Martin,
I'd tip my hat to you, but my bald head would get cold. What you wrote is
way above me, but I'm still impressed. I have 2 choices here in windows:
Gold wave, and Scanner Recorder. One of these should suit my needs.
Your point about sample rate versus time is something I'd never thought of,
but it's very interesting.
73 Steve KW3A
----- Original Message -----
From: "Martin McCormick" <[log in to unmask]>
To: <[log in to unmask]>
Sent: Wednesday, October 20, 2004 12:45 PM
Subject: Re: Scanner recording
> I don't know if this is any help, but I wrote a vox recorder
> program that runs under Linux and has been working well. It compiles
> with gcc. Is there anything in Windows such as /dev/dsp? This is one
> of the standard audio devices. It opens and closes like any file
> except that when read from, it produces an endless stream of 8-bit
> audio at 8,000 samples per second. When written to, it produces audio
> in your sound card at 8,000 samples per second.
>
> My program sits there, monitoring the sound card and looking
> for samples that are either above or below the normal value for
> silence which, in 8-bit audio is either 127 or 128. Usually, it kind
> of hovers around 128 due to noise. When audio comes in, it opens an
> output file called vx followed by numbers indicating the starting time
> such as vx0410201121.raw. It just keeps adding to the file every time
> there is more audio until you either kill the process or do a kill
> -HUP which makes it close that file and start a new one.
>
> If Windows doesn't have anything like /dev/dsp, then this
> isn't going to help you. Also, the program puts itself in the
> background when running. I am sure that Windows programs can do that
> also, but I bet the mechanism isn't the same.
>
> Mainly, I found that plus and minus 3 or 4 counts away from
> silence is a good sensitivity level and about a quarter of a second
> works well as a delay.
>
> I am not familiar with programming in MS Windows, but the
> concepts you have to deal with if you want to roll your own vox
> program should be the same. You need to continuously receive the
> audio stream from the sound card and only store the sound plus enough
> delay to make speech sound fairly natural. One neat concept to keep
> in mind is that the sampling rate is an excellent time base in your
> program. If you want 8-k audio, 1/2 second = 4,000 samples of audio,
> etc.
>
> Martin McCormick WB5AGZ Stillwater, OK
> OSU Information Technology Division Network Operations Group
|