Original Sound Unknown

[Gwenda Arguin]

Hearingis the most important organ of communication for humans. But the efficiency of this organ goes far beyond the mere transmission of linguistic information. Our sense of hearing helps us to orientate ourselves and we can filter out the relevant signal components from a multitude of sounds. A transfer of these abilities (localisation and separation) to technical systems could be applied in many areas of our lives. Especially the separation of the signals causes big problems. Using time-frequency representations, this project aims to develop methods for signal separation.

Licensing issue: These files are made available under the terms of the Creative Commons Attribution-NonCommercial-ShareAlike 2.0 license. The authors are Another Dreamer and Alex Q for music source signals and Hiroshi Sawada, Shoko Araki and Emmanuel Vincent for mixture signals.

The human ear perceives a multitude of different sounds every second. The proportion of interfering signals is particularly high in noisy environments and in crowds of people. Nevertheless, we can disentangle this mix of different sounds and concentrate on our conversation partner. We can also determine the origin of the sounds.

The difficulty of a technical realisation lies in the indeterminacy of the system. One knows only the sensor signals that can be recorded with a certain number of microphones. From this information the transmission matrix must be determined and the original signals reconstructed.

For about 10 years, intensive research has been conducted in this field. There are already various methods for the separation of unknown signals (blind source separation). However, the application of these methods is subject to various restrictions. Especially for real environments (reflections from objects, etc.) and an unknown number of signals there are almost no reasonable methods available.

The best results are obtained using time-frequency plots. Through this transformation, the convolution of the source signal with the specific room impulse response goes into a multiplication in the frequency domain. However, due to the instationality of the speech signals, only limited time periods can be considered. Especially the short-time Fourier transform is used very often. The disadvantage of this transformation is a rigid size of the frequency and time resolution. This problem can be solved by using other time-frequency representations, especially the Analytical Wavelet Packets. In addition, the influence of reflections can be reduced by clever preprocessing of the signals.

The project aims to improve signal separation methods and develop new algorithms. The improvement of separation results under real environmental conditions (reflections, unknown number of sound sources) is of special interest. In addition to a separation of the signals, a localization of the individual sources shall be enabled. The individual research topics within the project also result from these requirements.

We assembled approximately 230,000 National Ocean Service (NOS) bathymetric soundings from 39 lead-line and single-beam echosounder hydrographic surveys conducted from 1896 to 2005 in Norton Sound, Alaska. These bathymetry data are available from the National Geophysical Data Center (NGDC: ), which archives and distributes data that were originally collected by the NOS and others. While various bathymetry data have been downloaded previously from NGDC, compiled, and used for a variety of projects, our effort differed in that we compared and corrected the digital bathymetry by studying the original analog source documents - digital versions of the original survey maps, called smooth sheets. Our editing included deleting erroneous and superseded values, digitizing missing values, and properly aligning all data sets to a common, modern datum. We incorporated 3 multibeam surveys, and added an additional 6,992 single-beam soundings from the 2010 Northern Bering Sea bottom trawl survey to fill in where smooth sheet data was lacking. We proofed and digitized 312 cartographic features, comprised mostly of rocks and islets and also digitized 4,305 verbal sediment descriptors, and digitized or adapted 2,142 km of mainland and 837 km of island shoreline.

These data are not to be used for navigation. We consider this smooth sheet bathymetry, feature, shoreline and sediment compilation a fairly complete first draft. The Mean High Water (MHW) shoreline was hand-digitized off of georeferenced smooth sheets in a GIS.

IT Security and Contingency Plan for the system establishes procedures and applies to the functions, operations, and resources necessary to recover and restore data as hosted in the Western Regional Support Center in Seattle, Washington, following a disruption.

N2 - The paper presents the pch2csd project, focused on con- verting patches of popular Clavia Nord Modular G2 syn- thesizer into code of Csound language. Now discontinued, Nord Modular G2 left a lot of interesting patches for sound synthesis and algorithmic composition. To give this her- itage a new life, we created our project with the hope for being able to simulate the original sound and behavior of Nord Modular.

AB - The paper presents the pch2csd project, focused on con- verting patches of popular Clavia Nord Modular G2 syn- thesizer into code of Csound language. Now discontinued, Nord Modular G2 left a lot of interesting patches for sound synthesis and algorithmic composition. To give this her- itage a new life, we created our project with the hope for being able to simulate the original sound and behavior of Nord Modular.

All content on this site: Copyright 2024 Elsevier B.V. or its licensors and contributors. All rights are reserved, including those for text and data mining, AI training, and similar technologies. For all open access content, the Creative Commons licensing terms apply

3a8082e126