Content uploaded by Henrik von Coler
Author content
All content in this area was uploaded by Henrik von Coler on Jan 14, 2019
Content may be subject to copyright.
Proceedings of the 21st International Conference on Digital Audio Effects (DAFx-18), Aveiro, Portugal, September 4–8, 2018
TU-NOTE VIOLIN SAMPLE LIBRARY – A DATABASE OF VIOLIN SOUNDS WITH
SEGMENTATION GROUND TRUTH
Henrik von Coler
Audio Communication Group
TU Berlin
Germany
voncoler@tu-berlin.de
ABSTRACT
The presented sample library of violin sounds is designed as
a tool for the research, development and testing of sound analy-
sis/synthesis algorithms. The library features single sounds which
cover the entire frequency range of the instrument in four dynamic
levels, two-note sequences for the study of note transitions and vi-
brato, as well as solo pieces for performance analysis. All parts
come with a hand-labeled segmentation ground truth which mark
attack, release and transition/transient segments. Additional rele-
vant information on the samples’ properties is provided for single
sounds and two-note sequences. Recordings took place in an ane-
choic chamber with a professional violinist and a recording engi-
neer, using two microphone positions. This document describes
the content and the recording setup in detail, alongside basic sta-
tistical properties of the data.
1. INTRODUCTION
Sample libraries for the use in music production are manifold.
Ever since digital recording and storage technology made it possi-
ble, they have been created for most known instruments. Commer-
cial products like the Vienna Symphonic Library1or The EastWest
Quantum Leap2offer high quality samples with many additional
techniques for expressive sample based synthesis. For several rea-
sons, these libraries are not best suited for the use in research on
sound analysis and synthesis. Many relevant details are subject to
business secrets and thus not documented. Copyright issues may
prevent a free use as desired in a scientific application. These li-
braries also lack annotation and metadata which is essential for
research applications, if used for machine learning or sound anal-
ysis / synthesis tasks.
The audio research community has released several databases
with single instrument sounds in the past, usually closely related to
a specific aspect. Libraries like the RWC [1] or the MUMS [2] aim
at genre or instrument classification and timbre analysis [3]. Data-
bases for onset and transient detection which include hand labeled
onset segments have been presented by Bello et al. [4] and von
Coler et al. [5].
The presented library of violin sounds is designed as a tool for
the research, development and testing of sound analysis/synthesis
algorithms or machine learning tasks. The contained data is struc-
tured to enable the training of sinusoidal modeling systems which
distinguish between stationary and transient segments. By design,
the library allows the analysis of several performance aspects, such
1www.vsl.co.at/
2http://www.soundsonline.com/
symphonic-orchestra
as different articulation styles, glissando [6] and vibrato. It fea-
tures recordings of a violin in an anechoic chamber and consists of
three parts:
1. single sounds
2. two-note sequences
3. solo (scales and compositions/excerpts)
For single sounds and two-note sequences, hand-labeled seg-
mentation files are delivered with the data set. These files focus
on the distinction between steady state and transient or transitional
segments. The prepared audio files and the segmentation files are
uploaded to a static repository with a DOI [7]3. A Creative Com-
mons BY-ND 4.0 license ensures the unaltered distribution of the
library.
The purpose of this paper is a more thorough introduction of
the library. Section 2will explain the composition of the content,
followed by details on the recording setup and procedure in Sec-
tion 3. The segmentation data will be introduced in Section 4. Sec-
tion 5presents selected statistical properties of the sample library.
Final remarks are included in Section 6.
2. CONTENT DESCRIPTION
2.1. Single Sounds
Similar to libraries for sample based instruments, the single sounds
capture the dynamic and frequency range of the violin, using sus-
tained sounds. The violinist was instructed to play the sounds
as long as possible, using just one bow, without any expression.
Steady state segments, respectively the sustain parts, of these notes
are thus as played as steady as possible. This task showed to be
highly demanding and unusual, even for an experienced concert
violinist.
On all of the four strings, the number of semitones listed in
Table 1was captured, each starting with the open string. This
leads to a total of 84 positions. All positions are captured in four
dynamic levels which were specified as pp - mp - mf - ff result-
ing in a total amount of 336 single sounds. According to Meyer
[8], the dynamic interval interval of a violin covers a range from
58 . . . 99 dB.
3https://depositonce.tu-berlin.de//handle/
11303/7527
DAFX-1
Proceedings of the 21st International Conference on Digital Audio Effects (DAFx-18), Aveiro, Portugal, September 4–8, 2018
Table 1: Number of positions on each string
String Positions
G 18
D 18
A 18
E 30
Each item was recorded in several takes, until recording engi-
neer, the author and the violinist agreed on success. Although all
sounds were explicitly captured in both up- and down-stroke tech-
niques, these modes have not been considered individually in the
data set and thus appear randomly.
2.2. Two-Note Sequences
0 2 7 12 30
Position
Fifth, two strings
Fourth, low
Fourth, high
Fifth, one string
Figure 1: Violin board with positions for two-note sequences
For the study of basic articulation styles, a set of two-note
sequences was recorded at different intervals, listed in Table 2.
The respective positions on the board are visualized in Figure 1.
All combinations were recorded at two dynamic levels mp and
ff. Three different articulation styles (detached, legato, glissando)
were used and some combinations were captured with additional
vibrato. These combinations lead to a grand total of 344 two-note
items.
5 semitones on one string were captured in 8 pairs with 24
versions (2 dynamic levels, 2 directions, with and without vibrato,
3 articulation styles): 2·2·3 = 24.
Repeated tones were captured in 4 pairs with 6 versions (2
dynamic levels, legato and detached, the latter with and without
vibrato): 22+ 2 = 6
7 semitones on one string were captured in pairs with 20 ver-
sions (2 dynamic levels, two directions, detached only without vi-
brato, legato and glissando with and without vibrato): 2·2+24=
20
7 semitones on two strings were captured in 3 pairs with 16
versions (2 dynamic levels, two directions, with and without vi-
brato and two articulation styles [legato, detached]):24= 16
Table 2: All two-note pairs
5 semitones, one string
Two-note Note 1 Note 2
item no. ISO Pos. String ISO Pos. String
01-24 D4 7 G A3 2 1
25-48 A4 7 D E4 2 2
49-72 E5 7 A B4 2 3
73-96 B5 7 E F#5 2 4
97-120 D4 7 G G4 12 1
121-144 A4 7 D D5 12 2
145-168 E5 7 A A5 12 3
169-192 B 7 E E6 13 4
Repeated tones
Two-note Note 1 Note 2
item no. ISO Pos. String ISO Pos. String
193-198 D4 7 G D4 7 G
199-204 A4 7 D A4 7 D
205-210 E5 7 A E5 7 A
211-216 B5 7 E B5 7 E
7 semitones, one string
Two-note Note 1 Note 2
item no. ISO Pos. String ISO Pos. String
217-236 D4 7 G G3 0 G
237-256 A4 7 D D4 0 D
257-276 E5 7 A A4 0 A
277-296 B5 7 E E5 0 E
7 semitones, two strings
Two-note Note 1 Note 2
item no. ISO Pos. String ISO Pos. String
297-312 D4 7 G A4 7 D
313-328 A4 7 D E5 7 A
329-344 E5 7 A B5 7 E
2.3. Solo: Scales and Compositions
Two scales – an ascending major scale and a descending minor
scale – were each played in three interpretation styles, as listed in
Table 3. The first style was plain, without any expressive gestures,
followed by two expressive interpretations. Six solo pieces and
excerpts, listed in Table 4which mostly contain cantabile legato
passages were recorded. All compositions were proposed by the
violinist, ensuring familiarity with the material.
Table 3: Scales in the solo part
Item Type Interpretation
01 major, ascending plain
02 major, ascending expressive 1
03 major, ascending expressive 2
04 minor, descending plain
05 minor, descending expressive 1
06 minor, descending expressive 2
DAFX-2
Proceedings of the 21st International Conference on Digital Audio Effects (DAFx-18), Aveiro, Portugal, September 4–8, 2018
Table 4: Solo recordings
Item Composition Composer
07 Sonata in A major for Vio-
lin and Piano César Franck
08 Violin Concerto in E mi-
nor, Op. 64, 2nd move-
ment
Felix Mendelssohn
09 Méditation (Thaïs) Jules Massenet
10 Chaconne in g minor Tomaso Antonio Vitali
11 Violin Concerto in E mi-
nor, Op. 64, 3rd move-
ment
Felix Mendelssohn
12 Violin Sonata no.5, Op.24,
1st movement Ludwig van Beethoven
3. RECORDING SETUP
The recordings took place in the anechoic chamber at SIM4, Berlin.
Above a cutoff frequency of 100 Hz the room shows an attenua-
tion coefficient of µ > 0.99, hence the recordings are free of re-
verberation in the relevant frequency range. The recordings were
conducted within two days, taking one day for the single sounds
and the second day for two-note sequences and solo pieces. All
material was captured with a sample-rate of of 96 kHz and a depth
of 24 Bit.
Microphones
The following microphones were used:
•1x DPA 4099 cardiod clip microphone
•1x Brüel & Kjær 4006 omnidirectional small diaphragm
microphone with free-field equalization, henceforth BuK
The DPA microphone was mounted as shown in Figure 2,
above the lower end of the f-hole in 2 cm distance. Due to its
fixed position, movements of the musician do not influence the
recording. The B&K microphone was mounted in 1.5 m distance
above the instrument, at an elevation angle of approximately 45◦,
as shown in Figure 3.
Figure 2: Position of the DPA microphone
4http://www.sim.spk-berlin.de/refelxionsarmer_
raum_544.html
Figure 3: Position of the B&K microphone
Instructions
For each of the single-sound, two-note and scale items, a mini-
mal score snippet was generated using LilyPond [9]. Examples
for items’ instructions are shown in Fig. 4. The resulting 63 page
score was then used to guide the recordings. Although the isolated
tasks may seem simple and unambiguous, this procedure ensured
smooth recording sessions.
2
vib.
ff
2
(a) Two-note example with
vibrato and glissando
3
Ë
mp
mp
3
(b) Single-sound example
with upbow and downbow
Figure 4: Instruction scores for two-note aand single-sound b
4. SEGMENTATION
The segmentation of a monophonic musical performance into notes,
and even more into a note’s subsegments is not trivial [10,11].
During the labeling process, the best of the takes for each item
was selected from the raw recordings and the manual segmenta-
tion scheme proposed by by von Coler et al. [5] was applied using
Sonic Visualiser [12].
DAFX-3
Proceedings of the 21st International Conference on Digital Audio Effects (DAFx-18), Aveiro, Portugal, September 4–8, 2018
(a) Energy trajectory
(b) Peak frequency spectrogram
Figure 5: Sonic Visualiser setup for annotation of single sound 333
4.1. Single Sounds
Each single sound is divided into three segments, which are de-
fined by four location markers in the segmentation files5, as shown
in Table 5. The first time instant (A) marks the beginning of the
attack segment, the second instant (C) marks the end of the attack
segment, respectively the beginning of the sustain part. The end
of the sustain, which is also the beginning of the release segment,
is labeled with the (D). The label (B) marks the end of the release
portion and the complete sound. The left column holds the related
time instants in seconds.
Table 5: Example for a single-sound segmentation file
(SampLib_DPA_01.txt)
0.000000 A
0.940646 C
7.373000 D
8.730500 B
The definition of the attack segment is ambiguous in literature
[13] and shall thus be specified for this context: Attack here refers
to the actual attack-transient, the very first part of a sound with
a significant inharmonic content and rapid fluctuations. In other
contexts, the attack may be regarded the segment of rise in energy
to the local maximum. Often, there is still a significant increase in
energy after the attack-transient is finished. As the attack-transient
is characterized by unsteady, evolving partials and low relative par-
tial amplitudes, the manual segmentation process is performed us-
ing a temporal and a spectral representation. Figure 5shows a
typical Sonic Visualiser setup for the annotation of a single sound.
The noisiness of the signal during attack and release can be seen in
the spectral representation. How attack transient and rising slope
may differ, is illustrated in Fig. 6. The gray area represents the la-
beled attack segment, which is finished before the end of the rising
slope is reached.
Less ambiguous, the release part is labeled as the segment
from the end of the excitation until the complete disappearance
5The segmentation files are part of the repository [7]
0 0.511.522.5
0.00
0.02
0.04
0.06
0.08
t/s
RMS
RMS
Attack segment
End of rising slope
Figure 6: RMS trajectory of a note beginning with attack segment
(gray) and end of the rising slope (single sound no. 19)
34567
0.02
0.04
0.06
0.08
t/s
RMS
RMS
Release segment
Beginning of falling slope
Figure 7: RMS trajectory of a note end with release segment (gray)
and beginning of the falling slope (SampLib_19)
of the tone. As shown in Fig. 7, there is often a significant de-
crease in signal energy before the actual release starts. For items
with low dynamics, the release is also covering the very last part
of the excitation.
The ease of annotation varies between dynamic levels, as well
as between the fundamental frequency of the items. Notes played
at fortissimo show clear attack and decay segments with a steady
sustain part, whereas pianissimo tones have less prominent bound-
ary segments and a parabolic amplitude envelope. The higher SNR
in fortissimo notes allows a better annotation of the transients.
Tones with a high fundamental frequency have less prominent par-
tials, whereas the bow noise is emphasized. They are thus more
difficult to label, since attack transient are less clear in the spectro-
gram. The segmentation of high pitched notes at low velocities is
hence most complicated.
4.2. Two-Note Sequences
The two-note sequences contain the the segments note, rest and
transition with the labels listed in Table 6. Stationary sustain parts
are labeled as notes, whereas the transition class includes attack
and release segments, as well as note transitions, such as glissando.
All two-note sequences follow the same sequence of segments
(0-2-1-2-1-2). Figure 8shows a labeling project in Sonic Visu-
aliser for a two-note item with glissando. The transition segment
is placed according to the slope of the glissando transition.
DAFX-4
Proceedings of the 21st International Conference on Digital Audio Effects (DAFx-18), Aveiro, Portugal, September 4–8, 2018
Table 6: Segments in the two-note labeling scheme
Label Segment
0 rest
2 transition
1 note
(a) Energy trajectory
(b) Peak frequency spectrogram
Figure 8: Sonic Visualiser setup for annotation of two-note item
22
4.3. Solo
Solo items have been annotated using the guidelines proposed by
von Coler et al. [5]. Due to the choice of the compositions, only
few parts violated the restriction to pure monophony. Solo item
10, for example, starts with a chord, which is labeled as a single
transitional segment.
5. STATISTICS
This section reports selected descriptive statistical properties of the
sample library which are potentially useful when considering the
use of the data.
5.1. Single Sounds
Fig. 9shows the RMS for all single sounds, in box plots for each
dynamic level. The median for the dynamic levels is logarithmi-
cally spaced.
Table 7: Segment length statistics for the single-sounds
l/s µ/s
Attack 0.247 0.206
Sustain 5.296 1.118
Release 0.705 0.802
Statistics for the segment lengths of the single sounds are pre-
sented in Table 7and Figure 10, respectively. With a mean of
5.296 s, the sustain segments are the longest, followed by release
segments with a mean of 0.705 s. Attack segments have a mean
pp mp mf ff
−6
−4
−2
log(rms)
Figure 9: Boxplot of RMS for the sustain from the BuK micro-
phone
length of 0.247 s. Extreme outliers in the mean attack length are
caused by high pitched notes with low dynamics.
Attack Sustain Release
0
2
4
6
8
l[s]
Figure 10: Box plots of segment lengths for all single sounds
5.2. Two-Note
The two-note sequences allow a comparison of different articu-
lation styles. Figure 11 shows the lengths for detached, legato
and glissando transitions in a box plot. With a median duration of
0.72 s, glissando transitions tend to be longer than legato (0.38 s)
and detached (0.37 s) transitions.
detached legato glissando
0.5
1
1.5
Transition type
l[s]
Figure 11: Box plot of transition lengths for all two-note sequences
DAFX-5
Proceedings of the 21st International Conference on Digital Audio Effects (DAFx-18), Aveiro, Portugal, September 4–8, 2018
5.3. Solo
Table 8: Note statistics for items in the solo category
Solo item Number of notes l/s µ/s
1 8 0.698 0.745
2 8 0.721 0.768
3 8 0.728 0.776
4 8 0.707 0.753
5 8 0.724 0.771
6 8 0.774 0.848
7 104 0.695 0.661
8 75 1.074 0.899
9 89 0.911 0.923
10 63 0.735 0.690
11 76 0.689 0.707
12 56 0.615 0.740
For the solo category, the basic statistics on the note occur-
rences and lengths are listed in Table 8. All scales (items 1 - 6)
contain 8 notes, compositions (items 7-12) have a mean of 77 notes
per item. With a mean note length of 0.614 906 s, item 12 has the
shortest, and with 1.074 361 s, item 8 has the longest notes.
6. CONCLUSION
The presented sample library is already in application within sinu-
soidal modeling projects and for the analysis of expressive musi-
cal content. Overall recording quality proves to be well suited for
most tasks in sound analysis. Since the segmentation ground truth
follows strict rules and has undergone repeated reviews, it may be
considered consistent.
7. ACKNOWLEDGMENTS
The author would like to thank the violin player, Michiko Feuer-
lein, and the sound engineer, Philipp Pawlowski, for their work
during the recordings, as well as the SIM Berlin for the support.
Further acknowledgment is addressed to Moritz Götz, Jonas Mar-
graf, Paul Schuladen and Benjamin Wiemann for the contributions
to the annotation.
8. REFERENCES
[1] Masataka Goto et al. “Development of the RWC music database”.
In: Proceedings of the 18th International Congress on Acous-
tics (ICA 2004). Vol. 1. 2004, 553â556.
[2] Tuomas Eerola and Rafael Ferrer. “Instrument library (MUMS)
revised”. In: Music Perception: An Interdisciplinary Jour-
nal 25.3 (2008), 253â255.
[3] Gregory J Sandell. “A Library of Orchestral Instrument Spec-
tra”. In: Proceedings of the International Computer Music
Conference. 1991, 98â98.
[4] J.P. Bello et al. “A Tutorial on Onset Detection in Music
Signals”. In: IEEE Transactions on Speech and Audio Pro-
cessing 13.5 (2005), 1035â1047.
[5] Henrik von Coler and Alexander Lerch. “CMMSD: A Data
Set for Note-Level Segmentation of Monophonic Music”.
In: Proceedings of the AES 53rd International Conference
on Semantic Audio. London, England, 2014.
[6] Henrik von Coler, Moritz Götz, and Steffen Lepa. “Para-
metric Synthesis of Glissando Note Transitions - A user
Study in a Real-Time Application”. In: Proc. of the 21st Int.
Conference on Digital Audio Effects (DAFx-18). Aveiro,
Portugal, 2018.
[7] Henrik von Coler, Jonas Margraf, and Paul Schuladen. TU-
Note Violin Sample Library. TU-Berlin, 2018. DOI:10 .
14279/depositonce-6747.
[8] JÃ1
4rgen Meyer. “Musikalische Akustik”. In: Handbuch der
Audiotechnik. Ed. by Stefan Weinzierl. VDI-Buch. Springer
Berlin Heidelberg, 2008, 123â180.
[9] Han-Wen Nienhuys and Jan Nieuwenhuizen. “LilyPond, a
system for automated music engraving”. In: Proceedings
of the XIV Colloquium on Musical Informatics (XIV CIM
2003). Vol. 1. 2003, 167â171.
[10] E. GÃ3mezEtal.. “Melodic Characterization of Monophonic
Recordings for Expressive Tempo Transformations”. In: Pro-
ceedings of the Stockholm Music and Acoustics Conference.
2003.
[11] Norman H. Adams, Mark A. Bartsch, and Gregory H. Wake-
field. “Note Segmentation and Quantization for Music In-
formation Retrieval”. In: IEEE Transactions on Speech and
Audio Processing 14.1 (2006), 131â141.
[12] Chris Cannam, Christian Landone, and Mark Sandler. “Sonic
visualiser: An open source application for viewing, analysing,
and annotating music audio files”. In: Proceedings of the
18th ACM international conference on Multimedia. ACM.
2010, 1467â1468.
[13] Xavier Rodet and Florent Jaillet. “Detection and Modeling
of Fast Attack Transients”. In: Proceedings of the Interna-
tional Computer Music Conference. 2001, 30â33.
DAFX-6
