Jumping this up.
What sort of waves make the sound effects in old games? like a firebal or a sword hit?

Is white noise have a uniform random variable for every sample?
By attack you mean increasing the amplitude?

Quote:Original post by Dyamios
Now days people just use samples (wav files and what not) and plug them in to tracking and sampling software (like modplug and renoise).

All synthesis here, baby. Learn your oscillators and envelopes! What's the differnce between VCA and VCO? Find OUT!!!

Tony

Voltage Controlled Oscillator and Amp :)!

yeah, I made that easy. ;)

Come on now fellas, don't tell me y'all of all people are at least a little more familiar with the older techniques?

Well anywho, I stumbled across this thread while doing a search on old YM chip simulation VST's, and as a VST programmer myself and an avid vintage gamer, I just couldn't let this slide by. ;)


First off, the NES sound. I've actually taken a stab at simulating its sound mathematically verses by samples (a yet-unpublished work of mine), and as some of you more scrutinizing members may notice the NES has a unique sound compared to some other analog wave forms. Basically, the structure is like this.
2xPulse waveforms, with programmably-variable pulse width
1xTriangle waveform
1xNoise source
1xDAC source, for those awful (:]) sounding samples

You have to remember though that part of its signature "Lo-Fi" sound stems from the fact that indeed it has a considerably low bit-depth for generation. That makes the Pulse waves sound a little thin and makes the triangle wave have an almost square wave sound.

The most difficult thing I've found in emulating the chip is the noise source. I've found two important things through using ROMS designed to control the chip and reading:
1 you can achieve the noise sweeps sound by running a white noise source through a bit-depth reducer effect, or similar. I've tried running it through a low-pass filter, but that is not the same effect as that is far from what they used (no conventional filters were used in that chip).
2 There is an effect commonly referred to as "Frequency Count Divide" that I can't get my head around yet in simulation. Essentially though, through my simple understanding and keen ear ;) , It will sort of chisel away the noise into specific harmonics, making it psuedo-tunable. My goal is to try to simulate this with a delay line, but I've yet to implement it.


And that's a simple overview of the NES chip.


Now when you get into the Yamaha chips, you get into more complex pure digital stuff. It employs the technique of "Frequency Modulation", and without being too longwinded about it, imagine a vibrato so fast that instead of hearing the pitch vibrating, you heard a whole new timbre of sound. That's exactly what's going on here. By modulating the speed of one wave (typically sinusoid, definately when it came to that era of YM chips) by another of a speed still in the threshold of human hearing (best way I can put it), you begin to add harmonics to the fundamental Sine wave, creating those harsh metallic tones. To get into the physical mechanics of this would be to write a college term-paper, so I won't. But basically, those chips are quite a bit more complex than the NES set.



Also to clarify about Envelopes and Attack and what not, there are typically four stages to an envelope, the Attack, Decay, Sustain, and Release. All of them except Sustain control a time variable. The Attack phase is literally how long it takes an instrument or sound source to go from complete silence to full volume, so a zero attack setting would be similar to a piano strike or (as was said) a guitar being plucked, but having a longer one would be used for say if you had a sound of a string symphony and you wanted it to swell in, it would achieve that. The Decay setting is how long it takes to go from full volume to the Sustain option's setting, quite simple. Sustain is how loud the sound will be for as long as you have the key or control set to on, after the decay stage has finished. The Release stage is how long the sound will take to go from the current volume to silence once the key or control is released. Whenever you initiate an envelope, typical characteristic is that it will automatically go from wherever it's at to the release stage with whatever level it has whenever the key or control is released. So yeah, technical, but thorough.



Whoa, did I just type all that? :/


Well, I just thought I couldn't let the questions go answered incompletely. ;)


P.S. I could delve into more technical synthesis techniques for those old-school sound effects if anyone has some more specific ones they'd like to see how to do.