Electronic Guitar Tuner

1
Electronic Guitar Tuner

• K. Wettstein A. Wunderlich
• Senior Design Project
• Team 45
• May 3, 1999

2
Target Specifications

• For use with acoustic guitar.
• Accurate to 10 cents (musical term) for entire
  range of guitar pitches
• Easy to read output LED display

3
Background - Musical terms

• 1 octave 12 pitches
• octaves are numbered (lowest to highest)
  according to range of piano.
• 12 pitches A, A, B, C, C, D, D, E, F, F, G,
  G
• 100 cents defined between each pitch

4
Background - Guitar

• Acoustic guitar ranges from E2 (pitch E in
  pianos second octave) to C6.
• Plucked string Fundamental frequency and
  harmonics
• Guitar Spectrum fundamental frequency has lower
  power than some higher harmonics.
• Lower frequencies have more power in harmonics
  than higher frequencies.
• Demonstrates acoustic non-linearity of string

5
Guitar Frequency Spectrum
6
BackgroundFrequency-Pitch Conversions

• Frequency and musical pitch are related on a
  logarithmic scale
• f(octave, pitch) 16.352 2octave pitch/12
• or
• octave pitch/12log2(16.352/frequency)
• where pitch has values 0-11, starting with C.
• cents are represented by the first two decimal
  places of the pitch value.

7
Design Block Diagram

• 3 main components input, DSP, output

8
Design Input

• Microphone to receive signal
• Possibilities dynamic, condenser, pressure
  gradient microphone styles
• Decision dynamic
• Why? Frequency response and range, availability
• Amplifier To increase microphone amplitude
  (output Vpp 5-10mV)
• op-amp with gain of 100

9
Design Digital Signal Processor

• Motorola 56302 DSP Evaluation Board
• DSP board provides
• analog-to-digital conversion
• Fast Fourier Transform (FFT) of signal
• lookup tables for frequency conversion to pitch
  and cents
• digital output to LED display

10
Design Digital Signal Processor DSP Algorithm

• 1.) Ready State for continuous, real-time system
• check input amplitude
• below threshold, stay in ready state
• above threshold, proceed to FFT
• 2.) Take 1024 sample FFT. Store squared
  magnitudes in memory.

FOR MORE INFO...
See Handout for flowchart, DSP Algorithm.
11
Design Digital Signal Processor DSP
Algorithmcont.

• 3.) Start with ninth index, compare magnitudes to
  a new threshold, which is set above the level of
  the noise.
• 4.) Find first peak in FFT data. This
  corresponds to the fundamental frequency.
• 5.) Store this index value, use it for pitch and
  cents lookup tables.

12
Design Digital Signal Processor DSP
Algorithmcont.

• 6.) Write corresponding values to the Data
  register for Port B.
• 7.) Port B outputs values to display.

13
Design Output

• Digital high/low values from Port B of DSP board
• Output display
• hexidecimal display for note letters
• 1 LED indicating a sharp
• 10 LEDs 5 corresponding to sharpness in units
  of 10 cents, 5 corresponding to flatness

14
Verification

• Test signals (pure sine and square waves) from
  function generator.
• Tested final circuit with guitar and compared
  output to commercial guitar tuner.
• Verified correct identification of peak from
  memory contents.

15
Design Cost

• Parts cost
• Item Cost
• Motorola DSP Eval. 165.00
• Amplifier
  2.00
• 11 LED's
  5.00
• Hexadecimal Display
  2.00
• Microphone
  120.00
  
  Misc. Comp. 2.00
• Total Parts
  296.00
• Labor 3600.00
• Complete Cost 3,896.00

16
Summary

• Designed and implemented tuner.
• Accurate within approximately 10 cents.
• Accuracy could be increased by taking longer FFT.
• Lower sampling rate
• Easily Marketable