AI-Clear is an Artificial Intelligence-driven, Acoustics Fingerprints-based device that enables one to cleans the targeted voices from the directly measured signals that may have been contaminated by random and unsteady background noise. Applications of this innovative technology covers a wide range of industries ranging from the automotive, transportation, entertainment, manufacturing, military, homeland security, etc.
Every human being has a unique fingerprint. The same is true for the voices of every human being, the sounds generated by any machine, etc. By using AI-assisted deep learning process, we can train software to recognize a huge amount of characteristics of acoustics fingerprints, which can in turn make decisions automatically and continuously to discern who is talking. In this way, we can cleanse the directly measured signals and hear the target voice as if we were inside a fully anechoice chamber.
AI-Clear ⎯ Cleansing what you hear
Sound Viewer is a one-of-a-kind product that is available in the market and enables engineers to "see" sound and perform a comprehensive diagnosis and analysis of all vibro-acoustic characteristics of complex structures in non-ideal environment.
Sound Viewer is based on the theories of Smart SODAR (SOnic Detection And Ranging), modified HELS (Helmholtz Equation Least Squares) method, and BSS (Blind Sources Separation), making it a disruptive technology suitable for tackling a wide variety of complex airborne, structure-borne and aerodynamically generated noise problems under non-ideal test conditions.
The major functionalities of Sound Viewer include but not limited to: 1) locating a plurality of sound sources in 3D space simultaneously; 2) reconstructing all acoustic quantities on an arbitrarily shaped source surface and in 3D space; 3) performing 3D CAT (Computerized Axial Tomography) scan of the acoustic pressures; 4) evaluating STL (Sound Transmission Loss); 5) determining STP (Sound Transmission Path) through arbitrary material any arbitrarily shaped vibrating structures; 5) conducting order tracking analyses; 6) performing in-situ continuous machine health monitoring and in-line/end-of/line product QC (Quality Control) testing, etc.
The existing noise diagnosis tools can only provide a specific vibro-acoustic quantity at the measurement point, for example, the acoustic pressure spectrum and SPL (Sound Pressure Leve), the acoustic intensity spectrum and SIL (Sound Intensity Level), the normal surface acceleration, etc. The existing acoustic holography can the reconstruct the acoustic pressure distribution on a 2D plane. The existing source localization tools can only produce the bearing angles, but not the depth of a source.
In contrast, Sound Viewer can not only determine simultaneously the locations of a plurality of noise sources in 3D space (bearing angles and distances), but also reconstruct all the vibro-acoustic quantities such as the acoustic pressure spectra and SPL distributions, the acoustic intensity spectra and SIL distributions, the acoustic power spectra, the normal component of the velocity and acceleration on an arbitrarily shaped surface and in 3D space. In addition, Sound Viewer enables users to conduct a CAT scan of the acoustic pressure field in 3D space, evaluate STL (Sound Transmission Loss) spectra and levels, STP (Sound Transmission Path) through arbitrary material any arbitrarily shaped vibrating structures; conduct order tracking analyses, perform in-situ continuous machine health monitoring and in-line/end-of/line product QC (Quality Control) testing, etc.
Sound Viewer has the same frequency and dynamic ranges as those of measurement devices. For example, a measurement microphone has the frequency range of 20 – 20,000 Hz with 140 dB dynamic range, then Sound Viewer provides diagnosis and analysis results over the entire 20 – 20,000 Hz bandwidths with 140 dB dynamic range.
In particular, the spatial resolution of Sound Viewer in source localization is independent of the frequency.
Sound Viewer is suied for all kinds of complex airborne, structure-borne and aerodynamically generated sound fields. It is a comprehensive diagnosis and analysis system.
The hardware of Sound Viewer consists primarily of a 3D array of microphones with microphone tubes and an eight-channel data acquisition unit.
With a 3D array, users will be able to pinpoint the locations of a plurality of sound sources in 3D space simultaneously, including their ranges. Use of extendable microphone tubes makes the system versatile for tackling various noise issues, especially when measurement space is relatively confined, for example, inside an automobile passenger compartment.
Six channels are designated for microphone measurements. The remaining two channels are used as reference signals such as tachometer reading during order-tracking analyses of any rotating shafts, or input power signals during STL (Sound Transmission Loss) analyses, etc.
X-tractor is a one-of-a-kind product that is currently available in the market, which enables users to detect and extract the target information, for example, the sound pressure, structural vibration, etc., which are embedded in the directly measured data contaminated by random background noise and unspecified interfering signals.
The underlying theories of X-tractor are BBS (Blind Sources Separation), advanced signal processing techniques, patterns recognition, artificial intelligence, and neural network.
Using an hybrid approach, X-tractor enables users to perform data cleansing to eliminate random background noise and interfering signals, discern good products from defective ones, identify specific user-defined defects of target products, and predict the life expectance of the target product during in-line/end-of-line QC testing or the life expectance during continuous machine health monitoring.
X-tractor is an ideal tool for in-line/end-of-line QC testing, NDT (Nondestructive Testing), NDE (Nondestructive Evaluation), and MHM (Machinery Health Monitoring) to identify various product defects and structural failures in noninvasive in situ measurements on a factory floor, without the need of a special enclosure.
Currently, QC tests are often conducted under controlled environment such as inside an enclosure to acquire and analyze the characteristics of the measured signals, and compare them with those of pre-established benchmark database. Such an approach can be costly and greatly slow down the production cycles.
In contrast, X-tractor enables users to perform in situ QC tests, NDT, NDE, MHM, etc. without the need for a special enclosure.
In particular, X-tractor can not only discern good and bad products, but also depict types of user-defined defects in a product, and predict the life expectance of a machine. Therefore, X-tractor combines the procedures of testing and analyzing defects of products that are typically done in two steps under a controlled test environment for QC/NDT/MHM into a single step.
These salient advantages of X-tractor can significantly reduce the overall costs, increase the efficiency, and improve the product competitiveness for a company.
It is emphasized that X-tractor is uitable for testing products that are operated at both constant and variable speeds, for example, an electric motor, automobile transaxle, etc. during run-up or run-down speed processes.
The last but not least is the fact that X-tractor requires as few as one transducer, for example, a microphone or an accelerometer, to perform QC/NDT/MHM tests.
Before using X-tractor, it is necessary to train software by creating the database of the target signal characteristics of some given good and given bad products, respectively. Next, users can perform test runs to establish a second set of databases of the signals from some given good and given bad products under the actual working environment, respectively.
Once these databases are established, X-tractor perform in situ QC/NDT/MHM tests on a factory floor where random background noise level may be quite high.
The hardware of X-tractor includes:
1. Transducers to measure acoustic or vibration signals
2. One data acquisition unit with an USB cable and BNC cable
3. One USB power cable, 1.5 meters long
4. One laptop computer
5. One carrying case
X-tractor is easy to use, simple to install, convenient to carry, straightforward to understand.
(20190811) X-tractor Base Module (Single Channel)
(20190811) X-tractor Base Module (Duel Channel)
(20190811) X-tractor Advanced Module (Single Channel)
(20190811) (20190811) X-tractor Advanced Module (Duel Channel)
Copyright © 2024 Signal-Wise - All Rights Reserved.
Powered by GoDaddy