View Single Post
  #3  
Old 2011-10-03, 19:22
Dystopia's Avatar
Dystopia Dystopia is offline
Post-whore
 
Join Date: May 2004
Posts: 1,068
Send a message via AIM to Dystopia
Nice schematic there!

Yeah, the thing that was getting me is the low voltage coming from the guitar. Fuck it I'm hooking it up to an oscilloscope without power... I just couldn't wrap my head around not having a source voltage, but I now I see that it's the strings, and the pickup is just what, a transformer/coil? Doy! And the signal goes to the output without a voltage first going into the guitar, pushing it through, unless it is amplified internally with active pickups.. Okay, all making sense now.

So I just need to calculate the voltage and frequency output range of an non-amplified electric guitar. (F-range = about 60-1320 HZ for a 7-string, plus a little on the sides to make sure I get it it all... this really just has an effect on the sampling frequency, which would be an absolute minimum of 2.64kHz, which is kind of slow)

From there, I would need to divide the voltage signal into bits, and I'm thinking a 24-bit Analog to Digital converter, which would mean 2^24=16.77*10^6, or 16.7 million different possible "Sounds" that can be recorded into memory, which is the other thing I have to know... (memory type)

See, I know how to program an 8051 trainer to take information in (although the only form of INPUT information that we experienced was in the form of DIP switches), and since it was a trainer there was limited memory (the 8051's range of selectivity depends on the size of the the data-pointer, 16-bit we were using, meaning it could internally look up any instruction located from 0000000000000000 to 1111111111111111 [or 0000-FFFFhex].

Basically I'm gonna either need a PLC or a microprocessor that will move enough information fast enough to run recorded-music data through a digital to analog controller at the speed of real time so that when I put it through an op-amp to brush up the signal (which I might have to, I don't know) then that output signal will look exactly the same after being coded and decoded as it did when I first plugged it in, or at least as close as possible/enough to fool human ears.



So yeah, if there is any confusion what I mean is I'm going to have to program a microprocessor to go: "Take this byte + put it there, take second byte and put it behind that one" - basically doing that a few thousand/more times a second, filling up a memory storage (still not sure on what memory to use, we've only used small-scale memory and barely touched on how a few different types of memory work), and that process is strictly for record mode, which I will combined with other tracks (record bass while playing previously recorded drum-track, the timing of that processing that properly will be fun). The Assembly code would be something like, "MOV A@DPTR, MOV R0, A, INC DPTR, loop" or maybe something more efficient.

I also plan to do the same with a 5-string bass and a Simmons E-drum module.
__________________
t('~'t)
( ( ) )
Reply With Quote