Audio Waveform Rendering with Antialiasing Filters
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Digital synthesizers face limitations in generating arbitrary audio waveforms at different frequencies, leading to reduced fidelity and aliasing distortion due to discrete-time signal conversion, which complicates the synthesis of continuous waveforms with no upper bound on frequency content.
Innovation Solution
Developing techniques to author and render arbitrarily shaped waveforms by determining partial series, generating wave tables with varying numbers of partial waveforms, and selecting appropriate versions for resampling to maintain fidelity and avoid aliasing, including the use of antialiasing filters and partial series processing to limit and optimize the number of partials based on sample rate and volume.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If digital synthesizers use discrete-time signal conversion to generate audio waveforms, then the audio signal can be stored and reproduced digitally, but the fidelity is reduced and aliasing distortion occurs
Solution Approach 1:
The system performs preliminary band-limiting of the waveform before sampling to remove frequency components that would cause aliasing. By pre-filtering the continuous waveform to eliminate frequencies above the Nyquist limit, the system prepares the signal in advance to avoid aliasing distortion during the discrete-time conversion process.
Solution Approach 2:
The patent introduces an antialiasing filter as an intermediary component between the continuous waveform generation and the sampling process. This filter acts as a mediator that selectively removes problematic high-frequency components while preserving the essential characteristics of the waveform, thereby maintaining fidelity during digital conversion.
2Adaptability or versatility
If arbitrary waveforms with no upper bound on frequency content are synthesized, then waveform versatility is improved, but aliasing distortion increases due to discrete-time conversion limitations
Solution Approach 1:
The system extracts and removes the harmful high-frequency components from the arbitrary waveform that would cause aliasing during sampling. By selectively taking out frequency components above the Nyquist limit through band-limiting, the system preserves waveform versatility while eliminating the source of aliasing distortion.
Solution Approach 2:
The patent dynamically adjusts the bandwidth parameter of the waveform based on the sampling rate and desired pitch. By changing the frequency content parameters of the waveform to match the capabilities of the digital system, the system maintains waveform versatility across different pitches while preventing aliasing by ensuring no frequency components exceed the Nyquist limit.
3Adaptability or versatility
If waveforms are resampled at different rates to change pitch, then frequency flexibility is improved, but fidelity is reduced due to discrete-time signal conversion
Solution Approach 1:
The system performs preliminary band-limiting of the waveform at each pitch change before resampling. By pre-adjusting the frequency content to match the new sampling rate's Nyquist limit, the system prepares the waveform in advance to maintain fidelity during pitch transitions and avoid aliasing distortion that would occur with straightforward resampling.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
Described herein are techniques for synthesizing waves (e.g., audio waveforms) at various frequencies. For example, techniques described herein may be used to render musical notes by generating audio waveforms. According to some embodiments, the techniques generate a wave table for an audio waveform based on a set of partial waveforms. The system determines a set of notes with a corresponding entry in the wave table and determines a set of partials for each note. The system renders, for each note, an associated waveform comprising the associated number of partials for the note from the set of partial waveforms.