摘要:
The invention relates to a method for localizing and tracking acoustic sources (101) in a multi-source environment, comprising the steps of recording audio-signals (103) of at least one acoustic source (101) with at least two recording means (104, 105), creating a two- or multi-channel recording signal, partitioning said recording signal into frames of predefined length (N), calculating for each frame a cross-correlation function as a function of discrete time-lag values (τ) for channel pairs (106, 107) of the recording signal, evaluating the cross-correlation function by calculating a sampling function depending on a pitch parameter (f0) and at least one spatial parameter (φ0), the sampling function assigning a value to every point of a multidimensional space being spanned by the pitch-parameter and the spatial parameters, and identifying peaks in said multidimensional space with respective acoustic sources in the multi-source environment.
摘要:
The invention relates to a method for localizing and tracking acoustic sources (101) in a multi-source environment, comprising the steps of recording audio-signals (103) of at least one acoustic source (101) with at least two recording means (104, 105), creating a two- or multi-channel recording signal, partitioning said recording signal into frames of predefined length (N), calculating for each frame a cross-correlation function as a function of discrete time-lag values (τ) for channel pairs (106, 107) of the recording signal, evaluating the cross-correlation function by calculating a sampling function depending on a pitch parameter (f0) and at least one spatial parameter (φ0), the sampling function assigning a value to every point of a multidimensional space being spanned by the pitch-parameter and the spatial parameters, and identifying peaks in said multidimensional space with respective acoustic sources in the multi-source environment.
摘要:
The invention relates to a method for determining fundamental-frequency courses of a plurality of signal sources from a one-channel audio recording of a mixed signal, comprising the following steps: a) determining the spectrogram properties of the pitch states of individual signal sources using training data; b) determining the probabilities of the fundamental-frequency combinations of the signal sources contained in the mixed signal by combining the properties determined in a) by means of an interaction model; and c) tracking the fundamental-frequency courses of the individual signal sources.