Sonic Visualizer Tutorial

One of the many more important parts of
being a sound designer is to be able to

find a way to dig deep into analysing
digital sound, so there are many programs

out there which allow for composers or
sound artists or musicologist, even people

who are interested in discovering what is
happening in sound and music to analyze

that. So, we'll be looking at a top of the
spectrum analyzer. This is a program which

dissects the sound and breaks it down into
understanding all the different overtones,

the, the different frequencies that are
resonating as part of its spectrum. So

when we do that, we can use that for
creating the basis for our composition by

working on that analysis or you can use it
for musicological or theoretical goals,

where you want to discover some of the
internal patterns that are happening in,

in the sound. So, what I have here is a
website, sonicvisualizer.org. This is a

program, it's a free program for
Macintosh, Windows, and Linux that is a

spectrum analysis program but it has many
more features to it that, that can be used

for all sorts of applications. It's a
design at the Center for Digital Music at

Queen Mary University of London. And these
researchers are looking at broadly, the

questions of analysis of sound and audio
files for musicologists, for archivists,

for signal processing researchers,
composers, anybody who would like to find

more about what's happening inside the
sound of audio files. So, you can see here

from their site the opening page gives you
information about where funding for this

comes from. Over here, these are, are
important links, there's some great

tutorials here. So, if you go into the
documentation and then take a look here,

you will see there's a reference manual
which is really a very thorough manual

which takes you through all the different
components of that and then if there are

tutorials that you're looking at
specifically for musicologists guide to

Sonic Visualizer, which is many of the
applications that people at the CHARM

Center used as well. But then, there' s
some brief tutorials on individual

sections in that. So, I suggest as you
learn this, take a look at this page

because there's quite a bit of references
here. If you click on the reference menu

for a second, we would see then this whole
list of options, which take us through

very throughly all the different aspects
of this program and what it can do. So

lots of good information here. the first
thing in this tutorial is, is

understanding the difference between panes
and layers, which I'll come back to in

just a second so very good tutorial
information here. So, let's go to the

program. When you go to the Download page
here if we click on Download, then you'll

see the options here. You can download the
Linux, or the Macintosh, or the Windows

version of this. You can also go for the
source code, if you prefer. the important

part of this program is you will download
the Sonic Visualiser application and

whatever platform that is appropriate for
you. But then, the next thing you'll need

to do is to download a lot of the analysis
and effects plugins. These are the, in

some ways, the most important part of this
program, which allows us to, to get

involved with certain kind of analysis
detail in coming from these plugins. So,

and if we go over here, we will see this
other category of Vamp plugins. So, when

you click on this and we open that, we
will see this other page which describes

what these plugins are doing and so forth,
what the Vamp process is. if we click on

here, then get some plugins, find links to
downloadable Vamp plugins, it takes us to

a page where we can now download various
Vamp plugins for whichever architecture we

have. And I would suggest whichever
version platform you're working on, go

ahead and download all of these because
the will all be useful in giving

flexibility for how to use this program.
As you can see, there are quite a few of

these. Generally, when you download these,
you will install them by copying the

plugins, put, putting them into the files
that you download, putting them into

various locations on your computer and
you'll see here the various places where

you would put those. So, for example, I'm
using a Macintosh, if I go to the Main

Library folder on the hard drive, open up
the Audio folder and the plugins and

there's not one already created, create a
Wamp plugin, and then just copy all the

files into that and then when you reopen
the Sonic Visualiser application, they

will all be accessible from that. So, it's
not that difficult to install them but

this, this extra step of getting the
plugins is, is important for using this,

this program. So now, let's go to the
program. Once you've installed the

program, and you open it up, you will get
something like this. This is the a blank

screen. So, you'll see here the commands
at top, which we'll talk, take lot, take a

survey. over here, we'll have various
layers and panes that we can pull up and

then, you'll have an output volume here.
So, the first thing we might want to do is

to load in an audio file. And so, see that
laid out across our screen here. So

obviously, there's some standard kind of
transport and file functions that at the

top of here, plus we can change the
function of our cursor from a Grab to a

Select, to Edit, and so forth. So, we'll
come back to those. So, the first thing

would be, be to, to load some audio file.
I'm going to load one that I've just

recently did. This is an audio file of an
excerpt of a composition of my own called

Sundial, which was a piece written for
digital audio with a solo dancer a few

years ago. And this was constructed using
a Kyma system primarily where, where the

sounds I used in this using some granular
synthesis and some patches. So what I'm

going to do is we'll listen to one, I
think it's about one minute and twenty

seconds.
We'll just listen to this and then I'll

take some excerpts from this and we'll try
to transform them through going through

some of these categories. You will see
here, under transform, we can have

categories of, these are the effects
processors that you put into the Va mp

folder. So, whatever you put in there will
show up under this transformed list. So,

under classification different kind of
similarity features here. If you want to

look at the ones that are giving us
information about key internality, these

are the different analysis programs that
will affect that or we can go under sort

of the more low level features such as
amplitude or frequency coefficient and so

forth. Now, if we wanted to try to
identify notes in a chord or harmonic

changes information this would be routines
for that. time onset, for example, simple

percussion onset, which maybe we'll run in
just a second, so you can sort of find out

where there are beats in the piece, where
there are a certain kind of energy points

that we want to detect. And then different
ways of visualizing it, through. through a

spectrogram or a chromogram, different
kind of tools that are here. You can also

access the same ones here just by plugging
in a name, if you know what it is, you

need to get to quicker. So, it's the same
list, but just a different category or by

where these, if you know that it came from
a particular maker, such as from Queen

Mary, here is the list of ones they, they
made themselves. So different ways of

finding those. So, let's listen to the
following first and then, we will apply

some of these analysis effects to it.
Okay.

So, the file is a stereo file here. Once
again, we can Zoom in or out with this

button scroll down here or we can just
click and drag the center down here in the

Scroll command. We can also, if you want
to see more of the amplitude of various

spectrums, we can increase the, the size
of that so we can see a little bit more of

what's happening within the wave form.
maybe if I want to scroll out a little bit

more. And then let's go back, maybe
towards the beginning and just pull up

only a part of this. So, if I go back and
start at the beginning of this for a

second. Okay, I'm just going to randomly
select a part of this wave form. So, up

here, we can change to the Select tool an
d then, let's say, if we start dragging

about here, select all the way maybe down
here. I don't know where that is. And if

we want to listen to this loop, just this
section, we go up here to the Playback. We

can now turn on the loop Playback, and
also turn on the constrained Playback to

selection. You can do this with just the L
or the S command on your keyboard. so if

we do that, turn that on. So now, when I
hit Play or hit my Space bar, hit Play, it

will just continually loop this section.
Okay. So, if, if you want to find out more

what's going on inside this section, now
let's Zoom in a little more. And we can

sort of hear what, let me get that
centered. So now, if I want to slow it

down, to hear it more individually, we can
keep hitting the slow down. [MUSIC] Do it

again to even slow it down further. So,
we're getting a little bit more to

understanding what each articulation it
sounded like. Alright, now, let's apply

some of the analysis of this to see what
we can discover more that's inside here.

So, one of the first things would be to go
to the Pane window and we're looking at

the wave form now, but if I wanted to go
to a spectrogram the spectrogram will

analyze the, obviously, the spectrum of
each of the sounds and put it into a

second chart below that so I click on
that. We then get to see what are the

content of the various spectrum of sounds
occurring here. Now, on, on the left side

here, we do see the frequency range from
86 Hertz which we're seeing at the bottom

all the way up to 21 on the scale, but
we're probably topping out in our spectrum

around 18,000 is where this particular
recording was. Now, you'll notice as I

move my mouse over the various spectral
points, we do get to see an analysis of

what that's pointing to in the spectrum.
So, we can see what pitch it is or, or a

small range of where the pitch might be,
and then the frequency then the decibel

level of that as well. So, you're getting
a lot of information here. If I want to

maybe scroll into this a little bit
tighter and then, pull down to go down to

the bottom of the spectrum. Now, we can
see all the way down to 43 Hertz, we start

to see where the fundamental pitches are,
if we just put our mouse over, say, a

fundamental like that, we can start to see
its -thirteen db, it's at a frequency of

215 to 258, we're still pretty wide. If I
Zoom in more, I would be able to isolate

that to a closer range. but now, we're
getting more information. If I wanted to

Zoom in even further and, you know, pull
down, so now, when I start to look at this

particular spectrum here, I can see all
the overtones that are present in that or

all the parts of the spectrum and I can
start to identify where the amplitude and

the frequencies are for that, for example.
So, the, the spectrogram gives you a lot

of information about what is happening
inside the wave form that's, that's going

to report it for analysis purposes later.
let me kind of continue on with this. So,

as you get through these analysis forms,
you can click the X and that will, that

will go away. We could be adding other,
other kinds of panes I'll, I'll let you

explore that level, I was actually going
to skip around a little bit. If I do the

peak frequency spectrogram as next now,
I'll get rid of the spectrogram for a

second. Let's just look at, this is
another analysis of the, the spectrum but

it shows us where the peaks of, of, which
are the loudest partials in the spectrum.

So here, we start to get a little bit more
clear information about exactly what is

going on in these various points. And over
here, we can see a little a piano stuff,

which will help us understand where are
the pitches and partials are. So, if this

was close to, looks like an A, is that
what it is? No, that is like a G here. So,

we've got a G fundamental on the bottom
here but we're not really hearing anything

coming out of that point. The closest
loudest partial maybe starting here at 75

Hertz. Or if we look up there we can get a
clean indication of where that is. So

here, as I put my cursor over certain
ones, let me scroll a little b it more on

this. So now, we can start to really
pinpoint in as to where these markers are

and start to see where they 86 Hertz what
pitch it is, the, the decibel range. So

now, we're getting much more specific
information about the, where the peaks of

the spectrum are occurring. so, let's
listen to this once again to see if we

can. So, you can tell where the pitched
parts are, where you hear the, the female

voice, and then a lot of the sort of
granular noise sound, we can then identify

where those, it's, it's a spread of
different frequencies that are causing

that sort of noise sweep sound that's,
that's coming through it. So, we're, we're

getting a little bit different picture of
what's happening within the sound by, by

doing this kind of analysis. let's
continue on for a second. If we go over to

the transform, as I mentioned here, we can
get a lot more information about the

structure by looking at the transform. And
I can close this pane for a second so we

would see the information put on top of
this in a different layer. if I look

around and look for perhaps can we
identify a key in this maybe at this

point, if it runs a key, we can estimate
the key of the music. If there is all that

fitting into one particular key, then it
gives us sort of parameter, we can start

out with a tuning of a 440. And so, if we
click okay, then you start to see here, it

quickly analyzes these very sections and
see it says, you know, I was going to, I

was zooming in we have, right through
here, something that is an F sharp. It

maybe moves up to a G tonality here and
then down to a C sharp minor. So, it's

estimating whatever pitches are, are most
prevalent there and then trying to put

them into some key relationship. So, a
very useful kind of understanding for it

or figuring out is there a particular
modulation or a particular relationship of

keys in this section? Obviously, this
section, in my mind when I composed it,

was not written with any particular key
relationship but they do group themselves

into these sections. So, let's li sten to
this again with those, that information.

So, you can hear the C sharp change keys
from the G major here when we added one

more pitch into the, the vocalozation
there. So, we created a different kind of

a relationship. So, that's, that's a very
useful tool for figuring out key

relationships. So, there are many more out
here. I will let you do your own in

playing through with this. So, when you
download it let's go quickly maybe to one

of the on the time onset, for example. If
I pick, pick up the simple precussion

onset protector detector what it is going
to do is try and figure out are there

energy points that would identify beats
within this so we can then set some

threshold for that, but let's just leave
it fall for a second and see what it

discovers. And then, you'll see here that
it identifies peak points and, and, and

labels them with certain demarkations of
where the peak energies were and there are

quite a few on this one, obviously. And
then, we can start to look at each one

individually, the time value they occur is
of worth. But we can also move into

another kind of understanding where we do
say, note onsets or tempo, try to figure

out can it detect beats where the beats
are in this, and it will estimate a beat

and beat locator. And so, it does give us
some beats here. What I like about this

one it does give you a little percussive
sound on where it identified beats and

which allows you to, at a layer of
figuring out where those structures are.

So, we're getting more information about
the rhythmic structure of this particular

section, if I continue on into this, for
example, and go to see if it can estimate,

let's see, let's try another one. maybe it
can do let's try at bars or in beat

estimate locations and see what it gives
us on that. [MUSIC] So, I didn't really

find any bars identified in this category.
let's, moving them down to, how about if

it does maybe a tempo. Can it tell us
tempo on this? So, how fast is the tempo

moving through? So, once we get tempo
markings here. I don't see wh ere it gives

any information there. so you, you can
explore this where, there are all sorts of

other information, sometimes it will be
able to find that, sometimes not on that.

But there are tons of interesting kind of
information that you can pull out, extract

from these various kinds of tools that
you, you can pull into it. so I would, I

would really encourage you to work with
this program and to use it for many times,

for starting a composition. You may want
to start a composition where you take

another audio file or another musical
example and use that as the basis for your

sound design. But then, do extensive
analysis on that to pull out information

about your tempo, your, your rhythmic
units key signatures. different kind of

timbres that you want to focus on which
you can extract the information from the

spectrum and then use them in other
compositions. And then, if you want to

analyze other compositions and sort of to
see how other composers and other sound

designers have put together works, this is
a great program for unveiling that. So,

it's one that I suggest, you go through
the Reference Manual and go through all

the different units. But it does give you
a lot of great tools for doing sound

analysis. The proceeding program is
copyrighted by Emory University.