Are there any tutorials on how to write demostyle music?

[Louigi Verona]

Folks,

I'm interested not so much in tech aspects, as in musical ones - how do you write a lead melody, how do you think about arrangements, etc. Is there anything at all that explains how to write demostyle jazzy kind of leads that are known in tracker music?

[Saga Musix]

I have no definite answer, but I think many potential answers lie in 80s pop music and how it was written. That's without doubt where many demoscene musicians drew (and still draw) their inspiration from.

[Magnar]

Hi,
Interesting question.

Myself, I have devoted almost 35 years making demoscene music to the Commodore 64 + Commodore Amiga, and I would say that things have changed a bit from how I approach the lead melody and arrangement as of today compared to how I did back in the beginning.

When I composed  Jazzy-leads songs back in the 90s, I often started of with some groovy drum arrangement,  then continue to do a funky bass line to the drums. On top of that I try different chord progressions, and when I got all that nailed down first, I somewhat started on doing the lead parts. Often this started of with a intro, maybe only the drum and bassline. And then the chords comes in to play, and later the lead. So it built up like when you work with a more modern "looper-station" workflow more or less in a way. And then after the build-ups, you change things and create in the end a ending to the song.

Today, I often go around the other way! I maybe whistling a lead melody while making dinner, walking outdoor or the classic in-the-shower situation. I then often use my iphone to record it. When I got the melody recorded, I often also try to sing in some beats and bass line I think would fit the melody. When I later get to my Cubase DAW, I start to think how to build up chords arrangement that better fits the melody and try to make variations of bass line to that. When I have a more complex structure figured out, I start to compose it in the Amiga tracker. Often to begin with I have a very straight forward basic drum pattern that just follows the flow. But as soon as I got tracked the melody, chords and bass down I often work more intensive on the bass and drums to make them complement eachother better with the chords and the lead. At this point, this sequence might fit as the middle of the song, either as a chorus or the verse. Then I start to add content before and after the sequence. Like a intro, some fast song patterns and some slower song patterns to create variation in the whole song structure. And in the end, I compose the ending pretty much same way as always.

I hope this explains a bit of my workflow and how I do it. The Jazzy leads are more or less just things that comes in my mind as I work on the song. Playing it over and over again and doing small changes to the lead as it works it way through the patterns. It is very hard to explain it otherwise that it just "feels" like this is how it should sound together.

Before, I often also focused a lot of the technical parts of drumbeats, basslines and chords and tried to fit that with the lead. Today, I think I focus more on the melody and try to make it shine through more in the song than before. The technical stuff kinda took over a lot in my 90s songs, and the melody itself suffered.

Hope it helps.

Cheers Magnar

[Louigi Verona]

Hey Magnar!

Yep, that sounds good. But what about the melodic tips?

So, here's what I am searching for.

1. What kind of chord progressions does one typically use
2. Many 4 channel mods use samples of chords in interesting ways. You know, when a sample is a string pad that plays a whole chord. Are there any tips on how to do that? Writing something with those samples is actually quite puzzling, the way they switch them around and use different notes of chords.

[Magnar]

Hi,
In a tracker-program where you compose modules by replaying sampled waveforms, the common way to create samples of chords was to do a sample of the whole chord and all chords needed for creating the song. The reason for this was that the Commodore Amiga only had 4 channels 8-bit audio playback possibility by the "Paula" soundchip in that computer back in middle of 1980s. So, in the old days when the Tracker formats and SoundTracker/NoiseTracker/Protracker etc was publicly released to the masses, if you wanted to create a song and playback a chord, the best way would be to connect a "real" synthesizer/midikeyboard to your audio-sampler hardware and then sample the audio-output from that hardware sound-device of a chord. That would mean you needed to create and make a sample of every chord the song was in need for. Each chord sample would then only takes 1 audiochannel for the playback. That would leave the other 3 audiochannels on the Commodore Amiga for other playback such as; drums, bass, lead sounds. Which was the common setup of soundstructure for songs composed on the Commodore Amiga.

When the PC came around in the mid 1990s, with new and better soundcard possiblities that allowed more audiochannels than was possible on the Amiga, it then made sense to stop sampling chords and instead only use one sample of the root keynote and playback that sample in the amount of audiochannels necessary for creating the chord itself. This became a preferred way to do it for example when Fasttracker 2 on PC got released to the public masses etc. By doing only one sound sample to create many type of chords, it would save a lot of memory. 

So, to summarize: On the Commodore Amiga you would still need many audio samples for doing the variation of chords needed in the composition of songs. Every new chord spent in the song, would result usually in a new sample. But.... There is a "mathematical" way to avoid a new sample, and that is by changing the root note of the chord during sample playback. The most typical and most common samples of chords, I would say was C minor and C major. Those 2 chords could further be played with a F-root key and then you get a F-minor or F-major playback. If you used them as a G-root key, you get G-major or G-minor etc etc. So, by sampling a major and minor chords, you could make quite many chords out of only those two samples.

For composers that came from Commodore 64 and was known to create SID chips with Arpeggio chords, the mathematical way of looking on creating chords was "0-3-7" for a minor chord and "0-4-7" for a major chord. Basically, those numbers are the numbers of transpose you would need to shift from the root key in order to create the whole chord. This mathematical way of looking at a chord also means that it was very easy to use Minor and Major chords on Amiga to move these up and down and fit in different other chords playing along side the melody. For example, if you go from C major to A minor, that would be easily done with C 0-4-7 to A 0-3-7. But, you could also use E 0-3-8 instead of C 0-4-7. Or you could use C 0-4-9 instead of A 0-3-7 etc. Basically, you just need to move the root key inside a chord and use the next combination that fits in the scale to move up or down in a chord structure.

So, by using that mathematical solution, sometimes the C-major and C-minor chords was not the optimal chords to sample. Because, when you had to shift them too far in different root-keys, the sound quality of the chord would suffer. For example, if the C minor is played at C-3 key, and then afterwards E-2 key, thats last chord sample will sound very "low quality" due the fact that the original sample was done in C-3 key. Therefore, maybe was it better to sample instead a G-2 Minor sample that played first at F-3 and then G#2. Both which would give same result as the first C minor played at C-3 and E-2.

Ok, that was a lot of technical stuff in regards of why Chords was sampled the way they where. And also why the chord progression often was a selection of best-efforts in how to make it sound good as the root-key was shifted during the song progress. This might sound very puzzling to you, but I hope it make more sense now when you read all the above and it would answer your question nr.2.

For question nr 1, I would still say that a C-3 Minor and Major chord samples was the most common way to use. There was also some "jazzy" versions of these, C minor7 is very often used in Amiga songs with more jazzy flavour to it. Also that chord can be played back in very many root keys and still one chord sample sounds okay in the mathematical way I explained above to fit the chord structures.

So, hopefully, my answers above together with the explaination made you any wiser - or more confused.

/Cheers Magnar

[Louigi Verona]

No, no, this is very helpful. I mean, don't be afraid to be super technical with me, I am making music for over 20 years now.

My goal is to create a sort of tutorial on how to make something that sounds like your classic 4 channel mod. So, it is these subtle details that I'm looking for.

I understand what you are saying regarding the chords, will experiment with it a little bit. One thing you can do, btw, is have looped chords, which negates the problem of lower quality sample, because it then simply becomes a string, as opposed to a keyboard chord, and the change in quality is less noticeable.

[Magnar]

Well, in terms of looping - I always use looping on my samples. And fade them down with A0F / 60F etc commands. That way, even a Hihat gets lots shorter in memory.
Most of the samples are unnecessary big and for demostyle music to fit music with code and graphic in demos, the memory size of the song is a big thing.
For example, if the total production is to compete in a 64kb intro compo, the song must be very small.

Many times I keep on cutting down on my samples so much that they are barely resembling what they should, but by using volume controls to create delay effects or simply fade outs on looped area instead of having larger samples that contains a fade-out in the sample itself, that is something technical you can do that saves a lot of memory.

So yea, looping chords is ofcourse a way to keep the chord sounds like a very long sample without actually have to use that much memory.

/Magnar

[Louigi Verona]

I this your profile at modarchive?
https://modarchive.org/index.php?request=view_profile&query=82021

[Magnar]

Hi, yes. That's me. I might become a bit older than on that profile picture tho /Cheers

[LPChip]

When I make demostyle music, I just set one simple goal to myself: Create something that sound as awesome as it can be. Be a demo of what is possible. This mindset alone helps a great deal in starting, because you ask yourself: what is awesome?

And by answering this question, one would say suspense creates awesomeness. So what I tend to do is combine sounds together to create suspense. This can be an ambient feel, or a drone or even a great acid bass loop. But it must be something simple in the background that carries that moment in the tune.

When it comes to the lead, the first question is: is one really needed at that moment? If you have a great slow drum with an ambient feel, that alone can be good enough on its own. If you still want to add melody to that, ask yourself: how can I write a solo that goes with this? Like how a lead guitarist can play a solo.

Next thing I often try to do is keep my sections short and use transitions to go to something completely new.

Also, as we're creating a song to showcase what is possible, use effects to spice things up. This can be an arpeggio, but also delay, chorus, distortion, etc.. As long as it fits with your song and taht moment in time.

That said, keep in mind that demostyle is a very broad range because it has different interpretations to everyone. For example, some people think with demostyle: music used in demos, but more modern demos typically use pop/dance/electronic music mixed with demostyle elements which evolved into just pop/dance/electronic music being used in demos. Technically speaking they are not demostyle tracks.