I have one voice kit on hand. It is the rev2 version with the smaller HAT. After working with it for two weeks, here are my notes:
- You can get it directly by going to Target. I can't be more pleased to be able to get it from Target. - The kit is generally well-put together. The board looks very high-quality. It's called the Voice Bonnet. - The documentation is severely lacking. So extreme to the point that there isn't a single document detailing what each pin does on the HAT. This probably will be improved, but given what I saw for the v1: it was not great - not beyond the bare-bones. - The tutorials were easy to follow and worked just as expected. - There is nothing that is special about the hardware, it's just a DAC tacked on a RPi 0 HW. You can as well just get a DAC HAT/USB sound card and follow the same tutorials. You can do it by downloading their image on your RPi 0. - Now for that though, the kit price is reasonable for what you get in the box: A RPi 0 HW and a voice bonnet, and several other unimportant stuff. - This is not open hardware. - They have a stereo mic on the board, but you can't relocate them. - The driver is open source and can be built as DKMS, but until I asked on github, there was no indication whatsoever where the source is. For anyone who wonders the same, it's not published online, it's local on the Pi 0 image, / usr/local/src or something like that. Since it's GPLv2, I re-hosted it here: https://github.com/htruong/aiy-voicebonnet-soundcard-dkms-driver - As with every Google product ever, there is no support other than the project on Github. I emailed the developer of the driver that has an @google.com address more than a week ago, no replies. You can create issues and hope they will reply, but I'd say even an open source project gets better support than this. Another Google's attempt at making hardware. Google tried to make it sound like it's maker-friendly, promote AI and everything, but by the way it was executed, I'm not convinced. It sounds more like a gateway drug to Google's sound recognition API, there is nothing much of Machine Learning involved here. All you do is to send the voice data to Google and they return a transcription, or you send transcription and they send back voice. Everything is a black box. By the way, you get free 60 minutes per month of voice transcription API. If you actually want to do anything substantial with it, you can as well forget about it. Overall: 7/10. Works great as a raspberry with a nice sound card. And a cardboard box? On Monday, April 16, 2018 2:17:51 PM EDT Pete Soper via TriEmbed wrote: > I just ordered three of these from Arrow with shipping for $51.45. Had > to visit the page twice to get past transient error and see the price > and order button, so it may be they're being bombarded. > > https://www.arrow.com/en/products/3602/adafruit-industries?utm_source=eloqua > &utm_medium=email&utm_term=products&utm_content=arrow_aiy_sale&utm_campaign= > arrow_na_en_google-aiy-promo-control-a_apr2018 > > -Pete > > > _______________________________________________ > Triangle, NC Embedded Computing mailing list > > To post message: [email protected] > List info: http://mail.triembed.org/mailman/listinfo/triembed_triembed.org > TriEmbed web site: http://TriEmbed.org > To unsubscribe, click link and send a blank message: > mailto:[email protected]?subject=unsubscribe -- Huan Truong _______________________________________________ Triangle, NC Embedded Computing mailing list To post message: [email protected] List info: http://mail.triembed.org/mailman/listinfo/triembed_triembed.org TriEmbed web site: http://TriEmbed.org To unsubscribe, click link and send a blank message: mailto:[email protected]?subject=unsubscribe
