Sound Quality:
In my opinion, quality is a very important part of tracked music - a song can
have a great melody, with lots of harmony and counter-melodies, but if it has
low quality samples and bad mixing, it will be unpleasant to listen to, and
will likely be overlooked by most listeners (Compare my song *Escape from
Wrongful Incarceration*, which wasn't received particularly well, to its new
incarnation, *Courage Under Fire*, which scored a 10/10 on TiS).

Samples:
Samples are usually the culprit when a song is of low quality. There are a
few main things to look for when you are choosing samples.

First, there's the general quality - a 44khz/16-bit sample will sound better
than a 8khz/8-bit sample, obviously, but choosing the former will make the
filesize of your song much bigger. The conflict here, I guess, is whether you
want to go for quality or filesize (I personally always go for quality -
44khz/16-bit wherever possible), but I wouldn't recommend you ever go below
22khz/8-bit. Also, if a sample has a background hissing, which is audible
when the song is played, don't use it (unless it lends a nice effect to the
song). There's no excuse for using a hissy sample.

Next is the looping of the sample - depending on the type of sample being
looped, it can range from being stupidly easy (basic woodwind instruments) to
extremely difficult (a vibrato violin). I'm not going to give a lecture on
how to loop a sample, because it's very involved. All I'll say is that if a
sample has an audible loop that is detrimental to the song, don't use it.
Simple. Oh, I'll also say this: Pingpong loops aren't what they're cracked up
to be. And X-Fade (I think it only exists in FT2) is grandtastic if you learn
to use it right.

Finally, oversampling. And I am constantly surprised at the amount of people
that fall into this trap. Oversampling, aka clipping or peaking, occurs when
a sample has been recorded too loudly, or amped beyond all recognition
(*FUBAR*). It's easy to see when a sample has been oversampled, so there's no
excuse for using one. Here's the difference:

[I'll try to send you two pictures, one of a well recorded sample, and one
that peaks.]

Caption: *In case you can't tell, it's the one on the RIGHT that's
oversampled. The RIGHT.*

See? Big difference. And the difference in look is almost equivalent to the
difference in sound quality. Some instruments may survive the process, but
most of the time oversampling just makes a sample horrid to listen to.

Like I said, there's no real excuse for using an oversampled instrument, but
one that many people will use is "I needed to give the song more bass/more
impact" or "I needed that instrument to be louder, and it was already
normalised and at maximum volume." This is where our good friend Mr Mixing
comes in, which is the next category.

Mixing:
I spent my first three years of tracking putting very little effort into the
mixing of my songs (That's all I did. I didn't make songs, or even eat or
sleep; I just sat there, not putting effort into my mixing. For *three
years*), but as soon as I started to, the difference was immediately
noticeable (once again, compare *Escape from Wrongful Incarceration* to
*Courage Under Fire*).

Hmm. It occurs to me now that it's hard to write a manual for how to mix a
song. I'll start off with the basics, I guess, and maybe I'll be able to give
some more cryptic advice.

To begin with, you should almost *never* have a song that has every
instrument at maximum volume for the whole song (40 in FT2; 64 in IT) - in
fact I very rarely have *any* instrument at full volume - usually the loudest
I allow an instrument to go is 30 (48 in IT), so if I want a part of the
song to be particularly loud, *then* I go nearer to 40 (or 64). Anyway. It's
basically up to your personal taste which instruments you want to be at which
volumes, all I can really tell you is to make sure that there is contrast
between them, and that the end product is nice to listen to.

So there's your ground rule. Now come the little tricks you can use when you
want an instrument to be louder than the rest, in lieu of oversampling.

The first option's pretty obvious, but I'll tell you anyway - back up the
instrument. Don't double it up (that's the equivalent of oversampling), back
it up - have another (preferably similar) instrument playing the same tune.

The second option's not so obvious. Double up the instrument, but put some
space between them, panning-wise. Then flange the sound slightly by either
delaying one of the instruments a little bit (using the note delay command),
or pitch-bending it very slightly (the fine pitch-bend command) - there are
other commands you can use (such as the sample offset command), but those
two are the ones I usually use. This will make the instrument you're using
this effect on more prominent, without oversampling.

However, depending on the tracker you're using, you'll have to approach this
effect in different ways. If you're using IT, it's easy; you can just use the
track panning screen (F11) to set the panning of the note delay/pitch-bended
track, but if you're using FT2, you've already used the effects column to
apply the effect that flanged the instrument. So what you do is use the
default pan setting of the instrument to pan it where you want it to be -
this may mean you have to manually set the panning for the instrument for the
rest of the song, which can be annoying, but it's worth it.

That's about all I can help you with about mixing. A lot of it's personal
taste, and, to a lesser extent, trial and error. Play around; see what sounds
good. Just make sure that it *does* actually sound good.

The same goes for panning - all I'll say is that a song with interesting
panning (as long as it's not instrusive to the music) immediately catches my
attention, and it pays to be adventurous. Just don't overdo it. Much.

--


Gwilym Wogan
E-mail: vanguard_gw@hotmail.com
MP3.com Site: http://www.mp3.com/GwilymWogan
Trax in Space Site: http://artists.traxinspace.com/Vanguard
Homepage: http://vanguard.zerion.com