 |
PM |  | |
Quote | | |
Reply | | |
Full Topic | |
| Joined: | Jan 15th, 2008 |
| Location: | New Zealand |
| Posts: | 77 |
| Status: |
Offline
|
I should have clarified that Ricks' analysis was on one broadcast snippet, where there was an embedded message inside the middle of it, and a small padding of silence each side of the embedding. Not too disimilar to the metamorphosis of voices Ive heard on EVP, where an announcers voice has been altered to say something other than what he originally said.
The Speakjet chip used in speakjet "boxes" usually uses a PIC chip to interface the environment to the speakjet serial input, much akin to an A/D function, but where specific ascii commands are outputted to the speakjet i/p in response to environmental emf changes that are sensed by the front end of the PIC. Franks trial involved random analog voltage directly fed into the speakjet serial port, and sometimes sensible speech was outputted, so this sort of indicates that the random voltage was sometimes correlated to simulate the correct ascii commands and at the 9600 baud rate as well. The speakjet has a predefined range of ascii commands required for phoneme output as well as baud sync, length, pitch, etc. Ive got a speakjet chip, and am in the process of learning PIC programming.
regards,
JEFF