An audio signal classification method includes determining, according to voice activity of a current audio frame, whether to obtain a frequency spectrum fluctuation of the current audio frame and store the frequency spectrum fluctuation in a frequency spectrum fluctuation memory, and updating, according to whether the audio frame is percussive music or activity of a historical audio frame, frequency spectrum fluctuations stored in the frequency spectrum fluctuation memory, and classifying the current audio frame as a speech frame or a music frame according to statistics of a part or all of effective data of the frequency spectrum fluctuations stored in the frequency spectrum fluctuation memory.

Example embodiments relate to techniques for training artificial neural networks or oilier machine-learning encoders to accurately predict the pitch of input audio samples in a semitone or otherwise logarithmically-scaled pitch space. An example method may include generating, from a sample of audio data, two training samples by applying two different pitch shifts to the sample of audio training data. This can be done by converting the sample of audio data into the frequency domain and then shifting the transformed data. These known shifts are then compared to the predicted pitches generated by applying the two training samples to the encoder. The encoder is then updated based on the comparison, such that the relative pitch output by the encoder is improved with respect to accuracy. One or more audio samples, labeled with absolute pitch values, can then be used to calibrate the relative pitch values generated by the trained encoder.

A computer may train a single-class machine learning using normal speech recordings. The machine learning model or any other model may estimate the normal range of parameters of a physical speech production model based on the normal speech recordings. For example, the computer may use a source-filter model of speech production, where voiced speech is represented by a pulse train and unvoiced speech by a random noise and a combination of the pulse train and the random noise is passed through an auto-regressive filter that emulates the human vocal tract. The computer leverages the fact that intentional modification of human voice introduces errors to source-filter model or any other physical model of speech production. The computer may identify anomalies in the physical model to generate a voice modification score for an audio signal. The voice modification score may indicate a degree of abnormality of human voice in the audio signal.

The invention provides methods and devices for stereo encoding and decoding using complex prediction in the frequency domain. In one embodiment, a decoding method, for obtaining an output stereo signal from an input stereo signal encoded by complex prediction coding and comprising first frequency-domain representations of two input channels, comprises the upmixing steps of: (i) computing a second frequency-domain representation of a first input channel; and (ii) computing an output channel on the basis of the first and second frequency-domain representations of the first input channel, the first frequency-domain representation of the second input channel and a complex prediction coefficient. The upmixing can be suspended responsive to control data.

Innovations in phase quantization during speech encoding and phase reconstruction during speech decoding are described. For example, to encode a set of phase values, a speech encoder omits higher-frequency phase values and/or represents at least some of the phase values as a weighted sum of basis functions. Or, as another example, to decode a set of phase values, a speech decoder reconstructs at least some of the phase values using a weighted sum of basis functions and/or reconstructs lower-frequency phase values then uses at least some of the lower-frequency phase values to synthesize higher-frequency phase values. In many cases, the innovations improve the performance of a speech codec in low bitrate scenarios, even when encoded data is delivered over a network that suffers from insufficient bandwidth or transmission quality problems.

Innovations in phase quantization during speech encoding and phase reconstruction during speech decoding are described. For example, to encode a set of phase values, a speech encoder omits higher-frequency phase values and/or represents at least some of the phase values as a weighted sum of basis functions. Or, as another example, to decode a set of phase values, a speech decoder reconstructs at least some of the phase values using a weighted sum of basis functions and/or reconstructs lower-frequency phase values then uses at least some of the lower-frequency phase values to synthesize higher-frequency phase values. In many cases, the innovations improve the performance of a speech codec in low bitrate scenarios, even when encoded data is delivered over a network that suffers from insufficient bandwidth or transmission quality problems.

The present invention proposes a new method for improving the performance of a real-valued filterbank based spectral envelope adjuster. By adaptively locking the gain values for adjacent channels dependent on the sign of the channels, as defined in the application, reduced aliasing is achieved. Furthermore, the grouping of the channels during gain-calculation, gives an improved energy estimate of the real valued subband signals in the filterbank.

The present invention proposes a new method for improving the performance of a real-valued filterbank based spectral envelope adjuster. By adaptively locking the gain values for adjacent channels dependent on the sign of the channels, as defined in the application, reduced aliasing is achieved. Furthermore, the grouping of the channels during gain-calculation, gives an improved energy estimate of the real valued subband signals in the filterbank.

Innovations in phase quantization during speech encoding and phase reconstruction during speech decoding are described. For example, to encode a set of phase values, a speech encoder omits higher-frequency phase values and/or represents at least some of the phase values as a weighted sum of basis functions. Or, as another example, to decode a set of phase values, a speech decoder reconstructs at least some of the phase values using a weighted sum of basis functions and/or reconstructs lower-frequency phase values then uses at least some of the lower-frequency phase values to synthesize higher-frequency phase values. In many cases, the innovations improve the performance of a speech codec in low bitrate scenarios, even when encoded data is delivered over a network that suffers from insufficient bandwidth or transmission quality problems.

Innovations in phase quantization during speech encoding and phase reconstruction during speech decoding are described. For example, to encode a set of phase values, a speech encoder omits higher-frequency phase values and/or represents at least some of the phase values as a weighted sum of basis functions. Or, as another example, to decode a set of phase values, a speech decoder reconstructs at least some of the phase values using a weighted sum of basis functions and/or reconstructs lower-frequency phase values then uses at least some of the lower-frequency phase values to synthesize higher-frequency phase values. In many cases, the innovations improve the performance of a speech codec in low bitrate scenarios, even when encoded data is delivered over a network that suffers from insufficient bandwidth or transmission quality problems.