Hot answers tagged

13

This is a well-studied problem, dating back from the mid 90s (DARPA/NIST broadcast transcription challenges). Search for "speech/music segmentation" or "audio segmentation" and you'll find thousands of research papers. There are two broad approaches to solve this problem: Supervised classification Train a speech/music classifier, using a standard machine ...


9

In designing such transformations, one should take into account competing interests: fidelity to the human auditory system (that varies with people), including non-linear or even chaotic aspects (tinnitus) easiness of the mathematical formulation for the analysis part possibility to discretize it or allow fast implementations existence of a suitable stable ...


8

The very first thing to try would be to implement a rough speech/music classifier to detect a radio host/DJ; and to extract the amplitude envelope to detect fade outs and fade ins. Once you have implemented that and/or if this doesn't solve your problem because your stream is continuous (DJ-style mix), you'll certainly run into the song boundary / song ...


7

TL;DR: Subspaces are low-dimensional, linear portions of the entire signal space that are expected to contain (or be close to) a large part of the observable and useful signals or transformations thereof, with additional tools that allow us to compute interesting things on the data We are given a set of data. To manipulate them more easily, it is common ...


6

The matlab codes that implemented D. Ellis's algorithm are on their website: http://labrosa.ee.columbia.edu/projects/beattrack/


6

The Hilbert transform does NOT calculate an analytic signal. If we have a real-valued sequence $x(n)$, and the Hilbert transform of $x(n)$ is $x_H(n)$, the analytic signal, $a(n)$, associated with $x(n)$ is: $$a(n) = x(n) + jx_H(n)$$ Computing the magnitude of $a(n)$ will give you the instantaneous envelope of the original $x(n)$ sequence. This works ...


6

Depends to some degree on the instrument: electronic versus mechanical, string versus reed versus pipe, etc. Typical electronic keyboards have no coupling between keys, and the DSP mixer is usually very close to linear (other than quantization effects, etc., unless the level is high enough to cause clipping, or kick in an AGC). But guitars and pianos do ...


5

First, with the classic short-term Fourier transform approach, there are alternative to interpolation - in particular techniques making use of phase information to recover the instantaneous frequency (See this question) which can give you very accurately the position of a spectral peak without an increase of FFT size. The drawback, as you correctly said, is ...


5

As I commented on a previous post, the time-frequency analysis method known as "short term Fourier transform" $X$ is equivalent to a filter bank, analysing your signal $x$. For a given analysis window $w_n$, of size $N$, the filter at frequency $k/N$ is : $$ h_n=w_{−n}e^{j2\pi\frac{nk}{N}}$$ For usual analysis windows (Hann, Hamming, or even rectangle), ...


5

You probably need more than one algorithm for an accurate detection. Work on frequency domain (mostly via DFT and spectrogram) is the most often used initial transform. After that sophisticated probabilistic models originally used in speech recognition are applied, such as hidden Markov models, dynamic Bayesian networks, and conditional random fields. The ...


5

The two concepts are related to two different dimensions or aspects of music which might or might not be correlated. Onset detection is concerned with finding the points in time at which sounds start. Doing this does not require prior knowledge of the particular pitch (or fundamental frequency) of the sound. It may indeed rely on the property that at the ...


5

I've had a try at reproducing the effect, and I think these are some of the key elements: You need a high-resolution FFT (large size; windows overlapping in time so that this doesn't come at the expense of frame rate) and to discard all but the lowest-frequency bins. You can tell this from the video because there are only rarely any harmonics visible and ...


5

Since the notes from a guitar are no pure sinusoids, you should expect to see some harmonics, even when analyzing the dry signal without effects. E.g., the note E is the perfect fifth of the note A, i.e., it is the second harmonic. If you use distortion or modulation effects (chorus, flanger, and phaser) you get even more additional frequencies due to the ...


5

Yeah some of us can do it, you can speed up or slow down without affect the pitch, some guys call this applications of Time Stretch, there different ways to do it, you can do in frequency domain or time domain, you will need choose what is best for you, you will find some advantages and disadvantages of each. Time Domain: In Time Domain you can try some ...


5

Subspaces are a Linear Algebra concepts. The best representative example I can think of is the relationship of the XY plane to XYZ space, The former is a subspace of the latter. Any vector in the plane also lies in the space. Every vector in space has an orthogonal projection onto the subspace. So a set of vectors in your subspace can only reach vectors ...


4

Frequency or pitch? There are already tons of research papers and books on human pitch perception. But, IIRC, humans tend to be bad at accurately "extracting" frequencies unless they happen to be a pitch fundamental. And multiple frequency peaks within a "critical band" tend to be perceived as noise. So any method with "near human accuracy" may also have ...


4

A few things: Not all instruments have partials neatly defined by the $h_n = nf$. Kettle drums would be an obvious example, but bells are another. In this case, you will be way off looking in the "usual places" +/- 5%. Many instruments have only even or odd harmonics. Others are designed to omit some harmonics. The piano, for example, is designed to have ...


4

A pluck might produce significantly more broadband noise than a free string. An FFT of such noise would show more relative energy outside of all the FFT result bins that are related to (F0 or overtones of) a single pitch. Also, a free string has a more predictable decay rate in any FFT magnitudes related to the pitch across successive offset FFT frames ...


4

The up-sampling process will always change the signal in some measurable way. However, if it's done properly the changes are negligible and don't result it any audible difference. Most commercially sample rate converters (hardware or software implementations), do a really good job at this. Off course, if done badly, upsampling can result in clearly audible ...


4

If your square wave has a mean of zero (this is important!), then a simple accumulator can do the job. Its operation is described by $$y[n]=x[n]+y[n-1]$$ where $x[n]$ is the input (square wave) and $y[n]$ is the output (triangular wave). This is a simple Matlab/Octave script showing how it works: sq = [1,-1,1,-1,1,-1,1,-1]'*ones(1,5); sq = sq'(:); ...


4

Yeah, notation is not ideal. It is not - he assumes that each of the $M$ antenna elements is connected to its own RF chain, i.e., there are also $M$ receivers available. If you have fewer receivers you need to modify your $A$, it needs to contain the response your $K$ receivers observe given a wave from a certain direction. He doesn't put it but yeah, $F$ ...


3

You can try using a Pitch Track for monophonic signal, you need track low pitches! There is a wide range of algorithms to find Pitches in time domain (zero crossing, AMDF, YIN, MPM, ASDF, autocorrelation ....) or Frequency domain (shs, hps, cepstrum ....) Each one of them possess advantages or disadvantages ! With the Pitch Track you can convert one ...


3

If you can't tolerate any distortion, then divide by 2 the amplitude of the signals you want to mix. Please check this question


3

how much error is typical and what is the relationship between the degree of error and the harmonic number I think that's hard to answer. Do you want to include contrived weird instruments with terrible inharmonicity? Inherently inharmonic instruments that are manipulated to sound roughly harmonic, like bells and tympani? I started writing a program to ...


3

This depends highly on the signal and it's content. For narrow band signal, the loudness can fairly well be estimated through the equal loudness curves as published in ISO 226 (see for example) http://en.wikipedia.org/wiki/Equal-loudness_contour For wide band signal, things are more complicated. If the signals are stationary, you apply frequency weighting ...


3

the etymology of the word "note" as it applies to music is simply the notation, the note that a composer would make to paper to represent a particular action taken by the musician performing the music. like "taking notes". normally in audio-to-MIDI conversion, a musical note is something that can be represented with a pair of MIDI Note-On and Note-Off ...


3

there are a lotta people doing research regarding music synthesis and DSP. yes you can sample notes and different key velocities and play them back with different MIDI velocity parameters. if you wanted to somehow interpolate between a note played at $mf$ to $f$ and on to $ff$ and $fff$urther, you would need to find a way to phase align the waveforms. IMO ...


3

The CZT allows for a fairly general evaluation of the Z transform - the more general evaluation path looks like a spiral, so it has a radial component step size as well an angular step size.For spectral zooming, you're only using a subset of this. You're evaluating around the unit circle and only for a small set of frequencies. The Zoom FFT can be ...


3

The prime thing such algorithms aim to do is to make use of more information that you may have about the signal. In this case, the extra information is that you know the number of signals (sinusoids) present in your measurements. One pro for both is, therefore, when your measurements match the assumption, you get a more accurate representation of the ...


Only top voted, non community-wiki answers of a minimum length are eligible