If you're relatively new to PC music, it can seem like a maze of acronyms and mystifying computer-related terms — so here's a simple glossary to clear up any confusion.
Many musicians are familiar with multitrack recording, mixing, and mastering using dedicated hardware, but are moving across to PC-based systems and finding a whole new world of initially overwhelming technical PC terms, such as IRQ, partition, latency, buffers, BIOS, transfer rate, and so on. I've covered all these terms in some details in my extended PC Musician features over the years. You can find links to them by using the search criteria here: www.soundonsound.com/search.
However, for a PC novice there's probably just too much information to take in at once, and a popular topic request on the PC Music Forum has been for some sort of simplified introduction. So this month I've gathered together all of the most common terms you're likely to come across into one neat resource. Each one consists of a short paragraph explaining its practical significance to the PC musician.
- AGP (Advanced Graphics Port): This is a dedicated high-speed interface between PC memory and a graphics card and monitor. Before AGP existed, graphics cards were fitted into standard PCI slots (see PCI) and could cause audio clicks and pops if they didn't correctly release the PCI buss for other duties when requested, thereby preventing audio data from reaching a PCI soundcard. Nowadays all PCs use AGP graphics technology, which provides faster graphics performance and avoids such audio interference problems. As an aside, music applications don't use 3D graphics, so musicians don't need an expensive state-of-the-art graphics card. Far more useful is a good-quality 2D dual-head model that supports two monitor screens, since most music software will let you place the song's arrange page on one screen, and the software mixing desk on the other, which makes working with complex multitrack songs far more pleasurable.
- BIOS (Basic Input Output Subsystem): This is a small program stored in a motherboard chip that starts up your PC. Its duties include monitoring the various voltages in your PC, to check that everything's working OK; recognising all the fixed hardware devices on the motherboard, such as serial, parallel, USB, and FireWire ports; recognising the model and speed of CPU and its current operating temperature; detecting how much RAM is installed and what speed it is; registering the capacity of each hard drive; and so on. It also recognises the various expansion cards you've installed and can allocate them interrupts (see IRQ). Once the BIOS is sure that your hardware is working correctly, it will normally sound a brief single beep before passing control over to the Windows operating system. You can normally enter the BIOS Setup program by pressing the Delete key during boot-up. Doing this will let you change fundamental system settings and selectively disable unwanted motherboard devices, such as on-board soundchips.
- CPU (Central Processing Unit): This is the brain of the PC. Its clock speed determines how many software plug-in effects and soft synth notes can be run simultaneously in real time. The two most popular manufacturers of CPUs are AMD, with their Athlon, Opteron and entry-level Duron ranges; and Intel, with their entry-level Celeron, mainstream Pentium 4 and high-end Xeon processors. The two best choices for a musician are currently an AMD Athlon or Intel Pentium 4C processor of the highest speed that you can afford. Traditionally, buying Intel almost guarantees compatibility with the widest range of soundcards, while Athlon-based PCs are sometimes significantly cheaper for an equivalent amount of processing power, but have had a few compatibility problems in the past. Nowadays these have largely disappeared.
- CPU Fan: The CPU generates a significant amount of heat when it's running, so to ensure that it doesn't overheat it's normally fitted with a finned metal heatsink to help dissipate the heat, and some sort of fan that generally blows cool air onto the heatsink. The standard fans supplied with most CPUs can be quite noisy, which rarely goes down well in a recording studio, so to keep acoustic noise levels down you can buy a large variety of replacement heatsink/fan combinations (the combination is often termed a CPU Cooler). Additional system fans may be mounted in the PC case to encourage cool air to enter at the bottom of the front panel and to extract warm air near the top of the rear panel. Quiet versions of these are also available as replacement items.
- Dongle: A hardware device that plugs into a parallel or (more often nowadays) a USB port, acting as copy protection for a particular software application. Both Emagic's Logic Audio and Steinberg's Cubase sequencer range use dongles, and these applications will only run when their particular dongle is detected. If you have a desktop and laptop PC you can, in most cases, install the software onto them both and swap the dongle from one to the other as required, but you can never run more than one installation at once. Cakewalk's Sonar has no such copy protection.
- HyperThreading: A new feature found on Intel's latest Pentium 4 processors and on their more expensive Xeon range, HyperThreading (HT) lets the CPU act more like dual processors that can run two applications simultaneously, or run a single application significantly faster than on a standard processor. To gain this advantage, HT also requires a suitable PC motherboard running Windows XP, plus specially written software. Little music software has yet been so modified, apart from Steinberg's Cubase SX 2.0 and Nuendo 2.0, but fortunately you can nearly always simply leave HT enabled, even if the software isn't aware of it. Exceptions include Cakewalk's Sonar 2.2 and Tascam's GigaStudio, which won't run properly if HT is enabled.
- IRQ (Interrupt ReQuest): Nearly all hardware devices, including soundcards, graphics cards, an array of motherboard devices such as hard and floppy drive controllers, USB port controllers and FireWire controllers, and your keyboard and mouse, need to be interrogated periodically by the PC's OS to see if any new data needs to be processed. To do this, Windows stops what it's doing and sends an interrupt request to the device in question. Unfortunately there are rarely enough different interrupts to allocate one to each device, so some must share (IRQ sharing). This normally doesn't cause any problems, except when one of the devices in question prefers or even demands an interrupt all to itself. A few soundcards do this (although their manufacturers rarely advertise the fact) and, if they end up sharing an IRQ, can cause clicks and pops during audio playback and recording. For this reason, most musicians do their level best to ensure that their soundcard gets its own unique interrupt, to minimise possible problems.
- Monitor: Unfortunately, the word 'monitor' is used to describe a computer monitor screen and a monitor speaker, causing a lot of confusion among newcomers. In computer terms, a monitor screen is available in two main types. CRT (Cathode Ray Tube) models operate just like traditional television screens — using an electron beam that scans each horizontal line from top to bottom of the screen — except that they generally have a much higher-quality picture. However, they are being rapidly replaced by flat-screen TFT (Thin Film Transistor) screens as prices have come down. This type of screen provides noticeably sharper pictures with perfect geometry, far less heat generation, a smaller footprint on the often crowded desktop and no magnetic interference to cause hum with nearby guitars.
- Motherboard: The largest single component of any computer, this large circuit board has at its heart a 'chipset' that lets the different parts of the system 'talk' to each other. The main chipset used determines what family of processors can be plugged in — so, for instance, AMD Athlon-based PCs require a different motherboard to Intel Pentium-based ones. A few years ago, some early Athlon chipsets proved to be incompatible with a few soundcards, which left some musicians wary of relying on AMD-based PCs, but such problems have now almost completely disappeared.
- Partition: Hard drives are shipped as one huge storage device, but it's possible to divide them up into smaller and often more manageable chunks, called partitions. One huge advantage of creating a separate partition for your data (such as documents, songs and audio tracks) is that if you ever get a problem with your Windows installation, you can safely repair it or even erase it and install it from scratch without losing any of your work. Creating multiple partitions also enables you to install several different versions of Windows (or another operating system) on the same PC and choose which one to run each time you switch on. This may enable you to carry on running older software while still taking advantage of the new features of Windows XP, or to install a duplicate version of Windows solely intended for running music applications, to give it the best chance of optimum performance. This is a particularly good arrangement for a professional studio where reliability is of the utmost importance, or when children and games are associated with a family PC.
- PSU (Power Supply Unit): The PSU supplies various voltages to the computer motherboard, hard drives, CD drives and floppy drive, and is normally forgotten by most computer users — or it would be, if it were not for one thing: the PSU needs a cooling system to dissipate the heat it generates and to dispose of the extra heat generated by other system components such as the CPU and expansion cards. This nearly always means that the PSU requires an extractor fan, which creates unwanted acoustic noise in the studio. Consequently, a lively trade in replacement quiet PSUs using significantly quieter fans now exists. In some cases these incorporate a larger internal heatsink than normal, to share the heat-dissipation duties.
- Sustained Transfer Rate: The peak data-transfer rate of a hard drive is important to most general-purpose applications that suddenly require a lot of data to be loaded or saved, but the real-time nature of audio recording and playback means that the 'sustained transfer rate' is far more important, since this will determine how many audio tracks can be played back and recorded simultaneously by your hard drive. The easiest way for musicians to see how their hard drive performance compares with that of other users is to run the freeware DskBench utility written by Jose Catena, and available for download from www.sesa.es/us/dskbench/dskbench.htm. This also checks that your various drives are using Buss Master DMA, a technology that allows them to continue transferring data without involving your computer's CPU, leaving it to concentrate on running music software.
Windows JargonSome musicians might expect a section devoted to Windows in this jargon buster, but, as I explain in this month's PC Notes column, the latest Windows XP requires a minimal amount of tweaking to suit the special real-time requirements of PC music applications. However, there are still a few points worth explaining:
- Processor Scheduling: This should be set to 'Background Services' in the Advanced section of Performance Options, which you can find in the Advanced page of the System applet in the Windows XP Control Panel. This setting will ensure that your soundcard achieves its lowest usable latency setting.
- Power Schemes: Accessed via the Control Panel from the Power Management applet of Windows 98 or the Power Options applet of Windows XP, these are mainly relevant to those using laptops. Musicians should choose the Power Scheme named Always On to get maximum processing power for the whole time the laptop is running, whether from the mains supply or the battery. Switching to Home/Office Desk will give you more battery power when you unplug the mains supply, by activating various power-saving schemes, but you'll not be able to run as many plug-ins or soft synth notes. Using the Portable/Laptop setting will degrade processor performance when you're still plugged into the mains but will save power when you're word processing, with the result that the cooling fan will not come on so often.
- Visual Effects: While Windows XP looks very smart with its animated windows, fading or sliding menus and shadows, many musicians find these make it feel very sluggish. However, it's very easy to disable such effects using the Performance page in the Advanced section of the System applet in Control Panel. The only one I leave active is 'Show Windows contents when dragging', which makes it easier to see what you're doing when rearranging screen contents.
Sound Hardware Formats
- Soundchip: Although most modern PC motherboards incorporate a basic soundchip device that's perfectly capable of playing back system audio files such as the Windows startup and close-down sounds, MP3 files, general-purpose audio files, CD audio and even DVD audio tracks, few offer reasonable recording options and none offer 24-bit recording or playback. Using the built-in soundchip you're also likely to suffer from significant time delays between an incoming audio signal being recorded and monitoring it via your music software, and between playing a MIDI keyboard and hearing the output of a connected software synthesizer (see 'Latency' for more details). For these reasons, the musician nearly always finds it essential to buy some sort of sound device with higher audio quality and better audio drivers, and then to disable any motherboard soundchip. Such sound devices are available in various formats, each having its own strengths and weaknesses.
- PCI (Peripheral Component Interconnect): PCI is still the most popular format for musicians' soundcards. These are circuit cards that fit into expansion slots inside the PC. Most motherboards provide between four and six PCI slots, to house such hardware as soundcards, modems, MIDI interfaces, network cards, and so on, and to install them you must open up the PC to gain access to its interior. However, installing a PCI card only requires you to plug it into a suitable slot, tighten down one screw and then reboot your PC, to have the card recognised automatically, ready for its software drivers to be installed. PCI soundcards also require an IRQ (see 'IRQ' entry), and preferably an unshared one. If your PC already contains lots of expansion cards this may be difficult.
Simple soundcards have their various input and output sockets mounted on the backplate of the soundcard (although some require a second dummy backplate, housing more sockets, that will prevent a card being fitted where the dummy has been mounted). More advanced products have an external case (often a rackmountable one) for their sockets, the case being connected to the PCI card via an umbilical cable. The next generation of the PCI standard —PCI Express — has already been announced and will provide significantly greater bandwidth, as well as allowing hot-swapping of peripherals, like the USB format. However, even when PCI-X starts to appear, during 2004, original PCI slots will still continue to be fitted alongside PCI-X slots on motherboards, and this will be the case for several years to come.
- USB (Universal Serial Bus): The USB version 1.1 port, now confusingly given the label 'USB Full Speed' to differentiate it from the far more capable USB 2.0 (see later), first appeared on PCs and audio and MIDI peripherals in about 1999. Using a serial approach to sending and receiving data (ie. one bit after the next, in one long stream), rather than the parallel approach of PCI and PCMCIA, the USB port can use a much simpler cable and plugs. USB also allows users to plug and unplug external compatible peripherals whenever they like (termed hot-swapping or hot-plugging), even when the PC is powered up, although it's generally not advisable to do this with sound devices before exiting your music application. Since the peripherals are external to the PC, USB was initially seen by manufacturers as far easier for users, since it avoided the need to open up the PC.
Unfortunately, PC musicians experienced a wave of initial problems with USB audio peripherals, such as clicks, pops and high latency. These problems were largely due to the early USB controller chips in the PC, rather than to the peripherals, and have largely died out, but they've left quite a few musicians mistrusting USB 1.1 as a standard, which isn't entirely fair. Its 12Mbit/second bandwidth is perfectly adequate for devices such as mice, PC and music keyboards, printers, scanners, and the like, and perfectly adequate for MIDI and stereo audio devices down to latencies of 3ms. However, USB 1.1 runs out of steam when you ask it to record and play back stereo 24-bit/96kHz signals simultaneously. USB MIDI interfaces can also suffer from slightly higher timing 'jitter' than serial port or PCI-based interfaces.
USB 2.0 ports have been appearing on PC laptops for some time, and are now to be found on nearly all new desktop motherboards. The more correct nomenclature is 'USB Hi-Speed' — and with a huge 480Mbit/second bandwidth (compared with the 12Mbit/second of USB 1.1), this new standard lives up to its name and is far more suitable for multi-channel audio devices than the original USB standard. (The world's first USB 2.0 device is Edirol's UA1000, pictured on the previous page.) You can still plug USB 1.1 devices into USB 2.0 ports (and vice versa, although the latter won't operate well under the severely restricted bandwidth). If you're not sure whether your PC has USB 1.1 or USB 2.0 ports, take a look in Device Manager under the heading 'Universal Serial Bus controllers'. In the case of some chips, the description may include 'USB 2.0', but if not, write down the number of the chip and do a Google search for it.
- FireWire: More correctly termed IEEE-1394, due to its approval by the Institute of Electrical & Electronic Engineers, FireWire is another high-speed serial standard with a similar maximum bandwidth (400Mbit/second) to USB 2.0. It's particularly popular in the area of digital camcorders, allowing them to be digitally connected to computers for video editing, but both technologies support hot-swapping and similar maximum cable lengths of around five metres, and there's not too much between the two when it comes to deciding between them for audio devices. However, FireWire does have a reputation for causing fewer hiccups during data transfer, despite its slightly lower bandwidth, and also has peer-to-peer topology, which (for instance) lets you connect a VCR and digital camera without involving a PC. Still, when it comes to choosing an external audio peripheral, you should be guided more by the other features of competing products, since both USB 2.0 and FireWire are extremely capable.
- PCMCIA (Personal Computer Memory Card International Association): Due to size constraints, no PCI slots are available on a laptop PC, but you'll nearly always find a PCMCIA slot that can house a removable credit-card sized expansion device. The main advantage of a PCMCIA sound device over the laptop alternatives of USB and FireWire is that it's possible to make PCMCIA devices extremely compact, which may be ideal for a musician on the move. However, USB 2.0 and FireWire products are likely to feature a more versatile selection of inputs and outputs.
If you would like to find out more about specific topics mentioned here, archives of past PC Musician archived features are held on the Sound On Sound web site (www.soundonsound.com). Here's a list of the articles that are most appropriate for musicians who are new to PCs:
- Controlling Influence: Using Budget MIDI Hardware Controllers with a PC (SOS October 2001).
- PC Soundcards: A Practical Guide (SOS January 2002).
- One PC, Two Monitors: Setting Up & Using Multiple Monitors For PC Music (SOS May 2002).
- The Truth About Latency: Soft Synth Latency And Jitter Test (SOS September 2002).
- The Truth About Latency Part 2: Latency And Jitter Test (SOS October 2002).
- Clear Signals: Testing And Improving PC Soundcard Performance (SOS November 2002).
- The Bionic PC: BIOS Tweaks For Music Performance (SOS December 2002).
- Windows XP: One Year On (SOS February 2003).
- The Great Divide: Partitioning PC Hard Drives For Multi-boot Systems (SOS March 2003).
- Hitting The Slot: The Musician's Guide To PCI Slots And Interrupts (SOS May 2003).
- The Right PC For The Job: Specifying A PC For Your Needs (SOS June 2003).
- Click Tracking: Eliminating PC Audio Glitches (SOS July 2003).
- Laptop Options: Choosing A Portable PC For Music (SOS October 2003).
Each hardware device used by a PC needs a small piece of software, called a driver, to act as the glue between it and the Windows operating system. One of the major confusions faced by the new PC Musician is which type of soundcard driver to choose. Most are now shipped with a bewildering array of options, such as ASIO, DirectSound, EASI, GSIF, MME and WDM, and choosing the most appropriate one depends on which music software you're running.
- Latency: Before we discuss the different driver types, it's important to understand the concept of latency. Because Windows has so many tasks to perform, it can't ever devote itself to MIDI or audio tasks full-time. Instead, it multitasks, by doing a little bit of each required task in turn, to give the illusion that everything is happening simultaneously. While Windows is off doing other things, the soundcard or MIDI interface must have some pre-prepared data to access, so that we don't hear an audio interruption (a click or pop), or a MIDI interruption (uncertain note timing). This data is stored in small areas of RAM called buffers, and each time Windows returns to the music application it will prepare another buffer full of data to (ideally) last until its next visit. Unfortunately, using buffers means that there will always be a slight time delay between when a signal is recorded and when you are able to hear it emerge from the soundcard, or between playing a new note and hearing it emerge from a soft synth (see next section). The smaller the buffer, the shorter the time delay, or 'latency', but if you make the buffer too small Windows won't get back in time to keep it topped up, so you'll hear clicks and pops. The ideal size of a soundcard buffer is thus a compromise between stability and low latency.
- MME-WDM (MultiMedia Extensions Win32 Driver Model): Between them, these two standards incorporate the oldest and newest Windows standards, but if you run the latest Windows XP OS they are closely connected. MME was the very first soundcard driver format, first appearing way back in Windows 3.1, and it generally results in high latency, although it can work well with some stand-alone softsynths. WDM was first introduced in Windows 98, but really came into its own under Windows XP, where it provides better performance with USB and FireWire and generally much lower latency than MME with suitably written music applications such as Cakewalk's Sonar (see next section).
- DirectX: A set of low-level software routines included primarily for games in Microsoft's Windows operating system. DirectX interfaces between software applications and multimedia hardware such as graphics cards and soundcards. Its components include DirectDraw, Direct3D, DirectPlay, DirectInput, DirectMusic, DirectSound, and DirectSound3D.
- DirectSound: One of the components of DirectX. DirectSound drivers are generally capable of significantly lower latency than MME drivers and are a better choice for soft synths and playback of audio, but don't normally have recording options if you're running Windows 98/ME. (Make sure, if you choose DirectSound drivers in a music application, that the driver name doesn't have '(emulated)' at the end of its name, since this means that no properly written DirectSound drivers have been detected. If you choose an emulated driver it will perform extremely poorly and exhibit high latency.) WDM drivers automatically provide both MME and DirectSound support without the soundcard manufacturer having to write special code, so if your WDM-capable soundcard is running under Windows XP you'll get more options. However, in my experience choosing DirectSound over MME will generally allow lower latency.
- ASIO (Audio Stream Input Output): This driver format, introduced by Steinberg in their popular Cubase application, runs at a lower level (as it bypasses much of the Windows OS) and therefore typically manages lower latencies than both MME and DirectSound. Many musicians have managed to run their music software with latencies as low as 2ms using ASIO, although 6ms is a more realistic figure, and even 12ms is acceptable in most cases. Emagic's EASI (Enhanced Audio Streaming Interface) format is derived from ASIO and is capable of slightly better performance with Logic Audio. However, few soundcards now support this option.
- GSIF (GigaSampler InterFace): Designed specifically for just one application — Tascam's popular GigaStudio software sampler — GSIF drivers also work at a low level within Windows, providing guaranteed low latency of between 6ms and 9ms. If, like plenty of other musicians, you're interested in GigaStudio and its extensive collection of professional sample libraries, your soundcard must have GSIF drivers; you won't be able to run GigaStudio without them.
- Effect Plug-in: Modern PCs can easily perform the calculations required to add traditional effects, such as chorus, echo, equalisation and reverb, to audio files in real time, but it was the concept of the 'plug-in' that really revolutionised computer music processing. Allowing third-party developers to 'plug' additional functions into an existing host application means that one is no longer restricted to using the ones that are bundled with it, but can assemble a personal collection of favourites. The first plug-in format to appear on the PC is now called DirectX (since it uses the media-streaming components of Microsoft's DirectX). DirectX plug-ins can be launched from within nearly all music applications that support plug-ins. The other main format is Steinberg's VST (Virtual Studio Technology), originally developed for the Cubase VST sequencer, but again now widely supported by other music applications.
- Soft Synth: This is the generally accepted abbreviation for a software synthesizer. The first soft synths were stand-alone applications that could generate audio-waveform data in real time, either triggered from a suitable MIDI input signal played on a music keyboard, or from a set of notes played back using a software sequencer. Stand-alone versions of many modern soft synths are still released, but Steinberg's VST 2.0 specification added MIDI input capability to their VST plug-in effects, for automation purposes, and this also saw the creation of the VST Instrument (VSTi). This is a type of plug-in that accepts a MIDI input signal and generates an audio output signal. Many other music applications now support VST Instruments, and Cakewalk created the functionally-identical DX Instrument for their Sonar application.
- Soft Sampler: Strictly speaking, a soft sampler plays back pre-recorded samples, while a soft synth calculates its waveforms in real time, although the boundaries between the two are often quite blurred. In general, a soft sampler lets you load in and play back your choice of sampled sounds, whereas a soft synth relies on a preset collection of waveforms and an associated synth 'engine' to filter and otherwise treat these sounds. Soft samplers use two main approaches to sample playback: they either load the entire sound into system RAM or stream the sample files direct from a hard drive. The latter approach allows samples of almost unlimited length to be used, which avoids having to loop them for longer notes, although software buffers still need to store the start of each sample to ensure low latency when you play a new note. However, the extra overhead of streaming samples from the hard drive may result in you not being able to play back quite as many audio tracks as you otherwise might in your MIDI + Audio Sequencer (see next entry).
On the PC there are three main soft synth contenders: Steinberg's HALion, a VST or DX Instrument particularly popular with Cubase users; Native Instruments' Kontakt, a stand-alone application, VSTi or DXi; and Tascam's GigaStudio. The first two offer a choice of storing samples in system RAM or streaming them from the hard drive, while GigaStudio does streaming only but has arguably the most professional library of the three.
- MIDI + Audio Sequencer: The main music application that allows you to record and play back audio tracks, plus MIDI data recorded from external musical keyboards or entered using the PC keyboard. The three most popular on the PC are Emagic's Logic Audio, Steinberg's Cubase, and Cakewalk's Sonar. Logic Audio seems to be widely used in commercial studios, which may be useful if you intend taking song files in with you, but Emagic have discontinued development at version 5.5.1 on the PC, to concentrate on the Mac, on which platform version 6.3.1 has since been reached. This makes Logic Audio a far less attractive proposition for the PC user. However, Logic is renowned for its high-quality bundled plug-in effects and for its soft synths and soft samplers.
Cubase VST for Windows 98/ME reached version 5.1 revision 1, but has largely been superseded by Cubase SX, currently at version 2.0, written from the ground up on the PC to suit Windows XP. It supports ASIO drivers, VST and DirectX plug-ins and VST Instruments. Cakewalk's Sonar, now at V3.0, is PC-only, and also has many enthusiastic users. Version 2.0 supported low-latency WDM drivers, DX plug-ins and DX Instruments, plus VST Instruments and plug-ins via 'wrapper' utilities. However, V3.0 has a bundled 'VST Adapter' for this, and supports ASIO as well as WDM drivers. You should be able to download demos of all three applications to get an idea which is the best for you.