Eventually this 8-bit thing is going to be just as played out as steampunk. BUT! In the meantime: Someone has put together this fantastic NES-style arrangement of the final track from Dr. Horrible’s Sing-Along-Blog, Everything You Ever.

The arranger, on the original post, mentions that he’s actually done the entire thing and plans to release it en masse complete with sprites of the major characters doing their thing. I look forward to it!

Backseat Driver ©2009 Leif Chappelle

Another week, another chiptune! Well, that’s not entirely true. I’m officially one week off due to a bombastic weekend that involved much less time holed in my studio and more spent outside in the gorgeous weather. Not complaining, though! Anyways… With this ‘tune’s creation in mind, I do have another 8-bit ramble in me.

When writing chiptunes today, we often forget exactly what those hardware limitations were back when the games were actually being created. Yes, we have a whopping four tracks available to write in… But, there’s also those pesky things called sound effects to worry about. Not only does the music need to fit in there, so does everything else audible!

You’ve probably noticed at times, playing an NES game, and certain parts of the music cut out. There’s precisely the reason. Wonder why racing games never had in-race music? Likely both Square channels were being used for engine sounds: yourself, and others passing by. Most commonly used for sound effects was the Noise channel, since it creates assortments of white noise that can be shaped by filters into something vaguely recognizable: sword slashes, things hitting the ground, etc. Because of this, though, compositions had to be even more limited. Or, at the very least, compensate for the potential interruption.

The Noise channel also produces one of the most potent sources of percussion for a track. The problem being, if sound effects are getting mapped to it, what you end up with is a very spotty beat that keeps getting interrupted by player actions. As a result, not many tracks used percussion very often. It was saved for special moments that perhaps didn’t require very many noise-generated effects. Games like the Mega Man series however, used the tonal channels for pretty much all of their sound effects, allowing for full percussion and much more rhythmic tracks.

Thankfully, for those just interested in the pure creation of music using those limitations, all four tracks can be used to the furthest extent of their possibilities. In the track above, I used a variety of techniques — most just found by tinkering around — to create what could be construed as as drum kit, breathing, and wonky electrical chaos. The other trick, of course, is creating all of this with a single line of notes without any overlapping bits. Counter-rhythms and the like need to be built around the concept of taking turns playing, which leads to an interesting puzzle of creating the illusion of multiple percussion elements playing together.

It’s a puzzle I’m still working out, but I’m decently amused by what came out of this track. Enjoy!

Celluloid - ©2009 Leif Chappelle

After posting last week’s article on my exploration of chiptunes, commenter Kurt posted an excellent reply about something I had yet to cover in the first installment. He mentioned the compartmentalized nature of chiptunes and how they are less about the actual physical notes written and more about how they’re implemented. This is absolutely true. With this week’s installment in what I’m hoping will become a weekly or bi-weekly look at the world of lo-fi music creation, we’ll be looking at exactly how the modular nature of chiptunes and trackers can hinder or aid the writing process.

As with any tracker, chiptunes are created in a format that is optimized to be the most efficient way possible to create music using as little memory as possible. Since size was a huge constraint, this format of music storage and playback was indispensable while games were still hosted in a cartridge-based medium.

For example: While the mp3 attached to this article is 3MB in size, the entire song when saved as a .nsf (nintendo sound file) is all of 8KB.

The way this is achieved is because of how the music is generated and stored. For each sound file that is played back through the NES, there are three major categories by which the song is stored:

• Instrument data
• Pattern data
• Frame Order data

Within the instrument data, each modified wave form is given a set of parameters and assigned a hexadecimal identity. For example, the piece I wrote contains several major instruments: Vibrato Lead and Vibrato Echo, my two main melody/harmony instruments; Plain Triangle, my bare-bones bass synth; and a variety of drum settings. By applying filters such as volume curves, arpeggio, pitch bending, and duty cycle, with just a few parameters one can create an arsenal of different variations on the four channels available.

Within the pattern data, each of the four tracks are assigned cellular ‘patterns’, essentially one or more measures of music, that can be played back in any order. Based on how many rows per pattern are designated in the settings, each beat is given either a note value, a null value or left blank. Each beat can also be assigned an instrument, volume, and up to three effects. As a musical example, a whole measure of 4/4 time would be given 16 rows if you want the least possible value be 16th notes, or 32 rows for 32nd notes.

Finally, in the frame order data, each pattern is assigned to a frame (all four tracks playing a specific pattern). The order becomes a matrix of how everything will play back chronologically. For example, here is a snippet from the piece above during the 1st verse:

0A | 04 05 00 02
0B | 05 04 01 03
0C | 08 08 00 02
0D | 05 04 01 03

Now, to understand better, know that the numbers for each column relate only to the channel that is playing. The far-left numbers correspond to the frame number. In regular notation, this might be the measure number. The easiest channels to comprehend are the Bass (channel 3) and Drums (channel 4): The bass has a two measure pattern it is repeating: 00, then 01. So do the drums: 02, 03, etc. The melody, however (channel 1) has a slightly different pattern. Its first and third measures are different, but it returns to an identical phrase in between, making its melodic line an ABCB pattern. By acknowledging these repeating elements in the music, the composer can save on space and call back phrases without the need to create them anew.

I hope everyone is still awake! With all that said, go ahead and take a listen to the attached mp3 and see if you can’t identify how the compartmentalized nature of the format is being used and in what way. I’ll be back next weekend or the following one for another bit of passed-along explorations!

As an addendum, big props (which I’ve already given via email and twitter, but more doesn’t hurt!) to Jeremy Parish over at 1up for the plug on last week’s article! The traffic boost since then has been enormous and all the ‘herders appreciate it.

Maelstrom - ©Leif Chappelle 2009

You may be tempted to hit play, but hear me out first:

I’ve long been praising chiptunes on this blog — for good reason, mind — without actually taking the dive myself. There’s one thing that can be said about music, non-tangible as it is: It’s hard to really understand at its core until you get your hands dirty and do it yourself. You can appreciate it, you can jump up and down in excitement at a rockin’ chorus, but understanding comes from doing.

Does that sound too pretentious? I hope not!

Granted, the doing can be applied to performance just as much as it is creation. Singing your favorite song or melody can act as a form of performance that enhances understanding. Thing is, with stuff of a digital medium it’s a bit difficult to perform unless you’re a professional synthesizer operator. (Props to those folks.)

That said, this week I decided to take that dive and try my hands at tracking out some old-school chiptunes. It’s something I’ve always wanted to do, but lacked the knowledge or experience to wrap my head around it. Thankfully, I found a lovely little program called BoyScout that enabled me gain just that: a toolbelt filled with sound emulation from the late GameBoy.

Not really a looker, is it?

The above are two screenshots from a portion of the BoyScout interface, the left-most being a single cell of music data from one of the instruments, the right-most being the overall sequencer for all four tracks.

Knowing my game music history helped a bit with this one: The original GameBoy was an 8-bit game machine, meaning that it had the capacity for four channels of audio playing at once. Two channels were dedicated to square waves, one channel had a customizable waveform, and the fourth provided white noise. The channels were used universally between both music and sound effects, meaning that composers had to keep in mind that at any moment one of their precious channels could be taken over by a sword slash or jump.

That entire concept however brings with it an inspiration for polyrhythmic writing. Since the channels can change their settings on the fly, so long as the two lines don’t overlap, one can create multiple “voices” in a single channel to create the illusion of far more than four things happening at once.

And so the idea for an almost overwhelmingly polyrhythmic piece came to be. There are multiple repeating ideas that could be grasped onto as the concept of a down-beat, none of which entirely wrong, but none of which totally stable either.

There will certainly be more to come!