[mocaptoolbox] Re: Uncentered mccenter?
- From: Toiviainen Petri <petri.toiviainen@xxxxxx>
- To: "<a.r.jensenius@xxxxxxxxxx>" <a.r.jensenius@xxxxxxxxxx>
- Date: Mon, 9 Jan 2012 11:44:39 +0000
Hi Alexander,
mccenter calculates, for each of the three dimensions, the temporal mean of
each marker and subsequently the mean of these means across markers. This
yields a 3-component vector that is the mean point across all markers and all
time points. This vector is then subtracted from all data points.
Your example sets the mean for each marker to [0 0 0], which basically means
that the information about the relative positions between markers is lost.
I hope this helps.
Petri
Alexander Refsum Jensenius kirjoitti 9.1.2012 kello 12.17:
Hi again,
I have some problems understanding exactly what mccenter does. From the
description it sounds like it should center data around [0 0 0] for each axis.
For some reason that does not happen with the datasets I am working on. I am
not sure if this is a bug or if I misunderstand something. What I would imagine
is that mccenter should produce something similar to this (for a dataset a):
a_mean = mean(a.data);
a_center = bsxfun(@minus,a.data,a_mean);
Am I missing something?
Best,
Alexander
------------------------------------------
Petri Toiviainen
Professor of Music
Finnish Centre of Excellence in Interdisciplinary Music Research
Department of Music
PO Box 35(M)
40014 University of Jyväskylä
Finland
ptoiviai@xxxxxxxxxxxxx<mailto:ptoiviai@xxxxxxxxxxxxx>
http://users.jyu.fi/~ptoiviai
http://www.facebook.com/petri.toiviainen
------------------------------------------
Other related posts:
