DSP - Task2 - Basic Overview on DFT Properties
Properties of short-time DFT - spectral leackage
- Generate sinus signal with parameters: f=50Hz, fs=200Hz, duration 1s .
- Evaluate short-time DFT over this signal - ONE SPECTRAL LINE (function fft) - describe correctly axis (frequency)
- Evaluate short-time DFT over the sequence with length wlen=190 vzorký. Explain results.
- Evaluate short-time DFT over the sequence with length wlen=190
vzorký with application of weighting window (Hamming).
- Observe above points using other weighting windows (rectangular,
traingular, Hann, Hamming, Blackman)
- Observe lalso spectral characteristics of above mentioned windows
- Look at short-time magnitude spectrum of the first 512 samples of
signal vm0.bin with weighting by above
- Properties of short-time DFT - spectral resolution
- Generate sum of two sinusoidal signals with parameters: f1=400Hz,
f2=450Hz, fs=8 kHz, duration 1s .
- Try to find theoretically and then empirically minimal length of
the signal required for resolution of these two signals in
frequency-domain. Use also different weighting windows.
- Periodicity in DFT
- Evaluate spectrum (short-time DFT) of one period of voiced vowel in speech signal vm0-per.bin
- Observe, how the spectrum is changed in the case of adding zeros to the sequence.
- Observe, how the spectrum is changed in the case of the repetition of this period 2x, 3x, 4x.
- Observe, how the spectrum is changed in the case of 4-times repetition of the period and for the window length 480, 479, 470, 460, 450.
- Observe the periodicity in real signal (real part of voiced vowel) vm0.bin, use window length approx. 200 - 500.
- Estimate the basic period for signals vf5.bin a vm5.bin
Back to Digital Signal Processing - Home page
This page was created Thu Oct 26 09:56:01 CEST 2000