My own project with - OM13090: LPC54114 Audio and Voice Recognition Kit

[flo3590]

Hello everybody,

at the moment I work at a projekt to differ sounds. The sounds differ in amplitude and spectrum. In my opinion speech regonitation could be a solution.

So offline sound regonition offline on an microcontroller is not a big thing.

The only other solution I have found is a system from spansion/cypress  (http://www.cypress.com/file/252901/download) and other voice regonition chips.

For example the spansion system offers voice regonition and not sound regonition. So in this system you could write a textfile with word and the system can detect this words.

But this is not a system for me because I want to detect sounds. 

So I was realy happy at the moment I found this LPC54114 Audio and Voice Recognition Kit (First video). A system who could detect the door bell. Very similar to my system who should differ for example knoking hard/soft on a table. 

So what is my problem?

My Problem is that I have not enough information about the Voice Recognition Kit.

- Is the system "open" so I could feed them with my sound files? And build my own reconition system?

      If this is possible:

      - How many sounds the system can differ 

      - What is the sensory libary

or

-   Is the kit only a demonstration kit? And develop a reconition system is only possible with sensory (www.sensory.com).

Please help me before I buy this kit, try it out and recognize that the system is not "open" or could only differ 5 sounds or something else.

I hope you could understand my question. 

Thank you everybody for your help,

Flo

[flo3590]

Hello,

thanks for your answer.

It's poor that the kit is only demonstrate senory software and I couldn't design my own sound detection with it.

Regards,

Flo

[brendonslade]

Hi Flo,

The OM13090 kit isnt only for use with Sensory software. It is a kit designed to enable audio and voice applications, so it includes a digital microphone and an audio codec, both directly interfaced to the LPC54114 device. Examples are provided in both the supporting software packages (LPCOpen and SDK) that capture audio samples from either source and transfer them over USB. You can use these examples to have the board act like a generic USB audio device on your computer, and capture audio data that way.

Regards,

Brendon

[Clivetwo]

Hello,

I didn't find so much information about LPC54114 to make voice recognition. Some posts shows that Sensory has SDK for this, but in the OM13090 kit webpage I didn't find any documentation about this.

I'm looking for embedded voice recognition, someone has any explanation how this kit works or any other option?

Regards.

[brendonslade]

We are still partnering with Sensory, but only for larger volume opportunities. This is because each application requires custom voice trigger training, which is involves custom engineering work by Sensory. 

If your project is of this type then please contact our sales team or NXP distributor, or contact Sensory directly, and a demo binary for the board can be provided to you.

[flo3590]

Hi Brendon,

thank you very much for your answer. But I'm only interest at this one function. 

At the moment I write a thesis. In this thesis I build a prototyp to difference between sounds.

For example between:

Sounds from an angle grinder and a knocking person on a metal plate.

But this is a little bit difficult. Because sound recognition is very complex. 

So I'm not interrest at the kit itself. My hope was that I could build a regognition system with the nxp board (With sensory software).

But I couldn't find much information and sensory give me no reply.

So I have to develop my own solution.

Regards,

Flo

[Carlos_Mendoza]

Hi Flo,

Sensory does not normally offer source code licenses but may be prepared to do so for large customers. I would recommend you to contact Sensory directly to discuss this.

Hope it helps!

Best Regards,
Carlos Mendoza
Technical Support Engineer