Hello all, I am testing the new simple echo_suppress port as I see it was included in the new release of pjsip. It is running on an arm processor, 200 Mhz. Compilation is made with the --disable-floating-point option. Options in pjsua demo file for audio are: --add-codec=pcmu --auto-conf --quality=6 --ptime=10 --clock-rate=8000 --ec-tail=512 --ec-opt=2 Audio latencies set to default (100 ms I assume). I have tested different ec_tail lengths but I have same result: echo is present. It seems not to work for me. I have seen that in echo_suppress.c file there are some defines. I assume values predefined are correct. Cpu usage goes to 15% approx (so ok, compared to my test with AEC --of course, no way with AEC--). snd-test shows next results: root at at91sam9263ek:/home$ ./sndtest-arm-angstrom-linux-gnueabi 13:30:27.571 sndtest.c Frame interval: min=0.019ms, max=16.112ms 13:30:27.576 sndtest.c Jitter: min=5.908ms, avg=13.472ms, max=16.089ms 13:30:27.578 sndtest.c Frame interval: min=0.023ms, max=16.120ms 13:30:27.580 sndtest.c Jitter: min=5.903ms, avg=13.459ms, max=16.084ms 13:30:27.583 sndtest.c Test completed with some warnings Any extra parameter for tunning at run-time or compile-time to get any good result with ES and my audio hardware? Any tip will be welcomed. Best regards, Victor Librado Sancho ------------------------------------------------------------------------
