An incremental algorithm for signal reconstruction from short-time fourier transform magnitude
نویسندگان
چکیده
We present an algorithm for reconstructing a time-domain signal from the magnitude of a short-time Fourier transform (STFT). In contrast to existing algorithms based on alternating projections, we offer a novel approach involving numerical root-finding combined with explicit smoothness assumptions. Our technique produces high-quality reconstructions that have lower signal-to-noise ratios when compared to other existing algorithms. If there is little redundancy in the given STFT, in particular, the algorithm can produce signals which also sound significantly better perceptually, as compared to existing work.
منابع مشابه
Windowing Effects of Short Time Fourier Transform on Wideband Array Signal Processing Using Maximum Likelihood Estimation
During the last two decades, Maximum Likelihood estimation (ML) has been used to determine Direction Of Arrival (DOA) and signals propagated by the sources, using narrowband array signals. The algorithm fails in the case of wideband signals. As an attempt by the present study to overcome the problem, the array outputs are transformed into narrowband frequency bins, using short time Fourier tran...
متن کاملWindowing Effects of Short Time Fourier Transform on Wideband Array Signal Processing Using Maximum Likelihood Estimation
During the last two decades, Maximum Likelihood estimation (ML) has been used to determine Direction Of Arrival (DOA) and signals propagated by the sources, using narrowband array signals. The algorithm fails in the case of wideband signals. As an attempt by the present study to overcome the problem, the array outputs are transformed into narrowband frequency bins, using short time Fourier tran...
متن کاملSignal Reconstruction from Stft Magnitude: a State of the Art
This paper presents a review on techniques for signal reconstruction without phase, i.e. when only the spectrogram (the squared magnitude of the Short Time Fourier Transform) of the signal is known. The now standard Griffin and Lim algorithm will be presented, and compared to more recent blind techniques. Two important issues are raised and discussed: first, the definition of relevant criteria ...
متن کاملOn Signal Reconstruction from Fourier Magnitude
In this paper, three new algorithms for signal reconstruction from spectral magnitude are presented. The first algorithm reconstructs a signal from its discrete Fourier transform (DFT) magnitude and half of its samples using the decimation-in-time FFT algorithm, which results in a closed-form solution for the unknown data. The second is based on localized Fourier transform magnitudes and a sing...
متن کاملPhase initialization schemes for faster spectrogram-consistency- based signal reconstruction
In previous contributions [1, 2], we presented a fast algorithm for the reconstruction of a time-domain signal from a magnitude short-time Fourier transform (STFT) spectrogram, and showed experimentally that it could lead to substantial reduction in computation time compared to previous work by Griffin and Lim [3] when both algorithms are initialized using zero phase. We study here other strate...
متن کامل