by Craig Anderton
When hardware still ruled the earth, the Korg M1 workstation created quite a stir with its combination of sounds, multi-timbrality, a MIDI sequencer, and effects. And when MIDI sequencing took off, workstations became essential elements of many a MIDI studio because they could play back multiple sounds from a single, economical piece of hardware.
Today's software workstations can not only do pretty much anything their hardware ancestors could do, but a lot more - so let's discuss how to maximize the potential of these virtual instruments, from initial songwriting inspiration to final mix.
Software workstations are exceptionally useful for songwriting, because with one instrument, you can create 8, 16, or even more tracks. As a result, you can simply keep loading instruments into MIDI channels, create new MIDI tracks, and lay down overdub after overdub.
Fig. 1: Several techniques mentioned in this article are applied here to Native Instruments' Kontakt. Polyphony has been limited to save on CPU consumption (outlined in yellow; Electrik Guitar has been limited to 16 voices, Drawbar Organ to 32, and Fretless bass to 4). The blue outline shows that reverb has been added as a bus effect, to avoid putting a reverb in each channel. The red outlines show that each instrument has been sent to its own output channel.
However, as you add more instruments, CPU consumption will increase, sometimes dramatically. Many workstations allow adjusting polyphony for particular channels (Fig. 1), so take advantage of this minimize the number of voices that need to sound at once. For example, with many bass lines, you probably won't need more than two voices. Sounds with long decays, such as pads, tend to "eat" polyphony so restrict those as well?often the voices that are cut off are at a low enough volume, or masked by other notes, so that you won't notice they're missing.
Another way to reduce CPU consumption: Whenever multiple sounds use the same effect, use bus effects within the instrument (if present) instead of insert effects. For example, insert reverb on a bus and send signal to it from the instruments you want processed, rather than inserting reverb on all channels requiring reverb.
A big advantage to using a MIDI-based software workstation is that MIDI data is so malleable. If during the songwriting process you decide to change the key or tempo, it's much easier to do with a bunch of MIDI tracks than with digital audio. (Having said that, though, many workstations can also stretch digital audio loops with respect to timing and possibly pitch.)
A corollary MIDI advantage is that when it's time to mix, you can replace the sounds of individual tracks with individual instruments that may offer a better sound quality. For example, you can use the workstation to lay down a piano part, but then switch over to something like Ivory or another dedicated piano program to get the best piano sound possible. The same thing goes for drums, as you can use a program like FXpansion's BFD to replace the simpler drum sounds found in a workstation.
Also note that hosts with MIDI plug-ins get along very well with workstations. When you're laying down tracks in quick succession, rather than deal with quantizing or tweaking as you record, you can often use MIDI plug-ins to temporarily do quantization, scale velocities, and the like. After you're finished laying down tracks, then you can get into the editing process and tweak the MIDI data.
There's a growing trend with workstations and multitimbral samplers to include ever-greater amounts of content. Generally, this defaults to being installed in the root drive where the program lives; use enough of these products, and your main drive can run out of space pretty fast.
So, dedicate a big hard drive (250-500GB) just to content and samples. If your computer's motherboard has some unused connections for hard drives, you can mount the drive internally but if not, an external FireWire or USB drive will do the job. However, you will likely have to instruct the program where to look for the content; check any available preferences dialogs, as that's usually where you specify a path to the content.
With some workstations, you can create an alias/shortcut for the target samples in the original folder on your root drive. For example, if the workstation has a folder named "sounds," you may be able to create an alias for the drive containing your samples and put it in the "sounds" folder. Other workstations recommend against moving factory content to a different location, as any updates might be written to the current location, thus making it difficult to keep file structures "in sync" - although you can create a path to your own custom content. The instrument's documentation should mention any considerations involved in moving content.
Some workstations can stream long samples from hard disk, while others are restricted to what you can load into RAM. While using RAM is generally faster and smoother, you're not going to load a 30GB piano into a computer with 2GB of RAM. Streaming often needs to be enabled, sometimes for individual instruments within a multitimbral instrument, or sometimes for the instrument as a whole. Just remember that neither solution is without issues: Streaming lots of samples from a hard drive can limit the number of audio tracks you can stream from the same hard drive simultaneously (which is why a dedicated drive for content is helpful), while pushing RAM to the limit can cause instability problems because that same RAM is shared with the operating system and your host program.
I used to advise bypassing the included effects with virtual instruments, as you could likely do better by adding other plug-ins into the signal path hosting the instrument. But times have changed. With better computers, instruments can include far more CPU-intensive, and better-sounding, plug-ins (Fig. 2); some even include features like convolution reverb.
Fig. 2: SampleTank 2.5 comes with 33 effects, including convolution reverb. These can be different for each instrument in a multi-timbral setup, and can also be used as send (insert) or master effects.
However, the issue here isn't just about sound quality. Using effects included within the instrument makes projects more transportable and archivable: As long as you can load the instrument, you're loading the effects as well.
Workstations usually offer multiple outputs (Fig. 1), so you can take advantage of your host mixer's features to process individual sounds. Keeping in mind the above comments about effects, though, if you can do all your mixing and processing within the workstation, you again have a more ergonomic and transportable project. You can even save the workstation setup as a preset and import it into a different host, knowing that the sounds and mix will be as you intended.
If you really load up the channels of a multitimbral instrument, you may need to freeze tracks to free up CPU power. Freezing essentially disconnects the instrument from the CPU, replacing it temporarily with an audio track that makes much fewer demands on the CPU.
However, freezing works a little differently with multitimbral instruments compared to single-channel instruments; this varies from program to program. For example, you may be able to freeze one particular instrument of a multitimbral instrument or you may only be able to freeze the entire instrument. Check your host program's documentation for details.
Here are thumbnail descriptions of some common software-based workstations, listed alphabetically by manufacturer, along with screen shots for some of them.
Apple, EXS24. This is available only as part of Logic; while showing its age a bit, the EXS24 broke open the virtual sampler market.
Big Fish, Vir2 instruments. Based on the Native Instruments' Kontakt Player engine, these offer effects, mixing, streaming from hard disk, and other features derived from NI's flagship sampler. The screen shot shows Mojo, their horn section-based virtual instrument.
Cakewalk, TTS-1. Available only within Sonar, this basic (and CPU-friendly) workstation is useful for blocking out parts while songwriting. Note the window that does basic editing for one of the 16 available parts.
EastWest, Play. EastWest's proprietary virtual instrument engine hosts many of their sample libraries, and is compatible with 64-bit operating systems.
IK Multimedia, SampleTank-based products (SampleTank, Miroslav Philharmonik, Sonik Synth, SampleMoog). These are all characterized by large sound libraries, and offer a comprehensive set of effects.
Korg, Digital Legacy Collection. This set of virtual instruments includes a software M1, but it's not your father's M1: The sound is much cleaner, and it comes with all the M1's expansion card sounds. This screen shot shows the "Easy" page, so-called because of the ease with which you can make common edits.
MOTU, MachFive 2. The latest version includes advanced features such as beat-slicing, REX file importation, convolution reverb, the ability to import samples in just about any format, and a 32GB sound library.
Native Instruments, Kontakt 4. This ambitious sampler includes features not found elsewhere, like MIDI scripting (think of it as MIDI plug-ins you can write yourself). It also has highly developed slice-oriented "beat machine" functions.
Propellerhead Software, Reason. While not a workstation per se, the ability to ReWire it into any major host is compelling; and you can insert as many instances of the included synths and samplers (including the acclaimed NN-XT sampler, shown in the screen shot) as your computer can handle.
Sonivox, Muse. Based on GigaStudio technology, this isn't as editable as some of the "pure" samplers, but has a ton of sounds and fulfills the concept of a hardware ROMpler brought to software.
Steinberg, HALion. HALion is a traditional sampler that can stream samples from hard disk and offers multi-timbral operation.
Ultimate Sound Bank, PlugSound Pro. This workstation also handles loops and beats well, with excellent time-stretching options. Optional-at-extra-cost "virtual sound cards" are available for expansion.
Also note that several companies produce "application-specific" workstations, such as MOTU's Ethno Instrument and East-West's Ra (both dedicated to world and ethnic sounds); for orchestral work, there?s IK Multimedia's Miroslav Philharmonik, Garritan's Personal Orchestra, and HALion String Edition, among many others.
Craig Anderton is Editor Emeritus of Harmony Central. He has played on, mixed, or produced over 20 major label releases (as well as mastered over a hundred tracks for various musicians), and written over a thousand articles for magazines like Guitar Player, Keyboard, Sound on Sound (UK), and Sound + Recording (Germany). He has also lectured on technology and the arts in 38 states, 10 countries, and three languages.