Forum Discussion

Altera_Forum's avatar
Altera_Forum
Icon for Honored Contributor rankHonored Contributor
14 years ago

speech recognition

hey everyone

i am working on speech recognition project with nios2 and i need some help.

after recording the words from the audio codec i do software fft (with ffteasy code that based on numerical recipes algoritem).

my problem is that the power of the voice that i say the word has huge influence and because of this i cant recognize the speech.

i will be very greatfull if anyone can help me and give some solution to deal with this problam.

thanks

2 Replies

  • Altera_Forum's avatar
    Altera_Forum
    Icon for Honored Contributor rankHonored Contributor
    14 years ago

    hi

    did you get the data from audio codec and save in a memory ? i'm trying to take the right data and the left data from audio core to a dual port ram . how can I use two channel or ıf you use Microphone how can arrange the parameters in controller code? can u send me your code for examine ?

    about your question

    ın wm8731 datasheet there is some information about volume control and db calculatıon

    and in the cornell university final projects you can find the speech recognation project
  • Altera_Forum's avatar
    Altera_Forum
    Icon for Honored Contributor rankHonored Contributor
    14 years ago

    I'm also working on a speech recognition project. Does anyone know of a temp program in VHDL coding that can show me an example of getting the input voice (one word) from a Mic on a FPGA Altera D2E board?