Home -> Magazines -> Issues -> Articles in this issue -> View
Article Group: | |
The Gentle Art Of Transcription (Part 1) | |
Setting the SceneArticle from Electronics & Music Maker, April 1984 |
Part one of a survey of music-printing instruments.
David Ellis takes a look at the various transcribing devices that currently exist, from Yamaha's MP1 personal keyboard to alpha Syntauri's Apple-based system: with more to follow next month...
These days, just about everyone's doing it... transcribing music, I mean. It's not difficult to see why. Music notation has a splendid sort of arcane quality to it that evokes glazed expressions from those that don't understand it, and a certain awestruck regard from those that do. But let's be honest, there's an awful lot of re-writing, conscious or otherwise, that goes on in getting through to the real musical substance of many a complex score. And I'm guilty of that transgression myself.
So, wouldn't a more commonsense way of going about creativity be to work from the opposite direction, working tried and tested instrumental lines into the full score? Stravinsky did it, after all. Of course, most composers aren't multi-instrumentalists, and Stravinsky's efforts in the direction of pre-performing his compositions were generally restricted to piano and percussion parts. However, in the case of rock (or whatever) music, many musicians are multi-instrumentalists, so there's no reason why, rather than writing the score first (an ability which only a handful of rock musicians seem to develop - quite probably because it kills the improvisatory spirit), the starting point shouldn't be a performance, or at least a step-time equivalent derived from 'keying-in' of MCL, that's then transcribed by the computer for storing on cassette, disk, or in RAM. Once that's finished, the 'soft' score might then be subjected to editing, performance, or printing out as a 'real' score which even the most diehard of session musicians could appreciate, if not play.
You could argue that transcribing a musical performance into real printed notation is gilding the lily at best, superfluous at worst, and generally of limited value to anyone in the commercial field. But there are many instances where a printed version of a piece of music that only exists in the mind of the performers or on record would be extremely useful:
1 Copyright reasons: The Performing Rights Society in the UK encourage composers to write down the lead line of every song they have published. For greater copyright protection, the entire transcribed score can be kept locked away in the PRS's vaults.
2 Arranging: String and brass parts can be produced without recourse to outside arrangers. 'Intelligent' transcription might provide facilities for automatic transposition of parts, harmonising on the basis of other parts, and error-trapping that prevents lines from being written outside the natural range of the instruments.
3 Film scoring: Ditto orchestral parts for use with or without pre-existing synthesised parts. Film work frequently involves having to make changes at the last minute to accommodate an added or deleted split-second or two. With the 'soft' score facility, this becomes straightforward, as the synthesised parts can be edited in the normal sequencer fashion and a revised set of orchestral parts then printed out as and where necessary.
4 Educational purposes: Showing a musician what his keyboard performance looks like in notational actuality.
The problem with putting performance before scoring is that the interaction between eyes and ears stands a good chance of being relegated to the 4th Division; no MCL or steptime display is yet capable of showing all the vertical relationships that might exist in a piece of multi-part music with the same clarity as notation on manuscript paper. What this means is that the ears get the lion's share of sorting out the harmonic wood from the melodic trees. There's no harm in that, of course, but there's also no denying that people in general are not that good at remembering what tune was played five minutes ago. What we really need is a modicum of visual help applied in the direction of our meagre efforts at musical memory.
The first and most obvious place to look is at the screen display of the computer that has our score under lock and key. However, as things stand at present, high-resolution graphics just can't do justice to more than a handful of musical parts. In short, we run out of pixel power. Of course, you can head in the direction of Xerox's Mockingbird (see later in the series) for much classier musical graphics, but the computer that it runs on (the Xerox 1132) doesn't exactly grow on trees, and it isn't exactly a Sunday afternoon programming session to get a screenful of music looking anything like as good as what Mockingbird produces. On top of that, Mockingbird requires financial resources that zoom up into the six-figure region...
There are really two potential solutions to this predicament: first, use screen displays for editing the score part-by-part and a high-quality printer to allow visualisation of all the parts together; and second, start investigating easier and more efficient ways of getting parts together on graphics displays of limited resolution. The second option is obviously the more exciting, but it's somewhat difficult to know where to start. You could, for instance, have a go at translating traditional notation into a format that's suitable for the average graphics display by using Equitone, the alternative notation that was first proposed in 1958.
The obvious advantages of Equitone over traditional notation include a reduction in the number of lines (these occur only at intervals of an octave), a general saving of space, the eradication of tricky (for micros) note beaming, the ease with which multiple parts can be indicated on one stave, and the potential of turning the notation into a simple shorthand that could be used as the basis of a QWERTY keyboard MCL. And for greater visual effect, colour could be used to differentiate parts further. There are problems, too, but the beauty of Equitone is that it retains the same space and time axes of traditional notation. In fact, as the above example shows, if anything it makes them clearer. The fact that the traditional head-and-tail notation of duration is abandoned in favour of the horizontal position between beats and bar-lines, and that pitches are derived from their vertical position between octave lines, is surely a further point in its favour if, like me, you view the restrictions of conventional notation with some frustration.
Well, that's something of a personal hobbyhorse, and you'll not be surprised to know that an Equitone MCL is something that's processing away in the cavernous recesses ot my mind for a future issue, but for the more conceivable present, it looks as if we'll be obliged to tag along with the principles of conventional stave notation if we want to peruse all that makes a complex piece of music tick. And even supposing that we did have megabyte graphics displays readily available, or an ideal sort of alternative musical graphics, you'd still lose out on that useful manuscript paper feature of being able to flick from one page to another, compare parts, and generally inwardly digest what's going on over a long time-span. Of course, word processor users get accustomed to scrolling through text to check paragraphs and the overall feel of a piece, but you're still basically reliant on good, old-fashioned hard copy for the proof of the pudding. Plus, if we're being honest about it, there's something rather satisfying about seeing the fruits of your hard labour in the form of a printout. Text stored on a disk just doesn't seem the same.
The point about printing out musical graphics is that the formatting of the printed page can include several screens' worth of staves, and if you've access to an X-Y plotter like the Roland DXY-100 or one of the upmarket Hewlett-Packard jobs, then the sky's virtually the limit as far as quality and putting on extra bits and bobs is concerned. So, it's to some sort of printout rather than a display that those wanting lots of parts with the added luxury of a healthy tangible quality are obliged to turn, and what follows is a description of some of the transcribing options available on (in the broadest sense) computer music systems as diverse as the Yamaha MP1, Soundchaser Notewriter, Syntauri Composer's Assistant, Con Brio ADS-200, Fairlight CMI, Synclavier II, and, last but far from least, Xerox Mockingbird.
At just £480 or thereabouts, the MP1 is undoubtedly the cheapest way of getting into the transcription act, as it provides the wherewithal for turning a keyboard performance into monophonic printed lines of music complete with chord names, bar numbers, and so on. In fact, it's the sort of thing that's ideal for the songwriter who wants a hard copy version of his tunes, and there's also plenty of space on the print-out for writing in lyrics, expletives, and so on. The printer itself is an ingenious device that's really an X-Y plotter in miniature, and it's this fact that accounts for the very high quality of the printout - much better than with any dotmatrix printer. The main limitation to what the MP1 can do is the 2.25" wide paper roll, but an all-out pasting-up operation can help to get around this in part. That doesn't mean to say that "full orchestral scores can be produced with the ease of using an electric typewriter", as Yamaha's advertising copy so blithely puts it!
Like all good and obedient machines, the MP1 provides the options of transcription in both real- and step-time. If you go for the former, it's also possible to select timing assistance in the form of a whole gamut of auto-accompaniment features (which may or may not be to your fancy). As we'll see with the Synclavier and Notewriter, one problem with real-time transcription is making sure that your notion of accurate timing coincides with that of the micro's quartz crystal when it's doing the job of analysing your efforts. The MP1 takes this human limitation into account by providing a 'sensitivity' button that quantises your notes to the nearest beat and does a fair job of making a silk purse out of a sow's ear.
The Notewriter is a $99 program that's designed to be used in conjunction with the Soundchaser keyboard, Mountain Computer MusicSystem digital synthesis hardware, and the inevitable Apple II. In principle, it looks impressive: a real-time music transcriber with notes played on the keyboard (almost) simultaneously appearing in conventional music notation on the screen.
In practice, life isn't so kind, and the main problems lie with the eagerness of the micro to register the user's timing discrepancies and the slowness of the editing side of the program. The point is that to enter notes in real-time, you're obliged to play them against a metronome which, as a bleep from the Apple's speaker, is rather less than forceful. Still, with more practice, it is possible to optimise your interaction so that you and the micro roughly agree on the essential matters of your performance. But it'd take an extremely metronomic, not to say boring, individual who'd manage to play a long lead or bass line without making a balls-up somewhere along the line, so use of the editing facilities is virtually inevitable.
In fact, the real-time side of the program is utterly flummoxed by triplets and fast runs of notes, and the editor is the only way of getting those inserted into the score. It's unfortunate, then, that the Notewriter's editor is so difficult to use. If this was a nice, helpful step-time editor, then everything would be rosy, but, as it happens, even the simplest changes to notes involve learning a command syntax that makes Wordstar look positively Samaritan-like. Still, perseverance pays oft, and the printout isn't at all bad, even if it is monophonic and lacking in a certain calligraphic finesse.
What must be borne in mind is that the Notewriter program has the potential to do much better. In fact, a number of improvements have been (or are being) introduced, including pre-selectable transcribing accuracy to get over the problem of the performer occasionally jumping the beat (or the micro getting ahead of itself, if that makes you happier...) and the option of converting Notewriter files into the format that's required by the four-track performance software. The Soundchaser people are also working on a software package called Polywriter, which should provide some pretty sophisticated polyphonic transcribing features, but, of course, that remains to be seen...
Syntauri's transcribing program for their Apple-based system costs a good deal more than Notewriter ($295 at the last count), and it's also very much a non-real-time approach to musical transcription. In fact, the way the program works is by first analysing the multitrack note files generated and stored on disk from previous real-time keyboard performances with their Metatrak software, and then creating further files on a second disk-drive that represent the program's graphical interpretation of the notes. These can then either be displayed on screen and edited, or printed out on something like an Epson MX80 printer.
The program is really quite flexible. For instance, you can select which of the 16 Metatrak tracks you want transcribed, tailor the score analysis to your own brand of playing accuracy, and add text for the sake of dynamics, lyrics, chords, and so on. However, there are limitations: first, the display can only cope with a couple of bars at a time; second, the program eschews beaming (the joining of notes into groups) in favour of all note tails heading in the same direction (upwards); third, the analysis can only cope with 1000-note segments, so you're obliged to segmentise your compositions; and fourth, all the analysed parts end up compressed onto just two staves.
But, on the plus side, both the display and printout have a nice clean quality, which, if not 100% ideally suited to performing from, is certainly adequate for copyright purposes and less involved arranging.
The overall impression I've gained from using Composer's Assistant is that it's more concerned with strutting its stuff than actually being a helpful amanuensis (secretary!). Bearing in mind that the Metatrak software offers only real-time note entry, it's sad that Composer's Assistant doesn't provide the wherewithal for editing out duff notes and conversion of the analysed and corrected note files back to the format required for performance by Metatrak. That's what I'd like to see my assistant doing anyway...
Next month, from Fairlights to Mockingbirds.
Read the next part in this series:
The Gentle Art of Transcription (Part 2)
(EMM Jun 84)
All parts in this series:
Part 1 (Viewing) | Part 2
Technically Speaking |
Lab Notes: Blessed are the Seque |
Software Support - Hints, Tips & News From The World Of Music Software |
New ROMantics - The MT guide to CD-ROM |
Computers in Music - A Cheaper Option? |
The Programmable Digital Sound Generator (Part 1) |
Atari Notes |
PC Notes |
Software Support - Hints, Tips & News From The World Of Music Software |
Building A Bionic Sax |
Step-time Sequencing - for Wasp Synthesiser and Commodore 64 |
PC Notes |
Browse by Topic:
Computer Musician
Feature by David Ellis
Previous article in this issue:
Next article in this issue:
mu:zines is the result of thousands of hours of effort, and will require many thousands more going forward to reach our goals of getting all this content online.
If you value this resource, you can support this project - it really helps!
New issues that have been donated or scanned for us this month.
All donations and support are gratefully appreciated - thank you.
Do you have any of these magazine issues?
If so, and you can donate, lend or scan them to help complete our archive, please get in touch via the Contribute page - thanks!