[aprssig] An amusing aside (Text Pagers)

[Scott Miller]

> The aprs-tt would then either read back the packet to you (ie ai4px at 
> lat 34.xxx 80.xxx icon jogger, status off duty) or a simple "ai4px 
> packet accepted" in Majel Barret's voice of course. 
>  
> Back in the 1980's General Instruments made a text to speech chipset 
> that would speak when ascii words were fed to it.  Is there any 
> comparable chip available today?  Most things I see are the IVR store 
> and playback chips.

I found a (Winbond?) device a while back that did that.  I think Bob 
wound up using it on one of his satellite projects.  Not sure what the 
cost is.

The ADS-WS1 weather station, which uses the same basic hardware as the 
SR1 simplex repeater, does its voice generation with about 200 
pre-recorded snippets that it puts together as needed.  Alas, Majel 
Barrett is no longer with us (and would have been a bit out of my price 
range anyway) so the WS1 voice samples were done by Emily Zuzik.

http://www.youtube.com/watch?v=7vMIsYs59j8

> Dave VanHorn: very good idea of shortening the dead time between 
> digits.  It's one of those "why didn't I think of that" moments.

Works fine until you get a decoder that expects the Bellcore standard 40 
msec delay.  I'm pretty sure the SR1's decoder (Goertzel algorithm with 
7 filters) won't register a digit properly if it doesn't have sufficient 
dead time.  An immediate transition from one tone to another won't 
trigger an output, and the time constant of the filters means too short 
a gap won't really show up as a gap at all.

Scott
N1VG