Humdrum lab 5

From CCARH Wiki
Jump to: navigation, search

This lab is about the wikifonia data.

1 Update Humdrum tools

The beat command has some new options (-C and -U), so you will need to update humdrum extras:

   cd $(which beat | sed 's/humdrum-tools.*/humdrum-tools/')
   make update
   make

2 Basic information

How many files

   ls *.krn | wc -l
   6710

How many have lyrics:

    grep -l "\*\*text" *.krn | wc -l
    5460

How many have chords:

    grep -l "\*\*mxhm" *.krn | wc -l
    6282

How many have two or more verses:

    grep -l "\*\*text.*\*\*text" *.krn | wc -l
    2006

3 Bibliographic information

Who are the top 10 represented composer in the data:

  grep -h COM *.krn | sortcount | head -n 10
  132	!!!COM:	Unknown
  121	!!!COM:	Hungarian folk song
  119	!!!COM:	Traditional
  91	!!!COM:	Richard Rodgers
  75	!!!COM:	Irving Berlin
  67	!!!COM:	Hungarian song
  65	!!!COM:	Cole Porter
  46	!!!COM:	Harry Warren
  45	!!!COM:	George Gershwin
  40	!!!COM:	Harold Arlen

List the titles of all pieces where George Gershwin is the composer:

  grep OTL  $(grep -li  COM.*Gershwin *.krn) | sort -k2
  WF2190.krn:!!!OTL:	'S Wonderful!
  WF2191.krn:!!!OTL:	A FOGGY DAY
  WF2267.krn:!!!OTL:	A Foggy Day
  WF2186.krn:!!!OTL:	A Woman Is A Sometime Thing
  WF2192.krn:!!!OTL:	Bidin' My Time
  WF2178.krn:!!!OTL:	Blues
  WF2193.krn:!!!OTL:	But Not For Me
  WF2194.krn:!!!OTL:	By Strauss
  WF2195.krn:!!!OTL:	Clap yo' hands
  WF2185.krn:!!!OTL:	Do It Again!
  WF2196.krn:!!!OTL:	Embraceable You
  WF2197.krn:!!!OTL:	Fascinating Rhythm
  WF2268.krn:!!!OTL:	For You, For Me, For Evermore
  WF2198.krn:!!!OTL:	How Long Has This Been Going On
  WF2199.krn:!!!OTL:	I Got Plenty o' Nuttin'
  WF2103.krn:!!!OTL:	I Got Rhythm
  WF2200.krn:!!!OTL:	I Got Rhythm
  WF2222.krn:!!!OTL:	I Loves You Porgy
  WF2201.krn:!!!OTL:	I Was Doing All Right
  WF2202.krn:!!!OTL:	I Was Doing All Right
  WF2179.krn:!!!OTL:	I loves you Porgy
  WF2184.krn:!!!OTL:	I'll Build A Stairway To Paradise
  WF2203.krn:!!!OTL:	I've Got A Crush On You
  WF2204.krn:!!!OTL:	Isn't It A Pity
  WF2221.krn:!!!OTL:	It Ain't Necessarily So
  WF2205.krn:!!!OTL:	Let's Call the Whole Thing Off
  WF2220.krn:!!!OTL:	Liza
  WF2219.krn:!!!OTL:	Liza (All the clouds'll roll away)
  WF2269.krn:!!!OTL:	Love Is Here To Stay
  WF2206.krn:!!!OTL:	Love Walked In
  WF2223.krn:!!!OTL:	My Man's Gone Now
  WF2207.krn:!!!OTL:	Nice Work If You Can Get It
  WF2208.krn:!!!OTL:	Oh Lady Be Good
  WF2189.krn:!!!OTL:	SUMMERTIME
  WF2995.krn:!!!OTL:	Shoes With Wings On
  WF2183.krn:!!!OTL:	Somebody Loves Me
  WF2209.krn:!!!OTL:	Someone To Watch Over Me
  WF2210.krn:!!!OTL:	Soon
  WF2211.krn:!!!OTL:	Strike Up The Band
  WF2180.krn:!!!OTL:	Summertime
  WF2181.krn:!!!OTL:	Summertime
  WF2187.krn:!!!OTL:	Summertime
  WF2212.krn:!!!OTL:	Summertime
  WF2224.krn:!!!OTL:	Swanee
  WF2213.krn:!!!OTL:	That Certain Feeling
  WF2214.krn:!!!OTL:	The Man I Love
  WF2225.krn:!!!OTL:	The Simple Life
  WF2188.krn:!!!OTL:	There's A Boat Dat's Leaving Soon For New York
  WF2215.krn:!!!OTL:	They All Laughed
  WF2216.krn:!!!OTL:	They Can't Take That Away From Me
  WF2217.krn:!!!OTL:	They Can't Take That Away From Me
  WF2218.krn:!!!OTL:	Who Cares

4 Texture

How many contain more than one **kern spine (i.e., are polyphonic, probably piano):

   grep -l "\*\*kern.*\*\*kern" *.krn | wc -l
   58

WF5118.krn is an example:

Wf5118.png

This one is interesting because it has invisible chords in the top staff which are realizing the harmonic chords above the staff.

How many songs have chords (this takes a long time to calculate -- 70 songs per second = 95 seconds):

  for i in *.krn 
  do 
     extractx -i kern $i | serialize | ridx -H | grep " " | wc -l
  done | grep -v " ^0$" | wc -l
  365

How many songs do not have chords:

  for i in *.krn 
  do 
     extractx -i kern $i | serialize | ridx -H | grep " " | wc -l
  done | grep " ^0$" | wc -l
  6345

5 Duration

What is the duration of all songs if played back-to-back and at the specified tempo without repeats?

   gettime -T *.krn | tail -n 1
   286:50:23.1354 hours

What are the longest songs:

    gettime --simple -T *.krn | sort -k2 -nr | head -n 10
    WF6618.krn:	3120
    WF0181.krn:	3120
    WF0182.krn:	1864
    WF3616.krn:	1420
    WF6336.krn:	1134
    WF5131.krn:	909
    WF6068.krn:	785
    WF5004.krn:	696
    WF3226.krn:	671
    WF1249.krn:	664

The -k2 option means to sort by the second column of data. -n means to sort numerically rather than alphabetically, and -r means to sort by highest first.

What are the shortest songs:

    gettime --simple -T *.krn | sort -k2 -nr | tail -n 10
    WF2814.krn:	16
    WF2806.krn:	16
    WF2795.krn:	16
    WF2785.krn:	16
    WF2856.krn:	14
    WF2852.krn:	12
    WF2799.krn:	12
    WF6338.krn:	8
    WF5609.krn:	8
 The shortest song in VHV:
    cat WF5609.krn | pbcopy
Wf-shortest.png

6 Meter

What sort of meters are in the database and how much of each type?

    beat -Ca *.krn | beat -Ua  | extractx -s '$1-$'  | ridx -H | sortcount -p
    65.89	4	4
    15.44	3	4
    11.67	2	2
    3.24	2	4
    2.78	6	8
    0.52	12	8
    0.16	5	4
    0.14	9	8
    0.09	6	4
    0.06	3	8
    0.01	3	2
    0.01	7	4
    0     	2	8
    0     	7	8
    0     	9	4
    0     	10	8
    0     	17	16
    0     	1	2
    0     	5	8
    0     	1	4
    0     	4	8

The most common meter is 4/4, where 65% of the music is in that meter.

-C means extract the count of the meter (the top number).

-U means extract the duration unit from the meter (the bottom number).

-C and -U are output once for each measure, so using these are a simple way of counting the number of measures in the scores. If you add -F option with these two options, every data line will display the metrical information.

-a means to append the analysis to the end of the lines (keeping the original input score).

The extract option:

    -s '$1-$'

means to extract from one before the last spine to the last spine. $1 is one before the last spine, $2 is two before the last spine, and so on.

7 Chord labels

How many unique chord labels are there?

   extractx -i mxhm  * | ridx -H | sortcount | wc -l
   1399

What are the most common ones:

   extractx -i mxhm  * | ridx -H | sortcount -p | head -n 10
   7.21	C major
   6.14	F major
   4.94	G major
   4.83	G dominant
   4.07	C dominant
   3.55	D dominant
   3.28	B- major
   2.81	E- major
   2.3	D major
   2.25	F dominant

How many chord qualities:

    extractx -i mxhm  * | ridx -H | sed 's/[^ ]* //; s/\/.*//' | sortcount | wc -l
    80

Here are the 80 qualities:

  93061	major
  64155	dominant
  31402	minor
  27490	minor-seventh
  8594	major-seventh
  5861	dominant-ninth
  5733	major-sixth
  4138	diminished
  2943	min
  2912	7
  2816	minor-sixth
  2159	half-diminished
  2154	suspended-fourth
  1884	diminished-seventh
  1738	augmented-seventh
  1408	augmented
  1208	C
  1102	dominant-13th
  1082	min7
  1008	F
  1008	maj7
  967	dominant-seventh
  892	G
  878	minor-ninth
  705	D
  650	B-
  592	major-ninth
  355	E-
  352	A
  280	E
  249	power
  237	dominant-11th
  222	suspended-second
  202	minor-11th
  165	minor-major
  157	dim
  129	maj
  128	A-
  96	augmented-ninth
  84	9
  71	other
  66	B
  62	6
  62	major-minor
  58	sus47
  46	aug
  46	D-
  46	min9
  36	G-
  29	m7b5
  23	major-13th
  21	maj9
  19	min6
  19	none
  17	pedal
  16	dim7
  16	maj69
  15	F#
  12	major	B- major
  8	major	F major
  7	C#
  6	major	.
  5	minor	D minor	D minor
  4	minor-13th
  3	minor	G minor
  3	C-
  3	minMaj7
  3	D#
  2	minor	.
  2	major	F major	F major
  2	dominant	C dominant
  2	5b
  2	major	C major	C major
  1	major	E- major
  1	ma
  1	major	.	.
  1	major	G major	G major
  1	minor	.	.
  1	7sus
  1


  extractx -i mxhm  * | ridx -H | sed 's/ .*//' | sortcount
  48078	C
  45001	G
  38459	F
  33577	D
  24361	A
  23237	B-
  16334	E-
  15587	E
  8279	A-
  7869	B
  3487	D-
  3365	F#
  1701	C#
  1316	G-
  824	G#
  187	D#
  168	C-
  52	A#
  28	F-
  6	B--
  4	B#
  4	.	F
  3	B/D#
  3	E#
  3	C/G
  1	.	B-
  1	A--


What is the most common 3-note chord sequence:

    extractx -i mxhm  * | grep -v ^= | serialize | context -n 3 | ridx -H | sortcount | head -n 10
    1779	C major G dominant C major	(I V7 I)
    1334	F major C dominant F major	(I V7 I)
    1301	C major F major C major     	(I V I)
    1062	D minor-seventh G dominant C major	(ii7 V7 I)
    994  	G major D dominant G major	(I V7 I)
    939  	G dominant C major G dominant	(V7 I V7)
    863  	G major C major G major     	(V I V)
    857 	G dominant C major F major	(V7 I IV)
    812  	F major B- major F major	(V I V)
    781  	G minor-seventh C dominant F major	(ii7 V7 I)


8 Scale degrees

The key information is not present in the files. They need to be processed further for that. MusicXML input has key information, but it is often incorrect since people use it more for key signature information, and the "mode" part is usually left at "major". The finalis-tonic script can be used to add an approximate key.


  for i in *.krn 
  do 
     finalis-tonic $i | extractx -i **kern | deg -at | serialize | ridx -H | grep -v r
  done | sortcount

Output includes chords and some other junk, but basic counts are:

161963	1
135309	5
102726	2
102275	3 
86931	6
85740	4
50631	7
38827	7-
38765	3-
24511	6-
12031	4+
11188	2-
6041	5+
5822	2+
5803	1+
4907	5-
2822	1-
2669	6+
1874	4-
1222	7+
1126	3+

Looking at 5-note sequences:

  for i in *.krn
  do 
     finalis-tonic $i | extractx -i **kern | deg -at | serialize -f | grep -v '^[r=]' \
           | context -n 5 | ridx -H ; done | sortcount > /tmp/analysis-data.txt
  done
  head -n 25 analysis-data.txt
  6053 1 1 1 1 1
  4417 5 5 5 5 5
  2186 3 3 3 3 3
  1872 4 4 4 4 4
  1573 2 2 2 2 2
  1293 3 2 1 2 3
  1282 5 4 3 2 1
  1189 6 6 6 6 6
  1087 1 2 3 2 1
  915  2 1 1 1 1
  894  3 4 3 2 1
  890  3- 3- 3- 3- 3-
  883  3 3 3 2 1
  855  3 2 1 1 1
  852  1 1 1 1 2
  825  6 5 4 3 2
  806  5 5 5 5 4
  786  3 2 1 2 1
  785  5 5 5 5 6
  761  3 3 3 3 2
  761  7 7 7 7 7
  753  7- 7- 7- 7- 7-
  752  5 1 1 1 1
  726  1 2 3 4 5
  723  4 3 2 1 1


9 Key

Using entire contents of file to determine key:

  keycor *.krn | sed 's/.*: //' | sortcount -p
  22.33	C Major
  15.05	F Major
  11.87	G Major
  9.21	E- Major
  6.87	B- Major
  5.82	A Minor
  4.51	D Minor
  4.39	D Major
  4.16	C Minor
  3.44	G Minor
  3.03	E Minor
  2.11	A- Major
  1.78	A Major
  1.41	F Minor
  1.02	E Major
  0.64	B- Minor
  0.53	D- Major
  0.39	E- Minor
  0.39	B Minor
  0.29	B Major
  0.18	G- Major
  0.16	F# Minor
  0.16	C# Minor
  0.08	F# Major
  0.06	C# Major
  0.04	D- Minor
  0.04	A- Minor
  0.04	G- Minor
  0.02	G# Minor
  0.02	G# Major


Using only the first 8 bars of the music for analysis of key:

  for i in *.krn
  do
     myank -m 1-8 $i | keycor
  done | sed 's/.*: //' | sortcount -p
  19.46	C Major
  12.32	F Major
  10.2	G Major
  6.89	E- Major
  6.7	A Minor
  6.65	D Minor
  5.64	B- Major
  5.36	C Minor
  5.21	D Major
  4.3	G Minor
  4.1	E Minor
  2.41	F Minor
  1.72	A Major
  1.65	A- Major
  1.31	B Minor
  1.22	B Major
  1.04	E Major
  0.89	B- Minor
  0.56	E- Minor
  0.56	D- Major
  0.48	F# Minor
  0.45	C# Minor
  0.3	A- Minor
  0.23	G- Major
  0.12	F# Major
  0.08	C# Major
  0.07	G# Minor
  0.05	G- Minor
  0.03	G# Major
  0.02	D- Minor



Lab 1 (intro) Lab 2 (Essen) Lab 3 (searching) Lab 4 (JRP) Lab 5 (Wikifonia) Lab 6 (bar chart) Lab 7 (regular expressions) Lab 8 (chorck & cint)