IDMT-SMT-Bass-Single-Track

References

[1] Abeßer, Jakob; Dittmar, Christian; Kramer, Patrick; Schuller, Gerald:

Parametric Audio Coding of Bass Guitar Recordings using a Tuned Physical Modeling Algorithm, submitted to the 16th International Conference on Digital Audio Effects (DAFx), 2013

 

Database Content

The IDMT-SMT-BASS-SINGLE-TRACK dataset comprises of 17 bass lines from different music styles.

It is intended as a public evaluation dataset for:

  • retrieval of repetitive bass patterns

For each bass line, the pattern length (in seconds) and the begin of the first pattern appearance (in seconds) is annotated.

The patterns are in general no exact repetitions but instead contain occasional
pitch and rhythm variations.

  • bass transcription

Each note is annotated with the score-related parameters onset, offset and pitch.

  • spatial transcription / estimation of the fretboard position


Each note is annotated with the instrument-related parameters string number and fret number.

  • estimation of bass guitar plucking styles

Each note is played and annotated with one of the 5 plucking style classes:

  • Finger-style (FS) - alternate plucking of the string using the index and
    middle finger
  • Picked (PK) - plucking of the string using a plastic pick
  • Muted (MU) - plucking of the string using the thumb and index finger while simultaneously damping the string vibration using the palm of the hand
  • Slap-Pluck (SP) - picking of a string using either the index or the middle finger (causing a collision between the string and the upper frets)
  • Slap-Thumb (ST)- striking of the string using the thumb (causing a collision between the string and the upper frets)

 

  • estimation of bass guitar expression styles (6 classes)

Each note is played and annotated with one of the 6 expression style classes:

  • Normal (NO) - no expression, just “regular” bass note playing
  • Harmonics (HA) - flageolet tones
  • Dead-note (DN) - damped, percussive note
  • Bending (BE) - singular bending of the string during vibration
  • Vibrato (VI) - periodic bending and releasing the string during vibration
  • Slide (SL) - slide up or down after note is plucked

 

 

More audio datasets

Fraunhofer IDMT has compiled further audio data sets for various research areas such as instrument recognition, fingerings or game analysis.

Research topic

Audio and Visual Content Analysis

Extracting meaningful data from audiovisual content

Research topic

Environmental Sound Analysis

AI-based analysis of complex acoustic scenes and sounds

Research topic

Media Forensics

Trustworthy media content

Research topic

Industrial Sound Analysis

AI-based acoustic monitoring in production