Research 
TPSD Philosophical problem 
Before we take the first step for TP research, we have to make some language
shared with other researchers, i.e. standard protocol. I propose TP standard data(TPSD) form for TP research, and as extended one, unnotated timing standard data form for general musical research. There is no problem when converting MIDI timing data into 100% time base timing data. It's only converting time base. But we can not begin to analize without interpreting each tone as notated timing. Row timing data cannot be analized as a music. Whenever we talk about advance or delay of a tone, we do so as e.g. a front eighth note. Why do we take a tone as eighth notes regardless of having considerable timing fluctuation? Why do we take one as front and another as back? Do we 'naturally' listen to it as such and such? If we say that, our research is not serious. There is a Philosophical problem. Or there may be some ways of scientific research to approach it. If you do not so keenly respond to the problem, you may be acustumed to score music. 
Thus, we strike a wall before we begin to analyze TP. Then, once we seriously recognize that is a big philosophical problem, let us set it aside and proceed for the time being. In short, TPSD is not raw data but with interpretation. 
For the time being, we dismiss noteoff timing, and use only noteon timing. MIDI timing is like, for example, t=(measure, beat, tick)=(2,3,360) =(2nd measure, 3rd beat, 360/480 of a beat) "Time base=480" means 1 beat is divided into 480 and timing is expressd from 0 to 479 Accordingly, tick=360 shows just timing of 3/4 of 1 beat, or 3rd sixteenth note. 
pf, pb 
(1)Transform timing into 100% base; (2,3,360) to
(2,3,360/480*100)=(2,3,75) Time base is depending upon a sequence software, e.g. 120, 240, 384, 480, 960, etc. We are so acustomed to 100% ratio and easy to imagine each ratio that we use time base=100 for analizing. tf: timing of front eighth note tb: timing of back eighth note Sequence of successive n pair of eighth notes is shown as tf(1),tb(1),tf(2),tb(2), ,tf(n),tb(n) pf: proportion of front eighth note; pf(n)=tb(n)tf(n) pb: proportion of back eighth note; pb(n)=tf(n+1)+1tb(n) 
When time is divided into front and back, pf:pb is the proportion of divided length. As a proportion of a human's body does, 60:40 and 75:50 are different proportion of front and back eighth notes. dp: dp(n)=pf(n)pb(n); difference between front and back proportion sp: sp(n)=pf(n)+pb(n); sum of front and back proportion The value of dp shows how they bounces. dp>>0; pf>>pb; bouncing bounce dp>0; pf>pb; normal bounce dp=0; pf=pb; even bounce dp<0; pf<pb; reversed bounce 
TP standard data (TPSD) of 8bars1 is the following. Five numbers of TPSD (tf,tb,pf,pb,bl) are about a pair of (front and back) eighth notes, succesively showing timing of them (tf,tb), length (proportion) of them (pf:pb) and a beat's length (bl=pf+pb). One metronome beat is equivalent to '100'. For example, an utterly exact pair of eighth notes are shown in TPSD as tf, tb, pf, pb, bl 0, 50, 50, 50, 100 i.e. timing of them are (0, 50) and proportion of them is (50:50) and a beat's length is 100. If they bounce utterly exactly into 2:1, TPSD is 0, 67, 67, 33, 100 . A number of tf with '' implys 'an advanced timing' which means the front eighth note is advanced from the metric beat timing '0'. 
