-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathMasterarbeit.lof
140 lines (140 loc) · 10.1 KB
/
Masterarbeit.lof
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
\babel@toc {english}{}
\boolfalse {citerequest}\boolfalse {citetracker}\boolfalse {pagetracker}\boolfalse {backtracker}\relax
\babel@toc {german}{}
\babel@toc {english}{}
\babel@toc {german}{}
\babel@toc {english}{}
\defcounter {refsection}{0}\relax
\addvspace {10\p@ }
\defcounter {refsection}{0}\relax
\contentsline {figure}{\numberline {1.1}{\ignorespaces Structure of the thesis\relax }}{3}{figure.1.1}
\defcounter {refsection}{0}\relax
\addvspace {10\p@ }
\defcounter {refsection}{0}\relax
\contentsline {figure}{\numberline {2.1}{\ignorespaces Example spectrograms linear (a) and log-scaled (b)\relax }}{6}{figure.2.1}
\defcounter {refsection}{0}\relax
\contentsline {figure}{\numberline {2.2}{\ignorespaces Example spectrograms of a logarithmic frequency sweep\relax }}{7}{figure.2.2}
\defcounter {refsection}{0}\relax
\contentsline {figure}{\numberline {2.3}{\ignorespaces MFCCs of a logarithmic frequency sweep\relax }}{7}{figure.2.3}
\defcounter {refsection}{0}\relax
\contentsline {figure}{\numberline {2.4}{\ignorespaces Spectrogram of a guitar (a) and piano (b) sample\relax }}{8}{figure.2.4}
\defcounter {refsection}{0}\relax
\contentsline {figure}{\numberline {2.5}{\ignorespaces MFCCs of a guitar (a) and piano (b) sample\relax }}{8}{figure.2.5}
\defcounter {refsection}{0}\relax
\contentsline {figure}{\numberline {2.6}{\ignorespaces MFCCs mean and standard deviation of a guitar (a) and piano (b) sample\relax }}{9}{figure.2.6}
\defcounter {refsection}{0}\relax
\contentsline {figure}{\numberline {2.7}{\ignorespaces Melodic and timbral features of the song Layla by Eric Clapton\relax }}{9}{figure.2.7}
\defcounter {refsection}{0}\relax
\contentsline {figure}{\numberline {2.8}{\ignorespaces Rhythm features of the song Layla by Eric Clapton\relax }}{10}{figure.2.8}
\defcounter {refsection}{0}\relax
\contentsline {figure}{\numberline {2.9}{\ignorespaces Original scores, Rachmaninoff (a) and Beethoven (b)\relax }}{12}{figure.2.9}
\defcounter {refsection}{0}\relax
\contentsline {figure}{\numberline {2.10}{\ignorespaces Pitch extraction with Aubio\relax }}{12}{figure.2.10}
\defcounter {refsection}{0}\relax
\contentsline {figure}{\numberline {2.11}{\ignorespaces Pitch extraction with Melodia\relax }}{13}{figure.2.11}
\defcounter {refsection}{0}\relax
\contentsline {figure}{\numberline {2.12}{\ignorespaces MIDI transcription F\IeC {\"u}r Elise\relax }}{13}{figure.2.12}
\defcounter {refsection}{0}\relax
\contentsline {figure}{\numberline {2.13}{\ignorespaces Genre distribution of songs in various datasets\relax }}{18}{figure.2.13}
\defcounter {refsection}{0}\relax
\contentsline {figure}{\numberline {2.14}{\ignorespaces Extracted pitches, Spotify API (Spotipy)\relax }}{20}{figure.2.14}
\defcounter {refsection}{0}\relax
\contentsline {figure}{\numberline {2.15}{\ignorespaces Million Song Dataset genre distribution~\cite [p. 6]{msd5}\relax }}{21}{figure.2.15}
\defcounter {refsection}{0}\relax
\contentsline {figure}{\numberline {2.16}{\ignorespaces MapReduce algorithm~\cite {mapred1im}\relax }}{23}{figure.2.16}
\defcounter {refsection}{0}\relax
\contentsline {figure}{\numberline {2.17}{\ignorespaces Spark cluster scheme (according to~\cite [p. 46]{sparkbook1})\relax }}{24}{figure.2.17}
\defcounter {refsection}{0}\relax
\contentsline {figure}{\numberline {2.18}{\ignorespaces Spark application UI examples taken from the recommender system\relax }}{25}{figure.2.18}
\defcounter {refsection}{0}\relax
\addvspace {10\p@ }
\defcounter {refsection}{0}\relax
\contentsline {figure}{\numberline {3.1}{\ignorespaces Construction noise, first 100 song recommendations based on Musly toolkit (JS)\relax }}{33}{figure.3.1}
\defcounter {refsection}{0}\relax
\contentsline {figure}{\numberline {3.2}{\ignorespaces Chroma feature examples\relax }}{35}{figure.3.2}
\defcounter {refsection}{0}\relax
\contentsline {figure}{\numberline {3.3}{\ignorespaces Band-pass filtered audio, Sia - Chandelier\relax }}{35}{figure.3.3}
\defcounter {refsection}{0}\relax
\contentsline {figure}{\numberline {3.4}{\ignorespaces Thresholded chroma features, Sia - Chandelier\relax }}{36}{figure.3.4}
\defcounter {refsection}{0}\relax
\contentsline {figure}{\numberline {3.5}{\ignorespaces Processed chroma features, Sia - Chandelier\relax }}{36}{figure.3.5}
\defcounter {refsection}{0}\relax
\contentsline {figure}{\numberline {3.6}{\ignorespaces Workflow chroma feature extraction\relax }}{37}{figure.3.6}
\defcounter {refsection}{0}\relax
\contentsline {figure}{\numberline {3.7}{\ignorespaces Processing step 3 of chroma features in detail\relax }}{38}{figure.3.7}
\defcounter {refsection}{0}\relax
\contentsline {figure}{\numberline {3.8}{\ignorespaces 1D cross-correlation\relax }}{40}{figure.3.8}
\defcounter {refsection}{0}\relax
\contentsline {figure}{\numberline {3.9}{\ignorespaces 2D cross-correlation of beat-aligned and key-shifted chromagrams (audio snippets)\relax }}{42}{figure.3.9}
\defcounter {refsection}{0}\relax
\contentsline {figure}{\numberline {3.10}{\ignorespaces 2D cross-correlation of beat-aligned chromagrams (Sia / Pvris - Chandelier)\relax }}{43}{figure.3.10}
\defcounter {refsection}{0}\relax
\contentsline {figure}{\numberline {3.11}{\ignorespaces Filtered cross-correlation (high-pass)\relax }}{43}{figure.3.11}
\defcounter {refsection}{0}\relax
\contentsline {figure}{\numberline {3.12}{\ignorespaces Beat histogram examples\relax }}{44}{figure.3.12}
\defcounter {refsection}{0}\relax
\contentsline {figure}{\numberline {3.13}{\ignorespaces Rhythm pattern examples\relax }}{45}{figure.3.13}
\defcounter {refsection}{0}\relax
\contentsline {figure}{\numberline {3.14}{\ignorespaces Rhythm pattern extraction procedure as suggested by~\cite {rp_extract2}\relax }}{46}{figure.3.14}
\defcounter {refsection}{0}\relax
\contentsline {figure}{\numberline {3.15}{\ignorespaces Rhythm histogram examples\relax }}{47}{figure.3.15}
\defcounter {refsection}{0}\relax
\contentsline {figure}{\numberline {3.16}{\ignorespaces Detected onset examples (30 second song snippets)\relax }}{48}{figure.3.16}
\defcounter {refsection}{0}\relax
\addvspace {10\p@ }
\defcounter {refsection}{0}\relax
\contentsline {figure}{\numberline {4.1}{\ignorespaces Performance of various toolkits on a single computer\relax }}{55}{figure.4.1}
\defcounter {refsection}{0}\relax
\contentsline {figure}{\numberline {4.2}{\ignorespaces Feature extraction of the FMA dataset on the ARA-cluster (performance)\relax }}{57}{figure.4.2}
\defcounter {refsection}{0}\relax
\contentsline {figure}{\numberline {4.3}{\ignorespaces Feature file sizes\relax }}{58}{figure.4.3}
\defcounter {refsection}{0}\relax
\contentsline {figure}{\numberline {4.4}{\ignorespaces Workflow Spark\relax }}{59}{figure.4.4}
\defcounter {refsection}{0}\relax
\contentsline {figure}{\numberline {4.5}{\ignorespaces Lazy evaluation and caching optimization\relax }}{69}{figure.4.5}
\defcounter {refsection}{0}\relax
\contentsline {figure}{\numberline {4.6}{\ignorespaces Performance depending on the \#Executors spawned\relax }}{73}{figure.4.6}
\defcounter {refsection}{0}\relax
\contentsline {figure}{\numberline {4.7}{\ignorespaces Performance of different feature types\relax }}{74}{figure.4.7}
\defcounter {refsection}{0}\relax
\contentsline {figure}{\numberline {4.8}{\ignorespaces Performance ARA, full workload, (MFCC + Notes + RP)\relax }}{75}{figure.4.8}
\defcounter {refsection}{0}\relax
\contentsline {figure}{\numberline {4.9}{\ignorespaces Performance ARA, full workload, (JS + Chroma + RP)\relax }}{75}{figure.4.9}
\defcounter {refsection}{0}\relax
\contentsline {figure}{\numberline {4.10}{\ignorespaces Workflow of Merged DF approach\relax }}{76}{figure.4.10}
\defcounter {refsection}{0}\relax
\contentsline {figure}{\numberline {4.11}{\ignorespaces Performance of two subsequent song requests, all features\relax }}{77}{figure.4.11}
\defcounter {refsection}{0}\relax
\contentsline {figure}{\numberline {4.12}{\ignorespaces Performance of descending importance filter and refine, all features\relax }}{78}{figure.4.12}
\defcounter {refsection}{0}\relax
\contentsline {figure}{\numberline {4.13}{\ignorespaces Performance depending on \#Executors (36 CPU cores each)\relax }}{80}{figure.4.13}
\defcounter {refsection}{0}\relax
\addvspace {10\p@ }
\defcounter {refsection}{0}\relax
\contentsline {figure}{\numberline {5.1}{\ignorespaces Feature space example\relax }}{82}{figure.5.1}
\defcounter {refsection}{0}\relax
\contentsline {figure}{\numberline {5.2}{\ignorespaces Correlation matrix, 95 random songs, 19 genres (5 each), 1517-Artists\relax }}{82}{figure.5.2}
\defcounter {refsection}{0}\relax
\contentsline {figure}{\numberline {5.3}{\ignorespaces Cumulative distributions of distances\relax }}{83}{figure.5.3}
\defcounter {refsection}{0}\relax
\contentsline {figure}{\numberline {5.4}{\ignorespaces Impact of SKL scaling on the weighted sum\relax }}{84}{figure.5.4}
\defcounter {refsection}{0}\relax
\contentsline {figure}{\numberline {5.5}{\ignorespaces Correlation of features depending on SKL scaling\relax }}{84}{figure.5.5}
\defcounter {refsection}{0}\relax
\contentsline {figure}{\numberline {5.6}{\ignorespaces Scatter matrix, correlation 95 songs, 19 genres (5 each), 1517-Artists\relax }}{85}{figure.5.6}
\defcounter {refsection}{0}\relax
\contentsline {figure}{\numberline {5.7}{\ignorespaces Genre recall rate on 1517-Artists dataset\relax }}{88}{figure.5.7}
\defcounter {refsection}{0}\relax
\contentsline {figure}{\numberline {5.8}{\ignorespaces Scatter matrix, distances 1 random Rock\&Pop song, 1517-Artists, 4 genres\relax }}{89}{figure.5.8}
\defcounter {refsection}{0}\relax
\contentsline {figure}{\numberline {5.9}{\ignorespaces Scatter matrix, distances 1 random Electronic song, 1517-Artists, 4 genres\relax }}{90}{figure.5.9}
\defcounter {refsection}{0}\relax
\contentsline {figure}{\numberline {5.10}{\ignorespaces Scatter plots rhythm features / BPM for random Rock\&Pop and Classical songs\relax }}{91}{figure.5.10}
\defcounter {refsection}{0}\relax
\addvspace {10\p@ }
\defcounter {refsection}{0}\relax
\addvspace {10\p@ }
\defcounter {refsection}{0}\relax
\contentsline {figure}{\numberline {A.1}{\ignorespaces Distances 1 random song (Soundtrack), 5 genres (10 songs each)\relax }}{103}{figure.A.1}
\defcounter {refsection}{0}\relax
\addvspace {10\p@ }