Audio Codec Pitch Discrimination for Inter-Harmonic Noise Reduction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Transform-based audio codecs introduce inter-harmonic noise when processing harmonic audio signals, particularly at low delay and low bitrate, which affects the performance of these codecs, especially on highly tonal audio material, and existing solutions like LTPF and pitch-based PLC face challenges in accurately encoding and decoding pitch information, especially when harmonicity is not dominant.
Innovation Solution
An apparatus and method that discriminates between frames suitable for Long Term Post Filtering (LTPF) and those not, allowing for appropriate encoding and decoding decisions based on harmonicity, using a pitch estimator and signal analyzer to determine if pitch information should be encoded and used for LTPF or packet loss concealment, thereby optimizing the use of pitch information.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Object-affected harmful factors
If LTPF is applied to reduce inter-harmonic noise, then audio quality is improved, but device complexity increases due to additional processing requirements
Solution Approach 1:
The encoder performs preliminary analysis of the audio signal to determine harmonicity before encoding. This preliminary action allows the system to pre-decide whether LTPF should be applied, avoiding unnecessary processing complexity while ensuring LTPF is only applied when beneficial for reducing inter-harmonic noise.
Solution Approach 2:
The system changes the parameter of pitch lag encoding based on harmonicity detection. When harmonicity is detected, pitch lag is encoded to enable LTPF; when not detected, pitch lag encoding is skipped. This parameter change approach allows flexible adaptation to different audio signals without increasing fixed processing complexity.
2Object-affected harmful factors
If pitch information is encoded for LTPF, then audio quality is improved, but bitstream size increases
Solution Approach 1:
The system applies local quality by encoding pitch information only in specific frames where harmonicity is detected, rather than encoding it continuously. This selective encoding approach reduces the overall bitstream size while maintaining audio quality improvement through LTPF only when necessary.
Solution Approach 2:
The system uses partial action by encoding pitch lag only when needed for LTPF, rather than always encoding it. This partial encoding strategy reduces bitstream size while still providing sufficient pitch information for audio quality improvement in harmonic audio segments.
3Adaptability or versatility
If pitch detection is performed at decoder-side when pitch lag is not available, then adaptability is improved, but processing time increases
Solution Approach 1:
The encoder performs preliminary encoding of pitch lag information when harmonicity is detected. This preliminary action ensures that pitch information is available at the decoder without requiring additional pitch detection processing, thereby reducing processing time while maintaining adaptability to different audio conditions.
4Object-affected harmful factors
If LTPF is applied to highly tonal audio material, then audio quality is improved, but computational resources are consumed
Solution Approach 1:
The system applies local quality by detecting harmonicity in the audio signal and applying LTPF only to segments where it is beneficial. This selective application reduces computational resource consumption by avoiding unnecessary processing in non-harmonic segments while still improving audio quality where needed.
Solution Approach 2:
The system uses partial action by applying LTPF only to the extent necessary for reducing inter-harmonic noise in harmonic audio segments, rather than applying it universally to all audio data. This partial application strategy optimizes the balance between audio quality improvement and computational resource consumption.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
There are provided methods and apparatus and non-transitory memory units for encoding/decoding audio signal information. The encoder side may determine if a signal frame is useful for long term post filtering (LTPF) and/or packet lost concealment (PLC) and may encode information in accordance to the results of the determination. The decoder side may apply the LTPF and/or PLC in accordance to the information obtained from the encoder.