Audio Spectra Research Articles

Continuous speech recognition method for improving false alarm rates

A speech recognition method for detecting and recognizing one or more keywords in a continuous audio signal is disclosed. Each keyword is represented by a keyword template representing one or more target patterns, and each target pattern comprises statistics of each of at least one spectrum selected from plural short-term spectra generated according to a predetermined system for processing of the incoming audio. The incoming audio spectra are compared with the target patterns of the keyword templates and candidate keywords are selected according to a predetermined decision process. In post-decision processing, concatentation techniques, based upon a likelihood ratio test, for rejecting false alarms are disclosed. Post-decision processing can include also a prosodic test to enhance the effectiveness of the recognition apparatus.

Continuous speech recognition method

A speech recognition method for detecting and recognizing one or more keywords in a continuous audio signal is disclosed. Each keyword is represented by a keyword template representing one or more target patterns, and each target pattern comprises statistics of each of at least one spectrum selected from plural short-term spectra generated according to a predetermined system for processing of the incoming audio. The spectra are processed by a frequency equalization and normalizing method to enhance the separation between the spectral pattern classes during later analysis. The processed audio spectra are grouped into spectral patterns, are transformed to reduce dimensionality of the patterns, and are compared by means of likelihood statistics with the target patterns of the keyword templates. A concatenation technique employing a loosely set detection threshold makes it very unlikely that a correct pattern will be rejected.

Limitations of doppler spectral broadening in the early detection of carotid artery disease due to the size of the sample volume

In the present study the criteria of spectral broadening used in pulsed Doppler systems to diagnose minor disease in the cervical carotid arteries were evaluated in relation to the size of the sample volume. For this purpose the Doppler audio spectra as recorded in asymptomatic volunteers with an ATL Duplex scanner (relatively large sample volume) and a multi-gate pulsed Doppler system (relatively small sample volume) were compared with each other. The volunteers, randomly selected from the population registers, varied in age from 20–69 yr. Only in the common carotid artery at younger age, were normal spectra recorded, especially with the multi-gate system. In the internal carotid artery the number of spectra, supposed to be associated with low grade stenosis, increased significantly, especially when recorded with the Duplex scanner. More pronounced spectral broadening was found when the spectra were recorded distally in this artery. In the cervical carotid arteries normal spectra were found only when sampled in the middle of the bloodstream. These findings indicate that for the detection of vascular lesions in the cervical carotid arteries at an early stage of the disease on the basis of spectral broadening, small sample volumes and sampling in the centre of the vessel are required.

Local Doppler audio spectra in normal and stenosed carotid arteries in man.

Abstract In this study spectral analysis was performed of signals obtained with a focussing 5 MHz CW Doppler flowmeter along normal and stenosed carotid arteries. The position of the sound beam in the artery was determined using a Doppler imaging system which converts velocity signals into a map of the cervical carotid arteries. In normal carotid arteries maximum frequencies usually ranged between 2 and 4 kHz. The regular outline of the sonagram and the concentration of frequencies near the maximum frequency is indicative for plug flow and explains the smooth sound heard. In stenosed internal carotid arteries (degree: 65–74%) five identifiable zones could be detected. Upstream to the stenosis, generally normal maximum frequencies and sonagrams were found. Within the stenosis, increased frequencies with even spectral distribution and usually a regular outline of the sonagram were found, which is compatible with parabolic flow. Just distal to the stenosis, the outline of the sonagram became irregular and high maximum frequencies persisted but were associated with high amplitude low frequencies, producing a superimposed gruff quality sound. This quality is considered to represent artery wall vibration caused by high velocity turbulence. Approximately 2 cm distal to the stenosis lower maximum frequencies were found, but still higher than normal. The outline of the sonagram remained irregular. A fluttering quality sound was heard likely to be produced by low velocity turbulence. About 3 cm distal to the stenosis a normal distribution of blood velocities was found. It is suggested that the detection of flow disturbances along stenosed internal carotid arteries, which occur at relatively slight degrees of stenosis, may contribute to the early diagnosis of internal carotid artery disease.

A versatile audio spectrometer

A new method of obtaining audio power spectra is described. Advantages of the new method are speed of use and great frequency detail. The new method is based on the production of a which has a time continuous power spectrum. The composite signal method can be used with many types of signals such as speech, music, traffic noise, et cetera.