<html> <head> <meta content="text/html; charset=utf-8" http-equiv="Content-Type"> </head> <body bgcolor="#FFFFFF" text="#000000"> <div class="moz-cite-prefix">On 25.08.2015 13:58, Eytan Naim wrote: </div> <blockquote cite="mid:1a68760681c44202a08b293394d8d313@IL-EXCH02.marvell.com" type="cite"> <meta http-equiv="Content-Type" content="text/html; charset=utf-8"> <meta name="Generator" content="Microsoft Word 14 (filtered medium)"> <style></style> <div class="WordSection1"> Hi,<o:p></o:p> <o:p> </o:p> I’m trying to enable HFP on pulseaudio 6.0 using oFono and bluez 5 on a new embedded platform which doesn’t support alsa yet.<o:p></o:p> In order to confirm that audio is ‘played’ I record it using parecord tool. – This method worked just fine when enabled a2dp a short time ago.<o:p></o:p> <o:p> </o:p> After executing oFono & pulseaudio, a call is answered on the remote device- It is seemed that the relevant profile is chosen on pulseaudio, and parecord fails to record. </div> </blockquote> What is the source you are recording from? <blockquote cite="mid:1a68760681c44202a08b293394d8d313@IL-EXCH02.marvell.com" type="cite"> <div class="WordSection1"> <o:p></o:p> <o:p> </o:p> When executing ‘pacmd stat’ it seems that <o:p></o:p> Memory blocks allocated during the whole lifetime: 192, size: 699.9 KiB.<o:p></o:p> Is constant while pulseaudio prints to terminal the following message along the call with relatively fixed intervals:<o:p></o:p> <o:p> </o:p> ( 482.106| 0.013) D: [pulseaudio][modules/module-loopback.c:201 adjust_rates()] Should buffer 96 bytes, buffered at minimum 128 bytes<o:p></o:p> ( 482.118| 0.012) D: [pulseaudio][modules/module-loopback.c:226 adjust_rates()] [bluez_sink.98_D6_F7_34_98_E8] Updated sampling rate to 8000 Hz.<o:p></o:p> ( 482.132| 0.013) D: [pulseaudio][modules/module-loopback.c:197 adjust_rates()] Loopback overall latency is 610.13 ms + 0.00 ms + 25.00 ms = 635.13 ms<o:p></o:p> ( 482.146| 0.014) D: [pulseaudio][modules/module-loopback.c:201 adjust_rates()] Should buffer 96 bytes, buffered at minimum 0 bytes<o:p></o:p> ( 482.159| 0.012) D: [pulseaudio][modules/module-loopback.c:226 adjust_rates()] [bluez_sink.98_D6_F7_34_98_E8] Updated sampling rate to 8000 Hz.<o:p></o:p> ( 492.134| 9.975) D: [pulseaudio][modules/module-loopback.c:197 adjust_rates()] Loopback overall latency is 610.13 ms + 8.00 ms + 0.00 ms = 618.13 ms<o:p></o:p> ( 492.148| 0.013) D: [pulseaudio][modules/module-loopback.c:201 adjust_rates()] Should buffer 96 bytes, buffered at minimum 128 bytes<o:p></o:p> ( 492.161| 0.012) D: [pulseaudio][modules/module-loopback.c:226 adjust_rates()] [bluez_sink.98_D6_F7_34_98_E8] Updated sampling rate to 8000 Hz.<o:p></o:p> ( 492.175| 0.013) D: [pulseaudio][modules/module-loopback.c:197 adjust_rates()] Loopback overall latency is 610.13 ms + 0.00 ms + 25.00 ms = 635.13 ms<o:p></o:p> ( 492.189| 0.014) D: [pulseaudio][modules/module-loopback.c:201 adjust_rates()] Should buffer 96 bytes, buffered at minimum 0 bytes<o:p></o:p> ( 492.201| 0.012) D: [pulseaudio][modules/module-loopback.c:226 adjust_rates()] [bluez_sink.98_D6_F7_34_98_E8] Updated sampling rate to 8000 Hz.<o:p></o:p> ( 502.176| 9.975) D: [pulseaudio][modules/module-loopback.c:197 adjust_rates()] Loopback overall latency is 610.13 ms + 8.00 ms + 0.00 ms = 618.13 ms<o:p></o:p> ( 502.190| 0.013) D: [pulseaudio][modules/module-loopback.c:201 adjust_rates()] Should buffer 96 bytes, buffered at minimum 128 bytes<o:p></o:p> ( 502.202| 0.012) D: [pulseaudio][modules/module-loopback.c:226 adjust_rates()] [bluez_sink.98_D6_F7_34_98_E8] Updated sampling rate to 8000 Hz.<o:p></o:p> ( 502.216| 0.013) D: [pulseaudio][modules/module-loopback.c:197 adjust_rates()] Loopback overall latency is 610.13 ms + 0.00 ms + 25.00 ms = 635.13 ms<o:p></o:p> <o:p> </o:p> </div> </blockquote> This looks fine, the messages are from two module-loopback which are loaded to play back the audio coming from the phone and to transmit audio from your computer to the phone. But this is also where it is getting strange because you have no other sound device. The only sink available to play back audio to is the phone and the only source to record from (except the phone source) is the monitor source of the phone sink (See below for explanation of the monitor source). To avoid that situation you should add a null sink and a null source to your system using module-null-sink and module-null-source and make them the default sink and source. Then audio coming from the phone will be played back to the null sink (and you can record from null.monitor) while silence coming from the null-source is sent to the phone. <blockquote cite="mid:1a68760681c44202a08b293394d8d313@IL-EXCH02.marvell.com" type="cite"> <div class="WordSection1"> Please note:<o:p></o:p> I’m now familer with oFono at all – I simply run it as ofonod -d<o:p></o:p> Pacmd list output is attached while call is still ongoing. – there are 2 sources loaded and one sink.<o:p></o:p> <o:p></o:p> Also, could someone explain why are there two source available? Nexus 4 and Nexus 4 monitor? </div> </blockquote> The monitor source is created for each sink so that you have a source from where you can record what is played back to the sink. <blockquote cite="mid:1a68760681c44202a08b293394d8d313@IL-EXCH02.marvell.com" type="cite"> <div class="WordSection1"> <o:p></o:p> How can I confirm that the audio arrives at pulseaudio? By the log I would assume that the answer is positive, but according to ‘pacmd stat’ the value it seems not. </div> </blockquote> In general it looks like your setup is working. <blockquote cite="mid:1a68760681c44202a08b293394d8d313@IL-EXCH02.marvell.com" type="cite"> <div class="WordSection1"> <o:p></o:p> <o:p> </o:p> Thanks for your help,<o:p></o:p> Eytan.<o:p></o:p> </div> </blockquote> </body> </html>