Thunderbolt From The Blue
August 4, 2011
Solutions to problems are easy to find,
The Problem's a real contribution.
What's really a challenge to wring from your mind
Is a problem to fit a solution.
Apple's new connection standard, Thunderbolt, is probably seen by most folks out there as a solution in search of a problem. I mean, do we really need yet another funny-looking socket on the sides of our computers? Especially one being pushed by Apple, of all people?
The problem is aggravated by the fact that Apple has a tendency to "bless" and promote certain standards it finds suitable for its needs, often at the expense of standards that are rapidly finding wider acceptance—read, standards that are popping up on Windows computers. The IEEE came up with the 1394 standard, fair enough, and it was the subject of scholarly papers and earnest discussion, but while it was an interesting exercise in high-speed "smart" connectivity, it didn't really have any legs at all until Apple named it "FireWire" and started putting it on all their computers.
That said, at the time, FireWire was actually a solution for a very real problem: computers needed a hot-swappable way to connect devices that moved a lot of data. FireWire did the job, and over a decade later, it's still a very popular format for things like audio interfaces... although, if you look around the rest of the world, it's vanished from everything else out there (well, except for hard disks).
Nowadays all computers have USB 2.0 on them... thank heavens Apple agreed that was a worthwhile standard to support. Ditto Ethernet. But the newest and most popular connectivity options for really high-speed data transfer on Windows PCs, USB 3.0 and eSATA, are absent from Apple machines without third-party add-ons or surgery (Other World Computing will sell you an iMac with eSATA if you don't mind them voiding the warranty and cutting a hole in your iMac). Apple's LightPeak project finally saw the light of day as Thunderbolt, amid politics and grumbling about yet another standard... and peripherals for it have been very slow to come to market. There are a couple of hard disk arrays, one Apple display/hub that's a cool grand on top of the cost of your computer, and no audio stuff in sight at all.
So: should we care? Unfortunately, yes we should. We should care a LOT.
Audio professionals have a decades-long history of wanting more from computers. From a paltry 16 channels of MIDI to multiport interfacing, from two channels of 16-bit audio to 96 channels of 24-bit, from 32 kHz sample rate to 192 kHz and maybe even beyond, we've always pushed the limits. The thing is, we're now running up against the laws of physics... there is very little justification to store data words larger than 24 bits, and none at all for anything larger than 32. When you get up to 96 kHz sampling rate and beyond, the benefits are vanishingly small to go much higher. And only the very largest studios will realistically need more than, say, 48 simultaneous inputs for a single session.
Let's do the math. 48 channels of 32-bit/192 kHz audio is a data rate, flat out, of just under 300 Megabits per second, or 37 Megabytes per second. That's about 75% of the theoretical full bandwidth of FireWire 400; in a perfect world, in other words, you could have that kind of bandwidth for an audio interface using FireWire. In practical terms, FireWire has way too much overhead, too much housekeeping to do, to sustain a rate like that on a consumer computer, and if you put any other device on the FireWire bus, you may as well give up—if 75% of your bandwidth is going to bringing audio IN, what about moving it OUT onto a storage medium? Not gonna happen.
USB 3.0 and eSATA are a lot faster than that, but Thunderbolt is faster still, with a theoretical maximum of 10 Gigabits/second... about thirty times what FireWire could do. Not only that, but Thunderbolt can power some pretty hefty peripherals. 10 Watts may not seem like much, but it's enough to properly phantom-power mics on a good-sized multichannel interface; being based on DisplayPort technology, Thunderbolt was designed from the outset to power displays, which drink way more juice than most audio devices. USB has always been an anemic standard for audio; you can't phantom-power more than two mics on a bus-powered USB 2.0 interface without getting into trouble.
One more advantage of Thunderbolt: it's based on PCIe, the standard for audio cards that are prevalent in the Windows world but require an expensive Mac Pro on the Apple side. That means at least theoretically, PCIe-based peripherals should be able to run on Thunderbolt with fairly little adaptation and similar bandwidth, giving designers a leg up on getting useful products to market in a hurry.
Audio folks have complained about how expensive the Mac Pro is. To get PCIe, you need to spend a bare minimum of $2500, and a properly loaded system can be $4000 or more. Contrast that with the Windows world, where a PCIe-capable machine is actually your cheapest option. With Thunderbolt, all that PCIe zap becomes available to the much more affordable Apple Mac mini, iMac, and the various laptops... the MacBook Air was just re-released with Thunderbolt and dual USB 2.0 ports, suddenly making the little 2-pound machine a viable computer for live audio...
...IF anyone releases Thunderbolt audio peripherals.
So far there hasn't been a peep from anyone on the subject. AES in October? NAMM in January? We'll have to see. Audio tends to have much longer product cycle time than most other applications that require specialized hardware, and there's some question as to whether there will be the right synergy of:
- Manufacturers wanting to develop for a new interface
- Developers wanting to take advantage of the increased throughput
- Users wanting to invest in yet another standard for their studios
...in order to make Thunderbolt a going concern. Would it be a very cool thing to have? I think so; a USB MIDI controller and a Thunderbolt bus happily handling audio I/O and all my data storage, all hooked up to my laptop with just a couple of cables, would be convenient and high-quality all at once.
But will it happen? We're going to have to wait and see. Thunderbolt could be the standard that finally gives audio professionals (I'm not addressing video people or folks who are doing computer graphics rendering here, just audio) all the bandwidth they realistically need to do anything. It'd be ironic if the standard didn't fly.
Am I advocating "wait and see"? Of course I am. When there is nothing to hook to that little port on your Mac, waiting is all you can do.