WEBVTT

00:00:00.000 --> 00:00:00.570 align:middle line:90%


00:00:00.570 --> 00:00:03.810 align:middle line:84%
We have been looking at how to
record music related motion,

00:00:03.810 --> 00:00:06.140 align:middle line:84%
and to get data
streams from this.

00:00:06.140 --> 00:00:09.770 align:middle line:84%
And we've also seen
that motion and emotion

00:00:09.770 --> 00:00:12.120 align:middle line:90%
are closely connected.

00:00:12.120 --> 00:00:14.195 align:middle line:84%
So I'd like to
ask you Alexander,

00:00:14.195 --> 00:00:17.680 align:middle line:84%
do you think it's
possible to identify

00:00:17.680 --> 00:00:21.770 align:middle line:90%
human emotion from motion data?

00:00:21.770 --> 00:00:23.990 align:middle line:90%
I think the short answer is yes.

00:00:23.990 --> 00:00:26.950 align:middle line:90%
But it's very difficult to do.

00:00:26.950 --> 00:00:29.776 align:middle line:84%
But still I think it's important
to think about that we have

00:00:29.776 --> 00:00:30.900 align:middle line:90%
many different layers here.

00:00:30.900 --> 00:00:34.490 align:middle line:84%
So one thing is the actual
physical motion that we have

00:00:34.490 --> 00:00:37.100 align:middle line:84%
and that we can measure with
different types of motion

00:00:37.100 --> 00:00:38.340 align:middle line:90%
capture systems.

00:00:38.340 --> 00:00:41.040 align:middle line:84%
But then to understand more
about what actually happens

00:00:41.040 --> 00:00:44.570 align:middle line:84%
emotionally, how we would
perceive this in our brain,

00:00:44.570 --> 00:00:45.780 align:middle line:90%
it's a long way.

00:00:45.780 --> 00:00:47.510 align:middle line:84%
And it's many, many
layers in between.

00:00:47.510 --> 00:00:50.740 align:middle line:84%
So that's why I think, when we
work with this as researchers,

00:00:50.740 --> 00:00:52.180 align:middle line:84%
we need to do this
systematically

00:00:52.180 --> 00:00:54.950 align:middle line:84%
and to look at the
different layers separately.

00:00:54.950 --> 00:00:58.580 align:middle line:84%
And for example, go from the
continuous physical motion

00:00:58.580 --> 00:01:00.650 align:middle line:84%
to look at how we
perceive this as an action

00:01:00.650 --> 00:01:01.900 align:middle line:90%
with the beginning and an end.

00:01:01.900 --> 00:01:04.455 align:middle line:84%
And then from there again
it's possible to look

00:01:04.455 --> 00:01:07.340 align:middle line:84%
at kind of how we
experience these,

00:01:07.340 --> 00:01:10.111 align:middle line:84%
some of the more expressive and
emotional qualities in this.

00:01:10.111 --> 00:01:11.610 align:middle line:84%
And then obviously
here we also need

00:01:11.610 --> 00:01:15.490 align:middle line:84%
to combine both quantitative
methods and techniques,

00:01:15.490 --> 00:01:16.510 align:middle line:90%
but also qualitative.

00:01:16.510 --> 00:01:20.010 align:middle line:84%
Because of course the
interpretation of the data

00:01:20.010 --> 00:01:22.500 align:middle line:84%
is very much connected to
the cultural background,

00:01:22.500 --> 00:01:24.710 align:middle line:84%
the context that
this is situated

00:01:24.710 --> 00:01:26.420 align:middle line:90%
in, et cetera, et cetera.

00:01:26.420 --> 00:01:30.510 align:middle line:84%
So the long answer it's
difficult, and the short answer

00:01:30.510 --> 00:01:33.450 align:middle line:90%
is that yes, it's possible.

00:01:33.450 --> 00:01:35.010 align:middle line:84%
Ok, so then we
know that research

00:01:35.010 --> 00:01:39.520 align:middle line:84%
is going on, on how to identify
human emotion from motion.

00:01:39.520 --> 00:01:45.360 align:middle line:84%
But what do you think, is it
possible to identify emotion

00:01:45.360 --> 00:01:46.740 align:middle line:90%
from motion?

00:01:46.740 --> 00:01:49.670 align:middle line:84%
And in what way are
the two related?

00:01:49.670 --> 00:01:50.315 align:middle line:90%