# Tag Info

0

In order to remove the noise from the signal, you need to subtract the spectral content of the noise from that of the signal ( filtered_signal = unfiltered_signal - noise ). This technique is called spectral subtraction. This will only work well if the noise is rather uniform across the entire length of the signal. The steps required to do this are ...

1

If you want a strictly real result from an IFFT, then you have to force the input to be conjugate symmetric. That way, all the imaginary components will cancel out to (almost) zero (except for rounding “errors” or microscopic numerical noise).

1

For a real signal, the FFT is symmetric complex numbers in general. That is $X[((-k))_N] = X[((k))_N]$. When you did this c(round(n*1000/fs):end)=0 you have disturbed the symmetry and hence, the ifft of this new signal will no longer be real. See a simple example >> x=[1 3 5 6]; >> y=fft(x); >> y(1:2)=0; >> ifft(y) ans = -1....

2

The short answer is do not null out the "mirror frequencies" that are located above $f_s/2$ that match the frequencies you want to keep. If your FFT was generated from a real signal, then when you do the IFFT you will get the real signal back as long as you did not zero out those upper frequencies (as you did). The DFT (which the FFT computes) returns ...

0

ALSA breaks audio buffers into periods. One buffer has integer number of periods. The I2S silicon automatically streams out from or in to your audio buffers using DMA hardware. Typically the DMA size is scaled to an ALSA period size so that part of the audio subsystem is at least double buffered. Which means you are working on one buffer while the other one ...

0

I suspect that your problem is to do with having your speakers and microphone on different ALSA audio devices or subdevices. You need to find the available sound cards for recording and playback on your system. The following commands will tell you this information : aplay -l arecord -l You will need to define a default full duplex sound card using your ...

1

Typically the longer the buffer of audio, the more processing you can do. The reason for this is that most of the audio buffering is happening in hardware (silicon I2S controllers). When you are make less software calls for each buffer, there is less processing overhead. If you capture a larger amount of audio with each time slice, you will be able to ...

1

The main prerequisite for measuring anything meaningful is that the target parameter (THD) is much better in your measurement system than in the system under test. The higher the difference, the more accurate your result will be. If your measurement system is orders of magnitude better, that the system under test, the measurement error can be neglected. If ...

0

IQ sampling is needed when the sampling frequency is near or at the center frequency of the desired signal or signal band. Thus a second sample is needed to discriminate between spectrum above, and spectrum below the sampling frequency. Audio is sampled well above the center of the audible signal bandwidth, and above twice the highest audible or filter ...

0

PCM/PDM is used for storing audio samples, while IQ data format is used when transmission using digital communication techniques. An audio signal (of any any real world signal) when sampled is represented by bits on a computer. The bits represent quantized values of real numbers in a standardized format (for example IEEE 754 floating point standard). It ...

2

To complement @DanBoschen's answer: a real baseband signal is a purely in-phase signal. Its quadrature component is zero, so there is no need to sample it or represent it in any way. An interesting approach, though, would be to represent a stereo signal as quadrature. You could define the right channel as the in-phase signal, the left channel as quadrature, ...

3

It is because the audio signals are real and already at baseband. In contrast radio frequency signals are often represented as complex numbers once they are brought back to baseband. Real signals can be represented as a single stream of real numbers, while for complex numbers two streams of real numbers are required to represent them (as in $I+jQ$). When ...

1

Simple clipping : threshold = 0.5 If x > threshold x = threshold elseif x < -threshold x = -threshold end Real-world clipping can be significantly more complex than this, involving various time-constants, asymmetry, heating-effects,....

1

Yes it does exist and these are called "all-pass" filters in that over the band of interest the magnitude does not change (other than a possible fixed gain at all frequencies) but modify the phase response. The time delay at a specific frequency is the negative derivative of phase with respect to frequency, so the goal of these filters is to selectively ...

2

Python librosa library has a functionality you can use: librosa.effects.split(y=buffer, frame_length=8000, top_db=40) Split an audio signal into non-silent intervals. Given sampling rate of 8000 it will split the audio by detecting audio lower than 40db for period of 1 sec Or, you can trim the audio "silent parts" using: librosa.effects.trim(y=buffer,...

0

However, since I have taken the modulus, it must be impossible to go from spec back to audio correct? You need to resort to some approximations like the Griffin-Lim algorithm or various vocoders (WaveNet,WaveGlow,etc.). So does that mean that librosa.istft does NOT convert a spectrogram to a wav file? It's a bit more complicated than that. An FFT ideally ...

0

In addition to the distinction that frequency is an objective measure and pitch is a subjective one, it's also useful to note that the pitch of a note may not be directly related to any easy measure of frequency. Case in point: a bell, tuned to A440, will generally emit sound energy at roughly 880Hz, 1320Hz, 1760Hz, etc. -- in other words, for a pitch ...

3

Frequency is mathematically defined as the number of cycles per second. So it is a more strict word mathematically. It is represented numerically by the unit called Hertz. $f=1/T$, where $T$ represents the one-period length of a waveform. This makes frequency quantifiable. Pitch on the other hand, is a perceptual characteristic of a sound frequency, so it'...

1

Frequency is a mathematical/physical concept while pitch is a perceptual concept that correlates with frequency. Edit: or in wikipedias words: « Frequency is an objective, scientific attribute that can be measured. Pitch is each person's subjective perception of a sound wave, which cannot be directly measured.» https://en.m.wikipedia.org/wiki/Pitch_(music)

4

If you can get it: Zölzer, Udo (1997). Digital Audio Signal Processing. John Wiley and Sons. ISBN 0-471-97226-6 or Orfanidis might have to buy them used. is your interest music processing and synthesis? there are some nice online books. Look for anything Julius Smith. But there are others, but i can't remember any names to search for. Remembered ...

3

Many years ago, when I was studying audio DSP in my university, I had to read the following: Y. You, “Audio Coding: Theory and Applications,” Springer, 2010 here A. Spanias, T. Painter, V. Atti, “Audio Signal Processing and Coding,” Wiley, 2007 here. I was very dissapointed by both. Spanias & Painter provide a nice and thorough overview of MP3 coding (...

3

My personal favorite, though I don't usually do much audio signal processing, are Julius O. Smith III's Spectral Audio Signal Processing and Physical Audio Signal Processing. They're available online here and here. A while ago, he made the Physical book available in paperback on Amazon.

Top 50 recent answers are included