An OQO in the hand... Isn't this the portable computer you've always dreamed about?

Who wants a studio in a box when you could have a studio in your pocket?

When I wrote about the Sony Vaio C1XN, a computer small enough to fit into a reasonably-sized pocket, in February's Cutting Edge I didn't expect to be following up with a story of another computer around a quarter of the size and not made by Sony. The OQO, however, is roughly the same dimensions as a handheld organiser or PDA and features 256MB RAM, a 10GB hard disk, an LCD touch-screen with a 640 by 400 resolution, a 1GHz Crusoe processor and Windows XP. This is pretty impressive. But most importantly, in terms of connectivity the OQO features FireWire and USB, and is compatible with two standards for wireless connectivity: 802.11b (the standard Apple affectionately call Airport) and Bluetooth.

One potentially exciting application for the OQO is as a recording device, and although you wouldn't use the built-in A-D converters for serious work, you don't have to with the various USB and FireWire audio devices currently on market. Devices that could be ideal include Digidesign's M Box and, for multitrack recording, the Presonus FireStation, which supports MIDI, multitrack digital I/O, direct mic inputs and mLAN. Of course, I'm not sure how well the tiny hard disk would handle multitrack requirements — these drives aren't the fastest in the universe — but 10GB is enough to hold around two hours of 16-track recording, which isn't bad for a computer the size of your palm. However, with a screen that small, both physically and in terms of resolution, you're going to be hard-pushed to comfortably run any sequencer, although there's nothing to stop you plugging the OQO into a bigger monitor.

The relevance of these 'tiny but powerful' devices to musicians is that they're capable of acting as a pretty complete recording studio, with racks of virtual processors, synthesizers and samplers. Aside from valve-based processors, it seems that the only dedicated audio and music hardware left consists of I/O units and controllers, such as mixer control surfaces and keyboards (plus speakers, of course!). This could all sound a bit negative if you've been around the music and audio business before the days of the software revolution, but I don't think it needs to be.

The OQO includes all the connectivity a modern computer musician needs (from right to left): a FireWire connection, a USB port and a general-purpose expansion connector for external monitors and so on.

While it's a shame that some of the more traditional skills associated with audio (such as how to design an analogue EQ circuit) have become virtually redundant, this isn't something that's happened overnight. The writing was on the wall for analogue signal processing the moment the first digital mixing consoles arrived (at least, the first ones that sounded any good!), and this changeover has been going on for at least 15 years.

Modern DSP techniques were first theorized in the 1930s, but there was no practical application for the theoretical techniques since the hardware didn't exist at the time, which was still the case until the late '60s and '70s. The level of skill needed to program DSP techniques on standard processors ensures that the overall level of audio expertise doesn't diminish. However, what has changed is that digital audio processing is far cheaper than analogue techniques, and even basic Windows and Mac platforms are far ahead of the basic requirements for real-time digital audio processing.

Of course, these miniscule computers aren't only designed for musicians: they'll also have significance in just about every other industry. It's not just the fact that you can carry desktop computing power around in your shirt pocket that's important, it's the fact they'll all be able to take part in a mobile, ad hoc network. Put enough of them together in the same area and you effectively have a mobile communications network for free. You only have to look at the roller-coaster ride telecom shares have had recently to see what an economic earthquake this could cause.

A computer journalist once said of a big American computer company that its marketing skills were so bad that if it ran KFC, it would describe the fast food chain's business as 'selling warm dead chicken'. Literal descriptions rarely sell products — would you recognise a terminal for an asynchronous transfer mode network that uses 53-byte words in a hardware circuit-switched configuration giving guaranteed quality of service? Well, my mother's got one, I've got one, and I expect you have too: a telephone. (And don't worry, I'm doing only one of these examples to prevent this column from turning into a game!)

It's very easy to understand what a telephone does, and the benefits are obvious; but introducing a new technology alongside an established one can be a real problem. This is less of an issue when a new product is revolutionary and obviously better than any predecessors: CDs have been around for 18 years now, for example, and had obvious benefits compared to records; DVDs have already reached critical mass with the public and have obvious benefits over VHS tapes. However, it's not always so clear-cut.

When petrol-driven cars superseded horses and carts, it wasn't entirely obvious that these noisy, slow and unreliable contraptions would displace animal locomotion. Only the more technically minded and commercially brave would have recognized what today we would have described as a 'disruptive' technology — disruptive in the sense that it threatens the commercial viability of an existing technology. If your business is confronted by a disruptive technology, you have to decide whether to fight it or embrace it. Manufacturers of horse-drawn carts would, with hindsight, have been well advised to connect engines to their wheels.

Video and audio streaming is disruptive to the conventional broadcast industries as it offers very clear benefits, such as video and audio on demand, an infinite number of channels, very low distribution costs, worldwide audiences and so on. Yet in spite of this, the nascent video streaming industry is, at least on the face of it, in drastic decline, with virtually every streaming company either cutting back severely or out of business altogether. So was streaming just the wrong idea, or did it emerge at the wrong time?

The Windows Media Player: is this the future of home entertainment?

In my opinion, it's the latter. When you explain streaming to someone who might potentially benefit from it, you either have to make it sound like something we've got already, such as radio and television — in which case what's the point? — or you have to go into an entirely inappropriate level of technical detail that will completely perplex the vast majority of people who might ever use such a service, putting them off in the process.

The public opinion of streamed media is also pretty low. Most people think of streaming as the experience of watching a poor-quality, fuzzy and jerky picture on a computer screen. Actually, until April you could get much the same effect by watching ITV Digital! But seriously, when you access streamed media over a broadband connection, the picture is better, less fuzzy and not as jerky, although the way you actually access the content is no easier than using a 56k modem. To make streaming work for ordinary people, it needs to provide a great picture and be so easy to use that people don't even think about whether they're watching streamed video. Unfortunately, it seems we're some way from this situation.

ADSL in the form we have it is only an eighth as fast as we need for decent TV, and there's a growing feeling that it's only an intermediate fudge while we wait for genuine broadband in the form of Gigabit Ethernet to the home. There's even talk of 10-Gigabit Ethernet, which is more than a Gigabyte per second and enough for 400 channels of MPEG 2 compressed, high-definition TV and 7,000 channels of uncompressed CD audio. In fact, as MP3 files, you could download a sizable portion of the modern recorded music repertoire simultaneously, which is probably why it'll never be allowed to happen.

Bill's Corona     The streaming experience will be improved with the next generation of Windows Media. Codenamed Corona, Microsoft's new streaming system claims to offer at least three major benefits: multi-channel sound support, better compression (around 25 percent smaller files for a given quality) and — importantly — instant access to streams, without the usual delay associated with buffering. If you're new to streaming, buffering is a delay that you experience when you first select a stream. It's quite possible to stream without buffering — video conferencing wouldn't work with any more than a tiny delay — but where the data rate delivered by the Internet varies enormously from second to second, which it typically does, then it is best to wait for a while until all the packets of information for the next few seconds of audio or video have arrived. This can completely smooth out random variations in data rate, as long as the long-term average data rate is equal to the rate of the stream. So, how will Microsoft get around the issue of buffering? I don't know for certain, but I suspect smoke and mirrors. In his speech to the April 2002 WinHEC conference, Bill Gates said it would be possible by "being a lot smarter about bringing things across the network and caching them", before going on to say that "using this large storage we have and adding intelligence becomes increasingly important because otherwise people will say 'my experience isn't improving as much as all of these raw performance numbers suggest', and the fact is that they'll be right. Unless we conquer latency with clever local intelligence, we won't be able to pass through those benefits." What I think Gates is alluding to here is that you can give the impression of zero latency by intelligently caching the streamed material. When a user selects a web page with a choice of streamed media, the system could transparently begin buffering the material from all the available choices. By the time the user clicked on a choice, the material would already be in the buffer. It's really just like GigaStudio or HALion: the response from a hard disk is too slow to find and play samples in real time triggered from a keyboard, but if you load the first second or so of each note into RAM, then the disk has a chance to catch up with the selected note. This would work for playback of prerecorded files. It wouldn't work for live streaming, though. One way to get instant access with live streams is to decode all of them continuously, and then switch to a particular one when a user selects it. Since it's already being decoded, there would be no buffering delay.