Amazon has patented a method using machine learning to estimate reverberation in audio recordings. The technique involves training a model to generate clean speech and reverb-only portions, improving performance by acting as a regularizer. The method filters out low-quality audio files and provides inferences to users outside the network. GlobalData’s report on Amazon gives a 360-degree view of the company including its patenting strategy. Buy the report here.
Smarter leaders trust GlobalData
According to GlobalData’s company profile on Amazon, Dynamic premium pricing was a key innovation area identified from patents. Amazon's grant share as of May 2024 was 54%. Grant share is based on the ratio of number of grants to total number of patents.
Machine learning model for estimating reverberation in audio recordings
A recently granted patent (Publication Number: US12014748B1) discloses a computer-implemented method for audio processing using machine learning models. The method involves receiving training audio files and a request to create a machine learning model, training an algorithm to generate clean speech and reverb portions of audio files, determining direct to reverberant ratios, filtering out files below a reverberance threshold, and updating the model through training iterations. The model then generates inferences for input audio files based on the clean speech and reverb portions, transmitting the results to a client application or storage location. The method also includes comparing reverb and clean speech portions, minimizing loss functions, and converting audio files between waveform and time-frequency domains.
Furthermore, the patent describes a system implementing the method, with electronic devices in a provider network storing training audio files and executing machine learning operations. The system trains algorithms to generate reverb portions, determines direct to reverberant ratios, filters out inadequate files, and generates inferences for input audio files. Additional features include generating clean speech portions, minimizing loss functions, converting audio files between domains, and training multiple algorithms for different audio processing tasks. Overall, the patent outlines a comprehensive approach to audio processing using machine learning models within a networked environment, showcasing advancements in audio technology and computational methods for enhancing audio quality and processing efficiency.
To know more about GlobalData’s detailed insights on Amazon, buy the report here.
Data Insights
From
The gold standard of business intelligence.
Blending expert knowledge with cutting-edge technology, GlobalData’s unrivalled proprietary data will enable you to decode what’s happening in your market. You can make better informed decisions and gain a future-proof advantage over your competitors.
GlobalData, the leading provider of industry intelligence, provided the underlying data, research, and analysis used to produce this article.
GlobalData Patent Analytics tracks bibliographic data, legal events data, point in time patent ownerships, and backward and forward citations from global patenting offices. Textual analysis and official patent classifications are used to group patents into key thematic areas and link them to specific companies across the world’s largest industries.
