Apparatus and Method for End-to-End Adversarial Blind Bandwidth Extension with one or more Convolutional and/or Recurrent Networks

Simple SummaryContent extracted from patent full text and abstract with AI.

This patent describes an apparatus and method that uses neural networks, such as convolutional and recurrent networks, to convert narrowband speech signals into wideband speech signals. It does this by predicting and reconstructing missing higher-frequency components, making speech sound fuller and clearer. The solution works end-to-end and uses multiple specialized neural network modules to separately extrapolate parts of the audio signal and then combine them into a higher-quality output.

Use CasesContent extracted from patent full text and abstract with AI.

• Enhancing audio quality in telecommunication systems (e.g., VoIP, mobile calls) where only narrowband audio is transmitted.
• Improving sound quality in hearing aids or assistive listening devices by reconstructing lost frequency components.
• Restoring or upscaling archival audio recordings to higher fidelity for preservation or playback.
• Bandwidth extension in music streaming or audio playback devices to simulate wideband audio from compressed sources.
• Applications in speech recognition systems to improve accuracy by providing richer audio input.

BenefitsContent extracted from patent full text and abstract with AI.

• Significantly improves perceived speech quality from low-bitrate or narrowband audio sources.
• Operates automatically and end-to-end, reducing the need for manual audio engineering interventions.
• Flexible—can be implemented using various types of neural networks, such as convolutional or recurrent.
• Enables legacy or bandwidth-limited equipment to deliver wideband audio without hardware changes.
• Enhances user experience in voice communication systems by making speech sound more natural and intelligible.

Technical Classifications (CPCs)

Inventors & Applicants

Patent Abstract

An apparatus for processing a narrowband speech input signal by conducting bandwidth extension of the narrowband speech input signal to obtain a wideband speech output signal according to an embodiment is provided. The apparatus includes a signal envelope extrapolator including a first neural network, wherein the first neural network is configured to receive as input values of the first neural network a plurality of samples of a signal envelope of the narrowband speech input signal, and configured to determine as output values of the first neural network a plurality of extrapolated signal envelope samples. Moreover, the apparatus includes an excitation signal extrapolator configured to receive a plurality of samples of an excitation signal of the narrowband speech input signal, and configured to determine a plurality of extrapolated excitation signal samples. Furthermore, the apparatus includes a combiner configured to generate the wideband speech output signal such that the wideband speech output signal is bandwidth extended with respect to the narrowband speech input signal depending on the plurality of extrapolated signal envelope samples and depending on the plurality of extrapolated excitation signal samples.