• State Resolved
  • Locked Locked
  • Replies 2 replies
  • Answers 1 answer
  • Subscribers 99 subscribers
  • Views 632 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.

TCI6630K2L: Float values in FFTC

Kenneth Witham
Intellectual 350 points

Part Number: TCI6630K2L

Hello, 

I am a evaluating the FFTC in the TCI6630K2L chip. I have the example working which uses a 16 bit I / 16 bit Q integer values. the data we are working with is 32 bit float values. We could potentially scale these down to 16 bit float values since we do not need much precision. Would it be possible to have the FFTC process float values? If so, where is the option to tell it which data type you are giving it?

Thanks so much!

  • 0
    Guru 13202 points

    Hello!

    To my knowledge, FFTC operates on integer 16-bit IQ data, and I haven't heard about floating point support. One may look for floating implementation to cope with numbers growth during transform. For that FFTC has so called dynamic scaling, which maximizes dynamic range and prevents overflows. So if I were you I would let FFTC operate in its natural integer way with dynamic scaling, but then use accumulated exponent to convert to floats right after FFTC for further processing.

  • 0 in reply to Victor Kazmirenko
    TI__Mastermind 21025 points

    Correct.  FFTC operates on integers.  For more information on dynamic scaling mode, please download SPRUGS2C from ti.com search for "dynamic scaling".

      -dave