Skip to content

Latest commit

 

History

History
166 lines (137 loc) · 5.59 KB

README.md

File metadata and controls

166 lines (137 loc) · 5.59 KB
Raw

QAudioLib

An audio management library written in Qt

This library allows to manage stereo audio signals, 16bit PCM with a sampling rate of 44100Hz. It provides QAudioLibBuffer that overloads and improve the Qt QAudioBuffer, QAudioLibFreqBuffer that works with complex number to work in the frequency domain and Mathematics class that performs common signal processing operations like fft and convolution.

Build

This library is static, if you want to build and run the example benchmark TESTqAudioLib do the following:

#build library
cd QAudioLib
qmake *.pro
make

#build the benchmark
cd TESTqAudioLib
qmake *.pro
make

#run the benchmark
./TESTqAudioLib

This is an example of benchmark output:

Generate 1s square wave...
Done in 9ms

Calculate fft...
Done in 86ms

Calculate ifft...
Done in 101ms

Generate 5000 samples sin wave...
Done in 1ms

Calculate dft...(take a while)
Done in 627ms

Generate 1s sin wave...
Done in 7ms

Generate 1s white noise...
Done in 6ms

Lowpass 2000Hz filtering...
Done in 132ms

Play noise...
Got a buffer underflow! <--This is not a problem, it warns that the buffer is ended

Code documentation

Default format

Currently QAudioLib works with an own audio format: DefaultAudioFormat:

setByteOrder(QAudioFormat::LittleEndian);
setChannelCount(2);
setCodec("audio/pcm");
setSampleRate(44100);
setSampleSize(16);
setSampleType(QAudioFormat::SignedInt);

Use QAudioLibBuffer::getDefaultFormat() to get the default format.

QAudioLibBuffer

Constructors

QAudioLibBuffer(); //a void, no samples buffer
QAudioLibBuffer(const QAudioBuffer & other); //a copy of `other` buffer
QAudioLibBuffer(const QByteArray & data); //a buffer from bytes `data`
QAudioLibBuffer(int numFrames); //a zeros buffer with `numFrames` length
QAudioLibBuffer(qreal sec); //a zeros buffer of `sec` seconds length

S16S type

This type is provided by Qt and allows to manage left and right channels in the buffer, for example:

int position=0;
qint16 val=5;
mybuffer.data()[position].left=val;
mybuffer.data()[position].right=SHRT_MAX;

Data

const S16S* constData() const; //returns an S16S array, read only.
S16S* data(); //returns an `S16S array, read and write

See also Qt QAudioBuffer for inner functions.

Operators

QAudioLibBuffer allows to manipulate buffers through its operators:

QAudioLibBuffer operator<<(const QAudioLibBuffer &x) const; //appends buffers
QAudioLibBuffer operator*(const QAudioLibBuffer &x) const; //multiplicates buffers
QAudioLibBuffer operator*(const qreal x) const; //multiplicates for a costant
QAudioLibBuffer operator+(const QAudioLibBuffer &x); const //sums buffers

QAudioLibFreqBuffer

This buffer is really similar to QAudioLibBuffer, then I will not explain each function. It works with complex number

qcomplex type

qcomplex manage double complex numbers, to declare a complex number use qcomplex(re,im)

Constructors

QAudioLibFreqBuffer(int numFrames);
QAudioLibFreqBuffer(const QAudioLibFreqBuffer & other);
static QAudioLibFreqBuffer newByRealBuffer(const QAudioLibBuffer & x); //it is not a constructor but create a complex buffer from a real buffer

S64CD type

It is defined in the library, it is like S16S but complex

Data

S64CD* data();
const S64CD* constData() const;
int frameCount() const;
int byteCount() const;
void appendZeros(int n); //zero padding, used by FFT
void clear(); //clears the buffer

Frequency domain utility

qreal deltaf(); //returns the frequency increment between successive samples
qreal nyquistfreq(); //returns the nyquist frequency
QAudioLibFreqBuffer mod() const; //returns a new buffer of samples module
QAudioLibFreqBuffer phase() const; //returns a new buffer of samples phase
QAudioLibFreqBuffer conj() const; //returns a new buffer of samples conjugate

Operators

QAudioLibFreqBuffer operator*(const QAudioLibFreqBuffer &x);
QAudioLibFreqBuffer operator*(const qreal x);
void operator=(const QAudioLibFreqBuffer &x);

Mathematics

This class provides a list of static functions to generate waveforms and to analyze sounds

Sound generation

static QAudioLibBuffer sinwave(qreal f, qreal sec); //generates a new sinusoidal waveform (with rotation matrix) of frequency `f` and duration `sec`
static QAudioLibBuffer squarewave(qreal f, qreal sec); //generates a new sqaure waveform of frequency `f` and duration `sec`

Sound manipulation

static QAudioLibBuffer conv(QAudioLibBuffer &u ,QAudioLibBuffer &v); //does the convolution of two signals
static QAudioLibBuffer corr(QAudioLibBuffer &u ,QAudioLibBuffer &v);//does the cross correlation of two signals
static QAudioLibBuffer filter(QAudioLibBuffer &x, int N, char type, int f0, int f1=0); //filters the signal (type can be 'l': lowpass, 'h' highpass, 'b' bandpass) set f1 only for bandpass filters

Fourier transform

static QAudioLibFreqBuffer fft(const QAudioLibFreqBuffer & x); //does the fft of a complex signal, woks only if the length is a power of 2
static QAudioLibFreqBuffer dft(const QAudioLibFreqBuffer & x); //does the dft, it is very slow (uses at your own risk)
static QAudioLibFreqBuffer fft(const QAudioLibBuffer & x, bool zeropadding=true); //does the dft of a real signal, if signal is a power of 2 or if `zeropadding` is `true` uses the fft alghoritm othermise uses the slow dft. If `zeropadding` is true the signal lenght will change to become a power of 2.
static QAudioLibBuffer ifft(const QAudioLibFreqBuffer & x); //does the inverse dft, uses fft alghorithm only if the input singal is a power of 2 length.