Audio Fingerprinting for Pitch, Time, and Resampling Changes
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional fingerprint and signature generation techniques are not robust enough to detect media that has been pitch shifted, time shifted, or resampled, as they rely on adjusting audio samples to compensate for these alterations, which is inefficient and affects media identification accuracy.
Innovation Solution
Generate sample fingerprints and signatures that can adjust bin values and frame numbers to accommodate for pitch shifts, time shifts, and resampling ratios, allowing direct detection of altered media without the need for compensatory processing.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional fingerprint generation techniques adjust audio samples to compensate for pitch shifts, time shifts, and resampling, then media identification can be performed, but detection accuracy deteriorates and processing efficiency decreases
Solution Approach 1:
Instead of adjusting the audio sample to match a reference fingerprint, the patent inverts the approach by adjusting the reference fingerprint to match the altered audio sample. The system generates multiple candidate fingerprints from the reference audio at different pitch shifts, time shifts, and resample ratios, then compares these directly against the query fingerprint without requiring audio sample adjustment.
Solution Approach 2:
The patent applies parameter changes by systematically varying pitch shift values, time shift values, and resample ratios to generate multiple candidate fingerprints from the reference audio. This allows the system to account for various transformations applied to the media while maintaining high detection accuracy across different altered versions.
2Reliability
If conventional techniques adjust audio samples for pitch and time shifts, then media can be identified, but measurement precision deteriorates
Solution Approach 1:
The patent inverts the conventional approach by not adjusting the query audio sample but instead generating multiple candidate fingerprints from the reference audio at various transformations. This inversion preserves the integrity of the original audio sample while allowing flexible matching against transformed reference versions, thereby maintaining measurement precision.
Solution Approach 2:
The system performs preliminary action by pre-generating multiple candidate fingerprints from the reference audio at different pitch shifts, time shifts, and resample ratios before comparison. This preparatory step ensures that the correct matched fingerprint is already available in the candidate set, eliminating the need for iterative audio adjustments and preserving detection accuracy.
3Reliability
If conventional methods use compensatory processing for altered media, then identification is possible, but device complexity increases
Solution Approach 1:
The patent extracts the transformation parameters (pitch shift, time shift, resample ratio) from the complex audio processing pipeline and applies them directly to generate candidate fingerprints. This separation simplifies the overall system by removing the need for complex iterative audio adjustments while maintaining identification capability across transformed media.
Data Source
AI summary
Methods, apparatus, systems and articles of manufacture are disclosed to identify media that has been pitch shifted, time shifted, and/or resampled. An example apparatus includes: memory; instructions in the apparatus; and processor circuitry to execute the instructions to: transmit a fingerprint of an audio signal and adjusting instructions to a central facility to facilitate a query, the adjusting instructions identifying at least one of a pitch shift, a time shift, or a resample ratio; obtain a response including an identifier for the audio signal and information corresponding to how the audio signal was adjusted; and change the adjusting instructions based on the information.