CUTTING EDGE

The technology behind computers is advancing at an incredible rate, but are they really stable enough tools to replace an entire studio?

Although my neighbour is a self-confessed technophobe, he recently purchased a Windows XP computer and, being visibly terrified by its arrival, asked me to help with the setting up. No problem -- or so I thought. We got the computer up and running without a problem, but a few days later the doorbell rang and it was my neighbour telling me that his PC loudspeakers were broken.

Upon inspection, the sound system was indeed completely dead and I eventually concluded that the soundcard must be faulty, even though this is a pretty rare occurance these days. It certainly didn't appear as though there was a software problem, and the only package installed since I'd last seen the system was a chess game, which I thought unlikely to kill a 5.1 Creative Labs sound system.

Screen 1: An Internet radio stream with a glitch.

Since the chess game had been the only addition to the original installation, I decided to take a closer look at the program's documentation. I was astonished to find that almost every paragraph referred to "possible audio problems" -- it seems the program makes a sound when you move the chess pieces. But really, this is hardly rocket science when you look at programs like Nuendo, Sonar and Acid.

I think problems, such as those caused by the chess game, raise the question of whether general-purpose computers are suitable substitutes for a studio stocked with dedicated audio equipment. Although this debate has been covered more than once before in SOS, the rate of advance in both hardware and software is so great that the dynamic of the discussion is constantly changing. Here's my take on it.

To begin with, hardware devices in the studio today are computers. There are exceptions, of course, but most effects processors, digital mixers, and hard disk recording devices are computers in the sense that they make use of processors running software. Perhaps the biggest difference between these devices and general-purpose desktop computers is that the dedicated units tend to have an 'embedded' operating system. This OS is tailor-made to a specific device and manages one or more general-purpose processors, designed for embedded use, and DSP chips, which do the actual music processing.

What all this means, of course, is that aside from the way the software is stored (normally in Flash RAM these days), there's little difference in the overall principle between a stand-alone studio device and a desktop computer fitted with audio I/O. This is especially true with the availability of hardware controllers to use with PC and Mac-based audio applications, and it's easy to forget that the control surface on a digital mixing desk is completely remote from the audio signal path. This means there's no topological difference between such a digital mixing device and an audio application plus external controller.

So, at what stage will the two types of device be indistinguishable -- in performance, if not necessarily appearance? Well, the above example of the mixing desk and the sequencer suggests that the two classes of device are converging rapidly, but there are still reasons why I would chose a desktop computer or a dedicated device. This would depend on the nature of the job to be done. But there is one overriding consideration: stability. Most people would agree that dedicated studio devices are more reliable than desktop computers, but why? And will this always be the case? I'd argue that to find the answer to this we don't need to look any closer at the architecture of the two classes of equipment: we need to look instead at the way people use them.

You wouldn't type a magazine article on a virtual-analogue synthesizer, and while you'd never consider trying, it would actually be impossible without a QWERTY keyboard, a suitable display and a filing system -- not to mention word processing software especially written for the platform. The input device on most keyboard instruments has 61 keys, for inputting notes rather than letters. The only software you'd run on dedicated keyboard instruments is the operating system, and the only time you'd load new software is when there are upgrades available. This software comes directly from the manufacturer, and you can be certain it will have been tested on equipment that, apart from a few knocks and scratches, will be identical to yours. You therefore know that if it worked in the manufacturer's test lab, it will work in your machine.

Compare and contrast this with the way you use a computer. Most people only have one that gets used for a variety of different tasks, which is fair enough, although it does raise the question about suitability for use in a studio. For example, if the maker of an audio-editing package doesn't know what other software will be running on your computer, how can they possibly guarantee their software will run properly? They can't, of course, and neither can they know very much about what actually comprises your PC: there's an incredible variety of motherboards out there, and an ever-growing variety of processors, for example.

Now, computers are not fundamentally unreliable. They have few mechanical parts: disk drives and fans. Hard drives are no more reliable in 'embedded' systems than in general-purpose computers, and fans can easily be replaced. I suspect that general-purpose computers would be as dependable as dedicated processing units if they were tightly specified by the creators of music software programs, and the users refrained from installing other software. This approach has worked well for Avid, whose Xpress DV software will run on a wide range of platforms, but is demonstrably (statistically) more stable on the small range of PCs that Avid has tested and ratified.

You could say that this is an argument in favour of using Macs, because of the finite range of models and possible add-ons. But it's also an argument for using any computer platform sensibly. If it works, don't touch it. Don't upgrade it or add anything to it. If you want reliability and stability, and if you have other computing jobs to do: get another computer.

I've just had an NTL broadband Internet connection installed at home, which gives me 512kbits/s via a Motorola USB cable modem. Because I already have a digital set-top box for my TV, the procedure was a simple matter of adding a splitter to the cable, with one branch going to the TV and the other to the cable modem. After only having it for a week, it's clear that always-on broadband is more than just an Internet connection -- it quickly becomes a way of life, and if you want to test the speed of your connection, broadband or dial-up, visit www.bandwidthplace.com and you'll be guided through the process.

Screen 2: Copying to another USB device on the same buss.

Burst mode uses the full bandwidth of the USB connection to transfer large amounts of data in the shortest possible time. It's a bit like closing off the M1 to allow the passage of several abnormal loads that take up all three lanes. If, for any reason, part of the data is missing at the other end, the receiving device requests that this data is re-sent, and will continue to do so until all the missing data arrives. This kind of service is obviously essential for connecting storage devices where data would be ruined if even the occasional bit was missing.

Consider the graph in screen one, which shows the bandwidth usage of an Internet radio stream that suffered a glitch (the data not arriving in time) during playback. Screen two shows what happened when I was trying to listen to the same Internet radio stream while backing up this very article to a USB-connected hard drive -- an Iomega Peerless drive, to be exact. This highlights the big problem when burst mode is used by another USB device on the same buss -- all other transfers are obliterated.

Although isochronous sounds a bit like 'synchronous', an isochronous connection works very differently to a synchronous connection like S/PDIF, for example. With an isochronous connection, a certain quantity of data is guaranteed by the system to be delivered within a certain time. Audio transfers on USB are always isochronous and, although this mode of operation provides some safeguards, it's important to note that something has to give if you overload the buss.

An interesting test would be to recieve an Internet audio stream via a USB-connected cable modem, and listen to it on USB speakers. I haven't tried this yet, but I wouldn't be surprised if there were a few problems. It's also important to remember that the 12Mbits/s data rate quoted in the USB 1.1 specifications is actually the theoretical maximum -- in practice you'd be lucky to achieve this. Hopefully these issues will become a distant memory with the wider adoption of USB 2, which is already starting to appear on motherboards from major PC component manufacturers like Asus, for example.

The Internet Saves The Radio Star     I'm completely nutty about radio, and often spend the early hours listening to a Chicago blues station or some Latin American samba. Maybe I'm not typical listener, tuning into stations like www.rfvn.com, but this highlights one of the greatest things about content delivery on the Internet: there needn't be any such thing as 'the typical listener'. Indeed, I still believe that Internet access to radio- and television-style content is likely to rival more conventional broadcast models within a few years, although it seems as if everyone got their timescales wrong about this: take the losses posted by AOL Time Warner recently, for example. Half a million of us in the UK now have broadband access, but accessing broadband content is a mess. Take RealNetworks' RealPlayer, for example, which is a perfectly good media player featuring compression formats that are improving all the time. But, rather annoyingly, it tries to become the default player for all your media content. Actually, it does more than try -- it virtually forces you, even suggesting that you should 'correct' the situation when file types are assigned to be opened by other applications. Note the screenshot below: can you see a cancel or 'no thank you' option? This dialogue box wouldn't go away until I manually deselected about 10 media formats that it would otherwise try to play when selected. There's a lot more to be said about music delivery on this Internet, but for now, I find the process frustrating and intrusive. The very measures designed to attract my custom are putting me off.

Home | Search | News | Current Issue | Digital Editions | Articles | Forum | Subscribe | Shop | Readers Ads

Advertise | Information | Links | Privacy Policy | Support