• State Resolved
  • Locked Locked
  • Replies 2 replies
  • Subscribers 98 subscribers
  • Views 156 views
  • Users 0 members are here
Support feedback
Options
Options
Related

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

Question related with FFT

Wendy.Z
Intellectual 525 points

Hello, 

Now I'm testing FFT, but I think there's problem occured by the different Data Type.

The type of my recieved input signal(a counter up to 1024) is uint32_t, in the beginning I was using that "DSP_fft32x32" source code. but unfortunately there's NO output data. (to match with it, I used gen_twiddle_fft32x32). so now based on DSPF_fft_SPxSP modified my code. I stored all my input data into data[2*i], and set data[2*i +1]=0, my input data(32 bit int) is plotted correctly in the graph, just like sawtooth wave with zeros in the middle of 2 values, but when I transfered it into float(the data type needed in DSPF_fft_SPxSP), there came the problem. the figure seems like having a big distortion, at least the tendency of my values are totally different from the integer ones. 

I wanna ask, is this proper that I change the data type directly? for example, int32_t A; float B; B=(float)A;

if its not, how can I make it work? 

Thanks in advance

  • 0
    Intellectual 680 points

    Hello Wendy ,

    I have directly not answer to your Question , I am try to run a DSP_fft32*32 on C6678 . I want to ask you :

    whichFiles are included into your Programm ? (location and Filename ). I am confused which the lots off Files inside the DspLib Directory

    Thank you for reply

    Lopez

  • 0 in reply to Lopet azer
    Intellectual 525 points

    Hi Lopez, 

    I was using the function called DSPF_sp_fftSPxSP which you could easily find under the directory of <...\ti\dsplib_c66x_3_1_0_0\packages\ti\dsplib\src> 

    as you could see from the function, you need to set your data as the type of float.

    and btw, I solved my old problem. its fine to cast integer directly to float. I just changed int32_t to uint32_t and it worked. because of the data was too large that it flowed.

    Regards,

    Wendy