue with other phones? Because games are interactive, you compensate for the delay. With static media like music and videos you can. Bluetooth simply is slower than your screen.

I tried it parallel on my Huawei p30 pro and there is no delay

That would be odd, you cant bend the rules of physics. I wonder what kind of trick they implement. Bluetooth has a noticeable latency. Thats why gamer headsets use WiFi.

I dont use bluetooth.

1 Like

Ah okay! In that case it’s quite odd. I only have one game, Mario Run, no delay there. Can you reproduce it with that game as well?

Hi, I’m experiencing this issue too. Mario Run seems fine, but any other game I tried has a quite big sound delay.

Yep Mario runs good here too.

@Icey so… what specific (free) game has the problem? ;⁠-⁠)

I’m gonna say Kawaii Kitchen (since it’s one I worked on myself) and it’s quite noticeable since it has a lot of button smashing.

Also team fight tactics (tft) or magica.io or hunt Royale

So I’ve been tinkering al little bit trying to compare situations. I’m not any android expert or nothing alike, so bear with me because maybe what I found isn’t really relevant.

First, the difference between the Mario game and other games is that other games when they try to play a sound they try to get what android calls a “low-latency track” or “Fast track” (source). I don’t really know wich type of “track” uses the Mario game, it just seems like it isn’t a “low-latency” one (maybe is a “normal” one, maybe is an even faster if it exists).

Now, the thing is, that when I compare the system dumps of the audio of a game that is suffering from audio latency in the FP5 with my other device that doesn’t have audio latency, I find this:

Fairphone 5:

Thread throttle time (msecs): 8886
  AudioMixer tracks:
  Master mono: off
  Master balance: 0.000000 (0.000000)
  FastMixer thread 0xb40000735b001cb0 tid=2268  FastMixer command=MIX_WRITE writeSequence=10119237 framesWritten=728584992
            numTracks=5 writeErrors=0 underruns=87 overruns=7465
            sampleRate=48000 frameCount=144 measuredWarmup=4.09 ms, warmupCycles=10
            mixPeriod=3.00 ms latency=515.00 ms
  FastMixer Timestamp stats: n=5054011 disc=38 cold=0 nRdy=0 err=0 rate=1 jitterMs(ave=0.000226565 std=0.484077 min=-0.973333 max=0.957761) localSR(48000.5, 8.35823e-07) correc
tedJitterMs(ave=-1.35806e-05 std=0.0101412 min=-0.830885 max=0.866562)
  Simple moving statistics over last 98.3 seconds:
    wall clock time in ms per mix cycle:
      mean=3.00 min=1.67 max=4.35 stddev=0.52
    raw CPU load in us per mix cycle:
      mean=260 min=38 max=742 stddev=49
  Distribution of mix cycle times in ms for the tails (> ~3 stddev outliers):
    left tail: mean=2.28 min=1.67 max=2.33 stddev=0.16
    right tail: mean=3.92 min=3.83 max=4.35 stddev=0.11
  Fast tracks: sMaxFastTracks=8 activeMask=0xe3
  Index Active Full Partial Empty  Recent Ready    Written
      0    yes  376       0   293    full  1872  728486640
      1    yes  898       0   974    full  1152    9438480
      2     no  842       0   943   empty     0    5885424
      3     no  642       0   183   empty     0    5884704
      4     no  986       0   116   empty     0    5884704
      5    yes  732       0   388    full   288    3184416
      6    yes  621       0   369    full   288    3184416
      7    yes  393       0   411    full   288    3183840
  Stream volumes in dB: 0:-12, 1:-inf, 2:-inf, 3:-21, 4:-15, 5:-inf, 6:-inf, 7:-inf, 8:-inf, 9:0, 10:-19, 11:-21, 12:0, 13:0, 14:0
  Normal mixer raw underrun counters: partial=0 empty=293
  4 Tracks of which 4 are active
    Type     Id Active Client Session Port Id S  Flags   Format Chn mask  SRate ST Usg CT  G db  L dB  R dB  VS dB   Server FrmCnt  FrmRdy F Underruns  Flushed   Latency
    F5     1357    yes  13344   11265    1639 A  0x000 00000001 00000003  48000  3   1  0   -29  -8.5  -8.5     0  003099F0    288     288 f         0        0  520.00 t
    F7     1359    yes  13344   11281    1641 A  0x000 00000001 00000003  48000  3   1  0   -21     0     0     0  003097B0    288     288 f         0        0  520.00 t
    F6     1358    yes  13344   11273    1640 A  0x000 00000001 00000003  48000  3   1  0   -21     0     0     0  003099F0    288     288 f         0        0  520.00 t
    F1     1352    yes  13344   11217    1634 A  0x000 00000001 00000003  48000  3   1  0   -21     0     0     0  009007E0   1152    1152 f         0        0  538.00 t

Other device:

Thread throttle time (msecs): 33
  AudioMixer tracks:
  Master mono: off
  Master balance: 0.000000 (0.000000)
  FastMixer thread 0xe760ce40 tid=1414  FastMixer command=MIX_WRITE writeSequence=32058211 framesWritten=3077588160
            numTracks=5 writeErrors=0 underruns=1899 overruns=15202
            sampleRate=48000 frameCount=192 measuredWarmup=16.8 ms, warmupCycles=5
            mixPeriod=4.00 ms latency=21.00 ms
  FastMixer Timestamp stats: n=16023302 disc=1176 cold=4 nRdy=0 err=0 rate=1 jitterMs(ave=-0.000112635 std=0.14333 min=-3.74734 max=397.39) localSR(48003, 4.22684e-08) correcte
dJitterMs(ave=-8.31926e-05 std=0.00877857 min=-20.0409 max=397.348)
  Simple moving statistics over last 131.1 seconds:
    wall clock time in ms per mix cycle:
      mean=4.00 min=0.65 max=11.77 stddev=0.43
    raw CPU load in us per mix cycle:
      mean=703 min=81 max=4785 stddev=224
  Distribution of mix cycle times in ms for the tails (> ~3 stddev outliers):
    left tail: mean=1.05 min=0.65 max=1.43 stddev=0.21
    right tail: mean=7.49 min=6.15 max=11.77 stddev=1.12
  Fast tracks: sMaxFastTracks=8 activeMask=0xd9
  Index Active Full Partial Empty  Recent Ready    Written
      0    yes  612       0   163    full  1152 3076442688
      1     no    0       0     0    full     0          0
      2     no    0       0     0    full     0          0
      3    yes  732      16   302    full  1536   15216960
      4    yes  735      20   470    full   384    7549056
      5     no    0       0     0    full     0          0
      6    yes  383       3   114    full   384    7548480
      7    yes  652       0    55    full   384    7548672
  Stream volumes in dB: 0:0, 1:-inf, 2:-inf, 3:-27, 4:-13, 5:-inf, 6:0, 7:-inf, 8:-inf, 9:0, 10:-19, 11:0, 12:0
  Normal mixer raw underrun counters: partial=0 empty=163
  7 Tracks of which 4 are active
    Type     Id Active Client Session Port Id S  Flags   Format Chn mask  SRate ST Usg CT  G db  L dB  R dB  VS dB   Server FrmCnt  FrmRdy F Underruns  Flushed   Latency
    F5     1632     no   3072   11985    1659 I  0x004 00000001 00000001  48000  8   3  0  -inf   -10   -10     0  00000000    960       0 I         0        0       new
    F1       55     no   2688       9      22 I  0x004 00000001 00000001  48000  8   3  0  -inf   -18   -18     0  00000000    960       0 I         0        0       new
    F2       56     no   2688      17      23 I  0x004 00000001 00000001  48000  8   3  0  -inf   -18   -18     0  00000000    960       0 I         0        0       new
    F7     9850    yes   8221   71545   10190 A  0x000 00000001 00000003  48000  3   1  0   -27     0     0     0  00733200    384     384 f         0        0   29.00 t
    F4     9848    yes   8221   71529   10188 A  0x000 00000001 00000003  48000  3   1  0   -36  -8.5  -8.5     0  00733380    384     384 f         0        0   29.00 t
    F3     9825    yes   8221   71265   10163 A  0x000 00000001 00000003  48000  3   1  0   -27     0     0     0  00E83440   1536    1536 f       384        0   53.00 t
    F6     9849    yes   8221   71537   10189 A  0x000 00000001 00000003  48000  3   1  0   -27     0     0     0  00733140    384     384 f       192        0   29.00 t

As we can see, the FP5 has a latency of 515 ms in what is suposed to me the “low-latency track”, when the other device gets 21 ms.

Again, I can be totally in the wrong track myself, but it seemed very odd to see that difference. Maybe some kind of error in the android-platform configuration?

6 Likes

Nice find. I suggest you contact support with this detailed information, that’s the surest way to get a fix into one of the future updates.

1 Like

Okey, do you know what is the best way to contact them?

I opened a Ticket AT the official Support contact from fairphone parallel to this thread. I have sent Them a link to this thread in the Ticket with the hint to your, i hope, helpful discovery :slightly_smiling_face:

1 Like

Even if @Icey opened a ticket with a link to this topic I would advice you to open one as well, as they dont look really at linked topics unless you insist they have to :upside_down_face:

1 Like

Yeah i also thought about that. You can try of you want. When this Problem is solved, i have no bigger (i am a little shocked while typing this because Auto correction suggested Nigger instead of bigger O_o) concerns about keeping this really nice Piece of Phone

Talked yesterday with customer support. He wanted to discuss it with the concerning Departement and reply when there is something New. He also took a Look at this thread.

3 Likes