Assessment of Optical Music Recognition tools, OMR Assessment, (type 1)

Musicnetwork Assessment (type 1) of Optical Music Recognition

Version 1.0, Date: 09-09-2003

Download .zip version

Addedd description at the end of the web page in date: 03/03/2014
see more on: P. Bellini, I. Bruno, P. Nesi, ``Assessing Optical Music Recognition Tools'', Computer Music Journal, MIT Press, Boston, USA, Vol.31, N.1, pp.68-94, Spring 2007 new

__________

Original Image

OMR 1 (vote)

OMR 2 (vote)

OMR 3 (vote)

 

File

Vote

File

Vote

File

Vote

esempio1.gif

es1-omr1.pdf

 

es1-omr2.pdf

 

es1-omr3.pdf

 

esempio2.gif

es2-omr1.pdf

 

es2-omr2.pdf

 

es2-omr3.pdf

 

esempio3.gif

es3-omr1.pdf

 

es3-omr2.pdf

 

es3-omr3.pdf

 

esempio4.gif

es4-omr1.pdf

 

es4-omr2.pdf

 

es4-omr3.pdf

 

esempio5.gif

es5-omr1.pdf

 

es5-omr2.pdf

 

es5-omr3.pdf

 

esempio6.gif

es6-omr1.pdf

 

es6-omr2.pdf

 

es6-omr3.pdf

 

esempio7.gif

es7-omr1.pdf

 

es7-omr2.pdf

 

es7-omr3.pdf

 

 

 

 

Symbol

Name

Vote

Symbol

Name

Vote

Empty Note Head duration 4/4

 

Beam 1 (duration 1/8)

 

Empty Note Head duration 2/4

 

Beam 2 (duration 1/16)

 

Black Note Head duration -> 1/64

 

Beam 3 (duration 1/32)

 

Rest duration 4/4

 

Beam 4 (duration 1/64)

 

Rest duration 2/4

 

Augmentation Dot

 

Rest duration 1/4

 

Accent

 

Rest duration 1/8

 

Number 1

 

Rest duration 1/16

 

Number 2

 

Rest duration 1/32

 

Number 3

 

Rest duration 1/64

 

Number 4

 

Barline Single

 

Number 5

 

Barline Double

 

Number 6

 

Barline End

 

Number 8

 

Barline Start Refrain

 

Number 9

 

Barline End Refrain

 

Number 12

 

Sharp

 

Number 16

 

Flat

 

Slur

 

Natural

 

P (piano)

 

Double Sharp

 

F (forte)

 

Double Flat

 

Comma (breath)

 

Treble Clef

 

C

 

Bass Clef

 

Staccato

 

Baritone Clef

 

Fermata

 

Tenor Clef

 

Mordent

 

Contralto Clef

 

Turn

 

Mezzosoprano Clef

 

Grace Note

 

Soprano Clef

 

Trill

 

 Dwn

 Up

Hook 1 (duration 1/8)

 

Tenuto

 

 Dwn

 Up

Hook 2 (duration 1/16)

       

 Dwn

Up

Hook 3 (duration 1/32)

       

Dwn

Up

Hook 4 (duration 1/64)

       

 

 

Complete Music Symbols & Relationships

Vote

Note with pitch and duration

 

Note with accidentals

 

Groups of beamed notes

 

Rests

 

Time signature and time change

 

Key signature and key signature change

 

Symbols below or above notes (turn, trill, accent, staccato, etc)

 

Grace notes

 

Slurs (and ties) & bends

 

Augmentation dots

 

Clefs

 

Irregular notes groups

 

Number of measures

 

Number of staves

 

 

  1. On the basis of your experience of what you expect to have from an optical music recognition please give a vote from 1 to 10 (mini to max) about the overall quality of reconstructed images by means of the three OMR applications with reference to the original images you found in the folder. In this assessment, please take into account only the results of the reconstruction and not the quality of printing resolution.
  2. This part of the questionnaire has been set up for identifying the relevance of the basic music notation symbols in the context of typical OMR application and recognition phases. On the basis of your experience, what is the importance you give in the recognition of the set of basic symbols listed below? Please, give a vote (Vote column) from 1 to 10 where 10 means the highest relevance.
  3. This part of the questionnaire has been set up for identifying the relevance of the complete and composed music notation symbols in the context of typical OMR application and recognition phases.
    On the basis of your experience, what is the importance you give in the recognition of complete music symbols and relationships among symbols listed below? Please, give a vote from 1 to 10 where 10 means the highest relevance.
  4. On the basis of your experience as user of one of more music Editors we intend to estimate the typical work time to perform the data entry opertions. They are the typical actions for polishing a result of an OMR processing.
    To this end, we ask you to evaluate in term of time (seconds) the editing operations listed below.
    Please, fill the Time field, in seconds, with the average time of each operation and fill the music editor field with the name of the software tools you are referring to.

 

 

Operation

Music Editor

Time

Inserting a symbol

   
   
   
   
   
 

Searching a symbol in the score

   
   
   
   
   
 

Deleting a symbol in the score

   
   
   
   
   
 

Preparing the page for editing

   
   
   
   
   

 

 

 

 

 

Page

Average Time needed to edit from scratch

esempio1.gif

 

esempio2.gif

 

esempio3.gif

 

 

 

Original Image

OMR 1
(verification time)

OMR 2

(verification time)

OMR 3

(verification time)

 

File

time

File

time

File

time

esempio1.gif

es1-omr1.pdf

 

es1-omr2.pdf

 

es1-omr3.pdf

 

esempio2.gif

es2-omr1.pdf

 

es2-omr2.pdf

 

es2-omr3.pdf

 

esempio3.gif

es3-omr1.pdf

 

es3-omr2.pdf

 

es3-omr3.pdf

 

 

  1. On the basis of your experience as user of music Editors, how many minutes are needed to editing from scratch the 3 pages of music examples considered?
  2. On the basis of your experience as user of music Editors, how many minutes are needed to verify the correctness of the 3 pages of music examples considered, that consists in marking the errors in the page ?
  3. On the basis of your experience as user of music Editors, how many minutes are needed to perform the corrections of the 3 pages of music examples considered, that consists in polishing the score to mek it equal to the original.

 

Please state the music editor that you are supposed to use.

USED MUSIC EDITOR

 

Original Image

OMR 1
(polishing time)

OMR 2
(polishing time)

OMR 3
(polishing time)

 

File

time

File

time

File

time

esempio1.gif

es1-omr1.pdf

 

es1-omr2.pdf

 

es1-omr3.pdf

 

esempio2.gif

es2-omr1.pdf

 

es2-omr2.pdf

 

es2-omr3.pdf

 

esempio3.gif

es3-omr1.pdf

 

es3-omr2.pdf

 

es3-omr3.pdf

 

 

------------------
see more on P. Bellini, I. Bruno, P. Nesi, ``Assessing Optical Music Recognition Tools'', Computer Music Journal, MIT Press, Boston, USA, Vol.31, N.1, pp.68-94, Spring 2007 new
----

Integration notes

The selected seven images, used as test set, are monophonic sheets music and are mainly exercise of “solfeggio”. They were taken/scanned from Italian music theory books (We suppose that some book is available also as international publication).
For each image of test set, alphanumeric strings (i.e. title, page number, exercise number, etc…) were manually removed in order to evaluate the behavior of OMR systems only with music symbols (OCR capability was out of scope). The selected images allow coping with:

  • Different font size and shapes
  • The most frequent music symbols in the repertoire of classical music
  • Density and complex layout of musical symbols
  • Different tuplets (triplets, quadruplets, quintuplets and sextuplets)
  • Small musical symbols (appoggiaturas, cues, no alterations)
  • Different symbols of beats subdivision (begin refrain, end refrain, end score, simple/double barline, etc.).
  • Changes of time and key signatures

The following table lists for each image: book and reference, page(s) and exercise number(s).


Original Image/Link Book Page(s) Exercise(s)
esempio1.gif (Pozzoli 3° corso) [1] 10 9
esempio2.gif (Pozzoli 3° corso) [1] 14 12
esempio3.gif (Gentilucci)[2] 20 6
esempio4.gif (Curci) [3] 33 213-214-215
esempio5.gif (Pozzoli 3° corso) [1] 19 16-17
esempio6.gif (Pozzoli 3° corso) [1] 31 25
esempio7.gif (Gentilucci) [2] 11 3
 
Books/References
[1] Pozzoli, Ettore. “Solfeggi Parlati e Cantati 3° corso”, Edizioni Ricordi, Milano Italy (1973). ASIN B009LFPA0K
[2] Gentilucci, O. “13 Prove d’esame per la licenza di Solfeggio”, Edizioni Curci – Milano, Italy (1983) E.6032C.
[3] Alberto Curci, Tecnica Fondamentale del violino, Parte seconda, pag. 33 Copyright 1952 by Edizioni Curci, Milano. ISBN: 0215902513