Getting Started - Beginner Stadium (Cools)

-Continued

- Sampling: Theory

Let's talk about all the business that goes on before a sound ever
gets to your computer's memory. Sound in the air is continuously
changing, and when it gets converted to an electrical signal the
changes are still continuous. Your computer,however, can only store
numbers using a limited number of digits or precision. Continuously
varying sound is called an analogue signal. Once the computer grabs
the sound, it doesn't have enough precision to store all the
information about the sound in order to perfectly reproduce it. What
the computer has stored is called a digital signal representation.

Your sound card captures information about an analogue sound signal
by measuring its intensity at a given instant. This corresponds to
one single point on the waveforms we've been looking at. In order
to capture an entire waveform, the measurement process must be
repeated at a high rate, usually thousands of times a second. Since
the hardware has limited speed and memory capacity, there are only
so many points it can capture. Any information between those points
is lost forever. This process of capturing the sound in small
intervals is called sampling. To play back a sound, we just reverse
the process and convert the digital samples back to an analogue
signal. Of course, the new signal will probably retain some of the
staircase effect, so the reproduction won't be perfect.

There are four main things to consider when sampling. The sample
resolution and frequency, amplitude, and copyright (very important).

The sample resolution is another term for the number of bits a sound
is sampled at. All trackers can handle 8-Bit samples,and most modern
ones are able to use 16-Bit samples as well. Sampling in 16-Bit will
render the better quality sound all the time. 8-Bit samples can be
difficult to distinguish from 16-Bit samples, if they are recorded
with good hardware. But most people would advise 16-Bit samples all
the time.

The main problem with a lower resolution is that you are likely to
get some or a lot of noise, depending on the quality of your source.
The only trouble with 16-Bit samples is that they are twice as large
as 8-Bit ones. A good trick to use is to sample in 16-Bit, do all of
your editing in 16-Bit, compose with 16-Bit samples, then for the
release convert the all the samples to 8-Bit. You'll find you can
halve the size of your MOD this way (But make sure you keep a copy
of the 16-Bit version). The listener may lose a small amount of
quality, but this is usually masked by the mixing routine of the
player. This may also deter some rippers from using your samples.

More important than the resolution of the sample when determining
quality is the sample frequency. The sample frequency refers to the
number of "snapshots" of the incoming sound taken per second. The
higher the sampling frequency, the better the reproduction of the
sound is.

So just how many snapshots do we need? If you look at audio specs
much, you've seen CD sampling rates of 44.1kHz, or 44,100 samples
per second. That's a lot of snapshots! A well-known signal
processing theorem (Nyquist Theorem) says that to accurately
reproduce a signal, you have to sample at a rate at least twice
the highest frequency component in the signal. So the CD
sampling rate of 44.1kHz will capture frequencies up to
22.05kHz.

You might be wondering what happens if you don't sample at a high
enough frequency. Well, what you get is something called aliasing.
This sinister sounding term just means that since the sample points
aren't close enough together, it looks as though you sampled a lower
frequency that really wasn't part of the original signal. Alias
frequencies are like ghosts - poltergeists really - you can't see
them but they make a lot of noise. So by sampling at too low a
rate, not only do you miss some of the high frequencies; some
of them get thrown back into the mix as unwanted guests at lower
frequencies. They are audible as background noise and distortion.

Monitoring the volume of the incoming sound is vital to produce a
good quality sample. If your sampler uses oscilloscopes to "view"
the sound then make sure the waveform gets as close to the top and
bottom of the window, without flattening out (clipping). If your
sampler uses volume meters instead you want to get the sample as
near to 0 dB as possible, without going above.

- Practical

Okay, you're probably fed up of reading about sampling and actually
want to do some for yourself. First of all you need some sampling
hardware; on a PC virtually every sound card in existence can do
some sort of sampling. On an Amiga or Atari you're going to need
some extra hardware on top of the built in chips.

Sample editing isn't really that hard, it's mainly lot of trial
and error, searching for the precise point where a sound begins
and ends. It takes a long time before you'll be able to read a
waveform like a book.

This is where tracking scores 100% over MIDI. MIDI samplers will
rarely have an accurate, easy to see waveform display,and they
don't have mice either. One of the few reasons I use Windows 95
is for its sample editors on a nice high resolution screen.

Start by centralising, and then normalising the sample. Then,
starting at the end of the sample zoom in and look for a point on
the centre where you think your sample ends. Always work from the
end first, as any computer will find it easier to fill an area of
memory with 0s than shifting a large chunk of memory around. If
you’re using virtual memory this can speed up editing by a huge
amount. Zoom back out to the whole view, does it look like you've
marked the right place, if it does then mark from that part to the
end of the sample and delete it. Play back the sample. If it gets
cut off too soon, then either paste the cut part back in, or use
the undo function built into many sample editors.

Centralise, and normalise again. This is because the part you
chopped off may have been off centre and/or louder than the part
you want. Zoom in, and look for where your sample begins. Cut off
anything before that. Play the sample to check you cut off the
right part, not too much and not too little. If you cut off too
much, then simply paste the data back in.

Keep on cutting bits off, and playing the sample back, until it
sounds how you want it to sound. Do a final centralise and
normalise, and save the sound to disk. Give it a meaningful
name, if the sample is from a synthesiser preset or a Sample CD
then use the name they give it. If you use a DETAILS.TXT or
similar then update it to include this new sample.

Take the time to tune all your samples as accurately as possible.
To do this, play a long, clear, looped sample, then move to
another channel and tune ALL your other samples to this one
sample (so they all have the same reference). Many potentially
excellent modules have been spoilt because they were poorly
tuned. Of course, this doesn't count the cases where samples are
intentionally slightly sharp or flat for effect (which should be
a rarity instead of a rule).

- Overusing Voice Samples

An extremely common mistake made by even some experienced
trackers is finding a voice clip that they think sounds absolutely
great or hilarious, and sticking it into their latest song
approximately 87 times. People often do this with dance tracks.
This very frequently kills what would otherwise be some truly
great songs. No matter how funny or cool something sounds the
first time you hear it, there are only so many times you can
hear and still enjoy it. Also, music is about hearing a melody
or grooving to a cool dance beat, not hearing somebody say the
same thing over and over again, so your song shouldn't rely on
voice clips to sound good. If you delete the voice clips from
one of your songs and find that it sounds terrible without
them, that means that you relied too much on the voice clip
and don't have enough music. Using a truly funny or interesting
voice clip once or twice can make a good song great, but it
can't make a bad song good.

- Ripping

There are a number of very important points that should be
kept in mind when ripping samples. Look for samples that sound
clear and don't have any clicking sounds at the point where it
loops. If you're looking for a sample of a real instrument, make
sure it really sounds like the instrument or else it will sound
stupid. Also, the newer the sample, the better. And finally, if
you rip samples, it helps a lot if you e-mail the person who
made them to get permission to use them, but if you don't get
permission,at least thank the person in the Sample Text.
That's just basic politeness.

- Copyright

Following on from ripping comes copyrights. If you're not
planning to ever release a tune commercially then use samples
from wherever you like. It's extremely unlikely anybody will
bother chasing you when they know you won't be making any
money from it.

If however, you eventually want to be able to release your
music, then pay close attention. If you sample individual
sounds,such as a single bass note, you should be able to get
away with it (especially if you hear the sound in a few
commercial tunes). When you sample large and/or easily
recognisable parts of any tune, get the samples checked out
before you even think about releasing yours. After all, I
don't expect you fancy paying out large sums of money just
because of one simple little sample.

- Releasing

Once you've produced a tune you like, you'll probably want
other people to listen to it, give you feedback etc. The
most important thing to remember at the moment is NEVER to
publicly release your first couple of tunes. There are very
few people who are gifted enough to really make a quality
tune the first time - it's all practice and experience! Once
you have finished a tune, listen to it a couple of days
after... see whether you can view it from another point of
view. Get a couple of friends to listen to it and ask for
some constructive criticism. You know, what's good as well
as what's bad about the tune.

When you feel ready to release a tune, probably the best way
of doing so is via the Internet. There are a number of good
sites which will allow you to upload to them, but
unfortunately most of them are incredibly busy, which makes
them very slow. If you don't have to worry about the
telephone bill, then by all means use them. If you do, then
probably the best way of releasing is to post your tune to
alt.binaries.sounds.mods. You could also set up your own web
site if you have some web space, either on your own ISP, or
on a free site provider like http://www.geocities.com
or http://www.fortunecity.com