LINGUIST List 2.355

Sunday, 21 July 1991

FYI: Minnesota conf, Speech analysis program

Editor for this issue: <>

Directory

1. , Minnesota Conference on Language and linguistics
2. keith denning, Signalyze: a speech analysis program

		Seventeenth Annual
MINNESOTA CONFERENCE ON LANGUAGE AND LINGUISTICS
		October 24-26,1991
		University of Minnesota
		Celebrating the
25th ANNIVERSARY OF THE DEPARTMENT OF LINGUISTICS
		Hubert H. Humphrey Center
		301 19th Avenue South
		Minneapolis, Minnesota
SPONSORED BY: Departments of Linguistics, Anthropology, Classical and
Near Eastern Studies, Communication Disorders, Computer Science, East
Asian Languages and Literatures, English, French and Italian, Philosophy,
Russian and East Asian Languages and Literatures, South Asian and Middle
Eastern Languages and Literatures, Spanish and Portuguese, Speech
Communication; by the Center for Research in Learning, Perception, and
Cognition; and by the Institute for Child Development.
[The full program, including registration details, is available by
sending the message:
get minneapolis-91
to:
listservuniwa.uwa.oz.au
This file is NOT available from listservTAMVM1]

[The following notice descibes a commercial product. Although
we welcome such postings as a service to subscribers, the moderators
wish to make clear that LINGUIST does not endorse specific products.]
	re: Acoustic analysis of speech on the Macintosh
	SIGNALYZE 2.0 from InfoSignal
	Signalyze is not produced by a large company and does not
	advertise in magazines, but it offers more features than
	any of the common Macintosh speech analysis programs.
	It is also *much* cheaper than the MacSpeech Lab
	package everyone knows about.
	Signalyze 1.0 first came out in January of last year
	(1990). Since then, a much more powerful Signalyze 2.0 has
	been in the works. Its beta version is about to be released
	and the official release of Signalyze 2.0 is slated for
	September. The program is about 620 k compiled at this
	point. I'll shortly describe Signalyze 2.0 here.
	Signalyze's basic conception revolves around up to 100
	signals, displayed synchronously in HyperCard fashion on
	"cards". Everywhere it aims for top speed and ease of
	manipulation. Signalyze works entirely with the superior
	16-bit standard (unlike SoundEdit which is an 8-bit
	environment). 8-bit signals are automatically converted to
	the 16-bit standard. Signalyze 2.0 will have an extensive
	help feature. For international customers, Signalyze
	can be switched instantaneously into French or German.
	The program offers a full complement of signal editing
	features, quite a few spectral analysis tools, manual
	scoring tools, pitch extraction routines, a good set of
	signal manipulation tools, and extensive input-output
	capacity.
	* Signal editing: Cut, copy, paste, clear, taper
	transitions, averaged transitions.
	* Spectral analysis tools: Spectra and spectrograms at
	narrow-band, wide-band and extra wide-band, frequency zoom,
	settable pre-emphasis, inverse filter, log scale. There
	are also cepstra and cepstrograms and there will be an LPC.
	Spectrograms adapt to the current color/grayscale
	environment (256/16 colors and grayscales, B/W
	dithering). 256-grayscale spectrograms can be
	converted into 3D displays through the freeware program
	"Image". Color/grayscale palettes are user-adjustable.
	* Manual scoring: You can easily obtain and store
	durations, frequencies, amplitude differences (dB), and
	amplitude differences from spectra.
	* Pitch extraction: There are three routines, FFT-comb,
	autocorrelation and temporal structure analysis. All three
	have been heavily tested and improved since version 1.0.
	* Signal manipulation: Arithmetic transformations,
	recoding (up-down sampling, differentiation), splining,
	amplitude and RMS envelopes, zero-crossing
	* Input/Output.
	--File formats directly in Signalyze: Signalyze, MacSpeech
	Lab, AudioMedia/SoundDesigner II, SoundEdit/MacRecorder,
	SoundWave, three sound resource formats, ASCII-text. File
	formats through freeware program FileConverter (runs in
	background, converts all files of a folder at once): Same
	as above, plus AIFF and ADF (BLISS, CSRE) formats. TEXT
	format for manual scoring results: compatible with most
	spreadsheet and statistics programs. PICT format for saved
	spectrograms and other graphs: compatible with most
	graphics programs.
	--Sound I/O: Direct sound input from MacRecorder and
	similar devices, AudioMedia and AD IN, MacADIOS boards
	and devices. Sound output via Macintosh internal sound and
	MacADIOS boards and devices as well as through AudioMedia's
	8-bit audio driver.
	Price and ordering information for Signalyze 2.0:
	Individual licence US$350, site license US$500. In the
	U.S., Canada and Latin America, contact InfoSignal Inc.,
	3002 Belvidere S.W., SEATTLE, WA., 98126, USA,
	FAX/Answering machine: (206) 935-1618. Other countries
	contact InfoSignal Inc., Rue de la Dime 80, CH-2000
	NEUCHATEL, Switzerland, FAX/Answering machine:
	+41 38 33.11.53,
	Email: 76357.1213COMPUSERVE.COM.
	Eric Keller (the developer of Signalyze) holds a
	Ph.D. in Linguistics and is professor for computer
	science at the University of Lausanne, Switzerland.
	He's also published in the phonetic sciences for
	the last 12 years. Also, he is in constant contact with
	several major speech labs to verify improvements to
	the program. A preliminary version of Signalyze 2.0
	was class-tested this spring in a phonetics class
	of final-year linguistics students.
	To obtain more information, send your postal
	address to the Neuchatel address, above. You'll be
	mailed some more material, so you can see for
	yourself.
	***************************
	Best,
	Keith