2021/03/01 by Benedikt Boenninghoff, Robert M. Nickel, Boenninghoff, Benedikt +5
Computer Science · #Advanced Data Compression Techniques #Audio and Speech Processing (eess.AS) #FOS: Computer and information sciences #FOS: Electrical engineering #Machine Learning (cs.LG) #Music and Audio Processing #Sound (cs.SD) #Speech Recognition and Synthesis #Speech and Audio Processing #electronic engineering #information engineering
paper · pdf · doi:10.48550/arxiv.2103.01173
openalex publication_date 2021/03/01 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28
The detection of voiced speech, the estimation of the fundamental frequency,\nand the tracking of pitch values over time are crucial subtasks for a variety\nof speech processing techniques. Many different algorithms have been developed\nfor each of the three subtasks. We present a new algorithm that integrates the\nthree subtasks into a single procedure. The algorithm can be applied to\npre-recorded speech utterances in the presence of considerable amounts of\nbackground noise. We combine a collection of standard metrics, such as the\nzero-crossing rate, for example, to formulate an unsupervised voicing\nclassifier. The estimation of pitch values is accomplished with a hybrid\nautocorrelation-based technique. We propose a forward-backward Kalman filter to\nsmooth the estimated pitch contour. In experiments, we are able to show that\nthe proposed method compares favorably with current, state-of-the-art pitch\ndetection algorithms.\n