Sampling Rate Converter Integer Phase Tracking for Accurate Resampling
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional sampling rate converters face inaccuracies in phase determination due to the use of fractional numbers, leading to resampling errors and distortion in up-sampled signals, particularly when dealing with infinite fractional conversion rates.
Innovation Solution
The techniques represent the phase of the output signal using non-approximated integer numbers, comprising an input sample index, integer phase, and fractional phase, allowing for precise tracking and reduction of phase errors by using three components represented as integers.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If fractional numbers are used to represent phase in conventional sampling rate converters, then the device can handle arbitrary conversion rates, but phase determination becomes inaccurate leading to resampling errors
Solution Approach 1:
The phase representation is segmented into three separate integer components: input sample index (tracks latest input sample), integer phase (tracks intermediate samples), and fractional phase (tracks timing difference). This segmentation allows each component to be represented precisely as an integer, eliminating the accuracy problems of fractional representation while maintaining the ability to handle arbitrary conversion rates through their combination.
Solution Approach 2:
The invention transitions from representing phase as a single fractional number to representing it as a three-dimensional vector of integer components. This dimensional change allows the system to maintain precision while handling arbitrary conversion rates, as the combination of the three integer components can represent any phase value without approximation errors.
2Measurement precision
If integer representation with three components is used to track phase, then phase determination accuracy improves, but device complexity increases
Solution Approach 1:
The three integer components work together in a self-service manner where the input sample index identifies the latest input sample, the integer phase identifies the relevant intermediate sample, and the fractional phase identifies the timing offset. This self-organizing structure eliminates the need for complex fractional arithmetic while maintaining high precision phase tracking.
Solution Approach 2:
The invention changes the parameter representation from continuous fractional values to discrete integer values across three different parameters (input sample index, integer phase, fractional phase). This parameter transformation maintains measurement precision while simplifying the underlying computational structure, as integer operations are inherently more precise and can be implemented more efficiently.
Data Source
AI summary
In general, this disclosure describes techniques for changing a sampling frequency of a digital signal. In particular, the techniques provide a more accurate way to determining a relative timing between a desired output sample and a corresponding input sample using a non-approximated integer representation of the relative timing. The relative timing between the desired output sample and corresponding input sample may be represented using a first component that identifies a latest input sample of the digital signal used to generate intermediate samples, a second component that identifies an intermediate sample, and a third component that identifies a timing difference between the desired output sample and the intermediate sample. Each of the components may be recursively updated using non-approximated integer values.