US 8,571,858 B2Granted

Method and discriminator for classifying different segments of a signal

AssigneeFraunhofer-Gesellschaft zur Foerderung der angewandten Forschung e.V.

InventorsGuillaume Fuchs, Stefan Bayer, Jens Hirschfeld, Juergen Herre, Jeremie Lecomte, Frederik Nagel, Nikolaus Rettelbach, Stefan Wabnik, Yoshikazu Yokotani

Life of this patent

Estimating the term…

patentgrep

Loading patent…

Abstract

For classifying different segments of a signal which has segments of at least a first type and second type, e.g. audio and speech segments, the signal is short-term classified on the basis of the at least one short-term feature extracted from the signal and a short-term classification result is delivered. The signal is also long-term classified on the basis of the at least one short-term feature and at least one long-term feature extracted from the signal and a long-term classification result is delivered. The short-term classification result and the long-term classification result are combined to provide an output signal indicating whether a segment of the signal is of the first type or of the second type.

Claims17

Dependent claims

A method for classifying different segments of an audio signal, the audio signal comprising speech and music segments, the method comprising:

short-term classifying, by a short-term classifier, the audio signal on the basis of at least one short-term feature extracted from the audio signal to determine whether a current segment of the audio signal is a speech segment or a music segment, and delivering, at an output of the short-term classifier, a short-term classification result indicating that the current segment of the audio signal is a speech segment or a music segment;
long-term classifying, by a long-term classifier, the audio signal on the basis of at least one short-term feature and at least one long-term feature extracted from the audio signal to determine whether a current segment of the audio signal is a speech segment or a music segment, and delivering, at an output of the long-term classifier, a long-term classification result indicating that the current segment of the audio signal is a speech segment or a music segment;
and
applying the short-term classification result and the long-term classification result to a decision circuit coupled to the output of the short-term classifier and to the output of the long-term classifier, the decision circuit combining the short-term classification result and the long-term classification result to provide an output signal indicating whether the current segment of the audio signal is a speech segment or a music segment.

1 more element

A method for processing an audio signal comprising speech and music segments, the method comprising:

classifying a current segment of the audio signal, wherein classifying comprises:
short-term classifying, by a short-term classifier, the audio signal on the basis of at least one short-term feature extracted from the audio signal to determine whether a current segment of the audio signal is a speech segment or a music segment, and delivering, at an output of the short-term classifier, a short-term classification result indicating that the current segment of the audio signal is a speech segment or a music segment;
long-term classifying, by a long-term classifier, the audio signal on the basis of at least one short-term feature and at least one long-term feature extracted from the audio signal to determine whether a current segment of the audio signal is a speech segment or a music segment, and delivering, at an output of the long-term classifier, a long-term classification result indicating that the current segment of the audio signal is a speech segment or a music segment;
and
applying the short-term classification result and the long-term classification result to a decision circuit coupled to the output of the short-term classifier and to the output of the long-term classifier, the decision circuit combining the short-term classification result and the long-term classification result to provide an output signal indicating whether the current segment of the audio signal is a speech segment or a music segment;
dependent on the output signal provided by the classifying step, processing the current segment in accordance with a first process or a second process;
and
outputting the processed segment.

5 more elements

An audio signal processing apparatus, comprising:

a input configured to receive a audio signal to be processed, wherein the audio signal comprises speech and music segments;
a first processing stage, configured to process speech segments;
a second processing stage configured to process music segments;
a discriminator comprising:
a short-term classifier configured to receive an audio signal and to determine whether a current segment of the audio signal is a speech segment or a music segment, the short-term classifier comprising an output to provide a short-term classification result of the audio signal on the basis of at least one short-term feature extracted from the audio signal, the short-term classification result indicating that the current segment of the audio signal is a speech segment or a music segment, the audio signal comprising speech and music segments;
a long-term classifier configured to receive a audio signal and to determine whether a current segment of the audio signal is a speech segment or a music segment, the long-term classifier comprising an output to provide a long-term classification result of the audio signal on the basis of at least one short-term feature and at least one long-term feature extracted from the audio signal, the long-term classification result indicating that the current segment of the audio signal is a speech segment or a music segment;
and
a decision circuit coupled to the output of the short-term classifier and to the output of the long-term classifier for receiving the short-term classification result and the long-term classification result, the decision circuit configured to combine the short-term classification result and the long-term classification result to provide an output signal indicating whether the current segment of the audio signal is a speech segment or a music segment coupled to the input;
and
a switching device coupled between the input and the first and second processing stages and configured to apply the audio signal from the input to one of the first and second processing stages dependent on the output signal from the discriminator.

7 more elements

An audio encoder, comprising:

an audio signal processing apparatus comprising:
a input configured to receive a audio signal to be processed, wherein the audio signal comprises speech and music segments;
a first processing stage, configured to process speech segments;
a second processing stage configured to process music segments;
a discriminator comprising:
a short-term classifier configured to receive an audio signal and to determine whether a current segment of the audio signal is a speech segment or a music segment, the short-term classifier comprising an output to provide a short-term classification result of the audio signal on the basis of at least one short-term feature extracted from the audio signal, the short-term classification result indicating that the current segment of the audio signal is a speech segment or a music segment, the audio signal comprising speech and music segments;
a long-term classifier configured to receive a audio signal and to determine whether a current segment of the audio signal is a speech segment or a music segment, the long-term classifier comprising an output to provide a long-term classification result of the audio signal on the basis of at least one short-term feature and at least one long-term feature extracted from the audio signal, the long-term classification result indicating that the current segment of the audio signal is a speech segment or a music segment;
and
a decision circuit coupled to the output of the short-term classifier and to the output of the long-term classifier for receiving the short-term classification result and the long-term classification result, the decision circuit configured to combine the short-term classification result and the long-term classification result to provide an output signal indicating whether the current segment of the audio signal is a speech segment or a music segment coupled to the input;
and
a switching device coupled between the input and the first and second processing stages and configured to apply the audio signal from the input to one of the first and second processing stages dependent on the output signal from the discriminator,
wherein the first processing stage comprises a speech encoder and the second processing stage comprises a music encoder.

9 more elements

Dependencies are inferred from the claim text.

Facts

Application: 13/004,534Filed 11 Jan 2011 · continuation
Priority: 11 Jul 2008
Published: US 2011/0202337 A118 Aug 2011
Granted: 29 Oct 2013
Term ends: Not estimatedThe source records don’t support an estimate.

Classification6 codes

G10L 19/20
Vocoders using multiple modes: using sound class specific coding, hybrid encoders or object based coding
G10L 19/22
Vocoders using multiple modes: mode decision, i.e. based on audio signal content versus external parameters
G10L 25/00
Speech or voice analysis techniques not restricted to a single one of
G10L 25/51
Specially adapted for particular use: for comparison or discrimination