Date: Wed, 24 Dec 1997 09:20:29 -0800 From: Don Wilde <don@PartsNow.com> To: Amancio Hasty <hasty@rah.star-gate.com> Cc: multimedia@freebsd.org Subject: Re: Remote and Voice control Message-ID: <34A1445D.32FA@PartsNow.com> References: <199712132104.NAA06292@rah.star-gate.com>
next in thread | previous in thread | raw e-mail | index | archive | help
Amancio Hasty wrote: > Voice Control > > Don if you or anyone has a good pointer to limited voice control > lets hear it . > Sorry it's taken so long to get back to you, Amancio, but I can't give you pointers to existing code, so I'll have to explain it as best I remember. The little micro had 5 inputs with bandpass filters which it sampled as fast as it could. The filters were biased towards the low end of the spectrum, concentrated in the 2-6khz range, only one was higher, about 10khz. What the trick was is that we did five 8-bit integer FFT's on the streams to get multiple comparison samples. With our higher processing power, we ought to be able to do one FFT on the raw data and have a good comparison sample. You want to do this in the frequency domain once you have identified the beginning and end of your word. I believe I did mention that the VCR Voice unit needed to be trained for recognition? Since this is not connected speech, you just need to search for the beginning of audio in the sample and frame the command properly. Comparing the sample with a limited set of 5-10 commands should be no problem in realtime. Rotsa Ruck, Amancio. If I can do more, I'll be glad to. I have a USR voice modem, when I get my motherboard working again I'll plug it in and have a go at it. I only have a cheapie sound card, so I will have to do my work through the modem. -- oooOOO O O O o * * * * * * o ___ _________ _________ ________ _________ _________ ___==_ V_=_=_DW ===--- Don Wilde [don@PartsNow.com] [http://www.PartsNow.com ] /oo0000oo-oo--oo-ooo---ooo-ooo---ooo-ooo--ooo-ooo---ooo-ooo---ooo-oo--oo
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?34A1445D.32FA>