WEBVTT 00:00:00.000 --> 00:00:01.190 align:middle line:90% 00:00:01.190 --> 00:00:05.240 align:middle line:84% Now we are standing in the music-related motion capture 00:00:05.240 --> 00:00:07.690 align:middle line:84% laboratory here at the Department of Musicology. 00:00:07.690 --> 00:00:11.800 align:middle line:84% And in this lab we are studying music-created body movements 00:00:11.800 --> 00:00:15.770 align:middle line:84% with different types of technologies and methods. 00:00:15.770 --> 00:00:20.450 align:middle line:84% There are different methods when we study music-related body 00:00:20.450 --> 00:00:21.280 align:middle line:90% movement. 00:00:21.280 --> 00:00:23.560 align:middle line:84% We typically work with camera-based systems 00:00:23.560 --> 00:00:25.040 align:middle line:84% and also with sensor-based systems. 00:00:25.040 --> 00:00:28.090 align:middle line:84% And today we're going to show you different types of these 00:00:28.090 --> 00:00:31.630 align:middle line:84% systems-- how we work with them, and how we also analyse 00:00:31.630 --> 00:00:33.040 align:middle line:90% the data. 00:00:33.040 --> 00:00:35.170 align:middle line:84% First, now, we're going to look at one system 00:00:35.170 --> 00:00:40.150 align:middle line:84% that we call an optical marker-based motion capture 00:00:40.150 --> 00:00:41.460 align:middle line:90% system. 00:00:41.460 --> 00:00:43.940 align:middle line:84% And here we have one of these cameras. 00:00:43.940 --> 00:00:47.370 align:middle line:84% We have many of these cameras around the space. 00:00:47.370 --> 00:00:49.700 align:middle line:84% And they each send out infrared light 00:00:49.700 --> 00:00:51.970 align:middle line:90% being reflected on the marker. 00:00:51.970 --> 00:00:54.390 align:middle line:84% Something like this that Christian has on. 00:00:54.390 --> 00:00:57.470 align:middle line:84% So when we work with this, we typically take on a suit, 00:00:57.470 --> 00:01:01.755 align:middle line:84% as we're going to do, and then we work in the space and record 00:01:01.755 --> 00:01:05.170 align:middle line:84% the body movement, and then we analyse it afterwards. 00:01:05.170 --> 00:01:06.400 align:middle line:90% So let's try. 00:01:06.400 --> 00:01:09.850 align:middle line:84% So the idea, now, is that we put these reflective markers 00:01:09.850 --> 00:01:11.700 align:middle line:84% on all the body joints so that we are 00:01:11.700 --> 00:01:14.559 align:middle line:90% able to track them separately. 00:01:14.559 --> 00:01:16.100 align:middle line:84% And then from the front and the back, 00:01:16.100 --> 00:01:19.090 align:middle line:84% it's also important to be able to differentiate 00:01:19.090 --> 00:01:19.960 align:middle line:90% the different sides. 00:01:19.960 --> 00:01:22.420 align:middle line:84% So we typically put, for example, one here, 00:01:22.420 --> 00:01:24.500 align:middle line:84% but also a symmetrical one so it's 00:01:24.500 --> 00:01:28.750 align:middle line:84% possible to see that this is the backside of the body. 00:01:28.750 --> 00:01:31.070 align:middle line:84% This optical motion capture system 00:01:31.070 --> 00:01:34.270 align:middle line:84% is extremely sensitive to motion. 00:01:34.270 --> 00:01:37.250 align:middle line:84% so that means that we need to calibrate it each time we use 00:01:37.250 --> 00:01:39.500 align:middle line:90% it to really get good results. 00:01:39.500 --> 00:01:41.540 align:middle line:84% So this is where my reference system 00:01:41.540 --> 00:01:45.790 align:middle line:84% will be because I record the x, y, and z positions in space. 00:01:45.790 --> 00:01:49.030 align:middle line:84% So there are three axes that we record, 00:01:49.030 --> 00:01:51.410 align:middle line:84% and they are based in this frame. 00:01:51.410 --> 00:01:53.960 align:middle line:84% And then I have this wand with markers on it 00:01:53.960 --> 00:01:56.780 align:middle line:84% and with a very specific length, so it's 00:01:56.780 --> 00:02:01.310 align:middle line:84% possible to measure exactly how good and precise the system is 00:02:01.310 --> 00:02:02.290 align:middle line:90% at any point in time. 00:02:02.290 --> 00:02:06.850 align:middle line:90% 00:02:06.850 --> 00:02:11.520 align:middle line:84% So now we are ready to start real recording. 00:02:11.520 --> 00:02:16.310 align:middle line:84% Here we can see all the images from each of the cameras, 00:02:16.310 --> 00:02:19.100 align:middle line:84% and I also have a 3D representation 00:02:19.100 --> 00:02:21.430 align:middle line:90% of the space in question. 00:02:21.430 --> 00:02:23.430 align:middle line:90% And then we are ready to record. 00:02:23.430 --> 00:02:25.165 align:middle line:90% 1, 2, 3, go. 00:02:25.165 --> 00:02:38.030 align:middle line:90% 00:02:38.030 --> 00:02:39.720 align:middle line:84% And as we see here in this overview, 00:02:39.720 --> 00:02:43.570 align:middle line:84% we get all the dots from all the different cameras, 00:02:43.570 --> 00:02:48.300 align:middle line:84% and then they can be combined into a 3D image like this. 00:02:48.300 --> 00:02:51.340 align:middle line:84% And the nice thing about such a system like we have here, 00:02:51.340 --> 00:02:54.320 align:middle line:84% we're able to move around and look 00:02:54.320 --> 00:02:57.860 align:middle line:84% at the movements of the body from different angles 00:02:57.860 --> 00:03:01.135 align:middle line:84% and perspectives, and even zoom in on details. 00:03:01.135 --> 00:03:04.120 align:middle line:90% 00:03:04.120 --> 00:03:09.210 align:middle line:84% So now we're looking at the recording of your movements. 00:03:09.210 --> 00:03:11.240 align:middle line:84% And if I start the playback here, 00:03:11.240 --> 00:03:14.840 align:middle line:84% we can take a look at the different markers. 00:03:14.840 --> 00:03:18.670 align:middle line:84% And the nice thing about such a motion capture system 00:03:18.670 --> 00:03:21.630 align:middle line:84% is that we're able to rotate around and look at this 00:03:21.630 --> 00:03:24.280 align:middle line:90% from different perspectives. 00:03:24.280 --> 00:03:26.710 align:middle line:84% Another thing is that we can also turn on trajectory 00:03:26.710 --> 00:03:31.040 align:middle line:84% so we can actually look at how the movements are changing 00:03:31.040 --> 00:03:31.650 align:middle line:90% over time. 00:03:31.650 --> 00:03:35.590 align:middle line:90% 00:03:35.590 --> 00:03:38.880 align:middle line:84% And this also makes it easy to isolate out, say, 00:03:38.880 --> 00:03:41.040 align:middle line:84% the vertical component of movement if we're 00:03:41.040 --> 00:03:44.730 align:middle line:84% interested in that, and also to calculate the velocity 00:03:44.730 --> 00:03:48.120 align:middle line:90% and acceleration rotation. 00:03:48.120 --> 00:03:49.080 align:middle line:90% Let's try to do that. 00:03:49.080 --> 00:03:52.280 align:middle line:84% So we can, for example, take out one of these markers 00:03:52.280 --> 00:03:56.180 align:middle line:84% and look at a plot of the set position, which 00:03:56.180 --> 00:04:04.480 align:middle line:84% is the up and down movement of this particular marker-- where 00:04:04.480 --> 00:04:07.590 align:middle line:84% you can see this repetitive pattern over time 00:04:07.590 --> 00:04:10.510 align:middle line:90% in the up and down moment. 00:04:10.510 --> 00:04:14.330 align:middle line:84% So to try to summarise then, from your experience, 00:04:14.330 --> 00:04:17.950 align:middle line:84% what are the good things about working in a lab like this, 00:04:17.950 --> 00:04:20.920 align:middle line:90% and are there any bad ones? 00:04:20.920 --> 00:04:23.810 align:middle line:84% Well the main advantage of working 00:04:23.810 --> 00:04:25.840 align:middle line:84% with this type of motion capture system 00:04:25.840 --> 00:04:30.060 align:middle line:84% is the high accuracy you get from the recordings. 00:04:30.060 --> 00:04:32.260 align:middle line:84% You're able to record the markers down 00:04:32.260 --> 00:04:36.300 align:middle line:84% to a 0.1 millimetre position, which is state of the art, 00:04:36.300 --> 00:04:36.830 align:middle line:90% really. 00:04:36.830 --> 00:04:38.038 align:middle line:90% And also at very high speeds. 00:04:38.038 --> 00:04:39.510 align:middle line:90% Very high speeds. 00:04:39.510 --> 00:04:41.840 align:middle line:90% Up to 500 Hz is no problem. 00:04:41.840 --> 00:04:45.760 align:middle line:90% Even more is actually possible. 00:04:45.760 --> 00:04:49.370 align:middle line:84% The downside is really that you need a big space 00:04:49.370 --> 00:04:52.850 align:middle line:84% to be able to do it-- preferably a fixed location 00:04:52.850 --> 00:04:56.420 align:middle line:84% where the motion capture system can be attached because it 00:04:56.420 --> 00:04:59.380 align:middle line:84% takes a long time to set it up, and to calibrate, 00:04:59.380 --> 00:05:03.530 align:middle line:84% and also even to put on the suit like this. 00:05:03.530 --> 00:05:07.330 align:middle line:84% So to try to summarise then, this type of motion capture lab 00:05:07.330 --> 00:05:11.470 align:middle line:84% is, in many ways, the best you can do when it comes to getting 00:05:11.470 --> 00:05:13.860 align:middle line:84% very, very good and accurate results, 00:05:13.860 --> 00:05:16.295 align:middle line:84% but if we want to be more flexible when we work 00:05:16.295 --> 00:05:17.670 align:middle line:84% with music-related body movement, 00:05:17.670 --> 00:05:20.330 align:middle line:90% we need to use other techniques.