HW 7

1. 		a	Write a procedure “lsqlin(x,y)” that takes in x,y data 
			and returns “a,b”. 
			Test the procedure on some examples where you know the answer.
		b	Apply it to the data from butterfly effect homework:
			find the exponential separation rate, and
			plot the best straight line on your log plot of the data.

2.		While developing your program(s) for the following, do your testing on the
 		smallest usable audio extract. Only once you are convinced everything is
	 	working perfectly should you apply it to the whole audio track.
	
		a	Extract the precise time of every snare hit in the Chad Smith track.

		b	Generate a plot that illustrates the deviations from steadiness of his drumming, 
			and devise and compute some quantitative measures of any lack of steadiness, 
			both short-term and long-term.			
			Consider fitting a linear model by least-squares.


		c	Describe any noticeable glitches or sharp tempo-shifts that occur. 
			Are they associated with any identifiable events (musical or external) during the performance?

		d	Compare Chad Smith's steadiness with that of another drummer 
			(Billy Ward? Prof. Ringland's kid? - to be determined). 


As usual, the HW should be submitted as report containing everything in a single file.

Please present Problems 1 and 2 as separate "chapters" in your report.

I will continue to enforce the programming practice of 
"no data redundancy" by having the grader take
off points for each numerical datum that appears more than
once in a program. For example, if the maximum age in your
HW5 code is 110, you should set
amax:=110
and the number 110 should not appear anywhere else in
your code. As I said in class removes the danger
of mistakenly having different maximum ages in
different parts of the code, and also allows you
to change the maximum age very easily if you decide
you want to.