speechd-discuss
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Speech-dispatcher dropping audio after switch to optical out


From: Nolan Darilek
Subject: Re: Speech-dispatcher dropping audio after switch to optical out
Date: Thu, 28 May 2020 08:44:07 -0500
User-agent: Mozilla/5.0 (X11; Linux x86_64; rv:68.0) Gecko/20100101 Thunderbird/68.8.1

I seem to have fixed this. Added the following systemd unit to ~/.config/systemd/user/continuous-noise.service, enabled and started:


[Unit]
Description=Continuous noise

[Service]
ExecStart=/usr/bin/play -qn synth whitenoise vol 0.001

[Install]
WantedBy=default.target


Arrived at 0.001 after tweaking input and output levels such that I could lower it and make the white noise less audible. Initially I needed 0.01, but that introduced significant artifacts, and anything less still dropped speech. I increased my computer volume until 0.001 kept the connection active, then dropped the soundbar volume so I still had a bit of headroom and can get a reasonable volume range if I want it. I also tried brown noise, but that introduced artifacts into the sub channel that were even more audible. I still get a bit of white noise, but I'll just pretend I'm living in the 80s and deal. :)


Still open to better solutions if any exist, but for now this works. Thanks for reminding me that HDMI equipment does stuff like this.

On 5/27/20 3:22 PM, Nolan Darilek wrote:

Hmm, interesting idea, but I haven't made any headway with it. I tried:


* running `/usr/bin/play -qn`

* Running `paplay --raw /dev/zero`

* Downloading an actual "minute of silence" wave file and playing that


Nothing made any difference. Maybe it's smart enough to detect and not activate on absolute silence. I've looked through the manual and nothing jumped out at me as far as what to adjust.


I think you're onto something with this, since my line-in is noisier than HDMI. I'm actually getting noise from my backup drive, which is part of why I'd like to switch to HDMI.


Other things I've tried:


* Unloading the suspend-on-idle PulseAudio module. I've had issues with RPi HDMI output doing this, albeit with a longer delay, but unloading the module made no difference.

* Tweaking every number I found in /etc/speech-dispatcher/speechd.conf and /etc/speech-dispatcher/modules/espeak-ng.conf, within reason.


Maybe I need a nearly silent sound? Anyone know any incantations to, say, generate a quiet continuous stream of white noise and slowly bump up the volume until the thing stays awake?


Thanks for helping me debug this.


On 5/27/20 2:13 PM, Manuel Amador (Rudd-O) wrote:

I think your sound bar turns off when there is no signal (my receiver does the same). Quickest workaround is to play silence constantly.

On May 27, 2020 8:39:32 PM GMT+02:00, Nolan Darilek <nolan@thewordnerd.info> wrote:
FWIW, a few more details:


* It isn't specific to Optical. I ended up having to switch to HDMI 
because Optical can't do 5.1. Same result, though 5.1 audio works fine, 
including in games.

* I also set default-sample-rate = 48000 in ~/.config/pulse/daemon.conf 
thinking that might help. No luck.


Reverted to stereo and Line Out, and things are working again. I'd like 
to return to HDMI, so advice on either what to tweak or how to rapidly 
reset speech-dispatcher between tests welcome.


Thanks.

On 5/27/20 11:37 AM, Nolan Darilek wrote:
Hey folks, My soundbar includes an optical out, to which I switched in hopes of enabling 5.1 surround. Unfortunately, after having done so, speech-dispatcher appears to be dropping audio at the beginning of every utterance. As an example, I have Orca set to echo by word only. I hear "xample", "rca", "ord only", etc. when typing this. If I open the sound device by, say, playing an MP3, everything is fine. I just started a track on reduced volume so I could type this last paragraph and proof my email. What parameters can I tweak to fix this? In looking at the config, I see a PulseAudio min length, along with bunches of knobs in espeak-ng (the module I'm using.) I'd appreciate guidance on which to tweak so I'm not here all day. :) Also, how should I test a new value? What's the easiest way to restart speech-dispatcher from my user session? Can I just pkill it after tweaking the config and have it come back, or is it more involved? Thanks for any help.

--
Sent from my Android device with K-9 Mail. Please excuse my brevity.

reply via email to

[Prev in Thread] Current Thread [Next in Thread]