Welcome to Gaia! ::

The Physics and Mathematics Guild

Back to Guilds

 

Tags: physics, mathematics, science, universe 

Reply The Physics and Mathematics Guild
Harmonic Analysis

Quick Reply

Enter both words below, separated by a space:

Can't read the text? Click here

Submit

victorian doll

PostPosted: Sat Dec 15, 2007 8:01 pm


Hello. :]
I need some advice!
It is probably rather premature, but I really want to get into the mathematical side of music theory. Now, before anyone makes hasty assumptions, I must tell you that I am only in grade nine. ( mad d I know. )
So, my question is, can I start learning anything at my level, or do I need much more higher math to even begin?
PostPosted: Sat Dec 15, 2007 9:26 pm


All of the mathematical music theory I've seen requires at least college level mathematics: group theory, linear algebra, bits of differential geometry, etc.

Once you get beyond the basic idea of the well-tempered scale as having semitones differing by 12-th roots of 2, things apparently start getting very complicated very quickly.

Layra-chan
Crew


nonameladyofsins

PostPosted: Sat Dec 15, 2007 10:24 pm


I'm sure we can tell her something about fourier sums at least, that way she could understand modes and the break down of harmonics. Do you know any trig?
PostPosted: Mon Dec 17, 2007 11:33 am


I know about computing angles/sides of right triangles (sine, cosine, tangent).
That's just the very beginning of trig though. sweatdrop

victorian doll


nonameladyofsins

PostPosted: Mon Dec 17, 2007 5:40 pm


victorian doll
I know about computing angles/sides of right triangles (sine, cosine, tangent).
That's just the very beginning of trig though. sweatdrop


oh that's good, you know sine, cos and tan. Have you done functions? Do you know what y = sinx looks like?
PostPosted: Tue Dec 18, 2007 1:38 pm


poweroutage
victorian doll
I know about computing angles/sides of right triangles (sine, cosine, tangent).
That's just the very beginning of trig though. sweatdrop


oh that's good, you know sine, cos and tan. Have you done functions? Do you know what y = sinx looks like?


I didn't until a minute ago, when I found this.
So cool!!

victorian doll


Layra-chan
Crew

PostPosted: Tue Dec 18, 2007 11:32 pm


So the basic idea is that music (all sound, actually) is made up of circles (by which I mean sine waves). Different pitches correspond to different frequencies; higher pitches go faster around the circle, lower ones go slower.
If you go up exactly one octave, you increase the frequency by exactly a factor of 2; going up a semitone is equivalent to multiplying the frequency by the 12th root of 2.

Any sound wave can be decomposed into a sum of sine waves.
Musical sounds can be decomposed into a sum of sine waves where the frequencies of those sine waves are all integer multiples of some "fundamental frequency."

For example, when you hit the middle C on a piano, you get a sound with a fundamental frequency of approximately 440 Hz (meaning that the sound wave goes around the circle 440 times per second), but there are also waves of 880 Hz coming out, and 1320 Hz, and 1760 Hz, and so on. The different amplitudes (how big the circle is) determines the tone color, the timbre of the sound.

The standard chords are built to make other frequency matches, because when the frequencies of one note (not just the fundamental, but all of them) are simple fractions of the frequencies of another note, it sounds nice when the two notes are played together. For a perfect fourth, the ratio is almost 4:3; a perfect fifth is almost 3:2.
PostPosted: Mon Dec 24, 2007 1:12 am


Layra-chan

Any sound wave can be decomposed into a sum of sine waves.
Musical sounds can be decomposed into a sum of sine waves where the frequencies of those sine waves are all integer multiples of some "fundamental frequency."


Since all sounds can be represented by a sum of sine waves, it turns out that all functions inside an interval can be represented by an infinity sum of sine and cosine waves called a fourier series. A fourier series is just an extension of your basic sine and cosine sum to infinity. Thus all different types of sounds can be described by equations, that is trig or fourier series.

Layra: I know a lot of musical sounds require only finite sums, what kinds of sounds are described by infinite sums? I guess my question is, how intricate does the sound have to be such that the amplitude of cosines and sines whose frequency is greater than N don't die off to zero?

nonameladyofsins


Layra-chan
Crew

PostPosted: Mon Dec 24, 2007 3:54 am


poweroutage
Layra-chan

Any sound wave can be decomposed into a sum of sine waves.
Musical sounds can be decomposed into a sum of sine waves where the frequencies of those sine waves are all integer multiples of some "fundamental frequency."


Since all sounds can be represented by a sum of sine waves, it turns out that all functions inside an interval can be represented by an infinity sum of sine and cosine waves called a fourier series. A fourier series is just an extension of your basic sine and cosine sum to infinity. Thus all different types of sounds can be described by equations, that is trig or fourier series.

Layra: I know a lot of musical sounds require only finite sums, what kinds of sounds are described by infinite sums? I guess my question is, how intricate does the sound have to be such that the amplitude of cosines and sines whose frequency is greater than N don't die off to zero?


Actually, most periodic sounds require an infinite number of sines and/or cosines. Something as simple as a square wave requires an infinite number of sine waves. The amplitude dies off quickly (usually around the order of 1/n), but even so, we do need an infinite number.
PostPosted: Tue Dec 25, 2007 5:40 pm


Layra-chan
So the basic idea is that music (all sound, actually) is made up of circles (by which I mean sine waves). Different pitches correspond to different frequencies; higher pitches go faster around the circle, lower ones go slower.
If you go up exactly one octave, you increase the frequency by exactly a factor of 2; going up a semitone is equivalent to multiplying the frequency by the 12th root of 2.

Any sound wave can be decomposed into a sum of sine waves.
Musical sounds can be decomposed into a sum of sine waves where the frequencies of those sine waves are all integer multiples of some "fundamental frequency."

For example, when you hit the middle C on a piano, you get a sound with a fundamental frequency of approximately 440 Hz (meaning that the sound wave goes around the circle 440 times per second), but there are also waves of 880 Hz coming out, and 1320 Hz, and 1760 Hz, and so on. The different amplitudes (how big the circle is) determines the tone color, the timbre of the sound.

The standard chords are built to make other frequency matches, because when the frequencies of one note (not just the fundamental, but all of them) are simple fractions of the frequencies of another note, it sounds nice when the two notes are played together. For a perfect fourth, the ratio is almost 4:3; a perfect fifth is almost 3:2.


And now I know why a Tuba's C matches a Flute's C.


<3

AirisMagik


Layra-chan
Crew

PostPosted: Wed Dec 26, 2007 11:26 am


AirisMagik
Layra-chan
So the basic idea is that music (all sound, actually) is made up of circles (by which I mean sine waves). Different pitches correspond to different frequencies; higher pitches go faster around the circle, lower ones go slower.
If you go up exactly one octave, you increase the frequency by exactly a factor of 2; going up a semitone is equivalent to multiplying the frequency by the 12th root of 2.

Any sound wave can be decomposed into a sum of sine waves.
Musical sounds can be decomposed into a sum of sine waves where the frequencies of those sine waves are all integer multiples of some "fundamental frequency."

For example, when you hit the middle C on a piano, you get a sound with a fundamental frequency of approximately 440 Hz (meaning that the sound wave goes around the circle 440 times per second), but there are also waves of 880 Hz coming out, and 1320 Hz, and 1760 Hz, and so on. The different amplitudes (how big the circle is) determines the tone color, the timbre of the sound.

The standard chords are built to make other frequency matches, because when the frequencies of one note (not just the fundamental, but all of them) are simple fractions of the frequencies of another note, it sounds nice when the two notes are played together. For a perfect fourth, the ratio is almost 4:3; a perfect fifth is almost 3:2.


And now I know why a Tuba's C matches a Flute's C.


<3


It'd be kind of strange if they didn't, considering that both tuning and the musical scales were made so that transposition between instruments wouldn't be terribly difficult; otherwise an orchestra would never work.
PostPosted: Sun Dec 30, 2007 12:25 am


Oh my god xd Poor Victorian Doll. I think you all are going to confuse her ^^;; I read through your explanations and while it makes perfect sense to someone who has done trig for years, that must have been complete nonsense to her sweatdrop We didn't even explain why sine and a circle are related.

OK- first, let us have a simple trigonometry lesson. Is everyone fine with starting here? We'll present trigonometry first in its most basic form, and then in the form of a huge problem. 3nodding

The most basic form (which you have probably seen before): the triangle.

http://www.mathsrevision.net/gcse/trig.gif (yes- it's a crude drawing. I like it, though, because it was simple. This drawing courtesy of a google image search)

In this picture, you see an angle A, and you see a right triangle (it's extremely important that this triangle is a right triangle). Mathematicians learned something very amazing. That is this: if angle A is always the same, then the ratios of any two sides will always be the same. Since a ratio is nothing but a fancy word for a division problem, that means that so long as A doesn't change, you can divide any one side by any one other side and there is a certain number you can expect to get. Let me try and explain with two basic triangles that follow this pattern.

Triangle 1 will have an opposite side of 3, an adjacent side of 4, and a hypotenuse of (do you remember Pythagorean theorem?) 5. This makes sense. In this triangle, angle A would have a value of about 36.87 degrees

Triangle 2 will have an opposite side of 9, an adjacent side of 12, and a hypotenuse of (you guessed it!) 15. In this triangle, angle A also happens to be about 36.87 degrees (if you don't believe me, get a ruler and protractor and try to create these triangles).

You learn in geometry that if two triangles have 3 congruent angles then their sides all have a similar ratio, so this pattern is exactly what is taught in geometry. Now let us take the sine (opposite divided by hypotenuse) of each triangle. The first one yields 3/5. The next yield 9/15. Each of these has the same value of .6! Let us now take the tangent (opposite divided by adjacent) of each. The first yield 3/4, and the second 9/12. Again- each is .75

So we can see that the ratios of the sides will be the same regardless of their exact value. For this reason, we can state that the sine of 36.87 degrees IS ALWAYS .6 (regardless of the lengths of the sides). If you don't believe me, take a graphing calculator and type sin(36.87°)

Since an angle will always have the same sine, cosine, and tangent- we can therefore use these known values to solve common problems. Suppose we have a right triangle and we know one side and we know one angle (other than the right angle)- yet we know no other sides. Let us set up a sample problem and see just how easy this is to solve:

Triangle ABC consists of lines AB, BC, and CA as well as angles A, B, and C. Angle A is our right angle. Angle C has a known value of 30°. The length of BC (the hypotenuse) is known to be 16. We are asked to find the length of AB.

If we look at our trig triangle from earlier, we will see that line AB is opposite of angle C. Thus we are trying to find the length of the opposite side, and we know the length of the hypotenuse. Sine is defined as opposite divided by hypotenuse, and so we know that the sine of angle A is something (the length of the opposite) divided by 16 (the length of the hypotenuse). To write this algebraically:

sin(30°) = x/16

The thing is- we proved earlier that sin(30°) is a constant value! sin(30°) actually happens to be 1/2. So our algebra now looks like this:

1/2 = x/16

This problem should be child's play to you. x is clearly 8. We now know the value of AB. Why would this information be useful? What if BC (the hypotenuse) was a ladder of known length 16 feet and it had to be placed at a 30 degree angle in order to reach a tree branch? How else could we possibly know the height of this branch?

That is trig in it's simplest form. Given an angle and a right triangle we can determine the ratio of any two sides. Also, given one of these two sides we can determine the other since we know the ratio.

Now let us turn trig into a very annoying and difficult problem. Let us look at a circle with a radius of 1, and attempt to find any point on this circle.

At first, this problem seems impossible. How can we possible know where a single point is along a circle? We know that every point is a distance of 1 from the center, but we cannot know how far up it is without knowing how far out it is, and we cannot know how far out it is without knowing how far up it is. We're at a loss.

We solve this problem by defining points around the circle with angles.

http://www.wjagray.co.uk/maths/Lectures/Images/TrigCircle.GIF

The variable θ represents any angle (perhaps 30 degrees, perhaps 45 degrees, perhaps 52.839 degrees). As you can see, given an angle we can locate any one point along this circle. All we know about this point, however, is that it is 1 unit away from the center (since the circle has a radius of 1, all points on the circle are 1 unit from the center). We learn more about this point by drawing a vertical line straight down from it to touch the x-axis like so:

http://www.hyper-ad.com/tutoring/math/trig/images/trig_angl467.gif

Our radius, our vertical line, and the x-axis now form a triangle. Also- since the x-axis is horizontal and our line was drawn vertical, our line creates a right angle and we now have a right triangle. Since we have a right triangle, we know that the sine, cosine, and tangent of θ are constant. Our x coordinate (how far out our point is along the x-axis) can be determined by the length of the adjacent side. The y coordinate (how far out our point is along the y-axis) can be determined by the length of the opposite side. Knowing the length of the hypotenuse (1) and our angle (θ) we can use a problem similar to the branch and ladder one earlier to find our x and y.

However, we can simplify this just a bit. First, let us look at the equation we have been using prior:

Sine = opposite / hypotenuse

or:

sin(θ) = y / 1

(recall that the length of our opposite side is the same as our y coordinate)

When we divide by 1, the number will always remain the same. Thus:

sin(θ) = y

Using similar logic:

cos(θ) = x

So we can now find any point on a circle with radius 1 knowing only the angle. Given any angle θ, the coordinates of a point on the circle are:

(x, y) --> (cos(θ), sin(θ))

What do we do if the radius is more than 1, however? What if we have a circle of radius 5? Or 5000? Well, let us return to our algebra problem.

sin(θ) = opposite / hypotenuse

so:

sin(θ) = y / 5000

Multiply both sides by 5000:

5000sin(θ) = y

You'll find that for any radius, merely multiply this radius by the sine of θ and you'll receive the proper y coordinate. Multiply this radius by the cosine of θ and you'll receive the proper x coordinate.

Now we come to graphing. When we graph something, just what does the graph show? The graph shows what the value of one variable is at every given value of another variable. We have one point for every possible x showing what y would be given that x. These points join together to form a line of some form.

In this case, we wish to graph sin(x)

As you'll recall from earlier, sin(θ) = y for a point on a circle of radius 1. And so sin(x) is equal to the point on a circle of radius 1 at angle x.

This explains the strange drawing process of the website that you found. It drew the graph by moving to the right (increasing x) and as it did, it showed how high a point on the circle would be (the y of a point on a circle of radius 1) at that given angle. So at 90 degrees, we have a y of 1 (you're going straight up and touching as high as possible on this circle). At 270 degrees we have a y of -1 (you're touching as low as possible). At 0 and 180, you are on the x-axis and have neither gone up nor down. You have a y of 0. This explains why the sine graph comes back to cross over the x-axis.

In truth, though, a sine graph goes on forever. 360 degrees is an entire circle, however 720 degrees is two entire circles. At 810 degrees you will have gone two entire laps around the circle and now be at a value of 1. So a real sine graph goes up and down in a wave shape forever.

As was said earlier, if a pitch is higher, then you go around this circle faster, yes? That would mean that you go up faster, and then down faster, and then back up faster. This would mean that when you draw it out, your waves would be scrunched up close together. If you have learned about sound waves in any physical science class, this makes sense- as it would cause the sine wave to have a higher frequency (as a result of more frequently going a full lap around the circle)

Please tell me if you're confused. I tried to simplify a half-semester's worth of learning as best as I could ^^;;

spider_desu


Layra-chan
Crew

PostPosted: Sun Dec 30, 2007 7:42 am


I actually like the applet that Doll found; it gives a very straight-forward visual of the mapping from the circle to the graph of sin(x).
Reply
The Physics and Mathematics Guild

 
Manage Your Items
Other Stuff
Get GCash
Offers
Get Items
More Items
Where Everyone Hangs Out
Other Community Areas
Virtual Spaces
Fun Stuff
Gaia's Games
Mini-Games
Play with GCash
Play with Platinum