<html xmlns:o="urn:schemas-microsoft-com:office:office" xmlns:w="urn:schemas-microsoft-com:office:word" xmlns:m="http://schemas.microsoft.com/office/2004/12/omml" xmlns="http://www.w3.org/TR/REC-html40"> <head> <meta http-equiv="Content-Type" content="text/html; charset=Windows-1252"> <meta name="Generator" content="Microsoft Word 15 (filtered medium)"> <style></style> </head> <body lang="EN-GB" link="blue" vlink="#954F72"> <div class="WordSection1"> I just tried a totally different sound input 2x I2S mics on the Pi3 Gpio (Pi3 this time) Did the tsched thing and without but no luck. load-module module-udev-detect tsched=0 If you really want AEC on a pi via software just for now best I have been able to achieve is detailed here. <a href="https://community.rhasspy.org/t/2x-1-80-i2s-mems-mics-pi-3-5mm-ec-but-buy-adafruit-if-you-can/1015">https://community.rhasspy.org/t/2x-1-80-i2s-mems-mics-pi-3-5mm-ec-but-buy-adafruit-if-you-can/1015</a> Its Alsa has to have a single sound card clock. Even though listed as x2 soundcards its the same I2s clock on the Pi. USB soundcards are the same as one clock so synced with playback/capture Respeaker work but hate the drivers. If you have USB say PS3eye and Pi 3.5mm the speexdsp can not handle the unsynced clock drift and basically doesn’t work. But right sound source and the above so far is best method on the Pi, dunno why or what is happening with WebRTC but the results are bad. Stuart <o:p> </o:p> Sent from <a href="https://go.microsoft.com/fwlink/?LinkId=550986"> Mail</a> for Windows 10 <o:p> </o:p> <div style="mso-element:para-border-div;border:none;border-top:solid #E1E1E1 1.0pt;padding:3.0pt 0cm 0cm 0cm"> From: <a href="mailto:pulseaudio-discuss-request@lists.freedesktop.org">pulseaudio-discuss-request@lists.freedesktop.org</a> Sent: 23 May 2020 13:00 To: <a href="mailto:pulseaudio-discuss@lists.freedesktop.org">pulseaudio-discuss@lists.freedesktop.org</a> Subject: pulseaudio-discuss Digest, Vol 109, Issue 16 </div> <o:p> </o:p> Send pulseaudio-discuss mailing list submissions to pulseaudio-discuss@lists.freedesktop.org To subscribe or unsubscribe via the World Wide Web, visit <a href="https://lists.freedesktop.org/mailman/listinfo/pulseaudio-discuss"> https://lists.freedesktop.org/mailman/listinfo/pulseaudio-discuss</a> or, via email, send a message with subject or body 'help' to pulseaudio-discuss-request@lists.freedesktop.org You can reach the person managing the list at pulseaudio-discuss-owner@lists.freedesktop.org When replying, please edit your Subject line so it is more specific than "Re: Contents of pulseaudio-discuss digest..." Today's Topics: 1. Re: pa echo-cancel module and alsa (Monkey Bizness) ---------------------------------------------------------------------- Message: 1 Date: Fri, 22 May 2020 20:06:19 +0200 From: Monkey Bizness <monkey@infra-monkey.com> To: "pulseaudio-discuss@lists.freedesktop.org" <pulseaudio-discuss@lists.freedesktop.org> Subject: Re: [pulseaudio-discuss] pa echo-cancel module and alsa Message-ID: <2d4b5114769a3e27b327c945eff56abb6b1c7c10.camel@infra-monkey.com> Content-Type: text/plain; charset="utf-8" Hi Stuart, I've tried to specify sink_master=alsa_output.default which is the sink create by the alsa-sink module but wierdly my echoCanel_sink is slaved to the alsa_input of the usb webcam.... I'll probably do more testing and file a bug report is needed. Anyhow, I could work around the race for default sink by disabling the manual module-alsa-sink and adding the "tsched=off" option to load- module module-udev-detect tsched=offI just hope it doesn't mess with the usb mic. The capture was working well. If it is messed up i'll have to go back to setting the sinks manually or find a way to deal with it through autodetect module. Now the autodetected sinks use this option. The echo-cancel module uses the autodetected sink and source and default masters. This means i don't have to modify the alsa configuration.The set-default-sink and set-default-source are now correctly used. I can see some drops sink and drop source in the pa logs.I'll investigate into that and there must be some tweaking to do. webrtc seems to have some interesting additional features i want to try like noise suppression and mobile loud-speakerphone profile i want to try out.I couldn't find any docs on the speex module so i didn't really look into it So far so good. PS i just started with all the pulseaudio and aec so i'll do as much testing and reading. I'll come back with more results soon. Thanks for the advice. Monkey On Fri, 2020-05-22 at 13:28 +0000, Stuart Naylor wrote: > You are supposed to be able to define the master source & sink > > > > > My usual is > > > > >>>load-module module-echo-cancel use_master_format=1 > aec_method=webrtc aec_args=?analog_gain_control=0 > digital_gain_control=1 agc_start_volume=85? > source_name=echoCancel_source sink_name=echoCancel_sink > >>> > >>>set-default-source echoCancel_source > >>>set-default-sink echoCancel_sink > > > use_master_format=1 just seemed easier than the problems I had in > attaining a working defined sink/source > > I have forgot where I found the details for master_sink & > master_source parameters maybe its in the code but going to have a > good look again at PA webrtc. > > > > I also got into some problems as wanted to have asound.conf that > would use pulse. > > > > Using pulseaudio_alsa and > > > > >>># Use PulseAudio by default > >>>pcm.!default { > >>>type pulse > >>>fallback "sysdefault" > >>>hint { > >>> show on > >>> description "Default ALSA Output (currently PulseAudio Sound > Server)" > >>>} > >>>} > >>> > >>>ctl.!default { > >>> type pulse > >>> fallback "sysdefault" > >>>} > > > > But seemed to get stuck in this tail chasing of default sound on > reboots. > > > > As said this weekend going to have another good look at PA Webrtc AEC > on the Pi but I found when using the Pi output and usb mic the clock > drift was quite excessive. > > I am using soundcards with a singular clock and card for sink/source > and getting similar results which has me head scratching. > > You can wire up 2x I2s mics to a Pi quite easily and that would share > the same clock as the Pi output (doesn?t work pi4 sadly) > > > > I will look forward to how you get on as my SpeexEc results have been > quite good on the Pi and expected PA WebRTC AEC to be better but > actually seems worse. > > PS both of us are setting autogain and from speex that isn?t really a > good idea as it makes thing much harder for the adaptive FFT routine. > > AGC needs really long attack/delay values in fact with voice in > manner of speech its hard to use AGC and not just amplify noise. > > Really the gain shouldn?t be there are its problematic anyway. > > > > Stuart > > > > > Sent from > Mail for Windows 10 > > > > -------------- next part -------------- An HTML attachment was scrubbed... URL: <<a href="https://lists.freedesktop.org/archives/pulseaudio-discuss/attachments/20200522/5cf73dcc/attachment-0001.htm">https://lists.freedesktop.org/archives/pulseaudio-discuss/attachments/20200522/5cf73dcc/attachment-0001.htm</a>> ------------------------------ Subject: Digest Footer _______________________________________________ pulseaudio-discuss mailing list pulseaudio-discuss@lists.freedesktop.org <a href="https://lists.freedesktop.org/mailman/listinfo/pulseaudio-discuss">https://lists.freedesktop.org/mailman/listinfo/pulseaudio-discuss</a> ------------------------------ End of pulseaudio-discuss Digest, Vol 109, Issue 16 ***************************************************<o:p></o:p> <o:p> </o:p> </div> </body> </html>