respeaker

FAQ of ReSpeaker and Mic Array


This is the FAQ of ReSpeaker and Mic Array. If you have any other questions, feel free to post it here :) We are willing to help you!

Factory reset

Open the serial console or a ssh session and run firstbootMore detail

Wi-Fi config

We advise you to configure Wi-Fi via WEB-UI and if it can't be used, try command line tool wictl at the serial console.

How to change BING speech api recognize language

If you don't need to change the wake up words, just change text = bing.recognize(data) into

 text = bing.recognize(data,language="zh-CN") is fine. More details.

Got SD card warning message"Volume was not properly unmounted. Some data may be corrupt. Please run fsck"

If the files on the SD card are fine, ignore it. Otherwise, try to format it with sd card formatter.

ReSpeaker fail to find my Wi-Fi

Try factory reset first.

And the Wi-Fi Channel 12 is not supported by ReSpeaker. Make sure your router is not using that channel.

Bad flash from Arduino

Re-flash the bootloader on openwrt.

/etc/init.d/mopidy stop # stop mopidy if it's running, mopidy-hallo plugin will use SPI/etc/init.d/alexa stop # stop alexa if it's runningmt7688_pinmux set ephy gpiocd /etc/arduinoavrdude -c linuxgpio -p m32u4 -e -U lfuse:w:0xFF:m -U hfuse:w:0xD8:m -U efuse:w:0xCB:m -U flash:w:Caterina-ReSpeaker.hex -u -U lock:w:0xEF:m

Forgot the password of WebUI

Reset the juci password

orangectl passwd root 12345678 //replace 12345678 with the password you want to set

How to get audio source direction from respeaker microphone array?

ReSpeaker microphone array uses USB HID protocol.

When using Windows, follow the guide or use our HID tool.

How to support google speech or other Speach TO Text(STT) Engine?

Install speech_recognition library following the guide

How to change wake up word?

  • keywords.txt contains keywords and their threshold. For example, keywords.txtfrom here is

    respeaker /1e-30/alexa /1e-30/play music /1e-40/

    respeaker is a keyword, 1e-30 is its threshold. To improve sensitive, we can decrease the threshold, for example, 1e-50. We should know decreasing the threshold will increase False Acceptance Rate.

    If you want to add new keyword, you should firstly add the keyword to  dictionary.txt. The dictionary.txt is like:

    respeaker R IY S P IY K ERalexa AH L EH K S AHplay P L EYmusic M Y UW Z IH K

    The first part is a name (respeaker, alexa or music), the second part is its phonemes. You can find words in a large dictionary at here.

  • then change the code:

    if mic.wakeup('respeaker'):
  • more details issues#50issues#68

Failed to run Alexa with error"IOError: [Errno -9998] Invalid number of channels"

There is another application or alexa instance using the audio input device. Run /etc/init.d/alexa stop and /etc/init.d/mopidy stop to stop them. To disable mopidy to startup, run /etc/init.d/mopidy disable.

Failed to run python playmusic.py

It should be that mopidy is running in background and is using the USB device. try to run /etc/init.d/mopidy stop to stop mopidy and run your command again.

Python Library update

git clone https://github.com/respeaker/respeaker_python_library.gitpython setup.py install

JS or Node.js API

We can use the javascript programming of LinkIt 7688, such as

LinkIt Smart 7688 Developer’s Guide - https://labs.mediatek.com/en/download/ih80Qtjohttps://iamblue.gitbooks.io/linkit-smart-nodejs

For the limited on-board storage, we don't install node and npm by default. You install them using opkg.

Don't have a RPC connection

You need to reflash the firmware, following the guide

SFTP & FTP

We don't have a FTP on respeaker, just SFTP.

jerryyipApr 06,2017 17:38 PMAdd Comment

Comment

Popular Topics