A Sawtooth Waveform Inspired Pitch Estimator (SWIPE) has been developed for processing speech and music. SWIPE is shown to outperform existing algorithms on several publicly available speech/musical-instruments databases and a disordered speech database. SWIPE estimates the pitch as the fundamental frequency of the sawtooth waveform whose spectrum best matches the spectrum of the input signal. A decaying cosine kernel provides an extension to older frequency-based, sieve-type estimation algorithms by providing smooth peaks with decaying amplitudes to correlate with the harmonics of the signal. An improvement on the algorithm is achieved by using only the first and prime harmonics, which significantly reduces subharmonic errors commonly found in other pitch estimation algorithms.

/pdf/swipe-a-sawtooth-waveform-inspired-pitch-estimator-for-rx8xljapsb.pdf

Swipe: a sawtooth waveform inspired pitch estimator for speech and music

This chapter analyzes the problem of data cleansing and the identification of potential errors in data sets. The differing views of data cleansing are surveyed and reviewed and a brief overview of existing data cleansing tools is given. A general framework of the data cleansing process is presented as well as a set of general methods that can be used to address the problem. The applicable methods include statistical outlier detection, pattern matching,clustering, and Data Mining techniques. The experimental results of applying these methods to a real world data set are also given. Finally, research directions necessary to further address the data cleansing problem are discussed.

Data Cleansing: A Prelude to Knowledge Discovery

[서평]「Understanding Networked Multimedia applications and technology」

Robust fundamental frequency estimation in adverse conditions is important in various speech processing applications. In this paper a new pitch detection algorithm (PDA) based on the autocorrelation of the Hilbert envelope of the LP residual [1] is compared to another well established algorithm from Goncharoff [2]. A set of evaluation criteria is collected on which the two PDA algorithms are compared. In order to evaluate the algorithms in adverse conditions a suited reference database was constructed. This reference database consists of parts of the SPEECON speech database [3] where recordings of 60 speakers were selected and manually pitch marked. The recordings cover several adverse conditions as noise in the car cabin and reverberations of office rooms. The evaluation highlights the good performance of the new algorithm in comparison but shows, that low SNR conditions and strong reverberation are still a demanding challenge for future pitch detection algorithms.

https://www.isca-speech.org/archive/sp2006/papers/sp06_083.pdf

Evaluation of Pitch Detection Algorithms in Adverse Conditions

Data mining is a new technology that helps businesses to predict future trends and behaviours, allowing them to make proactive, knowledge-driven decisions. When data mining tools and techniques are applied on the data warehouse based on customer records, they search for the hidden patterns and trends. These can be further used to improve customer understanding and acquisition. Customer Relationship Management (CRM) systems are adopted by the organisations in order to achieve success in the business and also to formulate business strategies, which can be formulated based on the predictions given by the data mining tools. Basically three major areas of data mining research are identified: implementation of CRM systems, evaluation criteria for data mining software and CRM systems and methods to improve data quality for data mining. The paper is concluded with a proposed integrated model for the CRM systems evaluation and implementation. This paper focuses on these areas, where there is need for more explorations, and will provide a framework for analysis of the data mining research for CRM systems.

Data mining research for customer relationship management systems: a framework and analysis

In this paper we are presenting a new formulation for determining the fundamental frequency. The evaluation of the pitch is a very difficult problem mainly because of the great variability and irregularity of the speech signals. The algorithm we are presenting is original so far as it relies on the implicit calculation of the autocorrelation of the temporal excitation signal. We have tested our algorithm on the Bagshaw database, created at the Centre for Speech Technology Research at Edinburgh, which is primarily dedicated to the evaluation of algorithms estimating the fundamental frequency of speech. The results of our experiments show that our approach is very reliable.

https://hal.inria.fr/inria-00098759/document

An Efficient F0 Determination Algorithm Based on the Implicit Calculation of the Autocorrelation of the Temporal Excitation Signal

In the current paper, we propose a new pitch tracking technique based on a wavelet transform in the temporal domain. Our algorithm is designed to determine the pitch frequency of the speech signal using a simple voicing decision algorithm. The pitch period is extracted from the cepstrum excitation signal processed by a wavelet transform; then the pitch contour is refined by thresholding and correction algorithms without any post-processing. The results obtained show that the proposed algorithm provides very good pitch contours compared to those furnished by the Bagshaw database.

/pdf/on-the-use-of-wavelets-and-cepstrum-excitation-for-pitch-2heusebu3v.pdf

On the use of wavelets and cepstrum excitation for Pitch Determination in real-time

In this paper, we present a recent algorithm for pitch detection based on an implicit circular autocorrelation of the glottal excitation signal. This algorithm operates in real time without the use of any post-processing technique. This article focuses on the correction of the pitch contours estimated and on the reduction in classification errors in speech signals using simple voicing decision techniques. To evaluate the performance of our algorithms, we used the Bagshaw and Keele databases. We show in this study that the sum of the percentage of the unvoiced errors and the percentage of the voiced errors, for the male Bagshaw corpus, reaches a very good score of 14.67. For the female corpus, our results are also competitive compared to other algorithms using the same database. Concerning the Keele database, we succeed to obtain very good gross pitch error, voicing decision error and F0 frame error rates, respectively, 0.44, 0.65 and 1.55 % in the whole corpus.

An overview of the CATE algorithms for real-time pitch determination

This paper presents a novel speaking-aid system for enhancing esophageal speech (ES). The method adopted in this paper aims to improve the quality of esophageal speech using a combination of a voice conversion technique and a time dilation algorithm. In the proposed system, a Deep Neural Network (DNN) is used as a nonlinear mapping function for vocal tract vector transformation. Then the converted frames are used to determine realistic excitation and phase vectors from the target training space using a frame selection algorithm. Next, in order to preserve speaker identity of the esophageal speakers, we use the source vocal tract features and propose to apply on them a time dilation algorithm to reduce the unpleasant esophageal noises. Finally the converted speech is reconstructed using the dilated source vocal tract frames and the predicted excitation and phase. DNN and Gaussian mixture model (GMM) based voice conversion systems have been evaluated using objective and subjective measures. Such an experimental study has been realized also in order to evaluate the changes in speech quality and intelligibility of the transformed signals. Experimental results demonstrate that the proposed methods provide considerable improvement in intelligibility and naturalness of the converted esophageal speech.

Enhancement of esophageal speech obtained by a voice conversion technique using time dilated Fourier cepstra

In this paper, we present our solution for the KDD CUP 2003 task 2 competition. Our approach is based on a data cleaning methodology using Perl scripts. These scripts contain regular expression for automatically extracting relevant information from the 35472 LaTeX texts. These expressions were optimized by statistical investigations on the texts. Our solution has permitted us to obtain 144,087 associations.

Joseph Di Martino

Papers

An Efficient F0 Determination Algorithm Based on the Implicit Calculation of the Autocorrelation of the Temporal Excitation Signal

On the use of wavelets and cepstrum excitation for Pitch Determination in real-time

An overview of the CATE algorithms for real-time pitch determination

Enhancement of esophageal speech obtained by a voice conversion technique using time dilated Fourier cepstra

A data cleaning solution by Perl scripts for the KDD Cup 2003 task 2