13.07.2015 Views

Tutorials and Topics - Peabody Computer Music

Tutorials and Topics - Peabody Computer Music

Tutorials and Topics - Peabody Computer Music

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

Tutorial 26Frequency domain signalprocessing using pfft~You can easily convert between real/imaginary pairs <strong>and</strong> amplitude/phase pairs using theobjects cartopol~ <strong>and</strong> poltocar~:Cartesian to polar conversionTechnical detail: The amplitude values output by the left outlet of cartopol~depend on the amplitude of the signal you send to the pfft~ object. Due tothe way fftin~ <strong>and</strong> fftout~ automatically scale their window functions (inorder to maintain the same output amplitude after overlap-adding), themaximum amplitude value for a constant signal of 1.0 will be(FFT size / (sqrt(sum of points in the window/hop size))So, when using a 512-point FFT with a square window with an overlap of2, the maximum possible amplitude value will be roughly 362, with 4-overlap it will be 256. When using a hanning or hamming window <strong>and</strong> 2overlap, it will be approximately 325 or 341, <strong>and</strong> with 4-overlap, it will be230 or 241, respectively. Generally, however, the peak amplitudes in aspectral frame will most likely be only one-fourth to half this high for nonperiodicor semi-periodic “real-world” sounds normalized between -1.0<strong>and</strong> 1.0.The phase values output by the right outlet of cartopol~ will always bebetween -π <strong>and</strong> π.211

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!