Newsgroups: comp.speech
Path: pavo.csi.cam.ac.uk!pipex!uunet!mcsun!news.funet.fi!ousrvr.oulu.fi!phoenix!eha
From: eha@phoenix (Esa Haapaniemi)
Subject: AnimMan and others
Message-ID: <1992Oct2.123538.827@ousrvr.oulu.fi>
Summary: Amiga speech
Keywords: Amiga, animation, voice
Sender: news@ousrvr.oulu.fi
Organization: University of Oulu, Finland
X-Newsreader: TIN [version 1.1 PL6]
Date: Fri, 2 Oct 1992 12:35:38 GMT
Lines: 43

I'm somewhat interested in speach recognicion and computer generated speech.
As I'm mostly familiar with the programs on Commodore Amiga, and the
capabilities of it, I want to give some info to the new group.

Amiga: - Has 4 channel stereosound with up to 28 kHz sampling rate without
	 extra hardware. On accelerated Amigas that can be up to 56 kHz.
	 RS-323 port is capable of using many sound samplers available
	 for Amiga, and they are pretty cheap (PD documents available).
       - And I think most of you know that even the first Amiga had speech
	 synthetizing software on 1987. Now that has gotten a lot better...
	 The new SpeechToy2 is fully configurable text_to_speech program
	 that has a graphical interface, and suppports the new ECS and AGA
	 chips on Amigas.

Other software: There have been many different programs for Amigas to get
	 voice input/output. On -89 came out the first voice recognising
	 program that allowed computer controlling through spoken commands...
	 (for Amiga). As it wasn't very accurate, and was very limited in 
	 possibilities, it did't get much publicity. Mostly it was sold
	 as mouse/joystick replacement, and used by handicapped people to
	 control their equipment.
	 One demoprogram of Amigas capabilities in Speechsynthetizing is
	 AnimMan, that is a animated head, that speaks, learns many commands
	 trough some commercial soundsamplers, and does what it has been tought
	 to do (start programs, interface allowes really a lot of commands
	 to be given, and on multitasking machine it can run just well without
	 the need for specially written software). Also it can tell tales 
	 (user can write his own). And the program is really usable for such 
	 a small sice (about 300 k).
	 Also I have seen french, finnish and japanese speaking programs
	 on Amiga (all of them needeed some kind of text input, mostly
	 the direct text, and no phonems needeed). All of them (except
	 the french speaking program that one of my friends wrote) were
	 generating the speech totally by program. And the french spoken
	 program wasn't that good (though it could output the phonems
	 right from text) as it spoke like drunk with digitized sounds.

Maybe that's enough. Thanks for the bandwidth, and I hope to see more
quality software (PD or commercial) on my machine:)

Esa Haapaniemi
University of Oulu
Finlan
