tad@ssc.UUCP (Tad Cook) (03/14/91)
The Oregon Graduate Institute of Science and Technology is building a huge database of voices as part of a project to develop voice recognition for US West directory assistance. They want to be able to classify sounds according to regional differences, and they need thousands of samples of speech to do this. Call 800-441-1037 (I assume this is nationwide ... it may not be) and follow the voice prompts. They will ask your last name, where you are calling from, and where you grew up, and then ask you to pronounce several words and recite the alphabet. It takes about two minutes. Tad Cook Seattle, WA Packet: KT7H @ N7ENT.#WWA.WA.USA.NA Phone: 206/527-4089 MCI Mail: 3288544 Telex: 6503288544 MCIUW USENET:...uw-beaver!sumax!amc-gw!ssc!tad or, tad@ssc.UUCP [Moderator's Note: The number works from Chicago, and I gave my specimen yesterday. It does take just a couple minutes, and it was fun. I suggest everyone provide a specimen, and be sure to mention on the tape you read about their efforts in TELECOM Digest! PAT]
roeber@cithe1.cithep.caltech.edu (Frederick Roeber) (03/15/91)
In article <telecom11.202.1@eecs.nwu.edu>, tad@ssc.UUCP (Tad Cook) writes: > The Oregon Graduate Institute of Science and Technology is building a > huge database of voices as part of a project to develop voice > recognition for US West directory assistance. > They want to be able to classify sounds according to regional > differences, and they need thousands of samples of speech to do this. > Call 800-441-1037 (I assume this is nationwide ... it may not be) and > follow the voice prompts. They will ask your last name, where you are > calling from, and where you grew up, and then ask you to pronounce > several words and recite the alphabet. How available will this information be? Gee, I can think of lots of *really* *neat* *uses* for a database of people's voices connected with their names, current cities, and cities of birth. (Re: current cities: I think most people will call from home or work.) Maybe I'm too paranoid, but were I able to call, I think I'd leave out my name. Frederick G. M. Roeber | CERN -- European Center for Nuclear Research e-mail: roeber@caltech.edu or roeber@cern.ch | work: +41 22 767 31 80 r-mail: CERN/PPE, 1211 Geneva 23, Switzerland | home: +33 50 42 19 44
hakanson@ogicse.ogi.edu (Marion Hakanson) (03/16/91)
In article <telecom11.202.1@eecs.nwu.edu> tad@ssc.UUCP (Tad Cook) writes: > The Oregon Graduate Institute of Science and Technology is building a > huge database of voices as part of a project to develop voice > recognition for US West directory assistance. Hey, neat, the speech guys here will be glad to know folks all over are calling in. I thought I'd contribute a little second-hand info on the (telephone) technology involved. First, I hear it was a, um, challenge to get a USWest 800-number into our location (GTE territory). I'm not sure how that was resolved. > [Moderator's Note: The number works from Chicago, and I gave my > specimen yesterday. It does take just a couple minutes, and it was > fun. I suggest everyone provide a specimen, and be sure to mention on > the tape you read about their efforts in TELECOM Digest! PAT] Uh, it's not a tape. They first had a scheme which used a modified answering machine wired up to a DSP board in a workstation. This didn't work too well, but along came an off-the-shelf magic box (I think they have two now) which has a phone connector on one "side" and a SCSI connector on the other. I hear there's a PC inside of it, and some DSP magic (16-bit sampling?), and it's made by a company called Gradient Technologies. If you want more details, I'll have to track down the folks in the know and let them type at you. About the only thing I have to add is that they sure do fill up a lot of disks with their speech samples. They also use up a lot of Exabyte tapes for backups, and they soak up plenty of SPARC cycles, as well. Marion Hakanson Domain: hakanson@cse.ogi.edu UUCP : {hp-pcd,tektronix}!ogicse!hakanson
vances@xenitec.on.ca (Vance Shipley) (03/19/91)
In article <telecom11.202.1@eecs.nwu.edu>, tad@ssc.UUCP (Tad Cook) writes: > The Oregon Graduate Institute of Science and Technology is building a > huge database of voices as part of a project to develop voice > recognition for US West directory assistance. > Call 800-441-1037 (I assume this is nationwide ... it may not be) and I just tried it and it works from here! (Ontario, Canada) Vance Shipley
wegeng@arisia.xerox.com (03/19/91)
> The Oregon Graduate Institute of Science and Technology is building a > huge database of voices as part of a project to develop voice > recognition for US West directory assistance. Before I donate my voice to this project, I'm curious about one thing. Is information that is obtained by having such a large database of voices going to be kept private, or instead made available to the public? I think that good voice recognition systems would benefit the public, and would be happy to donate my voice to a project that was working on an "open" system. However, if the Oregon project is strictly for the use of US West, then I'm reluctant to participate. While I'm more or less a nice guy, I'm not so nice that I would donate my services to a for profit company such as US West. Of course, I would quickly reconsider if US West were to offer to compensate me for my participation. :-) What's the background behind this project? Don