=============================================================================
4. Techniques
-----------------------------------------------------------------------------
   One of the key points about producing a quality tune that I've found is
the amount of preparation you put in, before you even begin to start. This is
especially important if you intend to produce in a style unfamiliar to you.
Take the time to get good samples, and see how they could be made to fit
together. Listen to the style, you don't have to buy tons of new music, just
see what friends have lying around, and the radio can be a good source. Play
around with various ideas in your tracker, you needn't save them. Get hold of
a few MODs and see how they work.
   We're not talking about a few hours here, not even a few days. It may take
a few weeks or even months before everything's ready. But when it is you
should find that you're able to produce, fairly quickly, a quality piece.

-----------------------------------------------------------------------------

   - Spicing Up Your Percussion
  ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
   (Taken from CU Amiga May 1994 - Slightly edited to be more generic)

   Fat Beats
  ~~~~~~~~~~~
   There are a number of things you can do to add a Bit of life to your
percussion. One of the best ways to beef up a drum sample is to mix it with
another sample. You've probably already experimented with this, mixing kick,
snare and hi-hat samples, in order to fit your entire rhythm into one track.
However, to get a really kickin' sound, try mixing your percussion samples
with samples of tuned instruments. For instance, mixing a really deep
analogue-type bass sound with a kick drum produces a really heavy, squelchy,
dance floor sound. Similarly, try mixing snare and guitar sounds, for an
unusual and funky effect try adding Laser-type pulse sounds to 808 style
snares for an authentic Sheffield clunk and bleep sound.

   Echomania
  ~~~~~~~~~~~
   Another way to add a Bit of life to a rhythm track made up of individual
samples, is to echo the entire track. This is a quick way of funking up your
percussion, and you'll find you can create a great track with only kick,
snare, and open hi-hat when you use echo in this way.

   Bring Out Your Dead
  ~~~~~~~~~~~~~~~~~~~~~
   You've probably got quite a collection of hackneyed breakbeats, which are
instantly recognisable, and therefore pretty much unusable. One way round
this is to sample some more, but it theory at least, you always have to be
careful of the copyright laws when sampling other peoples material
   You could always buy a sample-compilation CD, but most of these are a tad
expensive for the casual user. On the other hand, it's quite possible to
breathe new life into a dead breakbeat.
   One method, is to apply some sort of sound effect to the sample,
preferably in stereo. Most sampling software nowadays has a range of effects
built in with which you can process you sample, but most of these produce
fairly unsubtle results when applied to percussion samples.
   So what's the alternative, if you can spare the memory and two tracks (a
stereo pair is what we're looking for here), is to use the tracker itself
to produce a real-time phasing effect.
   To do this, load the same breakbeat sample into two different sample
locations. For best results, pick a breakbeat that stretches over two bars
(32 lines of a standard 64 line pattern). Play the first instance of the
sample (at a reasonable rate!) on line 0 and line 32 of a 64 line pattern, on
one track. Do the same thing on track 2, but this time with the second
version of the sample. Now for the clever Bit.
   Fine tune the second version of the sample up or down one or two points.
Now when you play the pattern, you'll get a phasing effect, with the rhythms
moving in and out of the stereo field - great for trance techno type
extravaganzas. If you're feeling particularly adventurous, try playing one of
the samples an octave down from the other.
   If you can't spare the memory or two tracks for the rhythm, you can get a
similar effect in mono as follows. Load up the first and second breakbeat as
before, and resample or pitch shift the second by a few points, then mix them
together. The effect is a lot less subtle than the stereo version, but can be
just as effective in the right circumstances.

   On A Ragga Tip
  ~~~~~~~~~~~~~~~~
   Another way to squeeze the last Bit of life out of a dying rhythm is to
change the playing length and sample trigger positions from the normal start
of the bar. This is a technique much favoured by breakbeat and jungle techno
groups like SL2 and The Prodigy, and works best at fairly fast BPMs. Play
your breakbeat on lines 0 and 32, and adjust the tempo so that the rhythms
trigger in time, with no glitches. Now trigger the sample on the following
lines: 0,6,16,26,32,42,48 and 54. When you play this back, you'll have a
rhythm track that sort of rolls around the beat - perfect for just adding a
bassline and calling it your finished song!
   For a brutal stereo version of this, try playing the same sample on a
different track (on the opposite stereo channel) on the following lines: 0,
10,16,22,32,38,48 and 58. You might even go the whole hog and combine this
with the stereo phasing effect.

-----------------------------------------------------------------------------
   - The Zen of Tracking Advanced Tips and Tricks
  ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
   (Slightly edited for formatting)

   o Indian food for thought

   (Greets to Graham the Happy Scum for this idea in TW#86 "Wrecking Samples
with Impulse Tracker"... despite a few technical glitches.)

   You can get a very Indian-sounding "24-tone" scale in Impulse Tracker by
using this technique: (FT2 users will have to accomplish the same thing via
the "tone" setting)

   Load your sample twice. Look at the second one, and write down the
samplerate. Multiply that number by 1.0304 (NOT 1.0293, as in TW86) and put
the result in the "playback rate" field of the first sample. Now you have a
consonant tone in the second sample and a semitone above that in the first.
By playing the second at C-5 then the first at C-5 then the second at C#5
then the first at C#5... (and so on), you get a semi-tone chromatic, which is
pretty weird. : ) If you're really bold, you might get some cool Indian
sounding stuff going out of it. Good luck tracking it, though. It's a whole
new set of musical theory. : P

   o The Amigascene and you

   If you either release or listen to .MODs (not .XMs, .ITs or .S3Ms, etc),
then you're probably aware of the Amiga scene, which still uses the .MOD
format today. If so, keep this in mind: the Amiga plays music 1bpm faster
than PCs. For example, at speed '6' in a .MOD, a PC is playing it at 120 bpm
(I would assume, anyway), and an Amiga is playing it at 121 bpm.

   What this means to you, the listener, is that certain drum loops and riff
samples will sound off-kilter, rhythmically. So be a little more forgiving in
such circumstances. If you want to hear it as it was originally written on
the Amiga, put it in FastTracker (or whatever your favorite tracker is), save
it as an .XM (likewise with the favorites), and change all the tempos in the
song to their appropriate Fine-tempos (bpm), plus one.
   (Remember to do the reverse if you're producing a MOD on the PC that'll
probably get played on an Amiga. When the tune is finished convert all the
primary tempos to 1 less. This BPM thing sometime gets more extreme too,
maybe 2 or 3 BPMs out in certain circumstances - Cools)

   There are also some other effects that don't convert well from Amiga to
PC, which are apparent in chip tunes. For the best reproduction (though still
not perfect), look for a player called "Midas Player", since it handles
things a little better than most with .MODs.

   Radix has a few things to add:

   well.. In protracker the EFx command is used a lot... it actually changes
the waveform in the sample (only in the beginning), so in chiptunes,
chipsounds can get somekind of wavesequence sound, "weeeeeeeeoooong" ... that
does not work on any prog on PC I have seen anyway... and arpeggio on PC is
not that fun either :/ don't know really, but chip sounds sound better on
Amiga...
    another thing is that PC with a GUS can sound really awful while playing
a high and a low tone of the same sample at once... this is really lame..
like a C-3 and a C-7 (same sample) sound really untuned

   o Very Cool Reverb

   Sure, you have an echoed lead. But do you have a reverberated lead? This
sounds very cool indeed:

   Load the lead in your favorite sample editor (mine's CoolEdit), reverb it
however you like (I use a straight reverb, on the "last row seats" setting),
so that it's REAL deep.

   Now load the tracker. Create the echo track as usual (copy the lead,
offset it by a few rows, and change the volume to less than 50% of the lead)
.. Much nicer, eh?

   If you want a reverb that's not-as-deep to use somewhere else, you can
widen it for the echo track, creating this weird echoed attack kind of thing,
like this(FT2 Format, 1 is the lead, 2 is the reverb):

          01  C-5 01 40 000    C-5 02 08 840
          02  --- -- -- 000    C-5 02 10 8A0
          03  --- -- -- 000    C-5 02 20 880
          04  --- -- -- 000    --- -- -- 000
          05  F-5 01 34 000    F-5 02 08 8C0
          06  --- -- -- 000    F-5 02 10 860
          07  --- -- -- 000    F-5 02 20 880
          08  D#5 01 3C 000    D#5 02 08 840
          09  --- -- -- 000    D#5 02 10 8A0
          0A  D-5 01 30 000    D-5 02 08 8C0
          ..  ... .. .. ...    ... .. .. ...

   Of course, you don't need to keep retriggering the note. I just thought it
sounded cool with bouncing pan. In any case, I think a reverb-ed lead sounds
even better than an echoed version... Try it and see for yourself.

   o 'Phased' Leads

   A very cool effect for writing leads, which is commonly used by advanced
trackers is a phased synth string. (In fact, it's almost hard to call this an
'advanced' trick.) You can find samples that work for this is a lot of
different places (any good 'sweep' string sample will do), but the way that
they're used is the important aspect...

   It's quite simple, really. You just create an instrument with a volume
envelope typical of a lead... Something with a sharp attack, a moderate
length sustain, and an expontentially quieter decay (my ANSI art is
miserable, but I'll try):

           .  <-- Full volume here
           |\______
          /        \
          |<-- 60%  \_
              volume  \__
               here      \____
                              \________
                                       \ <-- 10% volume here (or less), and a
                                              moderate (300ish) fadeout.

   The total length of the envelope should be about twice as long as the
average length of the note (ie: an average length of a quarter-note should
have an envelope that lasts about as long as a half-note). Now, as you write
your lead, keep the notes in the same channel, and slide to them at a very
fast rate ('F', generally), like this:

          01  C-5 01 40 000   <-- This starts off the sweep
          02  --- -- -- 000
          03  --- -- -- 000
          04  --- -- -- 000
          05  F-5 01 34 3F0   <-- You slide to the note here
          06  --- -- -- 000
          07  --- -- -- 000
          08  D#5 01 3C 3F0   <-- And here...  See the effect?
          09  --- -- -- 000
          0A  D-5 01 30 3F0   <-- Etc.  Retrigger the note
          ..  ... .. .. ...       to 'start over' the phase.

   It's important, however, that you echo this lead in another channel, since
it will sound fairly flat otherwise.
-----------------------------------------------------------------------------
   - Sound & Sampling Explained
  ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
   (By Rubz)

   As you probably know from physics, sound is essentially made up of waves
travelling through the air - sound is merely vibration caused by some object
or another. Of course, that isn't entirely accurate, as sound can pass
through solids and liquids as well (in fact, the denser the medium, the
better the sound is conducted - that's why whales can communicate with each
other over distances of miles, because water is denser than air.) The medium
through which the wave is travelling doesn't actually move, either, or at
least not much more that it takes for one molecule to bump into the next one
(think of a Mexican wave at a football match, and you'll get the picture.)
The vibrations remain vibrations until they come into contact with something
that can hear, i.e. an ear (but *not* a microphone, because a microphone
merely captures some of the vibration and sends it down a wire.)
   The faster the vibration, the higher the frequency, the higher the pitch
of the sound; humans can hear from about 20hz to about 20,000hz (although the
more you abuse your ear by pumping high decibel sound into it, the less high
the frequency you can hear) There isn't much, musically speaking, in the
12Khz to 20Khz range - you would notice the difference if you compared a song
through 12Khz and 20Khz ears, but there wouldn't be much. It is claimed by
many that we are sensitive, although not actually aware, of sound well above
20Khz and below 20hz, and this is why professional equipment will have such a
wide frequency response.
   The intensity of the sound wave determines the loudness of the sound (the
harder you strike a drum, the bigger the oscillation of the skin, and hence
the louder the drum - the frequency is unaffected), and sound is
traditionally measured in decibels. Literally, 0 decibels (0 dB) is
equivalent to an sound pressure level of 20 microPascals, which is the lowest
possible level of sound that your average Joe will be able to hear. Clearly,
this is a relative figure, as everybody's hearing is slightly different. The
decibel scale is logarithmic, because that is the way our brain interprets a
change in sound level (for example, the brain reckons that 40,000 mPascals is
only twice as loud as 4,000 microPascals; the figure in decibels represents
our perception of it.)
   Now you are likely aware that computers operate entirely digitally (with
the only possible numbers at the lowest level being 1 or 0, one of two
states, on or off.) So how do we translate an analogue vibration into an
internal, digital, package of data? Well, imagine the sound coming into the
computer on a conveyer belt, and every few thousandths of a second the Bit
coming past is chopped off, and measured. Got it? That is essentially, the
way a computer samples a sound - a wave file on disk is essentially a large
stream of numbers, each representing the level that was measured in that
particular time interval. That time interval is what we are referring to
when we talk about sampling at 11.025Khz, 22.05Khz, 44.1Khz or even 48Khz -
the number refers to the number of times the knife comes down on the wave,
chops off a slice, and  measures it; accurate sound reproduction requires a
sampling rate of around 40Khz, CDs are done at 44.1Khz, and DATs at 48Khz.
Generally the sampling frequency is around twice the highest frequency that
can be represented, so if you sample at 22.05Khz, you are restricting the
discernable sound to between around 20Hz to 11.025Khz, which is why the lower
your sampling rate is, the lower the quality of your sound (of course,
sometimes you actually want it to sound that Bit rougher; also, if you know
that your sound won't use higher frequencies at all, then it is fine to
sample at a lower rate, and you'll be hard pushed to spot the difference.)
  But as you'll know, if you've used Cool Edit or something similar, you also
get the choice between sampling at 8 bits or 16 bits. So what difference does
that make? Well, if you know anything about binary numbers, you'll probably
be way ahead of me here, but just in case:
   A decimal number is made up of units, tens, hundreds, thousands, tens of
thousands and so on, in effect powers of 10 (10^0, 10^1, 10^2, 10^3, 10^4,
etc.) So when you write 3252 you are in effect saying 3 thousands, 2
hundreds, 5 tens, and 2 ones or 3 10^3s, 2 10^2s, 5 10^1s, and 2 10^0s (any
number to the power 0 is always 1.) Similarly, a binary number is made up of
ones, twos, fours, eights, sixteens (or 2^0s, 2^1s, 2^2s, 2^3s, 2^4s, etc -
2, because there are two possible states, 1 and 0). For example, the binary
number 1101 is in effect 1 2^3, 1 2^2, 0 2^1, and 1 2^0, or 8 + 4 + 0 + 1,
13.
   An 8 Bit number can represent 256 ((2^8) - 1) different states (0000,0000
through 1111,1111), and a 16 Bit number 65,536 ((2^16) -1) different states.
You remember earlier we said that when the computer measures the level of the
incoming wave on the conveyer belt, it stores it as a number. With an 8 Bit
sampling resolution, it has to choose that number from 256 possible states,
so if the wave happens to fall between 2 of those 256 numbers at that
particular time interval, the computer has to choose the nearest. You've
probably seen the same thing happen in primitive graphics packages - draw a
diagonal line, and you end up with a stepped line. 16 bits, therefore,
provide a lot clearer sound quality, as you have more levels to choose from;
even 16 bits, however, are not perfect, and studios commonly work with 20 Bitresolutions, which provide 1,048,576 different possible levels, or 24 Bit
resolutions, which provide 16,777,216 different levels.
   Similar to there being a relationship between sampling rate and the
frequency response of the sound, there is also a relationship between the
dynamic range (the possible variation in level of the sound) and the sampling
resolution (a 16 Bit resolution gives a dynamic range of 96dBs, or 6 times
the resolution.) Don't worry about why, just accept it. When we say a
dynamic range of 96dBs, we do not of course mean that the loudest possible
level is 96dB, we simply mean the range of possible levels is 96dB wide
(any amplifier can make something louder or quieter quite easily.)
   One thing you should ensure when sampling, then, is that your source is
within the dynamic range of at the resolution you are sampling at. As an
experiment, shout or scream into the microphone at 8 Bits, and then repeat at
16 Bits. When you look at the 8 Bit one, youll notice that the wave is cut
off at the highest possible point, it is just a straight line or block going
as high as the top of the screen. What this means is that there were sounds
at higher levels than the resolution allows, but the computer couldn't cope
with them because it was only sampling at 8 Bit; thus it assigned them to the
nearest level, which was the highest possible one. This is known as clipping.
Your 16 Bit sample will probably still have some clipping, but considerably
less. To get round this, either use a compressor, so that all sounds are
restricted to a certain dynamic range, or adjust your gain and input levels
(so if you know you are going to be recording a very loud noise, drop the
gain right down, to keep it all within the range.)
   Of course, if you are looking for weird effects and so on, you may wish to
try ignoring the guidelines for good quality sounds; things sampled at low
resolutions, frequencies or with clipping can sound interesting. It is
important that you understand what they mean, though, as you can only
properly experiment with something that you understand.                                                                                                                                                                                                                     !                                                                                                                                                                                                                                                                                                                                                                                                                                    

=============================================================================
