NanoPi Duo Sound Delay

[Johhny Blue]

Hi,

I am using the  Armbian NanoPiDuo image and I have enable sound through armbian-config.

It work well except for a small issue. When a file is sent to play, it misses out on about .5 seconds of the first part of audio.

So if I run speaker-test -t wav, instead of the sound playing "front left" it just play the work left, the continues playing "front left" correctly.

If I stop speaker test and run again straight away it work as planned.

If I wait a few seconds the original problem occurs.

To my basic knowlege, it seems like it is initialising the sound out.

Is anyone else having this problem?

Are there any settings I can adjust to fix the problem?

I know these are not completed versions, so I understand if it can not be fixed at the moment.

Thanks

V.

[Igor]

https://docs.armbian.com/User-Guide_Getting-Started/#what-are-testing-images
At the present situation, we are happy that sound works  and little less happy to find out for yet another (small) issue 

[guidol]

10 hours ago, Johhny Blue said:

It work well except for a small issue. When a file is sent to play, it misses out on about .5 seconds of the first part of audio.

If I wait a few seconds the original problem occurs.

Is anyone else having this problem?

 

For me - on my Duo -it looks like there is some energy-saving-management at the audio-out. when some seconds the audio isnt used it seems to shutdown the amplifier  

Did hear that problem while using espeak TTS, where the first word are missing on th first command....on the second command all word could be heard....
but after some seonds the same did appear to the sound output....

[Igor]

Try changing CPU governor.

[zador.blood.stained]

Most likely it's a delay in audio clock enabling, there were similar complaints previously that were solved by patching the driver (if this is on the mainline kernel).

[guidol]

1 hour ago, Igor said:

Try changing CPU governor.

@Igor I changed from ondemand to interactive or performance in  /etc/default/cpufrequtils

but after a reboot there is no difference.

First Sound is cut, second command after 3-5 seconds is OK but when the command is started 7-10 seconds after the first audio is cut again.

and this on the actual ARMBIAN 5.34.171111 nightly Debian GNU/Linux 9 (stretch) 4.13.12-sunxi
 

After 5-7 seconds I can hear a low noise like "click" - sound like shuting down the ampfiler.

 

So from espeak -s145 "Hello, how are you today"?" you will get only the "are you today?" in the first try

while getting the fill text when using the same command 3-5 seconds after the first one.

 

This is for "espeak TTS".

While using Internetradio with mpg123/mplayer there is a continuous audio-stream without problems because of no audio-break.

 

@zador.blood.stained How can I check if I do use the patched driver?

 

[zador.blood.stained]

41 minutes ago, guidol said:

How can I check if I do use the patched driver?

We don't have such patches in the kernel sources.

[zador.blood.stained]

Possibly related, though this patch was already applied a long time ago: https://www.spinics.net/lists/arm-kernel/msg555074.html

[guidol]

2 hours ago, zador.blood.stained said:

Possibly related, though this patch was already applied a long time ago: https://www.spinics.net/lists/arm-kernel/msg555074.html

Thats sounds exactly like the problem on the Duo - maybe this patch is missing or has be longer on the Duo?

[guidol]

On 10.11.2017 at 3:25 PM, guidol said:

First Sound is cut, second command after 3-5 seconds is OK but when the command is started 7-10 seconds after the first audio is cut again.

and this on the actual ARMBIAN 5.34.171111 nightly Debian GNU/Linux 9 (stretch) 4.13.12-sunxi
 

After 5-7 seconds I can hear a low noise like "click" - sound like shuting down the ampfiler.

 

So from espeak -s145 "Hello, how are you today"?" you will get only the "are you today?" in the first try

while getting the fill text when using the same command 3-5 seconds after the first one.

because I got the same "problem" with my Orange Pi R1 and the audio from the NAS-Expansionboard I did use a little workaround
First I do "send" a empty .wav to the audio-System, then sleep a second and then do the normal output - when the soundsystem is awake

root@opi-zero-r1:/home/guido# more ./speak.sh
espeak -w /tmp/espeak.wav -s150 ""; aplay -D plughw:0,0 /tmp/espeak.wav 2>/dev/null
sleep 1
espeak -w /tmp/espeak.wav -s150 "$1"; aplay -D plughw:0,0 /tmp/espeak.wav 2>/dev/null

Linux opi-zero-r1 4.13.14-sunxi #240 SMP Mon Nov 20 00:09:06 CET 2017 armv7l GNU/Linux