Patents Assigned to Audioshake, Inc.
  • REAL-TIME SOURCE-WISE ENERGY ESTIMATION SYSTEM
    Publication number: 20260321277
    Abstract: A method, system, and computer-readable medium for generating energy estimations for a plurality of component sources from a single input signal. A machine learning model processes the input signal, or a representation thereof, to generate per-source energy estimates without performing source separation. The model may output energy values directly or via intermediate representations. The techniques enable low-latency, real-time estimation of component source energy from a mixture signal.
    Type: Application
    Filed: April 7, 2026
    Publication date: October 15, 2026
    Applicant: Audioshake, Inc.
    Inventors: Fabian-Robert Stöter, Jessica Powell
  • Deep learning segmentation of audio using magnitude spectrogram
    Patent number: 12763084
    Abstract: A method, system, and computer readable medium for decomposing an audio signal into different isolated sources. The techniques and mechanisms convert an audio signal into K input spectrogram fragments. The fragments are sent into a deep neural network to isolate for different sources. The isolated fragments are then combined to form full isolated source audio signals.
    Type: Grant
    Filed: July 8, 2024
    Date of Patent: October 13, 2026
    Assignee: Audioshake, Inc.
    Inventor: Luke Miner
  • DEEP LEARNING SEGMENTATION OF AUDIO USING MAGNITUDE SPECTROGRAM
    Publication number: 20240412747
    Abstract: A method, system, and computer readable medium for decomposing an audio signal into different isolated sources. The techniques and mechanisms convert an audio signal into K input spectrogram fragments. The fragments are sent into a deep neural network to isolate for different sources. The isolated fragments are then combined to form full isolated source audio signals.
    Type: Application
    Filed: July 8, 2024
    Publication date: December 12, 2024
    Applicant: Audioshake, Inc.
    Inventor: Luke Miner
  • Deep learning segmentation of audio using magnitude spectrogram
    Patent number: 12057131
    Abstract: A method, system, and computer readable medium for decomposing an audio signal into different isolated sources. The techniques and mechanisms convert an audio signal into K input spectrogram fragments. The fragments are sent into a deep neural network to isolate for different sources. The isolated fragments are then combined to form full isolated source audio signals.
    Type: Grant
    Filed: May 6, 2022
    Date of Patent: August 6, 2024
    Assignee: AUDIOSHAKE, INC.
    Inventor: Luke Miner
  • Deep learning segmentation of audio using magnitude spectrogram
    Patent number: 11837245
    Abstract: A method, system, and computer readable medium for decomposing an audio signal into different isolated sources. The techniques and mechanisms convert an audio signal into K input spectrogram fragments. The fragments are sent into a deep neural network to isolate for different sources. The isolated fragments are then combined to form full isolated source audio signals.
    Type: Grant
    Filed: November 1, 2022
    Date of Patent: December 5, 2023
    Assignee: AUDIOSHAKE, INC.
    Inventor: Luke Miner
  • DEEP LEARNING SEGMENTATION OF AUDIO USING MAGNITUDE SPECTROGRAM
    Publication number: 20230093726
    Abstract: A method, system, and computer readable medium for decomposing an audio signal into different isolated sources. The techniques and mechanisms convert an audio signal into K input spectrogram fragments. The fragments are sent into a deep neural network to isolate for different sources. The isolated fragments are then combined to form full isolated source audio signals.
    Type: Application
    Filed: November 1, 2022
    Publication date: March 23, 2023
    Applicant: Audioshake, Inc.
    Inventor: Luke Miner
  • Deep learning segmentation of audio using magnitude spectrogram
    Patent number: 11521630
    Abstract: A method, system, and computer readable medium for decomposing an audio signal into different isolated sources. The techniques and mechanisms convert an audio signal into K input spectrogram fragments. The fragments are sent into a deep neural network to isolate for different sources. The isolated fragments are then combined to form full isolated source audio signals.
    Type: Grant
    Filed: October 2, 2020
    Date of Patent: December 6, 2022
    Assignee: AUDIOSHAKE, INC.
    Inventor: Luke Miner
  • DEEP LEARNING SEGMENTATION OF AUDIO USING MAGNITUDE SPECTROGRAM
    Publication number: 20220262375
    Abstract: A method, system, and computer readable medium for decomposing an audio signal into different isolated sources. The techniques and mechanisms convert an audio signal into K input spectrogram fragments. The fragments are sent into a deep neural network to isolate for different sources. The isolated fragments are then combined to form full isolated source audio signals.
    Type: Application
    Filed: May 6, 2022
    Publication date: August 18, 2022
    Applicant: Audioshake, Inc.
    Inventor: Luke Miner
  • Deep learning segmentation of audio using magnitude spectrogram
    Patent number: 11355134
    Abstract: A method, system, and computer readable medium for decomposing an audio signal into different isolated sources. The techniques and mechanisms convert an audio signal into K input spectrogram fragments. The fragments are sent into a deep neural network to isolate for different sources. The isolated fragments are then combined to form full isolated source audio signals.
    Type: Grant
    Filed: October 2, 2020
    Date of Patent: June 7, 2022
    Assignee: AUDIOSHAKE, INC.
    Inventor: Luke Miner
  • DEEP LEARNING SEGMENTATION OF AUDIO USING MAGNITUDE SPECTROGRAM
    Publication number: 20220108711
    Abstract: A method, system, and computer readable medium for decomposing an audio signal into different isolated sources. The techniques and mechanisms convert an audio signal into K input spectrogram fragments. The fragments are sent into a deep neural network to isolate for different sources. The isolated fragments are then combined to form full isolated source audio signals.
    Type: Application
    Filed: October 2, 2020
    Publication date: April 7, 2022
    Applicant: Audioshake, Inc.
    Inventor: Luke Miner
  • DEEP LEARNING SEGMENTATION OF AUDIO USING MAGNITUDE SPECTROGRAM
    Publication number: 20210035590
    Abstract: A method, system, and computer readable medium for decomposing an audio signal into different isolated sources. The techniques and mechanisms convert an audio signal into K input spectrogram fragments. The fragments are sent into a deep neural network to isolate for different sources. The isolated fragments are then combined to form full isolated source audio signals.
    Type: Application
    Filed: October 2, 2020
    Publication date: February 4, 2021
    Applicant: Audioshake, Inc.
    Inventor: Luke Miner