Voice response too fast when using google assistand SDK with Hotword #4

MyNameWasUnavailable · 2017-08-29T21:07:44Z

Myself and quite a number of others are having issues with this device when using the google assistant SDK.

Audio tests seem to work correctly using the steps in "Configure and test the audio" from the official sdk, however when running the assistant the playback is sped up considerably, to the point where it is unintelligible.

This issue occurs with both the Supplied raspbian image and a manual installation.

xiongyihui · 2017-08-30T01:32:18Z

Enabling pulseaudio will solve the issue.
It seems that google assistant's output is mono, but is played as stereo, so the speech voice is twice.

csprimary · 2017-09-03T15:58:06Z

Enabling pulseaudio with the command

pulseaudio &

before running the Google Assistant demo results in the following error

W: [alsa-sink-bcm2835-i2s-wm8960-hifi wm8960-hifi-0] alsa-sink.c: Resume failed, couldn't restore original sample settings.

being scrolled infinitely down the terminal.

Disabling pulseaudio results in the original fast playback issue. Have still not been able to resolve this.

MyNameWasUnavailable · 2017-09-03T22:17:28Z

I can confirm that my results are the same as the previous poster.

Can you please advise what the next steps would be?

xiongyihui · 2017-09-04T09:34:44Z

My pulseaudio works. Can you post the log of pulseaudio -vvvv

xiongyihui · 2017-09-06T09:13:43Z

try the 2 mic voice card with the latest raspbian image. The pulseaudio doesn't work by default. We must modify pulseaudio's configuration file /etc/pulse/default.pa. Find the two lines which contain module-alsa and change the lines as the following:

load-module module-alsa-sink device=hw:0,0
load-module module-alsa-source device=hw:0,0

and comment the following lines:

### Automatically load driver modules depending on the hardware available
#.ifexists module-udev-detect.so
#load-module module-udev-detect
#.else
### Use the static hardware detection module (for systems that lack udev support)
#load-module module-detect
#.endif

If the on-board codec is enabled, the seeed voice card should be hw:1,0, replace hw:0,0 to hw:0,0. Run arecord -l to check the voice card's hardware number.

Reboot to use the new configuration. Besides, pavucontrol and audacity can be used to debug audio R/W.

xiongyihui · 2017-09-06T09:23:04Z

Another way to solve the issue is to use Ubuntu Mate for RPi 2/3 which is much better desktop OS.

xiongyihui · 2017-09-06T10:03:33Z

@MyNameWasUnavailable @csprimary I got another solution. Run:

wget https://github.com/respeaker/seeed-voicecard/raw/master/asound.conf
sudo cp asound.conf /etc/asound.conf

The alsa configuration uses dmix plugin and dsnoop plugin. With the configuration, we can play and record with multiple applications simultaneously.

MyNameWasUnavailable · 2017-09-06T21:12:07Z

With regards to trying the latest raspbian image, do you mean the latest stretch image from Raspberry pi foundation? I had tried that also, however I had not made the modifications you mention above.

I will try the above solutions, most likely on the weekend, and report back.

aceycv8 · 2017-09-07T07:37:49Z

i've just done a fresh install and i'm having issues getting this to work also..

if i apt get pulseaudio and set it up with your settings that you mention above then i run the SDK - when i say something it hears what i say and searches, but then the response freezes and nothing happens.

if i uninstall pulseaudio google speaks back very fast but atleast it's working - using
wget https://github.com/respeaker/seeed-voicecard/raw/master/asound.conf
sudo cp asound.conf /etc/asound.conf

doesn't make any difference

xiongyihui · 2017-09-07T10:19:58Z

@aceycv8 When google assistant freezes using pulseaudio, can you open pavucontrol, check the Playback tab and Output Devices tab and post a screenshot here?
To install pavucontrol, run: sudo apt install pavucontrol

My pavucontrol looks like:

aceycv8 · 2017-09-07T10:37:08Z

I'm using the light version so its all cli - is there anything I can do to check through cli?

xiongyihui · 2017-09-07T10:47:16Z

You can use pactl list sinks to check output devices. and use pactl list sink-inputs to check applications which are playing.

The log is like:

(env) pi@raspberrypi:~ $ pactl list sink-inputs 
Sink Input #19
	Driver: protocol-native.c
	Owner Module: 7
	Client: 32
	Sink: 0
	Sample Specification: s16le 2ch 48000Hz
	Channel Map: front-left,front-right
	Format: pcm, format.sample_format = "\"s16le\""  format.rate = "48000"  format.channels = "2"  format.channel_map = "\"front-left,front-right\""
	Corked: no
	Mute: no
	Volume: front-left: 63571 /  97% / -0.79 dB,   front-right: 63571 /  97% / -0.79 dB
	        balance 0.00
	Buffer Latency: 77605 usec
	Sink Latency: 19708 usec
	Resample method: speex-fixed-1
	Properties:
		media.name = "ALSA Playback"
		application.name = "ALSA plug-in [python3.5]"
		native-protocol.peer = "UNIX socket client"
		native-protocol.version = "32"
		application.process.id = "4889"
		application.process.user = "pi"
		application.process.host = "raspberrypi"
		application.process.binary = "python3.5"
		application.language = "C"
		window.x11.display = ":0.0"
		application.process.machine_id = "143b13e352bc41dc89c2ddbba64df665"
		application.process.session_id = "c1"
		application.icon_name = "/usr/share/pixmaps/python3.5.xpm"
		module-stream-restore.id = "sink-input-by-application-name:ALSA plug-in [python3.5]"
(env) pi@raspberrypi:~ $ pactl list sinks
Sink #0
	State: IDLE
	Name: alsa_output.hw_0_0
	Description: seeed-voicecard
	Driver: module-alsa-sink.c
	Sample Specification: s16le 2ch 44100Hz
	Channel Map: front-left,front-right
	Owner Module: 5
	Mute: no
	Volume: front-left: 63571 /  97% / -0.79 dB,   front-right: 63571 /  97% / -0.79 dB
	        balance 0.00
	Base Volume: 65536 / 100% / 0.00 dB
	Monitor Source: alsa_output.hw_0_0.monitor
	Latency: 28574 usec, configured 40000 usec
	Flags: HARDWARE DECIBEL_VOLUME LATENCY 
	Properties:
		alsa.resolution_bits = "16"
		device.api = "alsa"
		device.class = "sound"
		alsa.class = "generic"
		alsa.subclass = "generic-mix"
		alsa.name = ""
		alsa.id = "bcm2835-i2s-wm8960-hifi wm8960-hifi-0"
		alsa.subdevice = "0"
		alsa.subdevice_name = "subdevice #0"
		alsa.device = "0"
		alsa.card = "0"
		alsa.card_name = "seeed-voicecard"
		alsa.long_card_name = "seeed-voicecard"
		alsa.driver_name = "snd_soc_simple_card"
		device.bus_path = "platform-soc:sound"
		sysfs.path = "/devices/platform/soc/soc:sound/sound/card0"
		device.string = "hw:0,0"
		device.buffering.buffer_size = "352800"
		device.buffering.fragment_size = "176400"
		device.access_mode = "mmap+timer"
		device.description = "seeed-voicecard"
		device.icon_name = "audio-card"
	Formats:
		pcm

aceycv8 · 2017-09-07T22:28:01Z

thank you xiongyihui - here is my output where it still freezes.

pi@VoiceAI:~ $ pactl list sinks
Sink #0
State: SUSPENDED
Name: alsa_output.hw_0_0
Description: seeed-voicecard
Driver: module-alsa-sink.c
Sample Specification: s16le 2ch 48000Hz
Channel Map: front-left,front-right
Owner Module: 5
Mute: no
Volume: front-left: 65536 / 100% / 0.00 dB, front-right: 65536 / 100% / 0.00 dB
balance 0.00
Base Volume: 65536 / 100% / 0.00 dB
Monitor Source: alsa_output.hw_0_0.monitor
Latency: 0 usec, configured 0 usec
Flags: HARDWARE DECIBEL_VOLUME LATENCY
Properties:
alsa.resolution_bits = "16"
device.api = "alsa"
device.class = "sound"
alsa.class = "generic"
alsa.subclass = "generic-mix"
alsa.name = ""
alsa.id = "bcm2835-i2s-wm8960-hifi wm8960-hifi-0"
alsa.subdevice = "0"
alsa.subdevice_name = "subdevice #0"
alsa.device = "0"
alsa.card = "0"
alsa.card_name = "seeed-voicecard"
alsa.long_card_name = "seeed-voicecard"
alsa.driver_name = "snd_soc_simple_card"
device.bus_path = "platform-soc:sound"
sysfs.path = "/devices/platform/soc/soc:sound/sound/card0"
device.string = "hw:0,0"
device.buffering.buffer_size = "352800"
device.buffering.fragment_size = "176400"
device.access_mode = "mmap+timer"
device.description = "seeed-voicecard"
device.icon_name = "audio-card"
Formats:
pcm

pi@VoiceAI:~ $ pactl list sink-inputs
Sink Input #1
Driver: protocol-native.c
Owner Module: 7
Client: 6
Sink: 0
Sample Specification: s16le 2ch 48000Hz
Channel Map: front-left,front-right
Format: pcm, format.sample_format = ""s16le"" format.rate = "48000" format.channels = "2" format.channel_map = ""front-left,front-right""
Corked: no
Mute: no
Volume: front-left: 65536 / 100% / 0.00 dB, front-right: 65536 / 100% / 0.00 dB
balance 0.00
Buffer Latency: 90000 usec
Sink Latency: 0 usec
Resample method: n/a
Properties:
media.name = "ALSA Playback"
application.name = "ALSA plug-in [python2.7]"
native-protocol.peer = "UNIX socket client"
native-protocol.version = "29"
application.process.id = "894"
application.process.user = "pi"
application.process.host = "VoiceAI"
application.process.binary = "python2.7"
application.language = "C"
application.process.machine_id = "811d63baac94455983b6a5a0609efcba"
application.process.session_id = "c1"
application.icon_name = "/usr/share/pixmaps/python2.7.xpm"
module-stream-restore.id = "sink-input-by-application-name:ALSA plug-in [python2.7]"

xiongyihui · 2017-09-08T01:45:41Z

The only difference I found is:

pi@VoiceAI:~ $ pactl list sinks
Sink #0
State: SUSPENDED
Name: alsa_output.hw_0_0
Description: seeed-voicecard
Driver: module-alsa-sink.c
Sample Specification: s16le 2ch 48000Hz

vs

(env) pi@raspberrypi:~ $ pactl list sinks
Sink #0
	State: IDLE
	Name: alsa_output.hw_0_0
	Description: seeed-voicecard
	Driver: module-alsa-sink.c
	Sample Specification: s16le 2ch 44100Hz

@aceycv8 When the assistant is freezing, are you able to play sound using other applications such as aplay, audacity?

aceycv8 · 2017-09-08T02:28:46Z

Yep i'm able to do a speaker test fine once assistant freezes -

i run speaker-test -c 2 -t wav -D plughw:0,0

speaker-test 1.0.28

Playback device is plughw:0,0
Stream parameters are 48000Hz, S16_LE, 2 channels
WAV file(s)
Rate set to 48000Hz (requested 48000Hz)
Buffer size range from 139 to 142663
Period size range from 69 to 71332
Using max buffer size 142660
Periods = 4
was set period_size = 2691
was set buffer_size = 142660
 0 - Front Left
 1 - Front Right

and audio comes through while assistant is frozen.

here are the pactl commands while assistant is frozen and the speaker test is running.

pi@VoiceAI:~ $ pactl list sinks
Sink #0
        State: SUSPENDED
        Name: alsa_output.hw_0_0
        Description: seeed-voicecard
        Driver: module-alsa-sink.c
        Sample Specification: s16le 2ch 48000Hz
        Channel Map: front-left,front-right
        Owner Module: 5
        Mute: no
        Volume: front-left: 65536 / 100% / 0.00 dB,   front-right: 65536 / 100% / 0.00 dB
                balance 0.00
        Base Volume: 65536 / 100% / 0.00 dB
        Monitor Source: alsa_output.hw_0_0.monitor
        Latency: 0 usec, configured 0 usec
        Flags: HARDWARE DECIBEL_VOLUME LATENCY
        Properties:
                alsa.resolution_bits = "16"
                device.api = "alsa"
                device.class = "sound"
                alsa.class = "generic"
                alsa.subclass = "generic-mix"
                alsa.name = ""
                alsa.id = "bcm2835-i2s-wm8960-hifi wm8960-hifi-0"
                alsa.subdevice = "0"
                alsa.subdevice_name = "subdevice #0"
                alsa.device = "0"
                alsa.card = "0"
                alsa.card_name = "seeed-voicecard"
                alsa.long_card_name = "seeed-voicecard"
                alsa.driver_name = "snd_soc_simple_card"
                device.bus_path = "platform-soc:sound"
                sysfs.path = "/devices/platform/soc/soc:sound/sound/card0"
                device.string = "hw:0,0"
                device.buffering.buffer_size = "352800"
                device.buffering.fragment_size = "176400"
                device.access_mode = "mmap+timer"
                device.description = "seeed-voicecard"
                device.icon_name = "audio-card"
        Formats:
                pcm
pi@VoiceAI:~ $


pi@VoiceAI:~ $ pactl list sink-inputs
Sink Input #1
        Driver: protocol-native.c
        Owner Module: 7
        Client: 6
        Sink: 0
        Sample Specification: s16le 2ch 48000Hz
        Channel Map: front-left,front-right
        Format: pcm, format.sample_format = "\"s16le\""  format.rate = "48000"  format.channels = "2"  format.channel_map = "\"front-left,front-right\""
        Corked: no
        Mute: no
        Volume: front-left: 65536 / 100% / 0.00 dB,   front-right: 65536 / 100% / 0.00 dB
                balance 0.00
        Buffer Latency: 90000 usec
        Sink Latency: 0 usec
        Resample method: n/a
        Properties:
                media.name = "ALSA Playback"
                application.name = "ALSA plug-in [python2.7]"
                native-protocol.peer = "UNIX socket client"
                native-protocol.version = "29"
                application.process.id = "944"
                application.process.user = "pi"
                application.process.host = "VoiceAI"
                application.process.binary = "python2.7"
                application.language = "C"
                application.process.machine_id = "8196dca91c114c3ca89d5fccff66bf84"
                application.process.session_id = "c1"
                application.icon_name = "/usr/share/pixmaps/python2.7.xpm"
                module-stream-restore.id = "sink-input-by-application-name:ALSA plug-in [python2.7]"

xiongyihui · 2017-09-08T04:10:38Z

speaker-test with -D plughw:0,0 will bypass pulseaudio. How about just run speaker-test?

If the speaker-test is using pulseaudio, it will show in the output of pactl list sink-inputs

pi@raspberrypi:~ $ pactl list sink-inputs
Sink Input #7
	Driver: protocol-native.c
	Owner Module: 7
	Client: 10
	Sink: 0
	Sample Specification: s16le 1ch 48000Hz
	Channel Map: mono
	Format: pcm, format.sample_format = "\"s16le\""  format.rate = "48000"  format.channels = "1"  format.channel_map = "\"mono\""
	Corked: no
	Mute: no
	Volume: mono: 17364 /  26% / -34.61 dB
	        balance 0.00
	Buffer Latency: 4787625 usec
	Sink Latency: 39143 usec
	Resample method: speex-fixed-1
	Properties:
		media.name = "ALSA Playback"
		application.name = "ALSA plug-in [speaker-test]"
		native-protocol.peer = "UNIX socket client"
		native-protocol.version = "32"
		application.process.id = "1065"
		application.process.user = "pi"
		application.process.host = "raspberrypi"
		application.process.binary = "speaker-test"
		application.language = "en_US.UTF-8"
		window.x11.display = ":0.0"
		application.process.machine_id = "143b13e352bc41dc89c2ddbba64df665"
		application.process.session_id = "c1"
		module-stream-restore.id = "sink-input-by-application-name:ALSA plug-in [speaker-test]"

xiongyihui · 2017-09-08T04:16:20Z

I just uploaded my pulseaudio configuration file default.pa. You can verify it.

wget https://github.com/respeaker/seeed-voicecard/raw/master/default.pa
diff default.pa /etc/pulse/default.pa
sudo cp default.pa /etc/pulse

aceycv8 · 2017-09-08T05:14:36Z

ok -

i've copied your default.pa down.

running speaker-test by it's self the first time didn't work and i had to ctrl-c to close it - then the second time it worked with a hissing noise.

So i ran assistant and exactly the same thing just freezes.

i've left assistant open and now re run speaker-test multiple times and i get no audio at all.

here is the pactl list sinks log while assistant is running and speaker-test is running ( but no audio at all )

pi@VoiceAI:~ $ pactl list sinks
Sink #0
        State: SUSPENDED
        Name: alsa_output.hw_0_0
        Description: seeed-voicecard
        Driver: module-alsa-sink.c
        Sample Specification: s16le 2ch 48000Hz
        Channel Map: front-left,front-right
        Owner Module: 5
        Mute: no
        Volume: front-left: 65536 / 100% / 0.00 dB,   front-right: 65536 / 100% / 0.00 dB
                balance 0.00
        Base Volume: 65536 / 100% / 0.00 dB
        Monitor Source: alsa_output.hw_0_0.monitor
        Latency: 0 usec, configured 0 usec
        Flags: HARDWARE DECIBEL_VOLUME LATENCY
        Properties:
                alsa.resolution_bits = "16"
                device.api = "alsa"
                device.class = "sound"
                alsa.class = "generic"
                alsa.subclass = "generic-mix"
                alsa.name = ""
                alsa.id = "bcm2835-i2s-wm8960-hifi wm8960-hifi-0"
                alsa.subdevice = "0"
                alsa.subdevice_name = "subdevice #0"
                alsa.device = "0"
                alsa.card = "0"
                alsa.card_name = "seeed-voicecard"
                alsa.long_card_name = "seeed-voicecard"
                alsa.driver_name = "snd_soc_simple_card"
                device.bus_path = "platform-soc:sound"
                sysfs.path = "/devices/platform/soc/soc:sound/sound/card0"
                device.string = "hw:0,0"
                device.buffering.buffer_size = "352800"
                device.buffering.fragment_size = "176400"
                device.access_mode = "mmap+timer"
                device.description = "seeed-voicecard"
                device.icon_name = "audio-card"
        Formats:
                pcm

and here is the pactl list sink-inputs at the same time.

pi@VoiceAI:~ $ pactl list sink-inputs
Sink Input #9
        Driver: protocol-native.c
        Owner Module: 7
        Client: 17
        Sink: 0
        Sample Specification: s16le 2ch 48000Hz
        Channel Map: front-left,front-right
        Format: pcm, format.sample_format = "\"s16le\""  format.rate = "48000"  format.channels = "2"  format.channel_map = "\"front-left,front-right\""
        Corked: no
        Mute: no
        Volume: front-left: 65536 / 100% / 0.00 dB,   front-right: 65536 / 100% / 0.00 dB
                balance 0.00
        Buffer Latency: 90000 usec
        Sink Latency: 0 usec
        Resample method: n/a
        Properties:
                media.name = "ALSA Playback"
                application.name = "ALSA plug-in [python2.7]"
                native-protocol.peer = "UNIX socket client"
                native-protocol.version = "29"
                application.process.id = "983"
                application.process.user = "pi"
                application.process.host = "VoiceAI"
                application.process.binary = "python2.7"
                application.language = "C"
                application.process.machine_id = "18fbcba7d1054997aab1e4a9d67e20e0"
                application.process.session_id = "c2"
                application.icon_name = "/usr/share/pixmaps/python2.7.xpm"
                module-stream-restore.id = "sink-input-by-application-name:ALSA plug-in [python2.7]"

Sink Input #26
        Driver: protocol-native.c
        Owner Module: 7
        Client: 29
        Sink: 0
        Sample Specification: s16le 1ch 48000Hz
        Channel Map: mono
        Format: pcm, format.sample_format = "\"s16le\""  format.rate = "48000"  format.channels = "1"  format.channel_map = "\"mono\""
        Corked: no
        Mute: no
        Volume: mono: 65536 / 100% / 0.00 dB
                balance 0.00
        Buffer Latency: 10922666 usec
        Sink Latency: 0 usec
        Resample method: copy
        Properties:
                media.name = "ALSA Playback"
                application.name = "ALSA plug-in [speaker-test]"
                native-protocol.peer = "UNIX socket client"
                native-protocol.version = "29"
                application.process.id = "1060"
                application.process.user = "pi"
                application.process.host = "VoiceAI"
                application.process.binary = "speaker-test"
                application.language = "C"
                application.process.machine_id = "18fbcba7d1054997aab1e4a9d67e20e0"
                application.process.session_id = "c3"
                module-stream-restore.id = "sink-input-by-application-name:ALSA plug-in [speaker-test]"

I then close assistant down and also close the speaker-test that was running and re run speaker-test multiple times with no sound at all

i then run speaker-test -c 2 -t wav -D plughw:0,0 and i get sound.

i run speaker-test again and nothing.

however if i run sudo speaker-test i get a hissing sound like the very first time.

xiongyihui · 2017-09-08T05:43:09Z

@aceycv8 would you mind to join the gitter seeed-voicecard chat room? It should be easier to discuss the issue in the chat room.

https://gitter.im/seeed-voicecard/Lobby?utm_source=share-link&utm_medium=link&utm_campaign=share-link

aceycv8 · 2017-09-08T05:48:53Z

just joined in

aceycv8 · 2017-09-09T01:05:52Z

@xiongyihui - i downloaded your Pi Image and set everything back up.

i tried the following and rebooted then ran assistant with no joy - fast playback

wget https://github.com/respeaker/seeed-voicecard/raw/master/default.pa diff default.pa /etc/pulse/default.pa sudo cp default.pa /etc/pulse

I then did

wget https://github.com/respeaker/seeed-voicecard/raw/master/asound.conf sudo cp asound.conf /etc/asound.conf

and ran assistant - this fixed my problem right away - i'm using assistant with the LED script of
python google-assistant.py

and so far so good,

Thank you for your help

now just to figure out how to get a beep to happen when you say "ok google"

thank you

MyNameWasUnavailable · 2017-09-12T04:45:47Z

I have been as time allowed today trying the suggestions in the gitter chat room, using the seed supplied image and my outcome is as follows:

*just using the asounf.conf results in chipmunk speed speech.

using the default.pa you uploaded, and using pulseaudio I get the following output:

ON_RESPONDING_STARTED:
{'is_error_response': False}
W: [alsa-sink-bcm2835-i2s-wm8960-hifi wm8960-hifi-0] alsa-sink.c: Resume failed, couldn't restore original sample settings.
W: [alsa-sink-bcm2835-i2s-wm8960-hifi wm8960-hifi-0] alsa-sink.c: Resume failed, couldn't restore original sample settings.

Output from pactl list-sink-inputs:

pi@raspberrypi:/env/bin $ pactl list sink-inputs
Sink Input #1
Driver: protocol-native.c
Owner Module: 11
Client: 6
Sink: 0
Sample Specification: s16le 2ch 48000Hz
Channel Map: front-left,front-right
Format: pcm, format.sample_format = ""s16le"" format.rate = "48000" format.channels = "2" format.channel_map = ""front-left,front-right""
Corked: no
Mute: no
Volume: front-left: 65536 / 100% / 0.00 dB, front-right: 65536 / 100% / 0.00 dB
balance 0.00
Buffer Latency: 90000 usec
Sink Latency: 0 usec
Resample method: n/a
Properties:
media.name = "ALSA Playback"
application.name = "ALSA plug-in [python3.4]"
native-protocol.peer = "UNIX socket client"
native-protocol.version = "29"
application.process.id = "1251"
application.process.user = "pi"
application.process.host = "raspberrypi"
application.process.binary = "python3.4"
application.language = "C"
application.process.machine_id = "aa0a4892531c42499224a4d1b2606e1c"
application.process.session_id = "c3"
application.icon_name = "/usr/share/pixmaps/python3.4.xpm"
module-stream-restore.id = "sink-input-by-application-name:ALSA plug-in [python3.4]"
pi@raspberrypi:/env/bin $

If I try and play some audio from another session, nothing plays, and I get another recurrance of the following message in the first session:
W: [alsa-sink-bcm2835-i2s-wm8960-hifi wm8960-hifi-0] alsa-sink.c: Resume failed, couldn't restore original sample settings.

output from pactl list sink-inputs

`pi@raspberrypi:~/env/bin $ pactl list sink-inputs
Sink Input #1
Driver: protocol-native.c
Owner Module: 11
Client: 6
Sink: 0
Sample Specification: s16le 2ch 48000Hz
Channel Map: front-left,front-right
Format: pcm, format.sample_format = ""s16le"" format.rate = "48000" format.channels = "2" format.channel_map = ""front-left,front-right""
Corked: no
Mute: no
Volume: front-left: 65536 / 100% / 0.00 dB, front-right: 65536 / 100% / 0.00 dB
balance 0.00
Buffer Latency: 90000 usec
Sink Latency: 0 usec
Resample method: n/a
Properties:
media.name = "ALSA Playback"
application.name = "ALSA plug-in [python3.4]"
native-protocol.peer = "UNIX socket client"
native-protocol.version = "29"
application.process.id = "1251"
application.process.user = "pi"
application.process.host = "raspberrypi"
application.process.binary = "python3.4"
application.language = "C"
application.process.machine_id = "aa0a4892531c42499224a4d1b2606e1c"
application.process.session_id = "c3"
application.icon_name = "/usr/share/pixmaps/python3.4.xpm"
module-stream-restore.id = "sink-input-by-application-name:ALSA plug-in [python3.4]"

Sink Input #4
Driver: protocol-native.c
Owner Module: 11
Client: 21
Sink: 0
Sample Specification: s16le 2ch 44100Hz
Channel Map: front-left,front-right
Format: pcm, format.sample_format = ""s16le"" format.rate = "44100" format.channels = "2" format.channel_map = ""front-left,front-right""
Corked: no
Mute: no
Volume: front-left: 65536 / 100% / 0.00 dB, front-right: 65536 / 100% / 0.00 dB
balance 0.00
Buffer Latency: 500000 usec
Sink Latency: 0 usec
Resample method: speex-fixed-1
Properties:
media.name = "ALSA Playback"
application.name = "ALSA plug-in [aplay]"
native-protocol.peer = "UNIX socket client"
native-protocol.version = "29"
application.process.id = "1322"
application.process.user = "pi"
application.process.host = "raspberrypi"
application.process.binary = "aplay"
application.language = "en_GB.UTF-8"
application.process.machine_id = "aa0a4892531c42499224a4d1b2606e1c"
application.process.session_id = "c5"
module-stream-restore.id = "sink-input-by-application-name:ALSA plug-in [aplay]"
pi@raspberrypi:~/env/bin $`

I'm assuming I must be doing something wrong, however I cant for the life of me figure out what it is!

MyNameWasUnavailable · 2017-09-12T05:40:13Z

ok, so I just rebooted the pi, and tried again without pulseaudio, and it worked!

I did receive the following error:

ALSA lib pcm.c:7843:(snd_pcm_recover) overrun occurred

However it did work correctly.

Looks like I am on the right track

MyNameWasUnavailable · 2017-09-13T02:49:29Z

Just a final follow up comment before I close this issue.

I was able to successfully install and configure this hat using Stretch, the files / instructions from the aiyprojects-raspbian project, and the drivers and asound.conf from this git.

It worked more or less first go after fiddling for a while with the sound levels etc. I am confident that the issue now is not with the hardware or drivers as such, simply a configuration issue on our parts, which it would appear in the official project, they have worked around.

I was also able to configure the button as a shutdown button, which also works.

ONly thing left is to integrate the LED's, which is the next thing I will be looking at.

gangaselva · 2017-10-18T03:32:31Z

I have followed the Google Assistant example. Exevrything works, but Google Assistant voice is very quiet through the headphone jack and I can barely hear it (headphone is at full on ALSA mixer). I have tried changing different mixer settings, but can't fix it. anyone know how to fix this? Seeed team please give me resolution to fix this issue. Thanks in advance

xiongyihui · 2017-10-18T09:59:08Z

You can try to save https://github.com/respeaker/seeed-voicecard/blob/master/asound_2mic.conf as /etc/asound.conf

gangaselva · 2017-10-18T10:16:34Z

I tried, but still same. any idea? ThanksGanga On Wednesday, October 18, 2017, 10:59:10 PM GMT+13, Yihui Xiong <[email protected]> wrote: You can try to save https://github.com/respeaker/seeed-voicecard/blob/master/asound_2mic.conf as /etc/asound.conf — You are receiving this because you commented. Reply to this email directly, view it on GitHub, or mute the thread.

xiongyihui · 2017-10-18T11:32:52Z

Is the sound OK when playing a wav or mp3 file?

gangaselva · 2017-10-18T18:02:17Z

Sound was really good, when I play wav or mp3. I did, speaker-test -t wav test. It worked without any issues. On Thursday, October 19, 2017, 12:32:54 AM GMT+13, Yihui Xiong <[email protected]> wrote: Is the sound OK when playing a wav or mp3 file? — You are receiving this because you commented. Reply to this email directly, view it on GitHub, or mute the thread.

xiongyihui · 2017-10-19T01:31:06Z

It seems a problem between alsa and google assistant.
Maybe have a try to use pulseaudio which has a util named pavucontrol to check the output sound level and amplify the level more than 100%. It will looks like:

gangaselva · 2017-10-19T05:35:40Z

Thanks for your support. I am running headless mode. How do I increase volume to 100%. On Thursday, October 19, 2017, 2:31:07 PM GMT+13, Yihui Xiong <[email protected]> wrote: It seems a problem between alsa and google assistant. Maybe have a try to use pulseaudio which has a util named pavucontrol to check the output sound level and amplify the level more than 100%. It will looks like: — You are receiving this because you commented. Reply to this email directly, view it on GitHub, or mute the thread.

xiongyihui · 2017-10-20T01:10:26Z

You can try https://github.com/GeorgeFilipkin/pulsemixer

gangaselva · 2017-10-20T10:14:10Z

I will try and let you know, if this resolve my issue. Did you guys figure it out actual root cause? It is so painful to try all sorts of options.

…

Sent from my iPhone

On 20/10/2017, at 2:10 PM, Yihui Xiong ***@***.***> wrote: You can try https://github.com/GeorgeFilipkin/pulsemixer — You are receiving this because you commented. Reply to this email directly, view it on GitHub, or mute the thread.

gangaselva · 2017-10-21T05:27:56Z

sorry, still same. please give me clear direction to resolve the issue. On Friday, October 20, 2017, 2:10:28 PM GMT+13, Yihui Xiong <[email protected]> wrote: You can try https://github.com/GeorgeFilipkin/pulsemixer — You are receiving this because you commented. Reply to this email directly, view it on GitHub, or mute the thread.

xiongyihui · 2017-10-21T09:24:20Z

The google assistant demo (hotword version) has issues with ALSA on Raspberry Pi. The gRPC version works well.
Using PulseAudio will solve the issue. You should try the latest raspbian with pixel desktop and setup PulseAudio, and then use pavucontrol or audacity to make sure PulseAudio is working.
If PulseAudio works, pulse will be the default ALSA device. You can run arecord -v -f cd s.wav to check the default ALSA device. Then try the hotword version google assistant.

gangaselva · 2017-10-25T08:37:27Z

Hi, I resolved the issue. Its work very well. Thanks for your support. On Wednesday, October 18, 2017, 10:59:10 PM GMT+13, Yihui Xiong <[email protected]> wrote: You can try to save https://github.com/respeaker/seeed-voicecard/blob/master/asound_2mic.conf as /etc/asound.conf — You are receiving this because you commented. Reply to this email directly, view it on GitHub, or mute the thread.

Support ubuntu-server with /boot/firmware/overlays

MyNameWasUnavailable closed this as completed Sep 13, 2017

This was referenced Jan 15, 2018

Respeaker 2-mic channel 1 has failed #24

Closed

Respeaker 2mic driver does not work on PI 3 B #26

Closed

giano mentioned this issue Jan 20, 2018

2mic not listed in aplay #19

Closed

jabulonboaz mentioned this issue Feb 9, 2018

This is not working - No Respeaker card gets installed #30

Closed

channelz mentioned this issue Mar 5, 2018

ReSpeaker 4mic for Raspbian Jessie: unable to record (arecord: main:722: audio open error: Invalid argument) #40

Closed

titusece mentioned this issue Nov 19, 2018

respeaker 2mic hat audio play is not proper #117

Closed

AndreV2890 mentioned this issue Jan 10, 2019

ReSpeaker_2_mic asoundrc configuration #127

Closed

treii28 mentioned this issue Jan 26, 2019

install.sh still not working after installing kernel from source #132

Closed

0xueping mentioned this issue Feb 2, 2019

kernel headers error，Respeaker 2-Mics Pi HAT driver cannot be install successfully #134

Closed

idreamerhx mentioned this issue May 20, 2019

Setup on arm archlinux, raspi zero w #156

Closed

KacperKorczak mentioned this issue Aug 20, 2019

kernel headers error, Echo error, Don't show in aplay, arecord, alsamixer #175

Closed

lcarado mentioned this issue Mar 5, 2020

Set default HDMI audio output #215

Closed

turmary pushed a commit that referenced this issue Sep 27, 2020

Merge pull request #4 from rotdrop/feature/support-ubuntu-server

86c198d

Support ubuntu-server with /boot/firmware/overlays

goderro mentioned this issue Sep 8, 2021

[Bug]: Failed to apply overlay '3 seeed-2mic-voicecard' (5.10.60+ kernel) #311

Open

StonedEdge mentioned this issue Oct 7, 2021

[Bug]: Code doesn’t compile on RetroPie v5.4.72-v71 #304

Closed

Voice response too fast when using google assistand SDK with Hotword #4

Voice response too fast when using google assistand SDK with Hotword #4

Comments

MyNameWasUnavailable commented Aug 29, 2017

xiongyihui commented Aug 30, 2017

csprimary commented Sep 3, 2017 • edited Loading

MyNameWasUnavailable commented Sep 3, 2017

xiongyihui commented Sep 4, 2017

xiongyihui commented Sep 6, 2017

xiongyihui commented Sep 6, 2017 • edited Loading

xiongyihui commented Sep 6, 2017

MyNameWasUnavailable commented Sep 6, 2017 • edited Loading

aceycv8 commented Sep 7, 2017

xiongyihui commented Sep 7, 2017

aceycv8 commented Sep 7, 2017

xiongyihui commented Sep 7, 2017

aceycv8 commented Sep 7, 2017

xiongyihui commented Sep 8, 2017

aceycv8 commented Sep 8, 2017

xiongyihui commented Sep 8, 2017

xiongyihui commented Sep 8, 2017

aceycv8 commented Sep 8, 2017

xiongyihui commented Sep 8, 2017

aceycv8 commented Sep 8, 2017

aceycv8 commented Sep 9, 2017

MyNameWasUnavailable commented Sep 12, 2017 • edited Loading

MyNameWasUnavailable commented Sep 12, 2017

MyNameWasUnavailable commented Sep 13, 2017

gangaselva commented Oct 18, 2017

xiongyihui commented Oct 18, 2017

gangaselva commented Oct 18, 2017 via email

xiongyihui commented Oct 18, 2017

gangaselva commented Oct 18, 2017 via email

xiongyihui commented Oct 19, 2017

gangaselva commented Oct 19, 2017 via email

xiongyihui commented Oct 20, 2017

gangaselva commented Oct 20, 2017 via email

gangaselva commented Oct 21, 2017 via email

xiongyihui commented Oct 21, 2017

gangaselva commented Oct 25, 2017 via email

csprimary commented Sep 3, 2017 •

edited

Loading

xiongyihui commented Sep 6, 2017 •

edited

Loading

MyNameWasUnavailable commented Sep 6, 2017 •

edited

Loading

MyNameWasUnavailable commented Sep 12, 2017 •

edited

Loading