Sound localization with xCORE Array Microphone Kits

XCore Project reviews, ideas, videos and proposals.
Post Reply
shaileshwankhede
Experienced Member
Posts: 65
Joined: Fri Dec 02, 2016 1:30 pm

Sound localization with xCORE Array Microphone Kits

Post by shaileshwankhede »

Hello,

We are working on a project to determine who is current speaker using xCORE Array Microphone Kits . Simply litting of led in the direction of sound will be first step. Do you have any sample or application note for this. If not specific to xmos implementation on other platforms?
Or any support engineer who can give right direction to proceed with this.

Thanks,
Shailesh


User avatar
mon2
XCore Legend
Posts: 1913
Joined: Thu Jun 10, 2010 11:43 am
Contact:

Post by mon2 »

Hi Shailesh. We are as well but only getting started. The subject is called "beam forming" in the audio world. XMOS has some very pricey tools as a solution. You can find some affordable alternatives using the same technology by XMOS from here:

https://www.seeedstudio.com/ReSpeaker-M ... -2719.html#

although been waiting for over 2 months for them to ship but Seeed's assembly house apparently gimped up some SMD LED mounting which is causing further delays. The new ETA is the end of this month for them to ship out the pre-orders.

This microphone array is what you are after and should be bundled with the following PCB to properly evaluate their platform:

https://www.seeedstudio.com/ReSpeaker-C ... -2716.html

https://www.seeedstudio.com/s/respeaker.html

This project was funded through KickStarter:



Then there is the solution from Conexant which was recently launched through Arrow (DS20921-EVK):

https://www.arrow.com/en/research-and-e ... everywhere

Our microphone kit by Conexant arrived yesterday so have not yet powered up the unit. We will pair up with the Raspberry PI3 by the end of this week to get a start on our reviews.

PS: Spansion had a great solution with their FM4 processors but Cypress purchased Spansion and hacked as they always do most of the acquired IP. The stand alone voice recognition unit which we have in the lab works very well and is speaker independent and can be expanded to other words - without training - just alter the support word templates. All activities for this product line have been suspended and the support FAE are buried in some pile of sand. Respectively, be cautious if working with Cypress products which may become EOL before you even launch your first widget.

====

Edit:

For the low price, consider to source one of the remaining kits from Spansion if this is the topic of your project:

http://www.digikey.com/products/en?keyw ... on%20voice

The remote demo works well so we plan to mate this demo with the XMOS microphone array for assorted testing.

Kumar
shaileshwankhede
Experienced Member
Posts: 65
Joined: Fri Dec 02, 2016 1:30 pm

Post by shaileshwankhede »

Thanks Kumar for elaborated response. We will try to buy Respeaker. Meanwhile will be continuing algorithm development for beam forming.
Need to check if any research publication or application note is there.

Thanks,
Shailesh
Post Reply