1- The study of orchestration and the comprehension of its nature and mechanisms:
Nothing can beat attending concerts and listening to reference recordings - as much as possible with score in hand - when it comes to the appreciation of the art of orchestration. Furthermore, following rehearsals of a good orchestra will enable the listener to discover the full richness and subtlety of orchestration as instrumental families or sections are often rehearsed separately, demonstrating their role in the global texture. One can later, of course, enroll in a good instrumentation and orchestration class taught by an experienced composer or orchestrator who will perform piano reductions of the scores selected for study and isolate the different instrumental parts, suggesting and demonstrating their possible orchestration. It goes without saying that performing an instrument and performing in an orchestra (even a modest one) will also greatly enhance the awareness of the orchestral possibilities as well as its characteristics, dynamics and the typical behaviour of each instrumental family or section.
2- Tools one can dream of:
In the best of all worlds, one would have unlimited access to a symphonic orchestra in a concert hall which would be happy to perform any excerpt from the repertoire in one's prescribed instrumental combinations. The scenario would unfold as follows: "Dear maestro, would you be kind enough to perform for us measures 14-16 of the first movement of Mozart's 40th symphony; we would like to hear the three following instruments: first the flute then the first clarinet followed by the first bassoon and than all three together so we can appreciate the blend." Or perhaps: "Dear maestro, this time could you perform measures 1-8 of the second movement of Haydn's Symphony No. 100; we would like to hear the first flute alone followed by the first violins and then both combined." Or maybe also: "We would like to hear measures 59-63 from the first movement of Debussy's La Mer; the solo oboe first and then combined with the first flute so we can follow the subtle colouration provided by the flute. We would then like to hear measure 62: first the two clarinets alone and then combined with the two oboes and the two flutes so we can study the timbral octave blend. Then please measure 63 (clarinets, first basson, third horn, viola and cello) in order to compare this mix of instrumental families with the preceding measure." The list could go on and on…
3- The more down-to-earth reality and the hopes raised by new recording technologies:
Having a symphony orchestra at one's disposal might well remain a mere fantasy… unless a billionaire patron becomes interested in the idea!
One could however imagine that the technological developments of the last decades in the multitrack recording process might provide access to orchestral recordings that allow us to isolate at will specific instruments or families in order to study their roles in the orchestral combinations. Unfortunately this is also more of a daydream since orchestral recording sessions take place in a concert hall to insure musical and acoustical cohesion and that the array of microphones used (principal, tree, section, ambient. etc.) enable at best to slightly modify the overall balance between instrumental families or bring up further front an instrument in particular, while contributing to the specific timbre of the orchestra in the overall room acoustics. The sophisticated art of the sound engineer and the "Tonmeister" can merge in a perfect symbiosis with a musical interpretation but isolating each musician separately would require them to be separated acoustically in sound-proofed booths in a time-delayed recording process (progressive track recording) that would render ensemble playing mechanical and even imprecise while depriving the musician from the possibility of modulating his performance according to the hall's real acoustics. Those multi-track delayed recording techniques are more suited for pop music as the smaller number of instruments makes it more practicable and even preferable; they allow each individual track to be transformed electronically and mixed in very creative ways that often become the acoustical trademark of a band. But symphonic music recordings aim more towards a faithful rendition of the original acoustical performance by focusing on the richness and combinational variety of the timbres.
4- The simulation techniques (midi protocol and sound sampling banks):
There is a wide range of sampled acoustical instruments of various qualities that can be controlled through the midi transmission protocol that have been on the market for more than two decades. The best ones allow a potentially convincing simulation of instrumental performance through a large array of sampled modes of playing and variations with programmed scripts. One might believe that these electronic instruments could easily serve as substitutes for real acoustical instruments. The reality is that even if convincing simulations are possible, each of these virtual instruments come with its strengths and weaknesses as well as its own limits and idiosyncratically erring behaviour. It is not really possible to automate much and the musician willing to use them must have a very clear idea of the result he is aiming for long in advance and then explore which of its different modes and techniques might potentially serve that idea, modifying and reprogramming the virtual instruments as needed. This amounts to the precise carving of each individual line and balancing it within the orchestral context, a prowess that only a few experienced musicians have.
For the past few decades several music notation programs have also started to offer diverse "automated interpretation" systems from a digitized music score. Despite some commendable efforts, orchestration students or researchers must be warned against those types of simulation. Although useful, their acoustical rendition of the instrumental world is so far from reality that their pedagogical potential is very limited, bordering on the counter-productive and amounting to a form of ear "de-training" rather than ear training.
5- The OrchPlay Music Library:
OrchPlayMusic has developed a rendering system for orchestral scores combining the best techniques gained from sound sampling with the expertise of professional performers: this is the OrchSim system.
Each OrchSim recording is not only rendering of the composer's prescriptions already in the score but also relies on the interpretation's traditions transmitted from generation to generation by the performers. For example, given a Haydn symphony, it is obvious that the score alone does not describe and prescribe all details contributing to a convincing interpretation of the piece; one is lacking information about the modes of attack, the different types of phrasings, the modulations of timbre and agogic, the goal-oriented fluctuations of rhythm, intonation, vibrato and tempo as well as the constantly varying balance between members of a section or instrumental families in order to prioritize specific musical layers. All those elements are nevertheless essential and the OrchSim system has been developed to implement them.
OrchSim's recording quality has been tested in university research labs by professional musicians and can be favourably compared to professional high profile orchestral recordings in terms of realism and subtlety of interpretation. The goal is to make the modalities of production recede in the background so that the listener can concentrate on the music itself in all its orchestral opulence.
All OrchSim recordings are produced in true multi-track format, enabling the selection of each instrument individually that can be heard from their specific location on stage and in the acoustics of the hall. The recordings are made in 24-bit audio format and then converted to the .OPL proprietary format that can be played by the OrchPlay software. Any instrumental combination can be selected and the balance can be modulated at will.
We really do our best to provide the next best thing to your own symphony orchestra and concert hall! … but maybe without - for the moment at least - the charming conversation with the maestro to voice your wishes… but we are working on that!
The OrchPlay Music Library currently comprises 100 pieces (from short excerpts to full movements) from the symphonic repertoire ranging from late 18th century to the present. We are planning more and more full movements and be able to offer a representative repertoire of close to 400 significant works within a few years.
6- The OrchPlay Music Library users:
-TEACHERS, STUDENTS, RESEARCHERS, MUSIC LOVERS
Orchestration teachers, students, researchers, sound engineers as well as music lovers will find here a tool of choice for pedagogy, research and symphonic music appreciation. The OrchPlay software is simple enough to be operated by children at primary school level but sophisticated enough to suit the needs of university researchers.
As the repertoire expands, the OrchPlay Music Library will become a very useful tool for performers as well. One will be able to record directly a performance (in several versions if desired) and then listen to it in the full orchestral context… and - why not? - send a copy to a teacher! OrchPlay as the new "classical karaoke" or worthy descendant of the famous "Music Minus One"!
OrchSim ties together in a coherent whole some of the best virtual instruments with the expertise of professional performers to propose a hyper-realistic rendition of symphonic works. The OrchSim team has developed this system in the last 7 years testing each step regularly with focus groups of listeners (music students at various level, professional musicians as well as sound recording specialists).
1- Integration, creation and harmonization of all virtual instruments, creating a large symphonic orchestra from the best available sample banks.
- The first step was to make sure that every virtual instrument (integrated or created from scratch) "behaves" like the acoustic instrument it is modelled on. This involves much more than simply selecting a sample bank with some accompanying scripts. We constructed a comprehensive taxonomy of all modes of playing incorporating their interactions. We had to "normalize" and "harmonize" a considerable number of individual samples in order to obtain better control of each of their parameters. This also implied the normalization of the different response curves of sample groups (sampled modes of playing) according to tessitura, dynamics and articulation. Each instrument of every instrumental family and section could then have a specific sound palette where all playing modes are balanced and enable very swift and convincing alternations from one to the other as well as complex combinations and variations over time.
- Relative volumes and dynamic ranges of each virtual instrument also had to match those of the acoustical models. This enabled the balance of each instrument within its family.
- We then balanced all instrumental families against each other according to the orchestral model, taking into consideration the spatial disposition.
- Relying on the structure and possibilities of the MIDI transmission protocol as well as the structure of our sample engine, we prepared a global mapping of all modes of playing and their variations for all instruments in order to enable their control in real time with minimal latency to ensure a very realistic orchestral rendering. Many basic "instrumental behaviours" were then scripted into primary routines accessible via meta-tools.
2- Establishing a link between virtual instruments and symbolic notation
- Note Entry: OrchSim assigns a specific command to our sample engine for every symbol (text of graphic) a musician can encounter in a score. We developed a comprehensive classification of all music symbols in order to provide logical categories corresponding to the experience of the performer. These categories were integrated into our music notation program and adapted to its own formatting.
- Interpretation of a musical score: OrchSim was then enriched with a whole new series of symbols (non-printable) to allow - as orchestral musicians would do - the modulation of most parameters of the "primary" information transmitted by the score into a real instrumental performance. The possibilities range from the placement of agogical and metrical stresses, the duration and variation of timbre and dynamic over the course of held notes to the control of the speed of attack, of parasitic noise, the position and pressure of the bow for the strings, the control of the "cuivré" and sordino aperture for the brasses, the vibrato as well as wind instruments key noises up to the choice of beaters and their strike position on cymbals, among many others.
Rendering of music scores and validation of the results
We first produced a series of renderings of representative works of the repertoire, from the classical to modern eras. Each excerpt was modelled on a selection of reference recordings and each individual part on the performance of a professional performer. As OrchSim developed, these renderings were gradually reworked and improved in search of higher quality and more convincing realism.
Over the years we submitted these renderings to many professional colleagues (composers, researchers, performers, sound engineers) for critical appreciation and collected their comments. Their feedback has been of prime importance for the improvement of what is now a hyper-realistic rendering process as well as the optimization of the notation system to efficiently control many aspects of the interpretation from a digitized music score.