Debian Codegraph @

upstream/debian-unstable/man1/midistats.1

raw
1 .TH MIDISTATS 1 "18 March 2024"
2 .SH NAME
3 \fBmidistats\fP \- program to summarize the statistical properties of a midi file
4 .SH SYNOPSIS
5 This is a long manual because the program extracts many different parameters
6 from the midi file depending upon the options selected. This manual
7 attempts to describe these parameters and how they are computed.
8 These parameters are formatted in a way so they can be read by other
9 applications such midiexplorer.tcl, runstats.tcl, and numerous Python
10 scripts.
11
12 Many of the options were designed specifically for analyzing the
13 percussion track of the midi file. They are described in a separate
14 section below (Other options).
15
16
17 .SH DESCRIPTION
18 If you run midistats without any options other than the name of
19 the midi input file, it will produce a table of values described
20 here. Each line of output starts with the name of the variable or
21 variable array and the associated values.
22 .PP
23 ntrks indicates the number of tracks in the midi file.
24 .PP
25 ppqn the number of midi pulses per quarter note.
26 .PP
27 keysig the key signature, followed by a major/minor flag, the number
28 of sharps (positive) or flats (negative) in the key, and the beat number
29 where the key signature was found.
30 .PP
31 trk is followed by the track number for which the following information
32 applies.
33 .PP
34 program is followed by the channel number and the General Midi Program
35 number.
36 .PP
37 trkinfo is an array of 19 numbers which indicates the statistical properties
38 of the track of interest. The following data is given:
39 0 the channel number,
40 1 the first program assigned to this channel,
41 2 the number of notes for this channel counting any chords as one note,
42 3 the total number of notes for this for this channel,
43 4 the sum of the MIDI pitches for all the notes,
44 5 the sum of the note durations in MIDI pulse units,
45 6 the number of pitchbend messages,
46 7 the number of control parameter messages,
47 8 the number of pressure messages.
48 9 the number of distinct rhythm patterns for each channel
49 10 the number of pulses the channel was inactive
50 11 the minimum pitch value
51 12 the maximum pitch value
52 13 the minimum note length in pulses
53 14 the maximum note length in pulses
54 15 the number of gaps in the channel
55 16 the entropy of the pitch class histogram for that channel
56 17 the number of notes whose pitch were the same as the previous note
57 18 the number of notes whose pitch changed by less than 4 semitones
58 19 the number of notes whose pitch changed by 4 or more semitones
59 (In event of a chords the maximum pitches are compared.)
60 .PP
61 After processing all the individual tracks, the following information
62 applies to the entire midi file.
63 .PP
64 npulses is the length of the longest midi track in midi pulse units
65 .PP
66 tempocmds specifies the number of times the tempo is changed in this
67 file.
68 .PP
69 pitchbends specifies the total number of pitchbends in this file.
70 .PP
71 pitchbendin c n specifies the number of pitchbends n in channel c
72 .PP
73 progs is a list of all the midi programs addressed
74 .PP
75 progsact the amount of activity for each of the above midi programs.
76 The activity is the sum of the note durations in midi pulse units.
77 .PP
78 progcolor: is a 17 dimensional vector where each component maps into
79 a specific group of MIDI programs. Some of these groups are, keyboard
80 instruments, brass instruments, wind instruments, and etc. More information
81 can be found in the midiexplorer documentation.
82 .PP
83 drums is a list of all the percussion instruments (channel 9) that were
84 used.
85 .PP
86 drumhits indicates the number of notes for each of the above percussion
87 instruments.
88 .PP
89 pitches is a histogram for the 11 pitch classes (C, C#, D ...B)
90 that occur in the midi file.
91 .PP
92 key indicates the key of the music, the number of sharps (positive) or
93 flats (negative) in the key signature, and a measure of the confidence
94 in this key signature. The key was estimated from the above pitch histogram
95 by convolving with Craig Sapp's model. The peak of rmaj or rmin (below)
96 indicates the key. A correlation less than 0.4 indicates that the pitch
97 histogram does not follow the histogram of a major or minor scale.
98 (It may be the result of a mixture of two key signatures.)
99 .PP
100 rmaj the cross correlation coefficients with Craig Sapp's major key model
101 for each of the 11 keys (C, C#, D, ...,B).
102 .PP
103 rmaj the cross correlation coefficients with Craig Sapp's minor key model
104 for each of the 11 keys (C, C#, D, ...,B).
105 .PP
106 pitchact is a similar histogram but is weighted by the length of
107 the notes.
108 .PP
109 chanvol indicates the value of the control volume commands in the
110 midi file for each of the 16 channels. The maximum value is 127.
111 It scales the loudness of the notes (velocity) by its value.
112 .PP
113 chnact returns the amount of note activity in each channel.
114 .PP
115 trkact returns the number of notes in each track.
116 .PP
117 totalrhythmpatterns is the total number of bar rhythm patterns for
118 all channels except the percussion channel.
119 .PP
120 collisions. Midistats counts the bar rhythm patterns using a hashing
121 function. Presently collisions are ignored so occasionally two
122 distinct rhythm patterns are counted as one.
123 .PP
124 Midistats prints a number of arrays which may be useful in
125 determining where the music in the track is a melody line or
126 chordal rhythmic support. These arrays indicate the properties
127 for each of the 16 channels. (The percussion channel 9 contains
128 zeros.) In the case same channel occurs in several tracks, these
129 numbers are the totals for all track containing that channel.
130 Here is a description of these properties.
131 .PP
132 programs: channel to midi program mapping
133 .PP
134 cnotes: the total number of notes in each channel
135 .PP
136 nnotes: the number of notes in each channel not including
137 those playing in the same time interval.
138 .br
139 nzeros: the number of notes whose previous note was the same pitch
140 .br
141 nsteps: the number of notes whose pitch difference with the previous
142 note was less than 4 semitones.
143 .br
144 njumps: the number of notes whose pitch difference with the previous
145 note was 4 or more semitones.
146 .br
147 rpats: the number of rhythmpatterns for each channels. This is a
148 duplication of data printed previously.
149 .br
150 pavg: the average pitch of all the notes for each channel.
151 .br
152 spread: the percentage of the track that each channel is active.
153 .PP
154 If some of the channels appear in more than one track, then
155 some of the above values may be incorrect.
156 .PP
157 In addition the midistats may return other codes that describe
158 other characteristics. They include
159
160 unquantized - the note onsets are not quantized
161 .br
162 triplets - 3 notes played in the time of 2 notes are present
163 .br
164 qnotes - the rhythm is basically simple
165 .br
166 clean_quantization - the note onsets are quantized into 1/4, 1/8, 1/16 time units.
167 .br
168 dithered_quantization - small variations in the quantized note onsets.
169 .br
170 Lyrics - lyrics are present in the meta data
171 .br
172 programcmd - there may be multiple program changes in a midi channel
173
174
175
176 .SH Other options
177
178 It is recommended that you only select one of the options
179 described here as the program was not designed to handle a
180 multiple number of options.
181
182 .PP
183 If you run midistats with the -CSV option, it will return the
184 results in a form of comma separated values that can be loaded
185 into a Python panda dataframe. Each line refers to one of the
186 16 midi channels. The following Python 3 code illustrates
187 how you would load the midistats output into a dataframe.
188
189 .br
190 import pandas as pd
191 .br
192 import io
193 .br
194 import subprocess
195 .br
196 cmd = ("midistats", "-CSV" , inputmidifilepath)
197 .br
198 process = subprocess.Popen(cmd, stdout=subprocess.PIPE)
199 .br
200 csv = io.StringIO(process.stdout.read().decode())
201 .br
202 df = pd.read_csv(csv)
203
204 where inputmidifilepath is the path to the midi file that
205 you are using. (eg. 'clean_midi/Zero/Chi sei.mid')
206
207 .PP
208 The MIDI file devotes channel 9 to the percussion instruments
209 and over 60 percussion instruments are defined in the MIDI
210 standard. Though there is a lot of diversity in the percussion
211 track, for most MIDI files only the first 10 or so percussion
212 instruments are important in defining the character of the track. The
213 program Midiexplorer has various tools for exposing the percussion
214 channel which are described in the documentation. The goal
215 here is to find the essential characteristics of the percussion
216 track which distinguishes the MIDI files. This is attempted
217 in the program midistats. Here is a short description.
218
219
220 .br
221
222 A number of experimental tools for analyzing the percussion channel
223 (track) were introduced into midistats and are accessible through
224 the runtime arguments. When these tools are used in a script which
225 runs through a collection of midi files, you can build a database
226 of percussion descriptors.
227
228 .SH OPTIONS
229 .PP
230 -corestats
231 .br
232 outputs a line with 5 numbers separated by tabs. eg
233 .br
234 1 8 384 4057 375
235 .br
236 It returns the number of tracks, the number of channels, the
237 number of divisions per quarter note beat (ppqn),
238 the number of note onsets in the midi file, and the maximum
239 number of quarter note beats in midi file.
240
241
242 .PP
243 -pulseanalysis
244 .br
245 counts the number of note onsets as a function of its onset time
246 relative to a beat, grouping them into 12 intervals and returns
247 the result as a discrete probability density function. Generally,
248 the distribution consists of a couple of peaks corresponding
249 to quarter notes or eigth notes. If the distribution is flat,
250 it indicates that the times of the note occurrences have not been
251 quantized into beats and fractions. Here is a sample output.
252 .br
253 0.349,0.000,0.000,0.160,0.000,0.000,0.298,0.000,0.000,0.191,0.000,0.000
254
255 .PP
256 -panal
257 .br
258 Counts the number of note onsets for each percussion instrument. The first
259 number is the code (pitch) of the instrument, the second number is the
260 number of occurrences. eg.
261 .br
262 35 337 37 16 38 432 39 208 40 231 42 1088 46 384 49 42 54 1104 57 5 70 1040 85 16
263
264 .PP
265 -ppatfor n
266 .br
267 where n is the code number of the percussion instrument. Each beat
268 is represented by a 4 bit number where the position of the on-bit
269 indicates the time in the beat when the drum onset occurs. The bits
270 are ordered from left to right (higher order bits to lower order
271 bits). This is the order of bits that you would expect in a
272 time series.
273 Thus 0 indicates that there was no note onset in that beat, 1 indicates
274 a note onset at the end of the beat, 4 indicates a note onset
275 in the middle of the beat, and etc. The function returns a string
276 of numbers ranging from 0 to 7 indicating the presence of note onsets
277 for the selected percussion instrument for the sequence of beats
278 in the midi file. Here is a truncated sample of the output.
279 .br
280 11 71 c1 1 11 51 f1 c1 1 11 51 e1 1 11 d1 c1 31 1 1 1 21 b1 d1 d1 51
281 .br
282 51 51 51 51 41 71 c1 11 1 11 41 11 1 11 41 11 1 11 41 11 31 51 41 11
283
284 .PP
285 -ppat
286 .br
287 midistats attempts to find two percussion instruments in the midi file
288 which come closest to acting as the bass drum and snare drum.
289 If it is unsuccessful, it returns a message of its failue. Otherwise,
290 encodes the position of these drum onsets in a 8 bit byte for each
291 quarter note beat in the midi file. The lower (right) 4 bits encode the
292 bass drum and the higher (left) 4 bits encode the snare drum in the
293 same manner as described above for -ppatfor. The integers are printed
294 in hexadecimal.
295 .br
296 02 88 20 02 a0 08 80 02 82 08 80 02 82 08 80 02 82 08 80 02 82 08
297 .br
298 80 02 82 08 80 02 82 08 80 02 82 08 20 a0 0a 08 80 02 82 08 80 02
299
300 .PP
301 -ppathist
302 .br
303 computes and displays the histogram of the values that would appear
304 when running the -ppat. eg.
305 .br
306 bass 35 337
307 .br
308 snare 38 432
309 .br
310 1 (0.1) 64 32 (2.0) 8 33 (2.1) 136 144 (9.0) 8 145 (9.1) 136
311 .br
312 The bass percussion code, the number of onsets, and the snare
313 percussion code and the number of onsets are given in the
314 first two lines. In the next line the number of occurrences of
315 each value in the -ppat listing is given. The number in parentheses
316 splits the two 4-bit values with a period. Thus 33 = (2*16 + 1).
317
318 .PP
319 -pitchclass
320 .br
321 Returns the pitch class distribution for the entire midi file.
322
323 .PP
324 -nseqfor n
325 .br
326 Note sequence for channel n. This option produces a string of bytes
327 indicating the presence of a note in a time unit corresponding to
328 an eigth note. Thus each quarter note beat is represented by two
329 bytes. The pitch class is represented by the line number on the
330 staff, where 0 is C. Thus the notes on a scale are represented
331 by 7 numbers, and sharps and flats are ignored. The line number is
332 then converted to a bit position in the byte, so that the pitch
333 classes are represented by the numbers 1,2,4,8, and etc. A chord
334 of consisting of two note onsets would set two of the corresponding
335 bits. If we were to represent the full chromatic scale consisting
336 of 12 pitches, then we would require two-byte integers or
337 twice of much memory.
338 .br
339 Though the pitch resolution is not sufficient to distinguish
340 major or minor chords, it should be sufficient to be identify some
341 repeating patterns.
342 .PP
343 -nseq
344 .br
345 Same as above except it is applied to all channels except the
346 percussion channel.
347 .br
348 .PP
349 -nseqtokens
350 Returns the number of distinct sequence elements for each channel.
351 The channel number and number of distinct elements separated by
352 a comma is returned in a tab separated list for all active channels
353 except the percussion channel. Here is an example.
354 .br
355 2,3 3,4 4,11 5,6 6,3 7,3 8,6 9,3 11,2 12,1
356 .br
357
358 -ver (version number)
359
360
361 .SH AUTHOR
362 Seymour Shlien <[email protected]>
363
364
365