In the following a clear description of the sector actors' needs and of the state of the art are presented.

General Needs and Opportunities

back to top

In the scenario, we have publishers and music consumers: theatres, orchestras, music schools, conservatories, recording studios, television networks, and libraries. This order has nothing to do with their relevance. In the following, the needs of each category are discussed.

Publishers -- Publishers need to distribute music at a lower cost and to a wider audience. The costs for distributing music for a medium size publisher (for classical music) are close to 1000 Keuro per year. For light/popular music this is close to be 5 times. Moreover, most of publishers have a large amount of music never used since it is totally unknown by most of the music consumers, and it is too expensive to be promoted. By using Internet, the music could be distributed at low costs or for free, thus limiting the cost for promoting new music and increasing the culture of music. This is very interesting for new artists, since a sort of free exposition of new composers could be provided. Therefore, by the WEDELMUSIC approach, a reduction of costs for delivering and promoting music could be reached for publishers. A valorisation of the huge music archives that are present in Europe will be more feasible and economically interesting.

Orchestras, Theatres, Recording Studios, Television Networks -- They rent and/or buy music sheets from publishers. Renting is less expensive than buying. In the renting price for music also the payment of copyright is comprised. Typically, the delivery of music (several kilograms of paper) is totally charged to them (directly or indirectly, obviously). The reception of music via Internet should reduce the cost of delivery for these music consumers. This means that music fascicles should be produced by the music consumers. The availability of laser printers capable of printing on both sides with binding facilities makes the result comparable to the format that can be acquired on the music shop as well as directly from the publisher. Moreover, in most cases, the archivist performs this work after that of cut and past with glue and scissors but at lower cost. The rented music sheets are frequently manipulated in deep during rehearsals and frequently they are no more usable when they are returned back to the publishers since they have been marked by musicians even indelibly. This means that publishers have to reprint them to loan them again to another music consumer with the corresponding expenses. Orchestras, Theatres, Recording Studios, Television Networks, need to manipulate music in several ways: from simple additions of instrumental and interpretation symbols to the deep changes performed for adjusting music for preparing operas and ballets. These needs have been highlighted in the MOODS project. For these music consumers, to have the possibility of manipulating music in symbolic format will be a strong improvement. In this case, the archivist, who manually prepares the music by cutting and pasting with glue and scissors, could be strongly facilitated in the work. Very complex and time-consuming changes could be performed in real-time if the music in symbolic format is available. By receiving the music in a symbolic format, the work of the archivist could change from that of simple cut and past activity to that of a technical organiser of music archive. This will reduce the mundane and time consuming work of archivists, conductors and music teachers in favour of experimentation of a more strictly coupled reaction between the changes decided and the response of the orchestra as well as of music class. The interactivity with music is strongly needed for theatres, orchestras, Recording Studios, Television Networks, as demonstrated in the MOODS project. Thus with the WEDELMUSIC approach large number of new functionalities will be available and exploitable; a reduction of costs for receiving music and for preparing final performance are foreseen.

Libraries -- The mission of libraries is two fold: (i) to maintain trace of all cultural activities, (ii) to provide free or paying information even if only for a temporary period. For this second case, libraries are strongly interested in providing a large amount of different music scores per authors and per versions (different executions and preparations) for reading and analysing it. For studying music, the analysis of single pieces and the comparative analysis among different pieces are needed. Libraries need to be able to manipulate scores in order to include them in interactive documents and notices provided to their users in the form of CD-r or WWW pages. As confirmed in a HARMONICA report on User needs several music manipulation would be appreciated: synchronisation of scores with recordings (asked by 41% of questioned users); comparison between different versions of the same score (32%); automatic transcription of a recording into a basic score (pitch and rhythm detection) (29%) for instance for searching melodies; excerpt capture (27%); time stretching without transposition (27%); possibility to store score excerpts (25%); decomposition of one score in order to listen to its different instrumental tracks (24%); personal tags and annotations on scores (23%); representation of music in a graphical form (envelope contour or sonogram) (19%).

In order to perform most of the new techniques of music analysis, a computer and the availability of music in symbolic format are mandatory. This service is presently too expensive to be supported by musicians, thus, the Libraries could be the right place in which such a service could be made available. With the availability of music in symbolic format the music production synchronously with the music visualisation is possible at any execution rate.

Moreover, the distribution of music in symbolic format is the only way for allowing the adoption of music in libraries for blind people. Partners, and specially libraries (IRCAM and SVB), are strongly interested in these aspects.

The presence of very large WWW archives (distribute or concentrated) could be considered the killer of the Libraries in the next years. On the contrary, most of the WWW archives are made by libraries and will be largely used at the first place in Libraries. Services similar to that based on WEDELMUSIC could be a reason for increasing their value.

Libraries as well as Music Schools and Conservatories can be regarded as Intermediate users. They could be a sort of self-service points in which the customers might select music by viewing and/or listening it. On these sites, it will be possible for the consumers to see the music, to try it, thus reducing the wrong buying. Libraries could play the role of music distributors for the publishers towards the real end-users. Libraries typically charge for membership, photocopies, SMT access, CD, video, etc., the distribution of music does not change much their role and organisation. Libraries are also seen from the publishers as a securer point from which the music can be delivered since they are much more involved in the music than music shops.

Libraries should transform their business figure. Their mission is not in conflict with playing the role of intermediate role between publishers and customers. Then, users of the Libraries could (i) download music, (ii) print music (direct or transposed as they need) by paying the publishers directly in the library. This process could be totally transparent for small libraries, thus avoiding the related effort for maintaining the service. Clients of libraries are typically students. They have a large number of needs for several kind of music. Libraries and Publishers share for a great percentage the same clients, the agreement could be strongly interesting for both, culturally and economically.

With WEDELMUSIC, an increment of the allowed functionalities, and thus of the corresponding services and profits is foreseen for libraries; even for those of music schools and conservatories.

Music Schools, Conservatories -- Music schools and Conservatories can be public or private and can be regarded as special kind of libraries. Their mission is to provide music courses and they have typically their own libraries. According to laws on copyright, it is possible to make copy of music sheets without paying the copyright if this is done for educational purposes. This means that teachers can make the copies and not that these can be freely distributed to students. Both Conservatories and Music Schools have specific needs regarding the manipulation of music. They are strongly interested in manipulating music: transposing, arranging, changing, preparing music for orchestra execution such as theatres and orchestras. They are also interested in analysing music in deep such as Libraries. Music in music schools has to be modifiable for satisfying the needs of several different musicians instruments (transpositions and rearrangements are frequent operations needed).

Moreover, they are also producers of music composers. The publishers could be strongly interested in distributing the new music produced in Music Schools via the same mechanism used for distributing commercial music, this could be a new social service. This mechanism could be a way for increasing the visibility of young composers. The interactivity with music is strongly needed for music schools, professors, conservatories, etc., as demonstrated in MOODS and MUSTUTOR projects.

The Maestro could ask to rearrange a music piece directly on a computer music editor. In music schools and conservatories, it is more frequent the presence of multimedia classrooms. An interactive music format could also be used for executing music via computer. A pupil studying a music piece of a quartet could set the computer to play three parts on four and, thus, he/she could be the fourth musician of the quartet with the computer. This feature could be strongly interested even for musicians of orchestras, theatres, etc. They could be interested in simulating the whole orchestra except for their parts. Velocity of execution could be controlled and thus this could be strongly useful to learn playing music. Pupils studying as orchestra conductors could be strongly assisted by the availability of an interactive music since they could arrange music and test the result in real-time with the available orchestra.

With WEDELMUSIC, a reduction of cost for receiving music and a new modality for teaching music are foreseen for Music Schools, Conservatories. They are also interested for the advantages foreseen for Libraries and Orchestras, Theatres, Recording Studios, Television Networks.

Opportunity -- According to MOODS music management of music, each music piece can be provided by the publisher in a so-called Neutral Version. The NV can be protected by using encryption techniques. The music consumer may load and modify the NV by means of one or more Additional Command Lists (ACL) including, editing, addition, deletions, transposition etc. In effect, the changes are not permanently performed, they are temporary applied, and they remain confined in a log file, that is the ACL. Different ACLs can be independently saved and loaded. Each ACL can be reapplied only on the corresponding NV, every time is needed. Several different ACLs can be produced by the consumer without changing the NV. Thus, the NV continues to be protected while the ACLs include the history of the changes performed for obtaining each modified version of the music piece. Therefore, for each NV, the music consumer can prepare several distinct versions with their corresponding ACLs. The protection of the NV can enforce limitations in the use time, and can be constrained to be loaded only on the computer of who has paid for the copyrights. A fine tuning of the allowed operations has also to be provided.

With this mechanism of NV plus ACLs, the publishers are satisfied since their copyright is maintained (see also results of Vienna HARMONICA forum). The NVs of music have not to be distributed without permission. Music consumers are also satisfied since they can use music of publishers in an interactive manner for producing their specific versions. NV and an ACL can be merged only if the correct permissions are provided. Each print of the produced music has to provide evidence of this. Therefore, specific watermark and fingerprint techniques have to be applied.

The copyright protection mechanisms have to maintain their efficiency even after music manipulation making hard and non-economic the violation of copyrights. This will be performed by using encryption with hardware support integrated with fingerprint and watermark techniques.

Moreover, there is the need of negotiable and adaptable licenses to satisfy specific and flexible necessities.

Each piece of music has to provide a specific copyright profile in which the permissions of exploitation are specified.

Protection mechanisms and the model have to be capable of defining the operations allowed for each piece of music and for each part included in.

For example, a music piece is comprised of several parts. For each part several different allowed functionalities could be set: print, edit, change notes, add instrumental symbols, delete instrumental symbol, execute, transpose, reformat, etc., and their detailed combinations. This means that different allowed operations could be defined for each piece of music, for each user, for each part in a main score and any combination of these. A musician could hold the permission to execute all parts but only the permission to manipulate a part.

In order to solve the above problems, several technical issues and innovations will be performed as regards of: music modelling, copyright protection.

Music Modelling

back to top

Most of music publishers have in their archives only scores on paper-based supports. Most of them are converting their archives from paper to images to survive on the market. There are several advantages in delivering music in symbolic format instead of the image format. Music in image format is strongly expensive to be sent on the network. An A3 sheet of music in symbolic format can be easily stored. In addition, the delivering of symbolic format is the only mechanism for assuring a certain degree of interaction and modifiability of music. The process of conversion will probably take several years and has been delayed for the lack of a standard symbolic music format. This means that for a very long period both music scores in image and symbolic formats will be present at the same time in the archives and thus on the market. The music delivering systems have to consider this fact: the availability of a uniform model, such as that proposed in WEDELMUSIC, for manipulating scores in image, symbolic and audio formats (as well as in their combination) is strongly needed.

Thus, there exist a large number of dialects. Among these, the most diffuse or discussed are: SCORE, FINALE (Enigma format), MIDI, others are SIBELIUS, NIFF, SMDL, and. As demonstrated by MOODS and CANTATE projects, these languages are too far from the needs required for distributing interactive music. In order to provide support for defining a new generation of applications based on music objects, such as the co-operative tuition of pupils in music schools, the music analysis, the co-operative work in orchestra, the building of virtual orchestras, the unification of the model for sighted and impaired people, a symbolic format capable of modelling all the music relationships among notation symbols in a formal approach is absolutely mandatory. This will avoid the production of several distinct applications for patching the lack of the existing models. In WEDELMUSIC, the model will integrate image scores and audio files and will include several symbols defining their relationships in formalised manner. Moreover, WEDEL format will integrate a support for protecting music objects.

Protection and Modelling

back to top

In the past European programme, several different projects have been developed addressing the problems of copyright, protection and delivering of music -- such as: AMIDE (copyright, electronic payment), COPEARMS (derived from CITED, copyright), DECOMATE (copyright, distribution material), ECUP (electronic delivering), ECUP+ (derived from ECUP for libraries), TECUP (electronic delivering, copyright), CANTATE (music delivering from publisher to libraries), HARMONICA (harmonisation of actions in the area of libraries, copyright, delivering), IMPRIMATUR (delivering, protection, copyright), MODE (music delivering, audio format), MUSICWEB (music delivering, educational), MUSE (copyright, watermark on audio), MUSTUTOR (OMR, music notation), MOODS (music distributed system, music model for distributed systems), OCTALIS (Offer of contents through trusted access links), TALISMAN (tracing author's rights by labelling image services and monitoring access network), and COPYSMART, COPYCAT, etc.

WEDELMUSIC considers their results and in particular those produced by: MOODS, HARMONICA, IMPRIMATUR, CANTATE, MUSTUTOR, OCTALIS, TALISMAN and MUSE.

Presently, several companies are distributing music to end users via their respective WWW sites. Most of the WWW sites are distributing music in audio format. Users are interested in manipulating music in audio. To protect the audio files, watermark techniques are under adoption for marking them with the owner ID-mark (identification code) and with the ID of the licensed user or distributor. By using these techniques, the publishers can protect their rights. Unfortunately these techniques have not been yet applied on MP3 format, which is the most large vehicle of unauthorised distribution of audio music. In WEDELMUSIC project specific watermark algorithms for marking audio files, with a special attention to MP3 files, will be analysed and defined.

On the Internet, there are also few examples of WWW servers for on-line delivering of music sheets (e.g., Musicsales, MUSICWEB project, etc.). Some of these are capable of sending images of music scores in PDF, PostScript or GIF formats. These solutions are a surrogate of the classical distribution of music sheets via music shops. The music acquired is not interactive and thus the related costs for using that music for preparing performances in orchestras, in music schools, etc., are high. Music consumers such as theatres, orchestras, music schools, music distributors, recording studios, blind people, and libraries need interactive music; that is, music that can be manipulated: arranged, transposed, modified, reformatted, printed on Braille, or described by speech, etc. This is a real need for preparing performances, studying music, analysing music, learning instruments, etc. Presently this is not possible since music scores are only distributed as music sheets. In order to be manipulated the music has to be in some specific symbolic format. Presently, no WWW server is delivering interactive music and as well as music in symbolic format (small tentative has been performed by MOODS, CANTATE, MUSICA and GUIDO projects but by using limited music models), and without protection and integration with image and audio. As discussed above several advantages are seen in delivering symbolic formats instead of the simple images of the music sheets. Moreover, the costs for distributing high resolution images of music sheets is presently prohibitive. In WEDELMUSIC project specific techniques for distributing music in symbolic formats will be derived considering the integration of symbolic format with images of music sheets and audio files, and a set of protection mechanisms.

The available mechanisms for protecting digital objects are under continuos restructuring. Typically, security mechanisms such as symmetric or asymmetric encryption techniques have been in use: triple DES, RSA, etc., also according to crypto regulations and laws which are defined at the European and national levels. In WEDELMUSIC project customisations on these techniques will be defined in order to set up a specific solution for music. The encryption algorithms will be used for protecting images of music sheets, audio files, and music in symbolic formats.

The solution proposed in WEDELMUSIC project is based on the concepts of Neutral Version, NV, and Additional Command Lists, ACLs, identified in MOODS project. According to MOODS, each music piece can be provided by the publisher in the so-called Neutral Version. The music delivered in images of score sheet has to be protected against duplication since these can be photocopied at low cost. The same problem is present when symbolic music is printed. In order to protect the rights (demonstrating the property of a music piece) the typical algorithms and mechanisms adopted for watermarking images and text cannot be used since they are too evident and disturb the music reading. For these reasons, in WEDELMUSIC project specific algorithms for introducing watermarks in (i) images of music sheets, and during (ii) the direct printing of music scores will be implemented. It will be possible to make distinction between the original version (Neutral Version, NV) and the manipulations performed (Additional Command Lists, ACLs) and by using digital signature and watermarking techniques for both audio and symbolic formats. The identification code of the owner will be maintained active and readable even after music manipulation: photocopying and audio duplication. It will be impossible or very expensive to use music without permissions.

Experimenting new Solutions and Model Validation

back to top

The WEDEL model, the protection support and the model that will be defined in the project have to be tested against real conditions in order to verify their suitability and effectiveness. For this reason they will be tested and validated by using a set of specific tools that will be capable of exploiting all the functionalities of the WEDEL model and format.

To this end, specific tools have to be built for assessing the model against real conditions:

The implementation and trial of these tools will make possible the reaching of the project objectives, verifying the adoption and the exploitation of new functionalities; and allowing also the dissemination of results. The innovation will be the possibilities of loading interacting music from the network in an integrated models in which several aspects will be enforced: audio, images of music sheets and music in symbolic format. The several above applications will access to the unified model according to their needs -- e.g., for proposing music on the screen, for printing music, for reproducing music, for manipulating music, for executing music, etc. for both impaired and sighted people. The experimentation will take into account all these facts and will be performed by skilled people.

During the experimentation several new functionalities due to the interactivity of the WEDEL music model will be tested.