source: doc/aubiopitch.txt @ 83a768e

feature/autosinkfeature/cnnfeature/cnn_orgfeature/constantqfeature/crepefeature/crepe_orgfeature/pitchshiftfeature/pydocstringsfeature/timestretchfix/ffmpeg5pitchshiftsamplertimestretchyinfft+
Last change on this file since 83a768e was 7f0fd1d, checked in by Paul Brossier <piem@piem.org>, 11 years ago

src/pitch/pitchyin.{c,h}: fix typo, developed
  • Property mode set to 100644
File size: 4.6 KB
Line 
1NAME
2  aubiopitch - a command line tool to extract musical pitch
3
4SYNOPSIS
5
6  aubiopitch source
7  aubiopitch [[-i] source] [-o sink]
8             [-r rate] [-B win] [-H hop]
9             [-p method] [-u unit] [-l thres]
10             [-s sil] [-f]
11             [-v] [-h] [-j]
12
13DESCRIPTION
14
15  aubiopitch attempts to detect the pitch, the perceived height of a musical
16  note.
17
18  When started with an input source (-i/--input), the detected pitch are
19  printed on the console, prefixed by a timestamp in seconds. If no pitch
20  candidate is found, the output is 0.
21
22  When started without an input source, or with the jack option (-j/--jack),
23  aubiopitch starts in jack mode.
24
25OPTIONS
26
27  This program follows the usual GNU command line syntax, with long options
28  starting with two dashes (--). A summary of options is included below.
29
30  -i, --input source  Run analysis on this audio file. Most uncompressed and
31  compressed are supported, depending on how aubio was built.
32
33  -o, --output sink  Save results in this file. The file will be created on
34  the model of the input file. The detected frequency is played at the
35  detected loudness.
36
37  -r, --samplerate rate  Fetch the input source, resampled at the given
38  sampling rate. The rate should be specified in Hertz as an integer. If 0,
39  the sampling rate of the original source will be used. Defaults to 0.
40
41  -B, --bufsize win  The size of the buffer to analyze, that is the length
42  of the window used for spectral and temporal computations. Defaults to 2048.
43
44  -H, --hopsize hop  The number of samples between two consecutive analysis.
45  Defaults to 256.
46
47  -p, --pitch method  The pitch detection method to use. See PITCH METHODS
48  below. Defaults to 'default'.
49
50  -u, --pitch-unit unit  The unit to be used to print frequencies. Possible
51  values include midi, bin, cent, and Hz. Defaults to 'Hz'.
52
53  -l, --pitch-tolerance thres  Set the tolerance for the pitch detection
54  algorithm. Typical values range between 0.2 and 0.9. Pitch candidates found
55  with a confidence less than this threshold will not be selected. The higher
56  the threshold, the more confidence in the candidates. Defaults to unset.
57
58  -s, --silence sil  Set the silence threshold, in dB, under which the onset
59  will not be detected. A value of -20.0 would eliminate most onsets but the
60  loudest ones. A value of -90.0 would select all onsets. Defaults to -90.0.
61
62  -m, --mix-input  Mix source signal to the output signal before writing to
63  sink.
64
65  -f, --force-overwrite  Overwrite output file if it already exists.
66
67  -j, --jack  Use Jack input/output. You will need a Jack connection
68  controller to feed aubio some signal and listen to its output.
69
70  -h, --help  Print a short help message and exit.
71
72  -v, --verbose  Be verbose.
73
74PITCH METHODS
75
76  Available methods are:
77
78  default  use the default method
79
80  Currently, the default method is set to yinfft.
81
82  schmitt  Schmitt trigger
83
84  This pitch extraction method implements a Schmitt trigger to estimate the
85  period of a signal. It is computationally very inexpensive, but also very
86  sensitive to noise.
87
88  fcomb  a fast harmonic comb filter
89
90  This pitch extraction method implements a fast harmonic comb filter to
91  determine the fundamental frequency of a harmonic sound.
92
93  mcomb  multiple-comb filter
94
95  This fundamental frequency estimation algorithm implements spectral
96  flattening, multi-comb filtering and peak histogramming.
97
98  specacf  Spectral auto-correlation function
99
100  yin  YIN algorithm
101
102  This algorithm was developed by A. de Cheveigne and H. Kawahara and
103  was first published in:
104
105  De Cheveigné, A., Kawahara, H. (2002) "YIN, a fundamental frequency
106  estimator for speech and music", J. Acoust. Soc. Am. 111, 1917-1930.
107
108  yinfft  Yinfft algorithm
109
110  This algorithm was derived from the YIN algorithm. In this implementation, a
111  Fourier transform is used to compute a tapered square difference function,
112  which allows spectral weighting. Because the difference function is tapered,
113  the selection of the period is simplified.
114
115  Paul Brossier, Automatic annotation of musical audio for interactive systems,
116  Chapter 3, Pitch Analysis, PhD thesis, Centre for Digital music, Queen Mary
117  University of London, London, UK, 2006.
118
119SEE ALSO
120
121  aubioonset(1),
122  aubiotrack(1),
123  aubionotes(1),
124  aubioquiet(1),
125  aubiomfcc(1),
126  and
127  aubiocut(1).
128
129AUTHOR
130
131  This manual page was written by Paul Brossier <piem@aubio.org>. Permission is
132  granted to copy, distribute and/or modify this document under the terms of
133  the GNU General Public License as published by the Free Software Foundation,
134  either version 3 of the License, or (at your option) any later version.
Note: See TracBrowser for help on using the repository browser.