Subject: Re: Sound Analysis Tools From: Thierry Rochebois <thierry(at)MAURY.IEF-PARIS-SUD.FR> Date: Fri, 17 Nov 1995 19:08:51 +0100>Since the Fourier transform is linear, the two are equivalent. As the Fourier transform is windowed, there is always a time/frequency resolution tradeoff. So, the interpolation is not as accurate as one wants. > As I recall, >it happened to be more convenient to subtract in the frequency domain. Well, I have to resynthesize the sound anyway. I just synthesize the sound in a way that allows substraction. It is not very hard. When I use every peaks (ie when I don't select the harmonic part), the direct substraction (in the time domain) validate the quality of the analysis (ie the accuracy of the interpolation). The error of the analysis equals the remaining signal. After extracting the harmonic part from the signal, I can study the signal with other time frequency/resolution tradeoff. This can be efficient to study transients (high time resolution, low frequency resolution). >The hard part was creating a properly interpolated window transform, >matched to the sinusoidal peak in complex amplitude and frequency of >translation. Yes! Properly interpolated means properly interpreted. The spectrum is convolved with the transform of the window. The good choice is to choose the window and to interpret its effects in the frequency domain. I think that the gaussian is the easiest window to interpret as its transform is a... gaussian. Maybe, one can choose another window which simplifies interpolation... Thierry PS I will be on weekend after this post. So, good week end and good night(to Europeans). +-------------------------------------------------------------+ | |\ | Thierry Rochebois Doctorant | + | IEF ARCEMA/TSI Bat 220 UPS 91405 ORSAY Cedex | | | thierry(at)ief-paris-sud.fr | | | http://maury.ief-paris-sud.fr:8001/~thierry/welcome.html | | | | | +-------------------------------------------------------------+ | \ \| +-------------------------------------------------------------+