Transcript Presentation Sample - DigitalSynth

Outline
• Digital music
• The power of FPGA
• The “DigitalSynth” project
– Hardware
– Software
• Conclusion
• Demo
2
Digital music
• Music has a powerful presence in our culture.
• Instruments have evolved from rocks and bones
to digital devices, becoming more and more
sophisticated.
3
Digital music
• Music has a powerful presence in our culture.
• Instruments have evolved from rocks and bones
to digital devices, becoming more and more
sophisticated.
• Digital instruments have been around for a long
time already.
• High performance = high prices?
4
The power of FPGA
• Newer families of FPGA are both powerful and
cheap. But are they suitable for digital
instruments?
• The main advantage is that the architecture is
easily reconfigurable.
• FPGAs have started to offer most of what
expensive chips do:
– high speed circuits,
– large memory blocks,
– multipliers, etc.
5
The “DigitalSynth” project
• The design is based on two components:
– hardware (the synthesizer with four-note polyphony,
capable of applying digital audio effects (delay,
echo, reverb, vibrato, flange), capable of playing
MIDI files and recording MIDI files),
6
The “DigitalSynth” project
• The design is based on two components:
– hardware (the synthesizer with four-note polyphony,
capable of applying digital audio effects (delay,
echo, reverb, vibrato, flange), capable of playing
MIDI files and recording MIDI files),
– software (the user interface that handles the MIDI
files, displays their music sheet and controls the
audio effects).
7
Block diagram
8
Oscillator unit
• The design uses oscillators to generate two
kinds of waves:
– sawtooth
– square
9
Oscillator unit
• There are four oscillators. Each of them
generates one wave whose type can be selected
by the user.
• The amplitude of the generated wave is
modulated  AM (amplitude modulation)
synthesis.
10
Amplitude modulation
• The ADSR envelope model is used for
modulation.
11
Amplitude modulation
• The attack phase starts when a key is pressed.
The decay phase follows.
The sustain phase lasts as long as the key is
pressed.
The release phase starts when the key is
released.
• Modulation is realized by multiplying the audio
data by the values of the ADSR envelope.
12
Amplitude modulation
13
Note selector unit
• There are two sources of control:
– the keyboard,
– the software application.
• When the user plays a MIDI file on the FPGA
board, the software application controls the
four oscillators.
In the rest of the time, the keyboard controls
them.
14
Keyboard unit
• The design uses 17 keys of the PS/2 keyboard.
Pressing a valid combination of one, two, three
or four notes will generate an equal number of
waves.
15
Polyphony arbitration
• There are 17 keys and only 4 oscillators, so a
polyphony arbitration rule is needed:
– if there are more than four keys pressed, the
“excess” of notes will not be rendered,
– if one of the four pressed keys is released, then a
new key can be handled.
• By keeping track of the status of the 17 keys,
the design determines which notes get to be
rendered by the four oscillators.
16
Polyphony arbitration
17
USB unit
• The design implements the Digilent Parallel
Interface Model.
• The software application controls the wave
type, the audio effects parameters and the
oscillators during MIDI file playing.
18
USB unit
• On the software side, USB communication
means calling the methods that are defined in
the “dpcutil.dll” dynamic library.
• On the hardware side, USB communication
needs implementing a parallel communication
protocol that uses an 8-bit data bus and
strobes.
19
Audio effects unit
• The design allows applying audio effects such
as delay, echo, reverb, vibrato and flange on
the sounds generated by the oscillators.
• The design generates audio samples with the
frequency of 48 kHz.
• These effects use samples generated at various
moments in the past, and for this purpose they
use a circular buffer.
20
Audio effects unit
• Writing in the circular buffer is done with the
frequency of 48 kHz.
• Reading is allowed in the rest of the time.
• Writing is done at a certain address which is
increased for every new sample of data.
• Reading at various offsets in relation to the
write address leads to accessing older data.
21
Audio effects unit
22
Audio effects unit
• Audio effects can be represented as a system:
23
Audio effects unit
• The delay effect requires reading data with a
constant delay.
• The echo effect is similar to the delay effect,
but its system has feedback, for more depth.
(The larger the feedback, the longer the effect
is perceived.)
• The reverb effect requires reading several data
with constant delays.
24
Audio effects unit
• The vibrato effect requires reading data with a
modulated delay. The modulation can be
sinusoidal or triangular. The system outputs
only the processed data.
• The flange effect is similar to the vibrato
effect, but its system has feedback for more
depth, and outputs a mix of the original and
the processed data.
25
SPI unit
• The result of the oscillators is processed by the
effect units or it can pass unmodified.
• In order to be output by the speaker, data must
be serially sent to the digital-to-analog
converter.
• The converter receives the samples and outputs
corresponding voltages. The speaker transforms
voltage values into sounds.
26
The software application
• “DigitalSynth” provides a user interface that
allows:
– opening format 1 MIDI files and displaying the music
sheet,
– playing the MIDI file on the FPGA board,
– recording a new MIDI file from the FPGA board,
– controlling the audio effects.
27
The software application
28
The software application
29
Conclusion
• FPGA architecture is suitable for digital
instruments (and more).
• DSP functions can be easily implemented.
• The Digilent Nexys board offers the possibility
to experiment and even innovate.
30
31