I am newbie in voice recognition and I want to build my first voice-controlled program in C. When I searched for my question I found some Matlab examples but they were very abstracted like [a,b]= lpc(x,p) but I am looking for a complete C program in both Linear Predictive Coding (LPC) and Dynamic Time Warping (DTW) to use them in my project (also learn implementation).

I read this article, they have built a voice-controlled system with STM32 microcontroller, using LPC to feature extraction (LPC coefficients) and DTW algorithm to compare them for matching. I like to do something like this and looking for C codes to run and learn.

  • I am afraid that the way this question is phrased makes part of it off topic for this DSP.SE. As far as the second part is concerned, involving LPC and DTW, can you please add more information about what you are trying to achieve? Are you trying to implement a complete speech recognition pipeline or do you want to use a ready made one? Are you trying to extract features simply for classification? It sounds more like what you are trying to do is recognise who is talking rather than what they are saying (?). – A_A Sep 19 '16 at 10:43
  • Hello,Thank you for reply I think my English is bad and couldn't explain good. I read this paper (ieeexplore.ieee.org/document/6092186/…) that talks about a complete voice recognition simple with STM32 microcontroller and using LPC and DTW algorithm. They extract features with LPC (LPC coefficients) and compare them with DTW lgorithm. I like to do that with my STM32F407VG(on Discovery board). I want to build a system that it's speaker independent and can recognize about 50 to 100 voice commands! – user3486308 Sep 19 '16 at 11:12
  • Thank you. Just to clarify, there was no problem in terms of language, my note was regarding the content of the question. Following the clarification, I think that the question is off topic for DSP.SE and would probably receive a more useful answer at Stack Overflow. In the meantime, please note that simply training a system to recognise a set of voice commands is relatively easy (e.g. see this library). But porting a specific library to your architecture might be the challenging step. – A_A Sep 19 '16 at 11:23

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.