WEBVTT 00:00:00.000 --> 00:00:02.400 align:middle line:90% [SQUEAKING] 00:00:02.400 --> 00:00:04.320 align:middle line:90% [RUSTLING] 00:00:04.320 --> 00:00:07.200 align:middle line:90% [CLICKING] 00:00:07.200 --> 00:00:10.147 align:middle line:90% 00:00:10.147 --> 00:00:12.230 align:middle line:84% MICHAEL SCOTT ASATO CUTHBERT: Hello, computational 00:00:12.230 --> 00:00:14.750 align:middle line:84% musicologist and music theorists. 00:00:14.750 --> 00:00:17.880 align:middle line:84% While you're working on your own representation formats, 00:00:17.880 --> 00:00:21.780 align:middle line:84% I want to introduce one representation format, indeed, 00:00:21.780 --> 00:00:24.170 align:middle line:84% the most commonly used representation 00:00:24.170 --> 00:00:27.630 align:middle line:84% format for common Western music notation scores. 00:00:27.630 --> 00:00:31.730 align:middle line:84% And that is MusicXML created by Michael Good around the year 00:00:31.730 --> 00:00:34.220 align:middle line:90% 2000. 00:00:34.220 --> 00:00:37.020 align:middle line:84% Looking over at the MusicXML website, 00:00:37.020 --> 00:00:39.930 align:middle line:84% we can see their version of Hello World! 00:00:39.930 --> 00:00:42.690 align:middle line:84% As you know, in programming traditions, 00:00:42.690 --> 00:00:44.720 align:middle line:84% it's become traditional to print out 00:00:44.720 --> 00:00:48.590 align:middle line:84% the word Hello, World to show how 00:00:48.590 --> 00:00:49.970 align:middle line:90% a programming language works. 00:00:49.970 --> 00:00:53.210 align:middle line:84% In music notation, it's become common, instead, 00:00:53.210 --> 00:00:58.460 align:middle line:84% to show how note middle C would be represented, 00:00:58.460 --> 00:01:00.060 align:middle line:84% whole note or something like that. 00:01:00.060 --> 00:01:05.220 align:middle line:84% And here's the example of that middle C whole note in MusicXML 00:01:05.220 --> 00:01:08.460 align:middle line:84% One of the things you can see in MusicXML, 00:01:08.460 --> 00:01:11.660 align:middle line:84% just like any XML or HTML, we have 00:01:11.660 --> 00:01:15.920 align:middle line:84% these things called tags which are enclosed in angle brackets. 00:01:15.920 --> 00:01:18.410 align:middle line:84% Then we have a content for the tag, 00:01:18.410 --> 00:01:22.160 align:middle line:84% and then the tag is closed with those same angle brackets 00:01:22.160 --> 00:01:24.950 align:middle line:90% with a slash in front. 00:01:24.950 --> 00:01:27.270 align:middle line:90% Tags can enclose other tags. 00:01:27.270 --> 00:01:31.110 align:middle line:84% So the score part encloses a tag called part-name, 00:01:31.110 --> 00:01:33.270 align:middle line:90% and then it's closed later. 00:01:33.270 --> 00:01:37.860 align:middle line:84% And tags can also have attributes, id equals P1. 00:01:37.860 --> 00:01:41.450 align:middle line:84% And we'll look at those in just a little bit. 00:01:41.450 --> 00:01:43.230 align:middle line:90% Let's scroll down to the note. 00:01:43.230 --> 00:01:48.720 align:middle line:84% So all of that is to set up what we might normally do for a note. 00:01:48.720 --> 00:01:51.000 align:middle line:90% So a note is a tag called note. 00:01:51.000 --> 00:01:53.910 align:middle line:84% Within it, there's a tag called pitch. 00:01:53.910 --> 00:01:56.460 align:middle line:84% Within that, there's a tag called step. 00:01:56.460 --> 00:01:59.700 align:middle line:84% And the step has a single letter. 00:01:59.700 --> 00:02:02.160 align:middle line:90% This is very similar to music21. 00:02:02.160 --> 00:02:06.890 align:middle line:84% In fact, I borrowed these concepts from here, 00:02:06.890 --> 00:02:10.289 align:middle line:84% so that the accidental would be encoded elsewhere. 00:02:10.289 --> 00:02:13.110 align:middle line:84% Then the octave is encoded as a single number, 00:02:13.110 --> 00:02:15.560 align:middle line:90% then the pitch is closed. 00:02:15.560 --> 00:02:18.720 align:middle line:84% After that, we have a duration tag. 00:02:18.720 --> 00:02:20.640 align:middle line:84% And we'll get back to that in a second. 00:02:20.640 --> 00:02:22.465 align:middle line:90% But it's duration 4. 00:02:22.465 --> 00:02:23.840 align:middle line:84% And then there's something called 00:02:23.840 --> 00:02:26.280 align:middle line:90% a type, which is different. 00:02:26.280 --> 00:02:29.690 align:middle line:84% And it says whole, so it says that this note's type is 00:02:29.690 --> 00:02:31.190 align:middle line:90% a whole note. 00:02:31.190 --> 00:02:36.420 align:middle line:84% We have the duration and the type encoded separately, 00:02:36.420 --> 00:02:40.250 align:middle line:84% because the type explains to notation software 00:02:40.250 --> 00:02:44.990 align:middle line:84% how to represent this note, what it should look like. 00:02:44.990 --> 00:02:51.060 align:middle line:84% And the duration, instead, tells it how long it is. 00:02:51.060 --> 00:02:56.450 align:middle line:84% So you could, in theory, put a duration of any number 00:02:56.450 --> 00:02:58.550 align:middle line:84% and a type of any number, and you 00:02:58.550 --> 00:03:00.500 align:middle line:84% could make things that looked very 00:03:00.500 --> 00:03:03.360 align:middle line:84% different from how they sounded or how they were represented. 00:03:03.360 --> 00:03:05.510 align:middle line:84% In practice, most notation software 00:03:05.510 --> 00:03:09.080 align:middle line:84% reads one or the other of these, so that's kind of too bad. 00:03:09.080 --> 00:03:14.750 align:middle line:84% The duration 4 does not, like in music21 represent necessarily 00:03:14.750 --> 00:03:16.500 align:middle line:90% four quarter notes. 00:03:16.500 --> 00:03:20.960 align:middle line:84% What it does is it refers back to the last tag 00:03:20.960 --> 00:03:23.640 align:middle line:84% within an attributes tag called divisions. 00:03:23.640 --> 00:03:28.640 align:middle line:84% And this says how many parts should a quarter 00:03:28.640 --> 00:03:30.330 align:middle line:90% note be divided up into. 00:03:30.330 --> 00:03:34.590 align:middle line:84% So in this case, because the only note is a whole note, 00:03:34.590 --> 00:03:37.200 align:middle line:84% we can just divide up the quarter note into one part, 00:03:37.200 --> 00:03:39.290 align:middle line:84% not divide it at all, and then say 00:03:39.290 --> 00:03:42.680 align:middle line:90% a whole note has four of these. 00:03:42.680 --> 00:03:46.350 align:middle line:84% But if we were going to later have an eighth note, 00:03:46.350 --> 00:03:48.270 align:middle line:84% then this would have to be at least 2, 00:03:48.270 --> 00:03:51.780 align:middle line:84% because the quarter note can be divided into two parts. 00:03:51.780 --> 00:03:55.880 align:middle line:84% If we might have 16th notes, it would need to be 4 because it 00:03:55.880 --> 00:03:57.600 align:middle line:90% can be divided into four parts. 00:03:57.600 --> 00:04:02.910 align:middle line:84% And if we were going to have 16th notes and triplets also, 00:04:02.910 --> 00:04:06.050 align:middle line:84% then it would need to be 12 because you can divide it 00:04:06.050 --> 00:04:08.770 align:middle line:90% into 12 different ways. 00:04:08.770 --> 00:04:12.190 align:middle line:84% In practice, these divisions usually end up 00:04:12.190 --> 00:04:15.190 align:middle line:90% being much higher numbers. 00:04:15.190 --> 00:04:19.700 align:middle line:84% Some people use 16, 3, 8, 4, which is a power of 2. 00:04:19.700 --> 00:04:23.890 align:middle line:84% I like using a high power of 2 multiplied by 3, 00:04:23.890 --> 00:04:25.940 align:middle line:84% multiplied by 5, multiplied by 7, 00:04:25.940 --> 00:04:30.190 align:middle line:84% so I can exactly represent triplets, quintuplets, 00:04:30.190 --> 00:04:32.210 align:middle line:90% septuplets, things like that. 00:04:32.210 --> 00:04:35.720 align:middle line:84% But you can change this division with each, 00:04:35.720 --> 00:04:38.620 align:middle line:84% I think, at least each measure, so that you can make sure 00:04:38.620 --> 00:04:42.010 align:middle line:84% that whatever you're going to be representing will be there. 00:04:42.010 --> 00:04:44.740 align:middle line:84% We can also see that there's a key signature, 00:04:44.740 --> 00:04:50.570 align:middle line:84% and that's marked in the number of fifths from C. 00:04:50.570 --> 00:04:52.220 align:middle line:84% So remember your circle of fifths. 00:04:52.220 --> 00:04:56.050 align:middle line:84% If you want a key signature of D major of two sharps, 00:04:56.050 --> 00:04:57.350 align:middle line:90% that would be 2. 00:04:57.350 --> 00:05:01.120 align:middle line:84% If you want B major, that should be 5. 00:05:01.120 --> 00:05:04.685 align:middle line:84% If you want F major, well, we're going the other way, 00:05:04.685 --> 00:05:06.310 align:middle line:84% so that's going to be negative 1, so we 00:05:06.310 --> 00:05:08.240 align:middle line:90% can represent keys that way. 00:05:08.240 --> 00:05:09.820 align:middle line:84% Time, you can probably figure out. 00:05:09.820 --> 00:05:12.100 align:middle line:90% Beats and beat type are 4/4. 00:05:12.100 --> 00:05:14.240 align:middle line:84% Could be called numerator and denominator. 00:05:14.240 --> 00:05:18.010 align:middle line:84% But it is one of these things about MusicXML 00:05:18.010 --> 00:05:26.660 align:middle line:84% that the format likes to show the musical meaning, 00:05:26.660 --> 00:05:28.850 align:middle line:84% not just how it's going to be laid out. 00:05:28.850 --> 00:05:30.320 align:middle line:90% We have the clef. 00:05:30.320 --> 00:05:31.880 align:middle line:84% Instead of saying it's treble clef, 00:05:31.880 --> 00:05:36.220 align:middle line:84% we say that it is a G clef on line 2. 00:05:36.220 --> 00:05:41.210 align:middle line:84% And so here is some of the basics of MusicXML 00:05:41.210 --> 00:05:45.980 align:middle line:84% You can go to musicxml.com and read more about this. 00:05:45.980 --> 00:05:49.420 align:middle line:84% We're going to look at a few other components of MusicXML 00:05:49.420 --> 00:05:50.360 align:middle line:90% before going on. 00:05:50.360 --> 00:05:56.170 align:middle line:84% But you're not necessarily need to learn everything about it. 00:05:56.170 --> 00:05:59.030 align:middle line:84% So if we're looking at how MusicXML works, 00:05:59.030 --> 00:06:02.140 align:middle line:84% we're going to look at a slightly more complex example 00:06:02.140 --> 00:06:04.010 align:middle line:90% than just a single quarter note. 00:06:04.010 --> 00:06:08.080 align:middle line:84% We have a vocal part and then a piano part 00:06:08.080 --> 00:06:11.530 align:middle line:90% represented on two staves. 00:06:11.530 --> 00:06:13.870 align:middle line:84% Here, we're using a divisions that's 00:06:13.870 --> 00:06:20.540 align:middle line:84% a higher number than 1, negative 3 for the key signature, 00:06:20.540 --> 00:06:22.178 align:middle line:84% and then we can explicitly encode 00:06:22.178 --> 00:06:23.470 align:middle line:90% that it's actually in minor So. 00:06:23.470 --> 00:06:24.640 align:middle line:90% We're not in E-flat major. 00:06:24.640 --> 00:06:26.955 align:middle line:90% We're in C minor, right? 00:06:26.955 --> 00:06:29.650 align:middle line:90% 00:06:29.650 --> 00:06:31.010 align:middle line:90% There's transposition. 00:06:31.010 --> 00:06:33.790 align:middle line:84% But we don't need to know that for this. 00:06:33.790 --> 00:06:37.360 align:middle line:84% The pitch element, the note element, I'm sorry to say, 00:06:37.360 --> 00:06:44.630 align:middle line:84% sorry, can also have ties on it and can have lyrics. 00:06:44.630 --> 00:06:48.010 align:middle line:84% We can say that this is an end of a syllable that ends "meil." 00:06:48.010 --> 00:06:50.570 align:middle line:90% And we end in extension. 00:06:50.570 --> 00:06:53.870 align:middle line:84% This is another thing you can do in XML-based formats. 00:06:53.870 --> 00:06:56.630 align:middle line:84% If you want to open a tag and immediately close it, 00:06:56.630 --> 00:06:58.740 align:middle line:84% you can just put a slash at the end, instead. 00:06:58.740 --> 00:07:01.300 align:middle line:90% 00:07:01.300 --> 00:07:06.170 align:middle line:84% We'll look at tied notes, even though it's just a little thing. 00:07:06.170 --> 00:07:10.840 align:middle line:84% And that is that there are two different parts 00:07:10.840 --> 00:07:14.110 align:middle line:84% in a tied note in MusicXML and for a number of things 00:07:14.110 --> 00:07:15.400 align:middle line:90% in MusicXML. 00:07:15.400 --> 00:07:21.380 align:middle line:84% There is the tie element which represents conceptually, 00:07:21.380 --> 00:07:22.940 align:middle line:90% is this note tie. 00:07:22.940 --> 00:07:26.660 align:middle line:84% And then there is the tied, with a D element. 00:07:26.660 --> 00:07:30.590 align:middle line:84% And that says, how do we draw the tie on the page. 00:07:30.590 --> 00:07:33.160 align:middle line:84% So you can draw a tie without having a note 00:07:33.160 --> 00:07:36.440 align:middle line:84% that sounds tied or is conceived of as tied. 00:07:36.440 --> 00:07:42.070 align:middle line:84% Or you can have sounds conceived of as tied but not drawn. 00:07:42.070 --> 00:07:44.390 align:middle line:84% And this shows up in other places, 00:07:44.390 --> 00:07:50.300 align:middle line:84% namely in accidentals because you can imagine a B-flat, 00:07:50.300 --> 00:07:53.170 align:middle line:84% you conceive of it as a B-flat, but you also 00:07:53.170 --> 00:07:54.890 align:middle line:90% put a flat sign in front of it. 00:07:54.890 --> 00:07:59.600 align:middle line:84% The next B in that same measure, you conceive of it is B-flat, 00:07:59.600 --> 00:08:01.850 align:middle line:84% but you don't put a flat sign in front of it. 00:08:01.850 --> 00:08:07.870 align:middle line:84% So there's a distinction between how we represent a note on paper 00:08:07.870 --> 00:08:09.260 align:middle line:90% and how we conceive of it. 00:08:09.260 --> 00:08:11.890 align:middle line:84% And this is very important, also. 00:08:11.890 --> 00:08:17.070 align:middle line:84% We can look over here at chords and see that a cord is simply-- 00:08:17.070 --> 00:08:18.300 align:middle line:90% here's a 3-note chord. 00:08:18.300 --> 00:08:19.350 align:middle line:90% We start with a note. 00:08:19.350 --> 00:08:24.170 align:middle line:84% It's a C in octave 4, And then we 00:08:24.170 --> 00:08:27.260 align:middle line:84% have the next note has this empty chord 00:08:27.260 --> 00:08:30.840 align:middle line:84% tag that says it's part of a chord of the previous note. 00:08:30.840 --> 00:08:32.780 align:middle line:84% So in MusicXML, if you want to know 00:08:32.780 --> 00:08:35.033 align:middle line:84% whether something is a chord, you 00:08:35.033 --> 00:08:36.450 align:middle line:84% won't know it from the first node. 00:08:36.450 --> 00:08:38.669 align:middle line:84% You'll only know it from the second note on. 00:08:38.669 --> 00:08:41.900 align:middle line:84% Here's an example where we have E-flat 00:08:41.900 --> 00:08:51.290 align:middle line:84% altered by negative 1 semitones, and then the next note G. 00:08:51.290 --> 00:08:56.840 align:middle line:84% Here's a little example of multi-part music, and especially 00:08:56.840 --> 00:09:01.850 align:middle line:84% this part where we have two different voices 00:09:01.850 --> 00:09:03.480 align:middle line:90% in the same measure. 00:09:03.480 --> 00:09:06.830 align:middle line:84% So we have two chords that can be represented as just 00:09:06.830 --> 00:09:10.500 align:middle line:90% a single set of note elements. 00:09:10.500 --> 00:09:12.930 align:middle line:84% And then we have a chord here and a chord here, 00:09:12.930 --> 00:09:15.270 align:middle line:84% but in between, we have this separate voice 00:09:15.270 --> 00:09:18.040 align:middle line:84% with separate stem directions and so on. 00:09:18.040 --> 00:09:21.630 align:middle line:84% One of the things that MusicXML does, 00:09:21.630 --> 00:09:23.770 align:middle line:84% is that when you get to a certain point, 00:09:23.770 --> 00:09:25.870 align:middle line:84% you're reading through all of your durations. 00:09:25.870 --> 00:09:27.820 align:middle line:84% They're all considered to be in a row. 00:09:27.820 --> 00:09:29.790 align:middle line:84% You get to maybe the end of the measure, 00:09:29.790 --> 00:09:34.290 align:middle line:84% and you can back up a certain number of divisions, 00:09:34.290 --> 00:09:36.070 align:middle line:84% and then you can go forward again. 00:09:36.070 --> 00:09:40.060 align:middle line:84% In this case, you would probably just back up 8. 00:09:40.060 --> 00:09:42.210 align:middle line:84% But a lot of notation software seems 00:09:42.210 --> 00:09:45.010 align:middle line:84% to only use backup to go to the beginning of the measure. 00:09:45.010 --> 00:09:46.900 align:middle line:84% And then if you need to skip forward, 00:09:46.900 --> 00:09:49.740 align:middle line:90% you use a forward again. 00:09:49.740 --> 00:09:53.800 align:middle line:84% And it is also customary, though not strictly required, 00:09:53.800 --> 00:09:57.510 align:middle line:84% I don't believe, but almost a lot of notation, 00:09:57.510 --> 00:10:01.930 align:middle line:84% things that read MusicXML, we call them consumers of MusicXML. 00:10:01.930 --> 00:10:03.960 align:middle line:84% A lot of them might have a problem 00:10:03.960 --> 00:10:07.690 align:middle line:84% or might crash if you don't include what's customary, again, 00:10:07.690 --> 00:10:09.720 align:middle line:84% then you end at the end of the measure 00:10:09.720 --> 00:10:13.770 align:middle line:84% by going forward until the very end. 00:10:13.770 --> 00:10:16.180 align:middle line:84% There's all kinds of other things on repeats, 00:10:16.180 --> 00:10:17.830 align:middle line:90% on sound suggestions, and so on. 00:10:17.830 --> 00:10:19.290 align:middle line:84% But one of the things you can see 00:10:19.290 --> 00:10:21.570 align:middle line:84% is that, compared to some of the thoughts 00:10:21.570 --> 00:10:24.610 align:middle line:84% we had on compact representations, 00:10:24.610 --> 00:10:27.670 align:middle line:90% MusicXML is very verbose. 00:10:27.670 --> 00:10:30.310 align:middle line:84% It takes up a lot of space on the screen. 00:10:30.310 --> 00:10:33.640 align:middle line:84% It takes up a good amount of space on disk, 00:10:33.640 --> 00:10:37.000 align:middle line:84% though not nearly as much as this video would ever take. 00:10:37.000 --> 00:10:41.140 align:middle line:84% So there is a compressed format, and they call it MXL. 00:10:41.140 --> 00:10:45.840 align:middle line:84% We call .XML or .musicxml, the file extension for standard 00:10:45.840 --> 00:10:51.550 align:middle line:84% MusicXML format, and .MXL if it's just gzipped. 00:10:51.550 --> 00:10:53.580 align:middle line:84% And right now, we're working on ways 00:10:53.580 --> 00:10:56.070 align:middle line:84% so that a gzip archive can include the score, maybe 00:10:56.070 --> 00:11:00.030 align:middle line:84% a compressed score, a transposed score, and also all 00:11:00.030 --> 00:11:00.880 align:middle line:90% of the parts. 00:11:00.880 --> 00:11:04.350 align:middle line:84% So new things are still coming out with MusicXML. 00:11:04.350 --> 00:11:08.490 align:middle line:84% And we will turn to learning more about MusicXML 00:11:08.490 --> 00:11:11.630 align:middle line:90% from the creator. 00:11:11.630 --> 00:11:19.000 align:middle line:90%