What is ESP-Skainet? (2021) | Learn Technology in 5 minutes

[@perspectivex]

Thanks for your review. It's good. Do you happen to have any idea if the engine can run continuously, without having to have a wake up word? I have a use case where the speech-recognition device (probably a Korvo) would operate in a quiet environment except for when various single-word commands would be given. Having to speak a wake up word every time is a waste of time and very very annoying. I briefly searched the github docs but nothing popped out clarifying if continuous mode recognition would be possible.

[@MAKERDEMY]

Hi Perspectivex Glad you liked the video. You are putting forth an interesting question. Essentially you are asking whether the system can keep listening forever and not repeat wake words every time. Sadly wake word is baked into the system and it is an essential trigger for detection and processing. This issue is still present in all modern Voice AI devices. All need a wake word. Having said that you can do modification in code to maybe simulate a wake word detection. These are outside the scope as ESP Skainet was not intended to be used like that. We hope this was helpful. Thank You, MAKERDEMY SUPPORT TEAM

[@perspectivex]

@@MAKERDEMY Thanks for the reply and info. It's too bad it's not just a simple programmatic switch to listen forever. I had thought about faking a wake word in code. I wonder how complicated that would be. I wonder how long after the wake word the software listens for speech input...if that time period was a variable that might also solve the problem as long as it could go at least 10-15min. before needing any kind of restart. "in all modern Voice AI devices" --> I've written some code using Google's speech api that let's you listen for a minute or so with no wake word, so an alternative is to use e.g. a Raspberry Pi + Respeaker mike + google api but I was really hoping to find something that can work completely offline, like the ESP Korvo apparently does. I think Pocket Sphinx also would work offline but it is much more complicated to code, I found.

[@38DDD]

Have you got a link to the ESP-Skainet IDE that understands English? I bought two KORVO kits a year ago and they only support Chinese! I get the feeling that ESP-Skainet is no longer supported!

[@MAKERDEMY]

Hi 38DDD ESP-Skainet is continuously supported by Espressif. You may want to check out this link: www.espressif.com/en/solutions/audio-solutions/esp-skainet/resources Based on the official resource: github.com/espressif/esp-skainet you can see that “Currently, Espressif MultiNet supports up to 100 Chinese or English speech commands” Please check out the same resource to get started with ESP Skainet. We hope this was helpful. Thank You.