FORMAT OF S.T.L. SPEECH FILES

	FIRST BLOCK - RUBBISH, LEFT FOR ASCII DESCRIPTION OF FILE
		    -NOT YET IMPLEMENTED.
	SUCEEDING BLOCKS OF 12-BIT WORDS.   ALL BLOCKS FULL.
	SAMPLES ARE SIGNED INTEGERS STORED AS PLINTH NUMBERS, THAT
IS,4000 IS ADDED TO EACH NUMBER AND THE INPUT NUMBER 4000 IS 
REPRESENTED  IN THE SAME WAY AS 4001, I.E AS 1.   THIS
ELIMINATES 0000 IN THE FILES.   THIS CODE IS USED FOR SIGNALLING
END OF SPEECH INPUT BETWEEN THE TWO COMPUTERS DURING REAL TIME
TRANSFERS.
	FILES:
		X2DC.SP - INPUT SIGNAL, 0-5KHZ, RECORDED WITH
			HIGH QUALITY MIC.  IT DOES HAVE A LITTLE
			MAINS HUM.
		X2OUT.SP - OUTPUT OF BACK-TO-BACK ANALYSIS/SYNTHESIS
			SYSTEM WITH CORRELATION PITCH EXTRACTION,
			FRAMED COEFFICIENT ANALYSIS, AND HAMMING
			WINDOWS ON THE FRAMES.   FRAME RATE 100/S
			WINDOW WIDTH 20 MSEC.  SYSTEM SAMPLE RATE
			11KS/S.
			NOTE FAILURE OF PITCH EXTRACT DURING THE /M/
			OF "CHRISTMAS".
UT BETWEEN THE TWO COMPUTERS DURING REAL TIME
TRANSFERS.
	x