Bundled Multi-Rate Feedback Autoencoder for Audio Bitrate Allocation
Find Innovative SolutionsGenerate 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
Engineering Contradiction Analysis
1Productivity
If the same number of bits is allocated to each frame, then encoding simplicity is maintained, but transmission bandwidth efficiency deteriorates
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.
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.
2Measurement precision
If more bits are used for each frame, then data fidelity is improved, but transmission bandwidth consumption increases
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.
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.
3Quantity of substance
If fewer bits are used for each frame, then transmission bandwidth is reduced, but encoding fidelity deteriorates
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.
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.
Data Source
Figure 1
Figure 2
Figure 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.