ECE 4760: Laboratory 3

Sound sequencer and synthesis.

You will construct a device with produces a a series of musical notes. The note sequencing will be determined by a Markov process. Synthesis of sounds will be controlled by your choice of additive, FM, or Karplus-Strong algorithms. A keypad will be used to set Markov and synthesis parameters.

Examples from 2012:
From Ranjay Krishna: Fur Elise, Star Wars, Harry Potter

Procedure:

You will build a syhtnesizer with audio output to amplified speakers. Analog output will be in the form of a PWM signal which will go to the audio input of the speakers. The PWM signal will be low-pass filtered enough by the speakers. The PWM signal will be mixtures of sinewaves generated using Direct Digital Synthesis (DDS), then added, multiplied and/or filtered. . For this lab, you will need to produce some waveform more interesting than a pure sine wave. This example shows how you can generate an FM modulated sinewave by DDS. This example shows how to generate additive synthesis. This example generates an audio sine wave and white noise. More explanation and examples are available on the DSP page. Use the timer0 fast PWM mode with the prescalar set to 1 to produce 62500 PWM pulses/sec, but only update the PWM at 8000 samples/second. A matlab program generates FM syntheisis for testing parameters and another matlab program generates power-law probabilities, normalized to 127.

You will need to get user input from a keypad with one of two following configurations. Some keypads have the connector on the top, as shown. Some have them on the bottom. A demonstration keypad scanning code for testing your connections is here. An example of a debounce state machine for a single STK500 pushbutton (not the keypad) is included here for reference.

Connector:  top:8-pin      top:9-pin   bottom
Pin 1  ---- row 1 2 3 A    col 1       col 1
Pin 2  ---- row 4 5 6 B    col 2       col 2
Pin 3  ---- row 7 8 9 C    col 3       col 3
Pin 4  ---- row * 0 # D    col 4       col 4

Pin 5  ---- col 1 4 7 *    row 1       row 1
Pin 6  ---- col 2 5 8 0    row 2       row 2
Pin 7  ---- col 3 6 9 #    row 3       row 3
Pin 8  ---- col A B C D    row 4       row 4
Pin 9  ----              (NO CONNECT--common)

(a) Each switch shorts one row to one column.
(b) Each pin should be connected to one bit of an i/o port.
(c) The i/o port pins will be used both as inputs and outputs. 
    When they are inputs, they should have active pullup resistors.
(d) On the 9-pin models, do not connect the common lead.


Hints on debugging are here. 

1. Write a C program which implements a sequencer and synthesizer with the following specifcations:
   • The synthesizer will generate notes with a fundamental frequency on a pentatonic scale:
     

     note names	C4	D4	E4	G4	A4	C5	D5	E5
     frequencies (Hz)	262	294	330	392	440	523	587	659

     (see note frequencies, scales, pentatonic)
     
   • The synthesizer must implement either FM, additive, or Karplus-Strong algorithms to make the generated sound more musical.
     There should be control over note dynamics including attack time and fall time.
     There should be at least two voices and the synthesizer should run at least 8000 samples/second.
     A few points will be reserved for making your code run at faster sample rates, or have more voices, or both.
     This example audio uses FM synthesis to generate two voices, each with its own modulation scheme and Markov process. The voices are tuned an octave apart.
   • The sequencer will use a first order Markov process to determine the order of the notes.
     One possible implementation is to build an 8x8 probability matrix in which the entries on a row of the matrix represent the probability of any of the eight notes being played next, given the note just played.
   • The keypad will be used to enter parameters. You must, of course, debounce the keypad.
     You should be able to choose one of 4 or more Markov matrices, one of which must generate a deterministic scale (play notes in ascending order). The others you can choose.
     You should be able to set at least 4 synthesis parameters (e.g. fm depth or attack time).
     When you set a parameter, its name and value should be displayed on the LCD.
     
   • The audio output should be free of clicks and other aritifacts.
   • Note frequency accuracy should be +/-0.5%.
2. Connect the circuit to a speaker audio input (green phone plug) and listen the output. This will aid in debugging and is required for the demo.
3. Demo this program to a staff member. Show that all specifications are met. At no time during the demo, should you need to reset the system.
4. Your written lab report should include the sections mentioned in the policy page, and :
   • How you generated the freqencies (DDS resolution, etc) and an measurement of how accurate they are.
   • How you generated the various time intervals required by this project.
   • The percentage of cpu time used to do the synthesis.
   • A heavily commented listing of your code.