Original 8-Bit Composers: How did they do it?

[Elysianhero]

Like any good nerd who grew up through the end of the 80's and beyond, I love 8-bit music. Sometimes, simply nothing beats it. Zelda, Contra, you know you love it. But, sometimes the theory behind it seems like an ancient science, or something blocked by the Chinese government. There is almost no information to be found on this old method of songwriting that has as much to do with composing as it did programming!

I've been trying to study 8-bit composers. Can any one in the community help me find some resources that would allow me to follow (in theory) the practices of a vgm programmer, what a day in the life would be like? It would be greatly appreciated.

[lazygecko]

One channel almost always plays a lead melody, the second channel will constantly alternate between assisting the first channel with echo notes, chorus, add harmonies or arpeggios etc, to keep things from getting stale.

[OverCoat]

Well I know a couple C64 guys made their music with BASIC... :3

[mickomoo]

I always imigined the process would be similar to that of composing non 8 bit music, but merely changing a few notes here or there to make it sound better as an 8 bit composition

[Yoozer]

But, sometimes the theory behind it seems like an ancient science, or something blocked by the Chinese government. There is almost no information to be found on this old method of songwriting that has as much to do with composing as it did programming!

And there you have the reason; composers had to build their own authoring tools.

As for "merely changing a few notes here or there" - nope. Even when you set out to create a cover version, you have to completely re-engineer the whole thing; when you create original work, you're already bound to the constraints.

Get yourself the old version of SIDPlay (not v2) - http://www.gsldata.se/c64/spw/sidplayw.html . Then, get the C64 High Voltage SID Collection. Try to find a SID file that's relatively recent (e.g. 1996 and onwards) - those are most likely to contain some of the advanced tricks.

In View -> Mixer you can mute individual channels (just make sure the emulation is set to Original, not reSID) - then you hear all the things going on.

An example;

http://www.theheartcore.com/music/sid_mute.mp3

The original track is "Mito" by Sebastian Kuehn (Rio). First everything at the same time, then the first channel, then the second, then the third, then the first again. You can hear some peculiar effects when the arpeggiator isn't used - it sounds very metallic. This goes way beyond the regular bleeps; it's audio-rate modulation of waveforms.

If you can't properly hear what's going on - slow it down. Play it back at half speed - you can use Audacity to tranpose wave files down.

[jmr]

Here you are:

http://theshizz.org/forum/index.php?/topic/31625-how-nes-music-was-made

A collection of interview snippets and anecdotes from various NES-era game composers.

[Kizyr]

Episode 33 of OCAD has an interview with Sivak, the creator of Battle Kid I/II (new NES games available at Retrozone).

I mention this because he went into some detail on how he composes the music for the game. He uses a few more modern tools so that it's not necessary to play-test the game to hear the music, but it still gives an idea of how it's more challenging than just putting together the notes. KF

[lazygecko]

Before there were actual interfaces for composing you had to also be a programmer to do music. But it's nothing you actually need to dablle in today unless you're a mega purist. Just trying to mimic the old techniques with modern equipment is enough to get "authentic" sounding chiptunes. The best middle ground would be something like FamiTracker which lets you make NES music in .nsf format using a tracker interface, but even though I know how to use it, I prefer doing all my stuff in FLStudio still.

Here are some common chiptune techniques you'll want to know about

Fake reverb - This is just a simple case of decreasing the volume of the note after it has been played for a while. On a tracker you'd just do this by inputting a number, in other tools you could just use volume envelopes, or the way I usually prefer doing it by just inserting a new note with a different volume, like so:

Chorus/echo - These are usually done together. First the echo is made by just playing the same notes as the first channel delayed a few steps in the second channel with a lower volume, but this usually results in some annoying phasing. You solve that by detuning the pitch on the second channel a bit, say 12-16 cents or so, which creates a nice chorus effect.

Next are some tricks for squeezing in some additional harmony into your music, since channels were limited back then composers had to get really creative with getting the music to sound fuller. The most familiar method is by playing very fast arpeggios which was widely used by European composers. But the most basic method of doing so can sound pretty harsh, so they would smoothen out the sound by making the arp pitch bend. In trackers this can get pretty cumbersome since it involves some pretty manual labour by inputting numbers, but most synths today support this in a quick and easy way by just enabling legato and adding a tiny bit of portamento.

The more subtle way of doing things is by mixing the fake reverb trick with different notes, usually played a few notes down on the scale. This works especially well with fast playing riffs, think it was a pretty common method among Konami composers and I know virt loves doing it.

By getting familiar with all of these techniques, mixed with the stuff I wrote in the earlier post which was the standard procedure for the Japanese way of chiptune composing, it's not as hard as you might think to compose music that sounds like Mega Man, Castlevania, etc.

Here is some stuff I made all in FLStudio

http://www.youtube.com/watch?v=xkcl8mPp0z0

http://www.youtube.com/watch?v=ZNi-Bd8WbfA

[Calum]

Yes, props to GeckoYamori! Very helpful!

I'd say the main concept is the fact that each channel can only play one note at a time and you only had a limited number of channels. Thus the arpeggio was the best way to create a sense of harmony and in terms of melody-writing you had to be precise and certain! There's more often than not a clear melody part, a clear bass part and a clear counter-melody/harmony part + drum/noise part.

Certainty certainly makes chiptune!

[Gario]

Good mini-tut on recreating chiptunage, Gecko. The downtuning bit on the chorus is something I didn't know about, so that's one more tool in my arsenal.

Some additional notes on that, if you really want to emulate the sounds of the NES and such. The good ol' NES was equip with the ability to play five channels, for most games - two variable channels that could play different ratios of saw/square waves, a channel that played a triangle wave, a 'noise' channel and a PCM channel (capable of converting SFX of up to... 8kbps, if I recall correctly). The square/saw waves allowed for a great deal of flexibility while the triangle channel doesn't allow for volume modulation nor timbre change.

That's why NES composers often played tricks on the system in order to create depth in their sound - they didn't have the channels to spare to create harmony and such otherwise. Other systems had more to work with (SNES, for example, had 8 channels, and they used sampled sounds to produce their effects, not waves), so these things should be kept in mind when emulating older VG music.

By the way, that's why I consider Silver Surfer to be the true apex of NES chiptune sorcery. The shit they did with that system... I mean, DA---------YUM.

EDIT: If you want a more visual example of how to use three channels, take a look at this.