By Craig Anderton
New York, NY (December 30, 2005)–64 Bits is the big buzzword in computers, but like so many buzzwords, it’s really shorthand for something deeper. First things first: There are two aspects to the 64-bit thang. To start, we’ll consider computers being able to crunch 64-bit data. This is just a logical progression; the 4004 processor handled 4 bits, the 8008 8 bits, and over the years we’ve moved on to 16-bit and 32-bit processing.
There are two main advantages to dealing with data 64 bits at a time. Speed comes into play, because the computer can crunch more stuff faster. The other is ability to address greater amounts of memory. 64-bit computers can address a terabyte of RAM (1,024 Gigabytes), as opposed to the 2GB limit of 32-bit machines. Granted, that may seem like overkill, especially given that memory isn’t free (although it’s getting there, right?).
But remember that 20 years ago the idea of having a Gigabyte of RAM in your computer was preposterous, so let’s not dismiss the possibility of needing dozens (or even hundreds) of Gigabytes of RAM in the not-too-distant future. If nothing else, this will allow reading sample libraries, video clips, and the like from RAM instead of burdening a hard drive with streaming.
And as if to amplify the “perfect storm” of 64-bit operation, we now have Microsoft’s Windows x64 operating system and Apple’s OS X Tiger designed for 64-bit operation, and a new generation of dual-core processors. Dual core simply means you have a multiprocessing system (two CPUs) on a single chip, but note that not all dual cores are created equal…yet. AMD’s (www.amd.com) devices communicate on-chip, whereas Intel’s Pentium D communicate via the Front Side Bus, which is slower than communicating on-chip. But AMD’s advantage may be short-lived, as Intel’s next-generation Conroe and Woodcrest chips will do true dual-core processing. (Of course, by then AMD will have gone on to something else, as the two companies play their Zen-like game of Infinite Leapfrog; and let’s not forget Apple’s Power Mac G5 Quad, with two dual-core processors.)
Note that when running a true 64-bit system, you do need an audio interface with 64-bit drivers, otherwise the interface can’t “talk” to the OS. Fortunately, there’s progress along those lines. Edirol, E-Mu, Creative Labs, M-Audio and RME have either production drivers or stable beta versions for many of their existing interfaces.
The issue of plug-ins is a little thornier, but Cakewalk’s Sonar (www.cakewalk.com) has a workaround called BitBridge. This allows using 32-bit VST effects within Sonar’s 64-bit edition. Eventually, though, as 64-bit systems become the norm rather than exception, there’s little doubt that 64-bit versions of plug-ins will appear.
Unfortunately, though, there are some things that just don’t work yet with a 64-bit hardware/software system. The 32-bit DirectX plug-ins choke (although there’s nothing in the DirectX spec that prohibits 64-bit operation, so it shouldn’t be hard to adapt DX plug-ins), and ReWire won’t support 32-bit clients. And we’re also waiting on Microsoft, who has yet to release x64-compatible codecs for MP3 import and Windows Media Video import/export, as well as Apple, which doesn’t have an x64 version of QuickTime. Surely, though, these will follow in the months ahead.
The other part of the 64-bit equation is 64-bit internal resolution in digital audio workstations. This isn’t a new concept; programs like iZotope’s Ozone (www.izotope.com) have been using 64-bit resolution to do audio calculations for a while. But this isn’t related to 64-bit hardware systems, aside from the superficial resemblance that they both use the catchphrase, “64 bits.” A 32-bit system running a 32-bit OS can still have a 64-bit audio path, with no complications.
So why would you want a 64-bit audio path? Well, this is one situation where less is not more. With 16-bit systems, if you did too many operations–even functions as basic as level changes–round-off errors would occur due to the limited “numeric dynamic range” of a 16-bit system. Accumulate enough round-off errors, and the sound quality suffered. The 32-bit floating point was a major improvement, but 64 bits gives just that much more headroom and dynamic range. The improvement is particularly noticeable with material like reverb tails that decay into nothing, even when you really turn up the volume at the end. Those “fizzing” decays of low-res systems are gone forever.
The other big difference occurs when you’re running projects with lots of tracks, plug-ins and soft synths. That’s a lot of calculations going on at once, and 64 bits of resolution can handle it without you having to worry too much about clipping and other issues that relate to limited calculation abilities.
And, of course, we like to put all this in perspective, so…is 64 bits a huge step forward? For audio, there’s no doubt that it removes the limitations of the not-very-limited-in-the-first-place 32-bit floating-point format. As there are still passionate arguments about whether mixing “inside the box” produces satisfactory audio results, 64-bit resolution should help people feel better about living in the digital domain.
As to 64-bit hardware and operating systems, although 64-bit systems aren’t necessarily twice as fast as 32-bit systems, they are indeed faster and as far as I’m concerned, faster=better. After all, playing or recording music is–or at least, should be–a real-time experience, and your tools should help facilitate that.
Check out Craig Anderton’s legendary Sound, Studio and Stage forum at www.harmony-central.com, as well as his sample CDs: Turbulent Filth Monsters, Adrenalinn Guitars and Technoid Guitars. For more information, go to www.craiganderton.com.