Music Object Oriented Distributed System
An activity within the: TETRApc-TTN
22-Sept.-1998, Teatro alla Scala, Milano
Partner responsible: DSI
Partners involved: DSI, SCALA ELSEL, SMF, RICORDI, SHYLOCK
Project and Consortium Coordinator:
Dipartimento di Sistemi e Informatica
Universita’ degli Studi di Firenze
Via S. Marta 3
50139 Firenze, Italy
email: firstname.lastname@example.org, email@example.com
Program: 22 Sept. 1998
(Project Coordinator: P. Nesi, DSI and Maestro C. Tabarelli, Teatro alla Scala)
16:40 -State of the Art (Maestro G. Dotto, BMG RICORDI)
16:50 -MOODS Architecture, aim, and acknowledgements (P. Nesi, DSI)
17:05 -Music that will be Executed (Maestro C. Tabarelli, Teatro alla Scala)
17:15 -The load and the execution of Serenade in G of Mozart
-A short explanation of what has been shown
17:25 -The load and the execution of La primavera (Le 4 Stagioni di Vivaldi)
17:35 -Archive Management and Format Conversions (Dr. S. Moro, SHYLOCK)
17:45 -The load and the execution of Traviata I, Verdi
-A short explanation of what has been performed
17:55 -The load and the execution of Traviata III, Verdi
18:05 -Conclusions and Future Work (P. Nesi, DSI)
Table of Contents
1 Welcome (P. Nesi, DSI, Maestro C. Tabarelli, Teatro alla Scala) *
2 *State of the Art (Maestro G. Dotto, BMG RICORDI) *
3 MOODS Architecture (P. Nesi, DSI) *
3.1 Main Features and Benefits of MOODS *
3.2 MOODS Components *
3.3 *Relationships among Lecterns *
3.4 *Editing Capabilities of MOODS *
3.5 MOODS During the Execution *
4 Acknowledgements (P. Nesi, DSI) *
4.1 The Musicians Involved in the Demonstration *
5 Music Executed During Demonstration (Maestro C. Tabarelli, Teatro alla Scala) *
6 Archive Management and Format Conversions (Dr. S. Moro, SHYLOCK) *
7 *Consortium and Project: Historical Issues *
8 Conclusions and Future Work (P. Nesi, DSI) *
The points marked with * will be summarized or omitted during the demonstration. They have been included in this document to give a more complete view of MOODS project.
We are here today to present the results of MOODS project.
This is the absolutely first public presentation of MOODS project.
MOODS is a project partially founded by the European Commission in the HPCN (High Performance Computer Networking) domain of the ESPRIT. The project is only one of the several projects using the HPCN technology Italy. The EC has established a network of TTN (Technology Transfer Node) to increase the diffusion of HPCN technology. In this case, project MOODS is an activity with TETRApc TTN, supported by CPR (Consorzio Pisa Ricerche, Pisa, Italy), CESVIT (High Tech Agency, Florence, Italy).
In this view, I am the MOODS Project/Consortium Coordinator and the Scientific Responsible of TETRApc TTN for CESVIT part and activities.
MOODS is a TETRApc ESPRIT IV HPCN Project n.25968
WWW site of MOODS: http://aguirre.ing.unifi.it/~hpcn/wwwmoods/wwwpag.html
The project is really innovative, and it could be very interesting to know the opinion of Maestro C. Tabarelli of the Teatro alla Scala that his hosting us for this presentation.
Explanation of the motivation of Teatro all Scala in participating in MOODS project (Maestro C. Tabarelli).
Before to explain what MOODS is, it is better to explain the current operative way of theatres and music publishers. To this end, I would like to pass this work at Maestro G. Dotto of CASA RICORDI.
The partners in the MOODS project have worked on a well-known problem: storing and managing via computer the enormous amount of musical information used by orchestras during rehearsals and public performance of concerts, operas or ballets.
The amount of information to be managed by orchestras in performance is huge. A main score is often comprised of more than 100 pages (for an opera, as many as 600 or more pages) in large format, and each musician, or every lectern, a specific derived single "part" (involving, for larger orchestras, over 100 musicians). When a chorus is present it can number over 100 choristers. And the performance time can range from a few minutes to over two hours.
When modifications or special written instructions are required, paper versions of main scores and individual parts are usually manipulated by the conductor or the archivist (before rehearsals) and by musicians (during rehearsals). The modifications are often quite significant and inserting them can be very time consuming, particularly in the case of operas and ballets, or of new symphonic works. Simple modifications consist in adding interpretation symbols (dynamics, expression marks, string bowings, mute on/off, etc.) or modifying other figures (ties, slurs, measure, etc.). Heavy modifications can also consist in omitting sections, moving score parts, adding new portions of score, arranging the music for different instruments, etc.
When different musicians use the same paper version of a music score, old modifications must be deleted (if possible) and new modifications are made on the same scores. If the marks are too heavy, the scores or parts must be replaced, at significant cost. And both processes are particularly time-consuming. On the other hand, sometimes various versions of a set of score and parts must be kept in an archive, since they carry the interpretation marks of important interpreters or because they can be required, in the same version, even years later. The costs of storing, maintaining or replacing multiple unique sets of parts can become exorbitant both for theater archives and publishers.
Modifications can range from minor indications or changes decided by each musician and inserted by pencil on the parts during rehearsals, to major cuts or alterations decided by the conductor (or composer, in the case of new works) for which the archivist must insert paste-overs, clip pages together, write in long passages of in place of rest, et cetera. Some such major changes may be decided before hand, but many may be decided during the rehearsal period, necessitating time-consuming interruptions of the rehearsal time or else costly extra time of archivists and copyists between one rehearsal and the next. Such major, on-site decisions are especially frequent with stage works like operas and ballets (which are also the works which involve the largest amount of musical material), where decisions for musical changes can derive, during rehearsals, from various needs – of singers, of stage directors, problems of acoustical space, etc.
The time spent in waiting for the adapted version of a score and parts can reach large amounts of the total time allotted for a rehearsal. This is a significant cost considering that in many cases musicians, dancers, choristers, supernumeraries, etc. are on an hourly wage and, in all cases, the total time allotted to rehearsals is fixed. It is clear that a distributed system of lecterns with instantly modifiable music would result in impressive cost and time savings.
Further, in many cases the logistics of renting music material (for publishers) or storing sets of music scores and parts (for opera-theater or orchestra archives) make it difficult, if not impossible, to retain copies of all the various versions generated of a given work by different conductors or theater productions. At best, a copy of a modified score may be set aside, but not the full set of parts. When a specific version or production is repeated some seasons later, the modifications have to be re-generated, creating redundant costs that can be significant. A solution that would make storage of multiple versions possible, and greatly reduce the amount of repetitive work involved, is an important goal of many performing organizations and publishers. MOODS offers just such a solution.
For publishers, this technology offers the added solution of creating a huge data-base over time from which to derive and store a potentially infinite number of personalized versions, and even of various, publishable "editions" of repertory works. The methodology of supply and distribution could also, over time, change radically, avoiding in select cases the "printed paper" stage of supply altogether, and transmitting digital music information directly to theaters and performing organizations, or supplying CD Roms with the stored data of any number of versions. This same system would offer exact accounting for copyright purposes (see project CITED of EC).
The technology for storing and recalling all various versions of musical works will be especially useful for schools of music. Students can thus examine (and perform) different interpretations of the same score.
Moreover, scores can be called up instantly on screen at any time, allowing for the programming of time-saving, complex rehearsal schedules.
MOODS consists in an integrated system of computer-based lecterns/stands for cooperative editing and visualization of music. MOODS is an innovative solution for automating and managing the large amount of information used by (i) orchestras during rehearsals and public performance of concerts, operas, ballets, etc. (ii) students of music during lessons in conservatories and schools of music, (iii) publishers during massive editing of music. The targeted end-users are theatres, itinerant orchestras, groups of musicians, schools of music, television network orchestras, and publishers of scores.
The main features and benefits of MOODS are:
Please consider that the prototype that you will see at work is the first prototype of very complex system. It is comprised of 8 microprocessor-based systems that work all together by mean of a high-performance.
In this phase, MOODS does not pretend to be a complete and robust product but only an advanced prototype for demonstrating the potentiality of the solution and for opening its adoption in many several other application fields.
It will be engineered after the project completion to make from it a set of derived products.
The look of the lecterns will be strongly improved to transform them into cold tools for the musicians.
MOODS project and prototype have been implemented by using technologies based on: HPCN, object-oriented paradigm, real-time technique, on the basis of the results of projects LIOO and MSLIOO of DSI. The implementation of MOODS prototype involves both hardware and software aspects.
A MOODS system is comprised of:
The configuration used at the demonstration at the Teatro alla Scala is comprised of 5 lecterns for musicians (DLIOO). These are of three different types and are differentiated for resolution of the LCD monitor. Two of them (those used for violins) can be used with both a simple pen and a trackball for adding symbols, the other support only the trackballs. Two different types of trackball are available. The lecterns are special purpose machines designed by ELSEL and are based on Pentium 133 Mhz.
The MASE station is a Hewlett and Packard Personal Computer with high performance. It is endowed of an high resolution monitor of HP for a total of 1600 x 1200 pixels.
The MASAE station is the core of the system and is a Hewlett and Packard workstation B32L, a RISC machine.
The Theatre Network is based on a support at 100 BT of Hewlett and Packard.
All the above machines are based on UNIX-like operating system.
The machine devoted for the Database is a high performance Personal Computer.
The other machines on the desks are only for the audio/video recording of the event.
Lecterns receive from Archivist Station (MASAE) all information for automatically visualizing the current score page by the means of network. On DLIOO lecterns the visualization is performed according to the specific musical instrument to which each DLIOO is assigned. On each DLIOO, different sets of interpretation symbols are provided depending on the instrument assigned. The modifications performed on DLIOOs in editor mode are sent in real-time to MASE/MASAE editors for visualization and saving and to the other lecterns.
DLIOOs are organized according to the orchestra hierarchy. First level DLIOOs are qualified to perform changes on the current score (main and interpretation symbols), while the others can be active or passive. The active ones have the same capabilities of their parent, while the latter (the PDLOO) can only visualize scores/changes performed by the former.
This hierarchical organization allows the first instrument (e.g., the 1st violin) to directly transmit to the other sub-lecterns (e.g., violins) the changes decided on the score. Sometimes, the 1st violins can be two: in this case a DLIOO child is qualified to perform a co-operative work on the score and changes are reflected to the other violins. In any case, changes are sent to MASAE workstation. This co-operative work among distributed lecterns is possible since a high reliable and fast network is provided.
Each modification performed on the current score by using MASE/MASAE is sent in real-time to the corresponding lecterns according to the association of each pentagram to the corresponding physical lecterns connected through the network (using the ONCM). Modifications on the main score can involve both classical symbols of music and interpretation symbols. From the MASE/MASAE it is also possible to build new main scores and modifying old scores without loading the scores on DLIOOs (off-line modifications).
Since the early 1970s, there exist on the market several professional programs for typesetting and editing music on computer: among the more recent higher-quality programs consistently used by publishers are Score, Sibelius, Finale, Encore, Graphire Music Press, etc. Many other non-professional programs for music are also available, but these are mainly based on executing music by means of a MIDI interface (Darms, Csound, QuickScore, CUbase, Cakewalk Express, Lime, MasterScore, MusicShop, Hightingate, Music Manuscript, Music Time, Notator, Overture, etc).
In recent years, almost all publishers have come to use the professional programs, instead of the traditional hand-processing technologies. Since 1991 the adoption of such programs has led to the computer-based production of a large quantity of scores. However, due to the constant evolution of music programs and to the fact that up to now there only exists a viable market for paper versions of scores, the electronic versions remain in publishers' archives (though they may sometimes appear in multimedia CDs).
There exist several European projects dealing with computerized storage and supply of music and music-related information, though these mainly regard library-related use: JUKEBOX (for managing music libraries of phonic format); MODE (techniques for delivering music in paper and phonic formats on Internet); PARAGON (the study of techniques and methods for managing archives of music on Internet); CITED (management of copyrights for selling information covered by copyrights on Internet); CANTATE (definition of a client-server architecture for distributing music in different forms); HARMONICA (a conglomerate of music-library projects); E-MOLL (a real scale development of CANTATE project).
MOODS project has taken into account the results produced by these projects and in particular by projects CANTATE, E-MOLL and HARMONICA. This has been is guaranteed by the presence of SHYLOCK among the partners of this project.
According to our analysis –interviewing publishers, conductors, archivists, teachers, and performing musicians, as well as by direct experiments – none of the professional editors in the market is completely satisfactory:
All of these programs are far too complicated to be of real, practical use to musicians in a live-time, actual rehearsal situation. They were designed as graphics tools to be employed by copyists and publishers rather than by musicians who need to quickly and efficiently insert modifications on scores in real time, using devices like a mouse, a trackball or a touch screen.
Music in the above main score has no sense, it is only a demonstrative view of MASAE user interface.
The mechanism of turning pages is driven by the MASAE lectern during execution. It should be noted that different instruments, and thus different lecterns have the page turning at different time instants; this is due to the different quantity of written music which has to be executed by each instrument.
During the execution, on the lectern of the director (MASE) the next page is constructed by starting from the left as soon as the beats on left are already executed. The screen shows two parts, one presenting the current page which is gradually reduced as soon as the measures and the pentagrams are executed. The DLIOOs presents the next page growing from top to bottom. In this way the modality adopted for reading scores by musicians is not changed.
As MOODS Project coordinator I would like to thanks:
I would like to personally thank Prof. Giacomo Bucci of DSI for his support and encouragements since the early phases of work at DSI. A thanks also to the several members of DSI, Department of Systems and Informatics that have followed with extreme interest the project; and, the Faculty of Engineering for providing space to host the rehearsals of MOODS in July.
A warm thanks to the several musicians that have tested MOODS prototype. Some of them have used MOODS yesterday for the first time, today are here for demonstrating its functionalities and usability.
From the point of view of musicians, the organization of demonstration as well as the rehearsals performed in Florence has been performed by the Scuola di Musica di Fiesole, SMF (Maestro Nicola Mitolo).
The execution of music at the demonstration are directed by Maestro Carlo Moreno Volpini.
The musicians involved in the demonstrations are:
The role of the Theatre Archivist is covered by: Timna Panfietti Monaco and/or Antonio Albino (DSI).
The role of MOODS configurator is covered by: Pierfrancesco Bellini (DSI)
The music used during rehearsals and execution were those selected at the begin of the project, and in particular:
Other available music pieces:
This part of the MOODS system provides two functions which are basic in a real-life use:
The score database in the theatre
As MOODS system will have to manage large amount of electronic scores, each possibility in more than one version (after modifying interventions during rehearsals) and formats, MOODDB provides the cataloging services of SHYLOCK's standard ARCHImusica - a software package especially devoted to musical libraries and archives -, customized within MOODS project with a versioning system integrated to the lecterns network. This could serve as the local database (at theatre). An Internet user interface (WWW) allows the archivist to use the system or from his/her office on Windows (3.1 to NT 4.0) or from the same MASAE workstation in the concert hall.
The score database at publishers
Also on the publisher's side there is the need of managing on computer the printed score catalogs, with an increasing number of entries enhanced with the score in electronic form, and possibly offering these scores directly via Internet.
The Theatre needs a score from the Publisher
MOODDB system allows the archivist of the theatre searching the score needed on the theatre local database or at the Publisher's one. As BMG-RICORDI has actually established its rental catalog on an ARCHImusica system with a SHYLOCK ARCHIweb module, a remote search via Internet on that catalog can be demonstrated, together with the retrieval of the needed electronic score. This application has been engineered after a prototype developed in another UE-cofunded project CANTATE (Computer Access to NoTation And TExt, 1996-1997).
Scores in different formats
As any text on computer, the format of data depends on the program which has been used to produce it. So is for music, where a number of musical writing programs have been used in the past 10 years to produce the most part of the published scores. MOODDB include a bi-directional conversion program from one of the most popular formats (SCORE by Leland Smith) to MOODS language and vice versa, allowing MOODS system to load and modify even scores available in that format. As other conversion programs exist on the market converting from another popular format (Finale by Coda Publishing) to SCORE, both formats are now suitable to feed MOODS system.
In the ‘70, the first programs for managing music on computer have been proposed [Abbott85], [Anderson91b], [Gordon85], [Gourlay86], [Loy85]. Those solutions were much more focussed on music editing, music generation, on storing music in audio formats for the eventual substitution of paper. Since the ‘80, researches on optical music recognition have been proposed in order to convert paper scores in "elettronic" format. In [Morita91], a system to follow the bat of director in order to provide at a computer generating music to follow the director’s interpretation has been also proposed. Therefore, the idea of "electronic lectern" is at least old such as the computer and first graphic editors. This is also demonstrated by the several attempts that have been performed in order built electronic lecterns, in Italy as well as in the rest of the world, from singles and from companies producing instruments. These have tried to insert small display for the visualization of music directly on their keyboard.
For the visualization of music there obviously were enormous technological and formal problem, due (i) to the low resolution of monitors; (ii) to the software modeling of music as visual language with complex rules for placing symbols; and (iii) to the architectural problem of a software itself.
Thus, the idea of electronic lectern, as many others, has been confined in the collective imagination for several years. It needed innovative technological solutions to be realized. This is demonstrated by the several attempts of which the national and international literature and the instruments builders are witness.
Recently, as regards point (i), the new technology has allowed the implementation of LCD TFT at high resolution. On music modeling and notation, several treatises have been written for several thousands of pages, without to reach standardization. The computer science and could be the right tools reach the objective, as demonstrated by the presence of more than 200 programs for music editing, executing, acquiring, that since the ’70 have been commercialized.
The architectural problems are the most complex to be solved; research technologies of software engineering and telematics with HPCN technology (High Performance Computer Networking), have allowed to set up a solution as demonstrated by the implementation of MOODS.
As regards historical issues of MOODS project, it is necessary to spend a few of words on reasons that have created the fundamentals for implementation of MOODS project.
On the basis of the results obtained by DSI in the research activity, that had produced LIOO prototype in January 1995 and MSLIOO in the 1996, some of the future partners of MOODS proejct started with the definition of a consortium.
In the first part of 1996, the TETRApc TTN (Techonlogy Transfer Node) was built: a consortium (founded by European Commission) comprised of CPR (Centro Pisa Ricerche), CESVIT e CSM. TETRApc TTN has as target the promotion and the diffusion of HPCN technology (High Performance Computer Networking) at the national level and joined with the others TTNs in Europe. Among the homeworks of TTN, there is the soliciting of project proposals with the adoption of HPCN technology. This role has stimulated CESVIT and DSI (scientific responsible for CESVIT in the TETRApc) to give a chance to projects LIOO/MSLIOO of DSI to become what is today MOODS. The evolution of projects LIOO/MSLIOO into a distributed system environment gave the real possibility to implement a system of lecterns for theatres, orchestras, schools of music, etc. The project had to be the typical operation in which the injection of new HPCN technology in a traditional system could produce new solution and market. This for demonstrating the usefulness of HPCN technology and to increase their diffusion for producing innovative systems.
In the 1996, the consortium was constituted: DSI (Dipartimento di Sistemi Informatica, Universita' degli Studi di Firenze), SCALA (Teatro alla Scala), RICORDI (BMG Ricordi, Casa Ricordi), ELSEL S.r.L., SHYLOCK Progetti, SMF (Scuola di Musica di Fiesole).
The three views of the potential end-users for MOODS system orchestra, music publishers and music scools were represented at high level, as well as the technological-industrial partners. In March 1997 the project was founded, the 16th of July 1997 started. The conclusion is planned for the 31st of October 1998. Since from its early formulation the coordinator has been Dr. P. Nesi of DSI.
[Abbott85] C. Abbott, Guest Editor's Introduction to the Special Issue on Computer Music, ACM Computing Surveys, Vol.17, N.2, pp.147-151, June, 1985.
[Anderson91b] D. P. Anderson and R. Kuivila, Formula: A Programming Language for Expressive Computer Music, IEEE Computer, pp.12-21, July, 1991.
[Blostein91] D. Blostein and L. Haken, Justification of Printed Music, Communications of the ACM, Vol.34, N.3, pp.88-99, March, 1991.
[Gordon85] J. W. Gordon, System Architectures for Computer Music, ACM Computing Surveys, Vol.17, N.2, pp.191-234, June, 1985.
[Gourlay86] J. S. Gourlay, A Language for Music Printing, Communications of the ACM, Vol.29, N.5, pp.388-401, May, 1986.
[Loy85] G. Loy and C. Abbott, Programming Languages for Computer Music Synthesis Performace Analysis, ACM Computing Surveys, Vol.17, N.2, pp.235-265, June, 1985.
[Morita91] H. Morita and S. Hasimoto and S. Ohteru, A Computer Music System that Follows a Human Conductor, IEEE Computer, pp.44-53, July, 1991.
We hope that the demonstration has been of interest for you.
It is obviously impossible to describe and to give a full demonstration of all MOODS features in so short time. This project started in 1994 internally at DSI and is become MOODS project in 1997 by adding several contributors. It has seen, in these years a considerable number of companies and people involved in. Thanks again to all of them.
From the early testing of MOODS, it has been considered strongly useful and interesting to be used in orchestras and music schools. MOODS functionalities leave free the artistic aim of Composer and of Conductor that can directly experiment new effects/arrangements with the orchestra in few seconds. It is our opinion that, MOODS opens a new era for expressiveness of the artistic aim of Conductors and Musicians establishing a direct feedback from requests to the result produced by the orchestra. MOODS opens the path to what in the next years could be a real revolution in the world of music publishing and in the approach of musicians toward the music to be executed.
MOODS project will be completed within the October 31st. Any formal activity related to MOODS project has to be always directed to its coordinator.
The completion of the project does not means that the effort spent for building this experiment/prototype will be destroyed. The results will be exploited.
Please don’t forget to compile and leave the questionnaire that you have in the documentation, for us, it is very important to know your opinion about the demonstration that you have attended.
As the spokesman of MOODS project, it is a pleasure to answer to any your questions.
Please contact the Project Coordinator by means of the references on the cover page of this document, if: