Re: [stella] Announcing "webTune2600"

Subject: Re: [stella] Announcing "webTune2600"
From: Doug Dingus <opengeek@xxxxxxxxx>
Date: Sat, 13 Aug 2005 17:17:17 -0400
On 8/13/05, B. Watson <atari@xxxxxxxxxxxxxx> wrote:
> On Fri, 12 Aug 2005, Doug Dingus wrote:
> > I did this long ago on the 8bitters. It's true @ 60Hz you will hear it. 
> You
> > hear it up to fairly high frequencies. What happens is the ear 
> translates
> > these into extra notes or it's heard as a pitch shift happening rapidly 
> if
> > the frequency is low enough.
> I was afraid of something like that...
> However... if I'm shifting at 60Hz, what's the extra note being heard? 

Actually you hear a coupla things. You get the 60Hz note, plus this wierd 
updownupdownupdownupdown pitch. 60 isn't fast enough to blend, so you hear 
both rapidly next to one another. (It's a cool effect, but not really 

> 60Hz tone? What if I use a tuning where 60Hz is an in-tune note (it's
> between a B flat and a B... I used to know a guy who tuned the B on
> his guitar to the 60Hz hum coming out of his crappy amp, and the rest
> of the guitar to that... it was the only way to make music instead of
> noise come out of that amp).

LOL!! If it's all in key, it works!

Bleah. Even if that works, I wouldn't be able to play a normal
> bassline. The 60Hz B's would clash with anything else in the bass.
> > One thing that does work fairly well is to alternate the pitches every
> > cycle. --with some pattern, 112112112, 121212121, etc... Let each tone 
> run
> > for exactly one cycle, then start the other one. I'm not sure the 2600 
> has
> > the muscle to actually get this done however.
> Not every CPU cycle... could use a hard-coded string of STA/STX and do
> it every other cycle. Of course, that's *all* the 2600 would be doing
> (no graphics, no game...)

I don't think that would get it done either. The cycle time varies with 
pitch. I pasted one cycle of one frequency in with another into an audio 
editor, then copied that to build up a second or two of audio. Too many 
timing problems at the 2600 speed. When I wrote cycle, I meant one audio 
cycle, followed by the other one. CPU cycles just don't match up.

Hmm, Manuel was talking about volume-only sound for playing 4-bit samples.
> How often do you have to update the volume register to make anything sound
> halfway decent? If every other scanline is enough (60 frames * 262 lines,
> divide by 2 = 7860 samples/sec), I could imagine doing a simple game
> with sampled audio using a 2LK and LOTS of ROM banks for sound samples.
> ...time passes...
> Ugh, I just tried converting a song to 4-bit samples. Either I'm doing
> the conversion wrong, or 4-bit samples sound like everything's going
> through a distortion pedal.
> What I did was convert a song to 8-bit unsigned samples, then write
> a little C program that strips off the bottom 4 bits of each sample
> (PC sound cards don't support 4-bit samples natively, so I'm trying to
> simulate them that way). The 8-bit version sounds OK, and my stripped
> version sounds like utter crap.
> I know next to nothing about digital audio sampling... is there a smarter
> way to convert than just removing the bottom 4 bits?

The other posts have this exactly right. I would trim the audio a bit 
farther below the limit however, but that's just me. You should be able to 
get AM Radio quality sound (a good AM radio that is) out of a 2600 that way. 
4 bits can sound fairly good actually. 

I've a question along these lines. The 2600 has two 4bit audio channels 
right? Using them both together could get more than 4 bits of sound. What's 
the maximum audio output without distortion? When they add together, is 
there room in the rest of the circut to handle both channels? I remember the 
8bitters having some limit on the 4 channels. If the sum of the 4 was too 
high, distortion would result. 

> B.
> Archives (includes files) at
> Unsub & more at
Archives (includes files) at
Unsub & more at

Current Thread