Inter-tone Noise Reduction for Decoded Tonal Signals
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Low bit rate speech-specific codecs, such as CELP, fail to effectively model tonal music signals with non-harmonically related tones, resulting in audible quantization noise in low-energy spectral regions, known as inter-tone regions or spectral valleys.
Innovation Solution
A system and method that reduces quantization noise in low-energy spectral regions of decoded tonal sound signals by applying an inter-tone noise reduction technique using spectral analysis and gain correction, specifically designed for music signals encoded with speech-specific codecs, which involves preprocessing, spectral analysis, and post-processing to enhance the decoded sound signal.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If a low bit rate speech-specific codec using a speech production model is used to encode music signals, then the bit rate is reduced, but audible quantization noise appears in the low-energy regions of the spectrum (inter-tone regions)
Solution Approach 1:
The patent applies preliminary action by performing noise reduction in the spectral domain before final signal reconstruction. The system identifies spectral valleys (low-energy regions) ahead of time and applies targeted gain reduction to suppress quantization noise in these specific regions, preventing the noise from being audible in the final decoded signal while maintaining the low bit rate encoding scheme
Solution Approach 2:
The patent applies local quality by treating different spectral regions differently. Instead of applying uniform noise reduction across the entire spectrum, the system specifically identifies inter-tone regions (spectral valleys) and applies gain reduction only to these low-energy regions, while leaving the tonal regions (spectral peaks) unaffected. This localized approach reduces quantization noise where it is most audible without degrading the quality of the actual musical content
2Reliability
If a pitch filter and all-pole synthesis filter are used to model the speech signal, then the speech quality is maintained, but the tonal structure with non-harmonically related tones cannot be properly modeled
Solution Approach 1:
The patent applies parameter changes by modifying the spectral parameters in the frequency domain rather than relying solely on the time-domain parameters of the speech production model. The system performs spectral analysis to identify the tonal structure and spectral valleys, then adjusts the gain parameters specifically in the inter-tone regions to reduce quantization noise. This parameter adjustment in the spectral domain allows the system to handle non-harmonic tones effectively while maintaining compatibility with the existing speech production model
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A system and method for enhancing a tonal sound signal decoded by a decoder of a speech-specific codec in response to a received coded bit stream, in which a spectral analyser is responsive to the decoded tonal sound signal to produce spectral parameters representative of the decoded tonal sound signal. A quantization noise in low-energy spectral regions of the decoded tonal sound signal is reduced in response to the spectral parameters produced by the spectral analyser. The spectral analyser divides a spectrum resulting from spectral analysis into a set of critical frequency bands each comprising a number of frequency bins, and the reducer of quantization noise comprises a noise attenuator that scales the spectrum of the decoded tonal sound signal per critical frequency band, per frequency bin, or per both critical frequency band and frequency bin.