Two days ago Ars Technica noted a new published report on why FireWire is doomed. And all I can say is, good riddance.
Look, I know the technical arguments. I know that FireWire uses less CPU time because it enables direct memory access. I know that it has faster sustained throughput than USB2, and that it carries more power to its devices. I know that it allows smarter daisy-chaining, since each connection can negotiate its own bandwidth needs and transfer details. Thanks to this fascinating page, I know that FireWire has a better connector design (at least in its original 6-pin plug) than USB does.
I know all of this, and I don't care. FireWire is an unholy pain.
If you hotplug a Firewire device, it can destroy either the computer or the device ("rendered permanently inoperable" is a nice turn of phrase). FireWire's power requirements are high enough that most laptops (even those that include the full-sized 6-pin port) will not provide bus-power while running on batteries, or even at all, requiring users to carry an extra power brick for each device. FireWire devices here at work sometimes conflict with each other--one editing rig in our studios will only recognize the second external hard drive if we first unplug the DV-CAM deck. Firewire is prohibitively expensive, in part because Steve Jobs got greedy about licensing fees. The cards that aren't expensive tend to be flaky when used for pro applications. The FW800 and FW400 standards use completely different connectors, meaning that we have to buy adapters and new cables to use them. And while the USB connector is not a great design, both its host and device sides are still better designed than the flimsy and shallow 4-pin FireWire plug (the little one that most PC laptops accept).
I suspect that it may survive in the audio and video markets for a while--it's still the best way that I know of to get media in and out of a computer without using PCI cards. I'm stuck with it for a while. But I can't wait for someone to come up with a better solution.
Fun fact from Wikipedia: The standard FireWire connector is based on the Gameboy Multilink cable, since it had proven reliable, solid, easy to use, and childproof.
Have you got a fever? And the only prescription...
...is more cowbell?
Rad Monkey Cowbells has the cure for what ails you. Not only do they offer the first ever electric cowbell, but they've got a preview up for the VLC800, which uses digital modeling to emulate 12 classic cowbell sounds. If only Gene Frenkle had lived to see this day.
(Yes, it is a joke. The same people created the Sonicfinger plugins, including "Virtual Studio Visitor" and the "Dead Quietenator.")
As a follow-up to my earlier post on how music companies should be selling (and we should be listening to) higher resolution, uncompressed recordings, CDM recently mentioned Korg's brand-new one-bit recorders. It sounds silly, but basically instead of running a set of filters to get a full multi-byte description of the waveform's state, these sample the waveform millions of times a second, checking only to see if it has gone up or down. The advantage is that they don't require filtering for noise that results from the Nyquist theory, which states that sampling may produce sine-wave artifacts at frequencies higher than 1/2 the sampling rate (thus the reason that CDs are set at 44.1KHz rates, which is slightly more than twice the 22KHz boundary of most human hearing). Instead, a one-bit digital-analog conversion is turned straight into voltage changes, for a theoretically cleaner sound--although they are vulnerable to extremely high frequency noise, well past the limits of perception but enough to mess with some older equipment.
Korg has a nice intro paper online to explain this in a little more detail, and to give context: they're basically selling these recorders as ways to hold onto mastered content in a completely lossless format. Sound on Sound reviewed the units in this month's issue, and they were impressed with them, although the mic preamps are apparently weak. I'm also unclear on why they're selling one of these in an iPod-style form factor, but I'm strangely tempted by them. Apparently you can get a whole 22 minutes of incredibly faithful audio per gigabyte of storage with one of these. I feel more exclusive just thinking about it.
Sound is one of the late senses. Only smell is slower, but its reactions often have a primal immediacy that belies its leisurely spread. Touch and taste are obviously close at hand (or tongue), and vision arrives with the speed of light. But sound takes its sweet time sauntering along, maybe stopping to radiate and reflect before it finally shows up, unashamed, and monopolizes the bean dip.
I remember that the first time I'd really seen with my own eyes how slowly sound travels (relatively speaking, of course) was during a summer camp in Indiana, watching kids in the batting cages from a half-mile off. It was close enough to see a batter hit the ball, but the crack of the bat wouldn't arrive for another half-second. Most of the time, we don't really pay attention to the lag involved in travelling sound waves, because the distances for our interactions are so short. But sometimes--say at that park, or when the drone of an airplane lags behind it in the sky--we can't help but notice.
Of course, no matter how tardy everyday sound can be, it only gets worse when you run it through a computer. The lag between input and output on a computer audio interface is called latency, and it's measured in milliseconds. Digital audio enthusiasts trade latency numbers the way car enthusiasts swap gear ratios. The lower latency, the more responsive the interface can be, and the less it will throw off a musician's timing. The closest analogy I can think of, for non-musicians, is actually playing a video game online, where the inputs are delayed slightly (I think they compensate for this on the client-side now, but I remember it from Quake). With small amounts of lag, the player might not even notice. As the lag increases, the game starts to feel a little "floaty" and players have to mentally compensate. When the lag becomes too high, keypresses become disconnected from the action onscreen, and it's impossible to play effectively. The amount of latency that can be tolerated in a game or a DAW varies from person to person, and musically it also depends on the instrument: some instruments have almost no natural latency (drums), while others involve physical mechanisms or note "bloom" (piano, bass guitar) and their players tend toward greater tolerance.
My Tascam US-122, one of the first USB audio interfaces, has pretty terrible latency. That's not why I'm selling it, but it's been noticeable and much more than newer interfaces. I get just under 30ms, compared to reports of less than 8ms with the Line 6 Toneports. 30ms is still not very much--.03 seconds---and it's enough to throw me off when I'm singing, although I don't notice it at all when playing bass. Like all modern computer soundcards, you can route the signal directly through to the outputs for almost-zero latency, but then you don't get the effects processing, and that's one of the main reasons I work with computer audio in the first place.
With all that said, I want to address a misconception that I hear from a lot of musicians with regard to computer-based effects. Many people say that even the slightest latencies throw them off. I've heard people say that even a 6ms delay is enough to disrupt the groove's timing. And frankly, it's all in their head. It's a form of snobbery to be able to say that you can hear such a short delay, one aimed at people like me who can comfortably play at much higher latencies (must be something wrong with us!). Here's why:
Remember the batting cages? As that example shows, sound moves through the air relatively slowly compared to the speed of light. In fact, when you work out the math, it travels about a foot for each millisecond. This is the physical latency of sound--for each foot from your amplifier or sound source, add a millisecond of latency. In other words, someone who complains about six milliseconds of latency is basically claiming that if they moved six feet further away from their amplifier, they'd be unable to play. Given the number of wireless systems and long cables typically used by rock musicians, I doubt that they could actually tell the difference in a blind test. It just sounds cool to say that your rhythm is so tight that even a .006 second difference throws it off.
Latency--physical or digital--is simply a natural part of musical performance. In an orchestra, performers need the conductor because they might be 40ms of distance apart. From the audience's perspective, they are all in time, but adjustments must be made from instrument to instrument in order to preserve that perspective. But musicians are nothing if not distrustful of technology, a perspective that Autotune and Pro Tools have done nothing to change. Digital latency is just an easy boogeyman. The next time someone makes a claim about their delicate timing, you might ask them to literally take a step back from that opinion.
In general, I think I've been more than fair to Digidesign, the people behind Pro Tools. I like their software, and I put up with its little oddities. But the upgrade process--that's got to improve, guys.
We ordered an upgrade to Pro Tools 7.3 a while back, since the Bank's 002 Factory rig came in just slightly before the upgrade grace period. After it managed to make its way from the distributor, through the security screening, and up to our office, I went to install it on the rig but found that they had sent us disks for 7.1.1 instead. A little piqued, I called them up to figure out the problem.
There are no disks for 7.3, said the distributor. You have to register the old upgrade, and then they'll send you an e-mail qualifying you for the new version, including the download link and registration key. They can't even send us a physical copy for our backups--I had to burn it from the web installer onto a CD-R myself. Classy.
I don't know who is more annoying here: Digi for putting a major distributor in this situation, or the distributor for failing to let us know about any of this when we placed the order.
A lot of midrange audio interfaces will boast lots of inputs, but then a closer examination shows that those are S/PDIF, a digital audio protocol that uses an RCA jack, but has to be connected to another digital source. For example, my M-Audio Firewire Solo is technically 4 ins and 4 outs, but 2 channels each way are run over the S/PDIF jacks. I just ignored those for a long time, since I don't record from anything with a digital output and I figured I never would. Standalone analog-to-digital converters can be extremely expensive, and I thought I was more likely to upgrade the interface than to buy a $1,500 preamp just for an extra input or two.
With that said, BSW is selling the now-discontinued Presonus Digitube, which can transmit 24-bit S/PDIF signals and includes a semi-parametric EQ, for $89, which strikes me as a pretty good deal. If you've got a small project studio and you'd like to add another input without having to replace your interface, this looks like a decent way to do it. I'm holding off until I have more to record than just myself, but it's tempting.
One of my coworkers was heading back to the Dominican Republic, and he saw the USB/MIDI keyboard I had on my desk. Apparently, good hardware is hard to come by down there, and he asked to buy it. It was near the holidays, so I sold it to him for the price plus about ten dollars. I'm not a good keyboardist, but I like to monkey around with samplers and synths, and I've gotten used to running Pro Tools with physical controls. So today I finally got around to getting a replacement.
The Axiom 25's a nice keyboard. I liked the little drum pads in the top-right corner, especially. But ah! No sooner do I get everything unpacked, but the drivers refuse to install. It seems that M-Audio has decided not to support anything less than Windows XP, even though the driver model for Windows 2000 should be entirely identical. Back to the store it went.
This isn't the first time that audio hardware's refused to work with the older operating systems that I keep around the house. Presonus's Inspire 1394, my first choice for a Firewire interface, is likewise incompatible--clearly it's not a problem with IEEE 1394 audio, since the M-Audio Firewire Solo works just fine. Why couldn't they write a Win2K driver, since it's practically the same kernel? Who knows?
I'll tell you what I do know: it gets really old listening to Guitar Center employees repeatedly comment "Maybe it's time for an upgrade" when I try to find a working exchange. Yeah, I really want to spend a lot of extra money on a completely new operating system with increased system requirements just to fumble around with MIDI--a technology that has existed since 1983, and has been run over USB practically from the protocol's creation. There's really no excuse for it not to work on almost everything.
There's every reason that musicians should be able to make digital music on the cheap. I used to record on the laptop where I'm typing this now--a 366 Celeron--because just streaming audio with a few plugins doesn't cause a lot of stress. Theoretically, USB takes more overhead than Firewire, but I could never tell a difference. And as I've pointed out, most entry- to mid-level hardware comes bundled with a decent sequencer. It's depressing to hear that relatively simple controllers and interfaces are becoming harder to use without upgrades.
Create Digital Music links to DSMIDIWiFi, a homebrew application that lets the DS act as either a MIDI controller or a synth over the wireless link. For some people this will be genuinely useful, and of course it's always nice to have something else in your bag of tricks. If nothing else, I can see using this to play with plugin parameters in Ableton or Phrazor, and the DS is actually small enough it could be disassembled and integrated into an instrument or another casing. DSMIDIWiFi also has the ability to add pitch-bend just by wiggling the stylus, which could be really expressive. On the other hand, I don't think it'll be challenging M-Audio's new wireless MIDI keyboard anytime soon--it's hard to play notes polyphonically on a DS, and there's no velocity sense capability. So for experimental musicians--or live musicians looking for an extra, non-traditional interface--this software will make sense. For most people, it'll be a novelty, and not much else.
Either way, the YouTube video that Tob, one of the DSMIDIWiFi programmers and creator of NitroTracker, put together to show off the program is pretty charming. In the future, all software should be demo'd by genial Scandinavians.
Ableton and Steinberg (makers of Cubase) seem to be the only audio companies that really understand the power of bundling. Buy any USB or Firewire interface or MIDI controller, and chances are you'll get either Ableton Live Lite or Cubase LE with it. And of course, both companies have a special upgrade pricing from those packages, saving $100-200 over the regular boxed edition. It's a good way to hook customers in. I've been tempted, although nowadays I'd be more likely to pick up Pro Tools M-Powered than Cubase or Live, just for compatibility and familiarity from my work environment.
Until I break down and actually buy a full DAW package, I'm working with the cheap stuff. Live Lite and Cubase LE are both good programs, each with its own limitations. Cubase LE doesn't allow MIDI plugins or use the latest VST plugin technology, but it does give users up to 48 audio and 64 MIDI tracks with 8 inputs/outputs, so normally it's my tool of choice. The fact that it's more traditional-looking and runs on lower-end hardware doesn't hurt. But unfortunately, due to poor impulse control the other day, I've temporarily uninstalled my copy, and it'll stay that way until I can find my old install CD or pull the files over from my other laptop.
So that leaves Ableton, recently upgraded to version 6. Live Lite only gives users four tracks, and won't let you use more than two of its proprietary plugins (called Devices) and one VST plugin at a time (between all four tracks). You can do a lot with four tracks, and Live Lite still boasts better routing and (surprisingly) a better GUI than Cubase. But the limitations on Devices and VSTs have been driving me crazy. I can get around the VST limitation by using Phrazor or another sub-host, but just adding EQ to vocal and bass tracks puts me up against the Device limit.
After some experimentation, here's a trick to get around the two-Device limitation without upgrading to the full version: instead of adding Devices one at a time to the track, create a Group out of the first plugin, then drag new Devices into that Group. Ableton counts a Group as a single Device instead of counting the individual elements, probably so they can include the enticing Group presets that they hope will incite you to buy the full version. You can drag Devices back and forth inside the Group to change their order, but be warned: once you've added something, trying to delete it again may cause Live to flash the nagware limitation window and cancel the deletion, although you can still delete the whole group and start over. Once you've got the effects stacked the way you want them, it might help to bounce the wet audio to a new track, freeing up the Device Group for more mixing. Treat it like one of the old four-track recorders, in other words.
I still think Cubase LE is the better bundle, and with the Tascam US-122 currently only $150 at Musician's Friend, it's hard to argue that it's not a bargain as a first interface. But with M-Audio's lower-priced boxes ruling the $200 range (and offering a Pro Tools M-Powered upgrade path), it's also easy to see why songwriters on a budget can end up with Live Lite as their main recording software. Being able to stack Devices with this trick goes a long way to making it a more useable solution, in my opinion.
Since Mile Zero is now the #11 on a Google search for "virtual pedalboard" (much to my dismay when I first started looking into using a laptop for my bass effects rig), and since I've had a few searches for information after first writing about it, I'd like to go into depth about the final results. As I've said, I'm not planning on using the laptop live for a variety of reasons, but it does make a fine recording setup.
As usual, I've worked on doing this on the cheap. There are plenty of ways to spend a lot of money for a laptop effects rig--Native Instruments makes Guitar Rig expressly for that purpose--but for many hobbyists like myself, spending a lot of money on that software seems extravagant, and might not go over well with significant others. I also see it as a personal challenge to do more with less.
I did end up purchasing one piece of software to run the pedalboard: Phrazor, which is technically a synthesizer workstation. Phrazor is built to let keyboardists and sequencer-based musicians easily run complex sets of virtual instruments and effects live, but it also hosts audio-based plugins. That makes it versatile enough to adapt for our purposes.
Phrazor provides 64 "Tracks," which are basically 8-slot effect racks. They can be chained or routed to other tracks, before being sent out through the audio channel. My pedalboard project contains one track for effects and another track for the Mobius looper plugin--the mix from the first track is sent to the second, so that the looper gets the complete effects mix.
Each track contains its own mixer. Click here for a screenshot of the effects track routing. There are two busses through the mixer, A and B. Each plugin receives an identical dry signal from the track's input (which can come from external or internal routing) along the A bus. The B bus is wired in series--it gets signal from a previous plugin, in a chain. So you use the A bus to mix multiple effects simultaneously, while the B bus feeds them into each other.
I realize that's tough to wrap your head around, so let's walk through the screenshot by way of illustration. Three effects--the compressor, Tube Screamer, and the Marshall Amp--feed into the B bus. Those effects are meant to be chained together. So when I activate the fuzz preset, my bass signal feeds first into the compressor. From there, it's sent to all plugins along the B bus, but most of them are muted. The first unmuted plugin is the Fuzz Plus, so I get my low-end distortion from there. At the same time, the signal also passes from the compressor into the Tube Screamer, which is muted but hooked into the signal chain (see the yellow arrow, indicating that its output is sent to the next plugin along bus B). The Tube Screamer sends to the Marshall, also muted and in series, before it finally emerges through the unmuted low-pass Filter. Because the last set of plugins have their A inputs muted, they get only wet signal from the previous plugin, and no dry signal from the main track Input.
By controlling the mixer's muting and signal flow, I can control which effects are heard and which ones are effectively bypassed. For my envelope pedal, for example, I mute everything except the GreenMachine Wah plugin. Phrazor stores the mixer state--along with any saved plugin presets--in the Track states to the lower right. It switches between those Track states in response to MIDI notes, which I trigger from the floor pedal (click here to see a screenshot of the remote states view). Technically, my floor pedal only sends program change messages over MIDI, but I use MIDI-OX to translate those into note messages for this purpose.
So that's effectively how the pedalboard itself works, but it doesn't explain how to record it. I have two audio workstations that came bundled with my recording hardware. The Tascam US-122 audio interface came with Cubase, which normally I prefer. Unfortunately, Cubase won't record post-effects, so I can't use it for this project. Instead, I host the Phrazor pedalboard as a VST inside of Ableton Live Lite, which came with my M-Audio O2 keyboard.
It barely works: Live Lite 5 only supports four tracks, and Phrazor needs three to run the pedalboard. Click here to see a screenshot of Live set up to record. "1 Audio" receives the bass signal from the interface and hosts Phrazor. In order to get MIDI messages to the plugin, we need a second track--"2 MIDI"--that passes input on to the first track. See how "1 Audio" has its Audio To control set to "Sends Only?" That allows it to feed to another track for recording ("4 Audio") instead of going to the Master output. If we recorded on "1 Audio" we would only be getting the dry signal run through the current Phrazor preset. Recording the MIDI signals as well would theoretically play back the control messages with the audio, but it's easier and more reliable to mix the output to an audio track during the song. The last audio track, "3 Audio," records directly from the microphone for vocals.
I've recorded a low-quality .mp3 of myself using the virtual pedalboard so that you can hear the results of all of this. You can click here to hear a walkthough of my effects board, including a breakdown of the elaborate distortion effect I wrote about in this post, and a familiar tune looped to show off how it all works in practice.