Bundled Multi-Rate Feedback Autoencoder for Audio Bitrate Allocation

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing data encoding methods for wireless devices are inefficient in terms of bit allocation and transmission bandwidth, often using the same number of bits for each frame, leading to increased header size and bandwidth usage.

Innovation Solution

A bundled multi-rate feedback autoencoder that dynamically allocates bits to different frames by designating a reference frame with a higher bitrate and predicted frames with lower bitrates, exploiting temporal redundancies to reduce overall bit usage.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If the same number of bits is allocated to each frame, then encoding simplicity is maintained, but transmission bandwidth efficiency deteriorates

Engineering Contradiction:
Improvetransmission bandwidth efficiencyVSAvoidencoding complexity
Core Design Contradiction:
ProductivityVSDevice complexity

Solution Approach 1:

The patent applies local quality by allocating different bitrates to different frames based on their importance. Reference frames are encoded at higher bitrates to preserve quality for future reconstruction, while non-reference frames are encoded at lower bitrates. This differential allocation optimizes overall transmission efficiency without requiring uniform complexity across all frames.

Inventive Principle:
Principle #3Local quality

Solution Approach 2:

The patent implements dynamic bitrate allocation where the encoding parameters change based on frame type (reference vs. non-reference). The system dynamically adjusts the number of bits allocated to each frame according to its role in the prediction structure, allowing the encoding complexity to vary adaptively rather than remaining static.

Inventive Principle:
Principle #15Dynamics

2Measurement precision

If more bits are used for each frame, then data fidelity is improved, but transmission bandwidth consumption increases

Engineering Contradiction:
Improvedata fidelityVSAvoidtransmission bandwidth
Core Design Contradiction:
Measurement precisionVSQuantity of substance

Solution Approach 1:

The patent applies local quality by concentrating higher bitrate allocation on reference frames where fidelity is critical for accurate prediction, while accepting lower fidelity in non-reference frames. This selective approach maintains overall data fidelity where it matters most while reducing total bandwidth consumption through strategic bit allocation.

Inventive Principle:
Principle #3Local quality

Solution Approach 2:

The patent changes the bitrate parameter dynamically based on frame type. Reference frames use higher bitrate parameters to preserve fidelity, while non-reference frames use lower bitrate parameters. This parameter variation allows the system to achieve acceptable overall fidelity with reduced total bandwidth usage compared to uniform high-bitrate encoding.

Inventive Principle:
Principle #35Parameter changes

3Quantity of substance

If fewer bits are used for each frame, then transmission bandwidth is reduced, but encoding fidelity deteriorates

Engineering Contradiction:
Improvetransmission bandwidthVSAvoidencoding fidelity
Core Design Contradiction:
Quantity of substanceVSMeasurement precision

Solution Approach 1:

The patent applies local quality by ensuring that reference frames maintain high fidelity through higher bitrate allocation, while non-reference frames accept lower fidelity. This selective quality preservation ensures that the most critical data (reference frames needed for prediction) maintains accuracy, while less critical data uses fewer bits, achieving overall acceptable fidelity with reduced total bandwidth.

Inventive Principle:
Principle #3Local quality

Solution Approach 2:

The patent uses dynamic bitrate allocation where the fidelity level adjusts based on frame type. The system dynamically determines which frames require high fidelity (reference frames) and which can tolerate lower fidelity (non-reference frames), optimizing the balance between total bandwidth usage and overall encoding quality.

Inventive Principle:
Principle #15Dynamics

Data Source

PatentEP4497133B1Bundled multi-rate feedback autoencoder
Publication Date: 2026.01.21 QUALCOMM INC
  • EP4497133B1 patent drawingFigure 1
  • EP4497133B1 patent drawingFigure 2
  • EP4497133B1 patent drawingFigure 3

AI summary

A method includes generating an input data state for each data sample in a time series of data samples of a portion of an audio data stream. The method also includes providing at least one input data state to a first bottleneck and at least one other input data state to a second bottleneck. The first bottleneck is associated with a first bitrate and the second bottleneck is associated with a second bitrate. The method further includes generating a first encoded frame based on a first output data state from the first bottleneck and a second encoded frame based on a second output data state from the second bottleneck. The first encoded frame and the second encoded frame are bundled in a packet.