Martin Russ looks at some of QuickTime 3's new features, and considers why the 'computer‑as‑complete‑recording‑studio' approach isn't always the best one... believe his luck.
I've mentioned software synthesizers a few times recently; in combination with a direct‑to‑hard‑disk recorder/MIDI sequencer, they dangle the tantalising possibility of a complete studio in a computer before many hi‑tech musicians. If you intend going this route, however, you will immediately require additional hardware on top of the computer, because for any serious work, you will need more than the computer's basic stereo audio ins and outs, which means buying at least a multi‑I/O PCI card, and perhaps even separate D‑A/A‑D converters to talk to the card. And then you'll probably still need a master keyboard. But even discounting these significant objections, perhaps you've still got the unshakable notion that you can replace everything with a computer?
Let's look at this idea more closely. Modern MIDI + Audio sequencers double as sample replay devices, so the only limit is the number of audio channels your computer can handle at once — and with a fast G3 processor you should be able to run quite a few simultaneously. But don't forget the software synthesis you need to produce sounds of your own, as opposed to pre‑packaged samples off a CD‑ROM — that will eat up quite a bit of processing power. Audio playback, especially in near‑real time, takes a huge amount of processing power all the time it is running — almost the exact opposite of a word processor or spreadsheet, where the computer is mostly twiddling its digital thumbs waiting for the next input from the keyboard. Worse still, although you might still get double‑figure polyphony in playback with this type of all‑in‑one setup, there's another complication, and it is all to do with the way computers work.
Have you ever wondered why it is that MIDI + Audio sequencers allow you to change the relationship between the audio playback timing and the MIDI timing? It's because it takes a finite amount of time for the audio information to come off the hard disk and appear at the audio output — and this time is dependent on factors like the hard disk, the processor speed, number of buffers, and so on. It all adds up to two things: one, you need to adjust your audio so that it matches the MIDI playback timing, and two, because of the time delay, you need to load the audio samples off the disk before they are played back. This has serious consequences when you try to play along with your backing in real time — because any delays in the software synthesizer or the sample replayer are going to affect notes from the master keyboard. Playing ahead of time is easy for computers, because they know what's coming up, but for most keyboard players it is a hard skill to master! Typical time delays can easily reach and pass the limit of human timing perception — about 10 milliseconds — and then you can start to hear the effect.
Optimists might comment at this point that we now have a software‑only solution, because if you can't play a keyboard live, then you might as well build up all of the playing from the on‑screen keyboard instead. Personally, though, I'm not the greatest fan of using a mouse for live note entry, and for solos, chords and improvisation, on‑screen note entry is simply not an alternative. Software synthesizers are very powerful tools for making sounds which would be impractical or impossible using analogue (or even digital) gear, but their place is as part of the synthesist's off‑line sound programming arsenal, not a replacement for live reality. And it's a braverman than I who takes a computer on stage with no way of recovering quickly from a crash...
You may think that faster processors will solve these problems, but this is not necessarily the case. Computers are built to respond quickly to events that are important to them — mouse clicks and other user interface functions, for example, have a powerful effect on the user's perception of the computer's speed, and these functions are usually assigned priority. The redrawing of on‑screen graphics is often high on the priority list, yet for real‑time music use it is probably one of the least important. Although the overall time taken up by these functions will decrease as processors get faster, the problem of priority will remain unaffected. It is possible to get into the inner workings of the operating system and rework it (and the Mac apparently has some advantages over the PC in this respect), but then you can end up with a computer whose real‑time performance is fine for the things that matter to a musician, like audio, but whose response to mouse clicks or keyboard commands is a little jerky at best.
Furthermore, computer operating systems are becoming increasingly sophisticated, which means that they perform more and more background tasks that you may not be aware of. Faster processors like the G3 and its successors are going to make Macs run ever faster, but you can be sure that there is plenty of operating system software ready to use up almost all of that power.
The upshot of all this is that real‑time applications and current general‑purpose computers aren't a good marriage — someone loses, and it is almost always the real‑time stuff. Dedicated hardware may be more expensive to develop and sell, but at least it is optimised for real‑time performance.
Quicktime 3
Everything that was promised for previous versions of QuickTime has finally arrived in version 3: true cross‑platform equality, broad importing options, video effects, customisable streaming for web delivery, and more. The downside is that there are two versions: the freeware replay‑only one, and the $30 Pro version which lets you do editing. Unlike previous versions, Movie Player now only plays — it does not let you assemble movies.
In addition to the video aspects of QuickTime 3, it is easy to forget that it is also a powerful tool for audio distribution. Adding a sound track to a movie (or a panorama) requires wide file‑format compatibility, and QT3 can import AIFF, WAV, AU, MPEG Layer 2 and now Digidesign's Sound Designer II files as well. The catch with the MPEG Layer 2 audio is that QuickTime 3 for Windows won't have this for a couple of months. Then there's the MIDI playback capabilities — and this is an often overlooked way of adding music to moving images. QuickTime 2 had a 0.5Mb QT Musical Instruments file containing a minimalistic set of General MIDI sounds licensed from Roland. QuickTime 3 has a much improved 2Mb GM set that also offers Roland GS MIDI compatibility, and makes QuickTime 3's MIDI playback overall rather less 'music‑by‑numbers' and a lot more creative.
QuickTime 3 also allows you to discover some of the joys and perils of software sound synthesis without spending a fortune (OK, you might have to splash out on $30 for QuickTime 3 Pro). All you need is QuickTime (just about any version) and the QTMA Atomic Editor. It's a kind of freeware version of the rather more primitive synth editor that was available in the Developer Version of QuickTime 2 and 2.5.
Charles Wiltgen's QuickTime support and resources pages (address below) have been a rich source of information for some time, and should definitely be your first stop after downloading QT3 and some demos from Apple's site. In the software area there's a very useful techie utility for OS installer fans, whilst David Van Brink's QTMA Atomic Editor is likely to appeal more to musical readers of this column. It allows you to create your own sounds for playback via the QuickTime Music Architecture (which is what the QTMA stands for). Since this is a beta version, there's no documentation, just the program, and it can take a while to figure out what is happening. You can edit sounds and create your own, and by importing audio samples you can make your own custom instruments for QuickTime, which can then be saved as a System Extension. Editing is of the classic single‑oscillator S&S (Sample and Synthesis) type, with a volume envelope, two additional envelopes, two modulation LFOs, and a resonant low‑pass filter.
W www.quicktimefaq.org/software/
Making Movies
Russell Clarke's Play It Cool 3 is a $15 shareware QuickTime and QuickTime VR movie player and simple movie editor that allows you to make stand‑alone self‑playing applications from QuickTime movies (which normally default to using Apple's Movie Player). The resulting application is double‑clickable, blanks out the whole screen to a background colour of your choice, fades smoothly between up to six pages of title text that you define, and then plays your movie. On my test movie, the resulting application was only a few hundred bytes bigger than the original 1.9Mb movie. You may be able to think of one or two ways of using this for promoting yourself, your music, or your band (did anyone say showreel, CV, business card, promo video...?) Nicely implemented and very useful. In a word: it's cool!
www.kagi.com/rwc/pic.html (365K)