Audio Fingerprinting Research Articles

Programa de atenção a alunos precoces com comportamento de superdotação: identificação e proposta de enriquecimento musical

O presente artigo tem por objetivo apresentar o processo de identificação musical e a proposta de enriquecimento musical que orientaram os trabalhos desenvolvidos pela oficina de música do Programa d Atenção a alunos Precoces com Comportamento de Superdotação da Faculdade de Filosofia e Ciências da Universidade Estadual Paulista - UNESP campus de Marília/SP, que acontece uma ves por semana matutino e vespertino. Por meio de testes psicométricos Primary Measures of music Audition (PMMA) (GORDON, 1986a), Intermediate Measures of Music Audition (IMMA) (GORDON, 1986b) e da analíse estatistica (Wilcoxon) pretende-se apresentar o processo de identificação musical que revelará os estudantes do PAPCS que apresentam precocidade e comportamento de superdotação musical e, a partir do modelo de enriquecimento, segundo Renzulli (2004), atrelado as metodologias da área da Educação Musical, pretende-se apresentar a proposta de enriquecimento musical que embasou os trabalhos da oficina de música entre 2014 e 2015 no PAPCS, com o objetivo de atender todas as crianças identificadas. Ressalta-se que este trabalho é parte de uma pesquisa maior.

(636) - In Vitro Acoustic Fingerprint of an Axial Cone Bearing Left Ventricular Assist Device

(636) - In Vitro Acoustic Fingerprint of an Axial Cone Bearing Left Ventricular Assist Device

A low-complexity audio fingerprinting technique for embedded applications

This paper proposes a new optimized audio-based fingerprinting technology for embedded applications. The target use case is related to TV content synchronization and its numerous applications for Social TV and second screen applications. The proposed technology can be used for automatically identifying the program being watched by capturing the sound of the TV set. It can also be used to know which program is being watched and to precisely estimate the timestamp of the currently broadcast moment with respect to the beginning of the program. This is very useful for second screen applications where notifications (e.g. quizzes, additional information, commercials) have to be sent to viewers with a perfect synchronization relatively to the broadcast TV program. The robustness of the proposed technique is first evaluated on a large music database and then by considering a realistic use case where this technology is embedded in a smartphone.

Compensating for ear-canal acoustics when measuring otoacoustic emissions.

Otoacoustic emissions (OAEs) provide an acoustic fingerprint of the inner ear, and changes in this fingerprint may indicate changes in cochlear function arising from efferent modulation, aging, noise trauma, and/or exposure to harmful agents. However, the reproducibility and diagnostic power of OAE measurements is compromised by the variable acoustics of the ear canal, in particular, by multiple reflections and the emergence of standing waves at relevant frequencies. Even when stimulus levels are controlled using methods that circumvent standing-wave problems (e.g., forward-pressure-level calibration), distortion-product otoacoustic emission (DPOAE) levels vary with probe location by 10-15 dB near half-wave resonant frequencies. The method presented here estimates the initial outgoing OAE pressure wave at the eardrum from measurements of the conventional OAE, allowing one to separate the emitted OAE from the many reflections trapped in the ear canal. The emitted pressure level (EPL) represents the OAE level that would be recorded were the ear canal replaced by an infinite tube with no reflections. When DPOAEs are expressed using EPL, their variation with probe location decreases to the test-retest repeatability of measurements obtained at similar probe positions. EPL provides a powerful way to reduce the variability of OAE measurements and improve their ability to detect cochlear changes.

Performance differences between electroacoustic and electric alone cochlear stimulation using complex tests in noise. A pilot study

Objectives: Hybrid cochlear implants enable very interesting performances in bisyllabic words recognition, mainly in noisy conditions. In this paper, we studied patients with implants performing some very difficult timbric tasks in quiet and noisy conditions.Methods: We studied three adult patients with hybrid implants (group A-hybrid implants) and three adult patients with traditional ones (group B-traditional implants). We made the following tests (in quiet and noise): bisyllabic words recognition; voice kind recognition; common melodies identification and recognition; musical instrument identification and recognition.Results: Mean results in quiet: test 1: 35% A, 47% B; test 2: 92% A, 42% B; test 3: identification scores 88% A, 100% B, recognition scores: 37.5% A, 50% B; test 4 identification scores 50% A, 20% B, recognition scores 30% A, 20% B. As for tests in noise: test 1: 28% A, 46% B; test 2: 75% A, 25% B; test 3: identification 81% A, 87.5% B and recognition 56% A, 50% B; test 4: identification 60% A, 30% B and recognition 50% A, 20% B.Conclusion: Patients with hybrid implants seem to have better results, mainly in noisy timbric tasks. This could be explained by the better quality of low frequency stimulation and could be an encouragement for further studies and a larger diffusion of this kind of implant in the future.

In Operando Structural and Acoustic Analysis of Lithium-Ion Batteries

Previously, we presented proof-of-concept results that demonstrated the promise of electrochemical-acoustic analysis as a scaleable method for in operando battery diagnostics. This technique is based on the redistribution of mass (density) and rearrangement of chemical bonds (modulus) that are fundamental to the operation of all closed, fixed-volume batteries. Using standard ultrasonic non-destructive testing hardware, we showed that for all closed batteries, each cell type (i.e., each chemistry and construction combination) has a unique acoustic fingerprint during cycling. These fingerprints display strong correlations between the acoustic properties of batteries and their state of charge (SOC). Here, we report on results gathered at beam line 6-BM-A at the Advanced Photon Source at Argonne National Laboratory. Simultaneous energy-dispersive x-Ray diffraction (EDXRD) and acoustic analysis measurements were made on several common types of commercial lithium-ion batteries, including LCO, LFP, and NCA. By comparing the in operando structural and acoustic data, we examine and discuss the physical underpinnings of these robust electrochemical-acoustic correlations.

An Inverse Model of Acoustics in Batteries for Determining Structural Changes, State of Charge and State of Health

There has been recent interest in using acoustic techniques as a diagnostic tool for batteries. Studies have shown that significant changes in the acoustic fingerprints of batteries occurs during cycling [1,2]. These shifts can be correlated with electrochemical and physical phenomena within the cell. Ideally, the goal is to determine battery properties based on the acoustic fingerprint of a cell without a prior understanding of the state of health or charge of the cell. Furthermore, we would seek to understand acoustic behaviour as cells degrade or approach catastrophic failure. Here we present modeling developed to understand acoustic and electrochemical interactions in a cell. A forward model of a cell subject to acoustic signals is presented, which captures some critical characteristics of the system. Then, an associated inverse method is presented, which seeks to back out critical material characteristics relevant to cell state of charge, state of health and potential failure modes. 1. A. G. Hsieh et al., Energy Environ. Sci., 8, 1569–1577 (2015).2. S. Bhadra, A. G. Hsieh, M. J. Wang, B. J. Hertzberg, and D. A. Steingart, J. Electrochem. Soc., 163, A1050–A1056 (2016).

Real-Time Collection, Visualization, and Processing of Electrochemical-Acoustic Data from Battery Cycling Experiments

In all batteries, the distributions of density and elastic modulus change as the cells are charged and discharged. As speed of sound in a material is a strong function of both its density and modulus, the behavior of an acoustic wave passing through a battery will therefore change during the cycle. In an earlier publication, electrochemical-acoustic analysis was demonstrated both experimentally and computationally as a simple-yet-powerful technique utilizing this phenomenon, providing unique in operando characterization of the physical dynamics within an operating battery in real-time.To date, testing of a wide range of battery chemistries and form factors has shown that the instantaneous acoustic fingerprint of a cell is strongly and consistently correlated to its state of charge (SOC), and the time-resolved acoustic behavior of a cell during cycling is unique to both its chemistry and form factor. This technique quickly generates large data sets (particularly with multiplex testing) that are rich with information and require proper strategies and algorithms for analysis, not only to determine useful SOC correlations but also to perform the inverse analysis and deconvolute the dynamic acoustic waveforms to extract layer-by-layer physical information. Additionally, spectral analysis and signal processing can be applied to these data sets in order to identify time-evolving physical phenomena, which would enable the use of this method for determining the state of health (SOH) of the cell. The development of such strategies and algorithms will allow the simultaneous analysis of a large number of cells, for example in an advanced battery management system, and will be the focus of this talk.

Application of audio fingerprinting to neonatal EEG.

A clinical neurophysiologist must recognize patterns in EEG signals to evaluate the health of a patient's brain activity. Rare or unusual patterns may take time to correctly identify. The ability to automatically assist this recall would be beneficial in ensuring that appropriate measures could be taken in a timely fashion. Audio fingerprinting is a method used to identify songs using only a snippet of the song. Fingerprints are extracted from a sub-section of the song and matched against a database of previously computed fingerprints. In this paper, a fingerprint quantization technique is implemented on neonatal EEG data to attempt to identify sections of EEG data when only seeing a sub-section of the data. The impact of signal distortions is investigated and results from a database of one hour recordings from 40 newborns are presented.

Monophonic Piano Music Transcription

This paper proposes a method for computational monophonic piano music transcription, which detects the pitches of piano music and thus to identify the corresponding musical notations. This computational music transcription method consists of two main algorithms, which are Onset Detection Algorithm and Pitch Detection Algorithm. The Onset Detection Algorithm involves sound wave filtering and sound wave segmentation. And the Pitch Detection Algorithm involves period determination, frequency computation and musical notation identification. These proposed algorithms adopt time-domain method and they are built based on the observation of characteristics of piano sound signal. The program is fast and simple to use, and able to output result with 88% accuracy. However, this music transcription method is limited to specific sound input only, that is the monophonic piano music with slow or average speed up to 120 crotchet beats per minute. It is because the performances of the algorithms are dependent on the threshold values set in the program. Therefore, further investigation and research have to be carried out in order to improve the performance of the program.

Fast Audio Fingerprinting System Using GPU and a Clustering-Based Technique

In this paper, we present our audio fingerprinting system that detects a transformed copy of an audio from a large collection of audios in a database. The audio fingerprints in this system encode the positions of salient regions of binary images derived from a spectrogram matrix. The similarity between two fingerprints is defined as the intersection of their elements (i.e. positions of the salient regions). The search algorithm labels each reference fingerprint in the database with the closest query frame and then counts the number of matching frames when the query is overlaid over the reference. The best match is based on this count. The salient regions fingerprints together with this nearest-neighbor search give excellent copy detection results. However, for a large database, this search is time consuming. To reduce the search time, we accelerate this similarity search by using a graphics processing unit (GPU). To speed this search even further, we use a two-step search based on a clustering technique and a lookup table that reduces the number of comparisons between the query and the reference fingerprints. We also explore the tradeoff between the speed of search and the copy detection performance. The resulting system achieves excellent results on TRECVID 2009 and 2010 datasets and outperforms several state-of-the-art audio copy detection systems in detection performance, localization accuracy and run time. For a fast detection scenario with detection speed comparable to the Ellis' Shazam-based system, our system achieved the same min NDCR as the NN-based system, and significantly better detection accuracy than Ellis' Shazam-based system.

Investigation of vibration impacts on HVAC transformer from HVDC system under monopole operation

When the high voltage direct current (HVDC) system operates in monopole ground-return mode, DC current flows through the neutral of AC power transformer with the neutral grounded which exacerbates the vibration of the transformer. During the field test of Yibin-Jinhua HVDC system in China, the vibration detection system is deployed to detect the vibration signal of the 500 kV transformer in Luzhou substation, which is close to the grounding substation. The spectrum of the transformer vibration signal is analyzed by means of the wavelet transform, through which the vibration acoustic fingerprint and three typical characteristic parameters, i.e., the distribution percentage of spectral energy, the energy ratio of odd and even harmonics, and the waveform distortion ratio, are extracted respectively. When the HVDC system current increases from 0 A to 5 kA, the impact of DC bias on the operation condition of the 500 kV transformer is investigated through the analysis of the variation of vibration acoustic fingerprint and three characteristic parameters. It is indicated that the energy radio of odd and even harmonics of the 500 kV transformer vibration signal decreases significantly with the increase of the current. In addition, the distribution percentage of spectral energy and the vibration acoustic fingerprint change significantly. This paper demonstrates that the effect of DC bias on the operating condition of the transformer can be accurately assessed by the proposed indices.

Scalable Discovery of Audio Fingerprint Motifs in Broadcast Streams With Determinantal Point Process Based Motif Clustering

In this paper, we study the scalable discovery of audio repetitive patterns/motifs in long broadcast streams, where two segments are said to be repetitive if their audio fingerprints are close to each other. In this task, as we are confined to handle limited variability, we can adapt an audio hashing technique, originally proposed for searching a given music clip in music tracks, to successfully devise a linear complexity similarity matching method with a new step of repeated interval formation. This is the first contribution of this paper. As the similarity matching is super fast and thus coarse, there are false alarms in the large number of pair-wise matches generated, which constitute a major source of noise. We propose applying subset selection to the original set of pair-wise matches based on determinantal point processes (DPPs), as a filtering step, to reduce the noise. The selected subset of pair-wise matches is then subjected to motif clustering. We successfully apply DPP-based subset selection to improve motif clustering, which has a nice property that favors both quality and diversity. This is the second contribution of this paper. The proposed method is thoroughly evaluated on a 9-hour real-world audio stream and is compared with several reference methods. The bootstrap technique is used for the significance test. It is shown that the similarity matching is computationally very efficient (above 100 times faster than real time), and the filtering step with DPPs can significantly improve the precision of motif discovery, without sacrificing the recall performance.

Audio Signal Identification and Search Approach for Minimizing the Search Time in Audio Fingerprinting Using Template Matching

Audio or music fingerprints can be utilize to implement an economical music identification system on a million-song library, however the system needs great deal of memory to carry the fingerprints and indexes. Therefore, for a large-scale audio library, memory imposes a restriction on the speed of music identifications. So, we propose an efficient music identification system which used a kind of space-saving audio fingerprints. For saving space, original finger representations are sub-sample and only one quarters of the original data is reserved. In this approach, memory demand is far reduced and therefore the search speed is criticalincreasing whereas the lustiness and dependability ar well preserved. Mapping audio information to time and frequency domain for the classification, retrieval or identification tasks presents four principal challenges. The dimension of the input should be considerably reduced; the ensuing options should be strong to possible distortions of the input; the feature should be informative for the task at hand simple. We propose distortion free system which fulfils all four of these requirements. Extensive study has been done to compare our system with the already existing ones, and the results show that our system requires less memory, provides fast results and achieves comparable accuracy for a largescale database.

Detecting ice noise in Arctic ambient noise recorded on a drifting vertical line array

From April to September 2013, a 600-m Distributed Vertical Line Array (DVLA) with 22 hydrophones drifted from the North Pole toward the Fram Strait. The DVLA recorded low-frequency ambient noise (1953.125 Hz sampling) for 108 minutes six days per week. Spectral analysis of the ambient noise data indicates the presence of acoustic transients, which include airgun and drilling sources as well as ice shear-induced broadband events and FM tones. Similar ice shear signatures in the Beaufort Sea were highly correlated with environmental variables leading to rifting [Kinda et al., J.Acoust. Soc. Am. 138, 2034–2045 (2015)], demonstrating the potential of acoustics to monitor rapid sea ice processes and pack ice breakup. Motivated by this study, a wavelet-based audio fingerprinting method is applied for identifying ice noise in the drifting DVLA data. The algorithm blindly groups spectral signatures, allowing the categorization of previously unrecognized signatures. The size of each category at points along the dr...

Removing effects of ear-canal acoustics from measurements of otoacoustic emissions

Otoacoustic emissions (OAEs) are the acoustic fingerprints of the inner ear—when carefully measured in healthy ears, their spectra, although highly individualized, remain stable over time. Thus, OAE changes usually indicate changes in cochlear function, e.g., due to efferent modulation, aging, noise trauma, and/or exposure to harmful agents. In humans, however, the reproducibility of OAE measurements is compromised by ear-canal standing waves at relevant frequencies. We show that even when stimulus levels are tightly controlled using methods designed to avoid standing-wave problems (forward-pressure-level calibration), distortion-product (DP)OAE levels vary by ~10–15 dB near half-wave resonant frequencies, depending on probe insertion depth (deep versus shallow). We propose a method, derived from a tube model of the ear canal, that separates the initial outgoing OAE pressure wave at the eardrum from reflected OAEs trapped in the residual ear-canal space. The emitted pressure level (EPL) represents the load-independent OAE level that would be recorded in an ideal, reflectionless canal. When DPOAE levels are converted to EPL, variability across insertion depths decreases by ~10 dB near half-wave resonant frequencies. EPL may provide a simple way to reduce confounding OAE variability across subjects and to improve the reliability of OAE measurements for detecting cochlear changes.

Robust Quad-Based Audio Fingerprinting

We propose an audio fingerprinting method that adapts findings from the field of blind astrometry to define simple, efficiently representable characteristic feature combinations called quads. Based on these, an audio identification algorithm is described that is robust to noise and severe time-frequency scale distortions and accurately identifies the underlying scale transform factors. The low number and compact representation of content features allows for efficient application of exact fixed-radius near-neighbor search methods for fingerprint matching in large audio collections. We demonstrate the practicability of the method on a collection of 100,000 songs, analyze its performance for a diverse set of noise as well as severe speed, tempo and pitch scale modifications, and identify a number of advantages of our method over two state-of-the-art distortion-robust audio identification algorithms.

Robust and Efficient Multiple Alignment of Unsynchronized Meeting Recordings

This paper proposes a way to generate a single high-quality audio recording of a meeting using no equipment other than participants’ personal devices. Each participant in the meeting uses their mobile device as a local recording node, and they begin recording whenever they arrive in an unsynchronized fashion. The main problem in generating a single summary recording is to temporally align the various audio recordings in a robust and efficient manner. We propose a way to do this using an adaptive audio fingerprint based on spectrotemporal eigenfilters, where the fingerprint design is learned on-the-fly in a totally unsupervised way to perform well on the data at hand. The adaptive fingerprints require only a few seconds of data to learn a robust design, and they require no tuning. Our method uses an iterative, greedy two-stage alignment algorithm which finds a rough alignment using indexing techniques, and then performs a more fine-grained alignment based on Hamming distance. Our proposed system achieves $>$ 99% alignment accuracy on challenging alignment scenarios extracted from the ICSI meeting corpus, and it outperforms five other well-known and state-of-the-art fingerprint designs. We conduct extensive analyses of the factors that affect the robustness of the adaptive fingerprints, and we provide a simple heuristic that can be used to adjust the fingerprint’s robustness according to the amount of computation we are willing to perform.

Using spindle noise to monitor tool wear in a turning process

A tool condition monitoring system can increase the competitiveness of a machining process by increasing the utilised tool life and decreasing instances of part damage from excessive tool wear or tool breakage. This article describes an inexpensive and non-intrusive method of inferring tool condition by measuring the audible sound emitted during machining. The audio signature recorded can be used to develop an effective in-process tool wear monitoring system where digital filters retain the signal associated with the cutting process but remove audio characteristics associated with the operation of the machine spindle. This study used a microphone to record the machining sound of EN24 steel being face turned by a CNC lathe in a wet cutting condition using constant surface speed control. The audio signal is compared to the flank wear development on the cutting inserts for several different surface speed and cutting feed combinations. The results show that there is no relationship between the frequency of spindle noise and tool wear, but varying cutting speed and feed rate have an influence on the magnitude of spindle noise and this could be used to indicate the tool wear state during the process.

Robust audio fingerprinting using peak-pair-based hash of non-repeating foreground audio in a real environment

In this paper, we propose a high-performance audio fingerprinting system used in real-world query-by-example applications for acoustic audio-based content identification, especially for use in hete...