Soft Sampling

Oct 1, 2001 12:00 PM, By Dennis Miller & Geary Yelton

With the enormous power residing in today's computers, much of the workload of traditional hardware devices has shifted to the desktop. One of the most common applications handled by the personal computer is sampling. Last year, EM examined a large number of software synthesizers and found that modern computers perform admirably as megavoice synthesis engines (see “Going Soft” in the July 2000 issue). In this roundup, we will survey 11 of the most popular software samplers for the Mac and PC platforms (see the table “Software Sampler Specifications” for a complete list of the programs).

We won't cover every soft sampler out there; although the soft-sampler market is modest compared with the soft-synth world, it still has too many programs to explore. We're going to skip programs, such as Native Instruments' Battery, that fall into the drum or rhythm category. We'll also pass over programs that play only samples and provide no significant editing, effects, or modulation options, as well as those that require dedicated hardware, such as CreamWare's Pulsar series. This roundup includes representative examples from every other category of sampling software available.

Software samplers come in several categories. Some work only as plug-ins — for example, IK Multimedia's SampleTank, Emagic's EXS24, and Nick Whitehurst's SamplerChan. Others, such as NemeSys's GigaStudio and Digidesign's Soft SampleCell, operate standalone only, whereas a third group — including Native Instruments' Reaktor, Speedsoft's Virtual Sampler, and Koblo's Stella9000 — function as plug-ins or standalones. Among the plug-ins, most are VST Instruments, but using FXpansion's VST — DX Adapter (www.fxpansion.com) on the PC, you can load nearly all VST devices in a DirectX host, such as Cakewalk's Sonar. (Among the Windows programs, only CreamWare's Volkszämpler and SamplerChan do not work in that configuration.)

You'll also find many prices for sampling software, from a low of $39 to a high of $699. Logically, the plug-ins are far less expensive than the standalone programs, and in most cases, you get what you pay for. We were pleasantly surprised that several less-expensive programs turned out to be real bargains, offering far more features than we expected.

In researching this article, we quickly discovered that many software-based “samplers” aren't samplers at all, at least not in the sense to which we're accustomed. Of the group, all but three lack an internal recording feature, which makes sense because most users have other audio-recording software on their systems. That also reflects a primary difference between the hardware sampler and new software samplers. Yet it's curious that the term sampler has been adopted for this breed of software when sample player or sample processor might be more appropriate.

COMMONALITIES

Traditional hardware samplers can record sound and play it back under MIDI control. Between recording and playback, they provide a means to edit the sound — inserting loop points, for example — as well as a way to assign sound to individual MIDI notes and Velocity ranges. All the audio and MIDI assignment data held in a sampler's RAM can then be stored on a disk or some other permanent medium and later retrieved to RAM. Soft samplers have many features in common with hardware samplers, yet many also take new approaches to sample manipulation and playback.

The programs in this roundup share many features. All but one program loads samples into RAM, which limits you to your computer's available memory. GigaStudio streams samples directly from your hard drive, so a 400 MB piano sample isn't a problem. (Virtual Sampler 2.7, now in beta, also has that capability.)

The range of effects included with the programs varies widely. Overall, plug-in samplers have a major advantage in the effects category because they allow you to pass their audio output to whatever third-party effects you can access through their hosts. That capability opens up the world of VST and DirectX effects and gives you a massive number of sound-shaping options.

You'll find support for the most common audio-file formats that are native to both platforms, with the Mac applications far more generous in supporting Windows WAV format than Windows apps are in supporting AIFF. Many programs can load the ubiquitous Akai format, SoundFonts, or both. When all else fails, you can use a file-translator program to convert files from almost any format imaginable to something that works for the software you choose (see the sidebar “Have It Your Way”).

Happily, most samplers are multitimbral and multichannel, though we were amazed at the vastly different approaches used to create a multitimbral configuration. In some cases, setting up a multi is almost automatic, but in others, that operation is buried beneath layers of options. Most of the programs we cover have only a few (sometimes, only one) screens, which simplifies navigation by putting the main working elements at your fingertips.

Many programs ship with various sample libraries. SampleTank XL wins the prize in that category with 2 GB of files ranging from standard orchestral instruments to special effects. NemeSys generously adds its renowned 600 MB piano sample to the GigaStudio bundle (talk about multisamples!), and many other programs include large sample or patch libraries to get you started.

Here, then, is our roundup of software samplers. There are tremendous advantages to using your computer as a sampler, but sticking with a dedicated hardware unit is a good idea too (see the sidebar “Pros and Cons”). We tried to cover each program's most significant features and mention likes and dislikes in each case. We highly recommend that you check out the demo versions available for nearly every program and see for yourself whether they provide the features you need.

EXS24 1.0

Emagic's EXS24 is a sampler plug-in that works as an Audio Instrument object within recent versions of MicroLogic AV and Logic Audio Silver, Gold, and Platinum. Depending on your computer's speed and available RAM, you can open as many as 24 simultaneous instances of EXS24, each with 64-note polyphony, without add-on hardware. Opening a Logic Audio file that contains EXS24 Sample Instruments automatically loads the appropriate samples.

EXS24 supports any audio hardware that works with Logic Audio. Users of Digidesign's TDM hardware can open as many as 32 instances of EXS24 using the Emagic System Bridge (ESB) TDM. EXS24 has no sample-recording abilities, but you can use Logic Audio to record and edit samples for use in EXS24.

Using Logic Audio, you can automate any parameter changes you make in the EXS24 plug-in window. When you record on an Audio Instrument channel assigned to an instance of EXS24, changing an EXS24 knob or slider produces MIDI data recorded for subsequent recall.

The heart of EXS24's user interface is the plug-in window's Editor view, which displays the main panel (see Fig. 1). With an appearance that emulates three-dimensional hardware, this view is the coolest-looking visual interface of the programs we surveyed. The window provides access to most sound-shaping capabilities, with a resonant filter, two envelope generators (EGs), and a pair of low-frequency oscillators (LFOs) as well as controls for tuning, panning, sample start time, glide, and Velocity response. Some sliders split into upper and lower halves, allowing you to specify modulation ranges. Poly, Mono, and Legato switches let you dictate how many notes will play and whether envelopes will retrigger when you press a key. At the top of the window is a pop-up menu for selecting from a list of Sample Instruments stored on your hard disk.

The lowpass filter has a choice of rolloff, with buttons to select 24 dB Classic, 24 dB Fat, 18 dB, or 12 dB modes. The Drive knob allows you to overdrive the filter for distortion effects, and the Key knob provides continuously variable keyboard tracking. At high resonance settings, the filter can self-oscillate.

The ADSR generators are “hardwired” to control amplitude and filter cutoff. You can modulate either envelope's attack time with Velocity; a split slider's two halves determine the minimum and maximum Velocity values.

Both LFOs offer a selection of seven subaudio waveforms. LFO1 is polyphonic, providing each note with independent modulation, and it restarts its cycle every time you play a note. LFO2 is monophonic; it runs continuously, simultaneously modulating all voices to which you assign it. That combination of LFOs offers maximum flexibility.

Switching from the Editor to the Controls view reveals a panel of fields and sliders and affords an alternate approach to editing the parameters displayed in the Editor. Values are shown in percentages or real-world numbers — decibels, milliseconds, and hertz. (In Editor view, you can see parameter values only as relative positions of sliders and knobs.) Although the Controls view doesn't look as spiffy as the Editor, you can quickly see precisely what's in it.

OTHER EDITORS

In Editor view, pressing the Edit button opens EXS24's Instrument Editor. Unlike the plug-in's Editor window, the Instrument Editor is shared by all instances of EXS24. In that window, you can assign samples to Zones, organize Zones into Groups, and assemble Groups into Sample Instruments. In each Zone, you can specify one sample's note or note range and enter its start, end, and loop points. Furthermore, the Instrument Editor lets you stipulate the volume, panning, Velocity range, and polyphony of all the Zones assigned to a Group.

If you put an Akai sampler disc into your computer's CD-ROM drive and select Akai Convert from the Instrument Editor's Instrument menu, a window will display the disc's contents. From there, you can easily convert programs, volumes, partitions, or entire discs into EXS24's native format. Additionally, you can listen to individual samples directly from the Akai disc. EXS24 automatically converts SoundFonts when you move them to Logic Audio's Sample Instruments folder and SampleCell files when you move their aliases there.

From the Instrument Editor, you can also open Logic Audio's Sample Editor for a selected sample; there, you can graphically position start, end, and loop points. You can also assign effects plug-ins and perform destructive sample editing in that window.

The small 58-page, spiral-bound manual has a brief tutorial. A well-rounded, 165 MB library of Sampler Instruments is also included to get you started. After registering EXS24, you'll get a CD-ROM with an additional 425 MB of samples, including duplicates from the 165 MB library. Emagic embedded EXS24's code in Logic Audio rather than on the EXS24 installation disc; the installer simply turns it on.

For users of programs other than Logic Audio, Emagic makes EXSP24, a sample player that runs with any VST 2.0 host on the Mac and PC. It converts SoundFonts, but not Akai or SampleCell files, into its native format. The Instrument Editor is missing, and you can't create Instruments without EXS24 for Logic Audio.

GIGASTUDIO 160 2.2

If third-party sample-library support were the main criterion, NemeSys's GigaStudio would win the prize as the mother of all software samplers. Dozens of dedicated sample libraries are available for the standalone application, ranging from multi-CD timpani collections to extraworldly ambiences. In addition, GigaStudio has won the support of many PC sound-card manufacturers that have developed custom drivers providing near-zero latency for GigaStudio users.

GigaStudio comes in two versions: the top-of-the-line GigaStudio 160, which offers as many as 160 notes of polyphony, and the slightly scaled back GigaStudio 96, which has 96 notes. With significantly fewer features, the original GigaSampler series is also available at greatly reduced prices.

GigaStudio's opening screen is optimized for creating multitimbral and multichannel setups. Unlike most hardware samplers, GigaStudio lets you assign more than one patch to a MIDI channel. It also supports as many as four streams of 16 channels each, so you can easily layer four patches on one channel. Each patch can contain multiple layers, providing even more options for creating extremely rich textures.

LOAD IN

At the top of the Main Mixer screen are 16 channel slots for loading your samples and controls for adjusting each channel's volume, pan position, and tuning (see Fig. 2). You can reprogram the sliders to generate any MIDI controller, mute or solo a channel, and adjust the channel's routing to the internal effects engine and your sound card. Clicking on the small arrow to the right of the patch name opens the Info screen, in which you can pick any loaded Instrument, jump to the Instrument, and view performance parameters for that channel.

The middle of the screen has buttons to access the three other main work areas: Loaded Instruments, MIDI Control Surface, and Distributed Wave. GigaStudio's basic structural element is called an Instrument, which is a set of keymapped zones that contain one or more samples. The Loaded Instruments screen displays all the Instruments that have been precached, which is a process GigaStudio uses to achieve near-zero latency (in brief, it loads a small chunk of each sample into RAM). Assign, at most, 16 controllers per channel in the MIDI Control Surface screen and bring new samples into the program in the Distributed Waves window.

An Explorer-style window in which you can access all the WAV files on your system is at the bottom of the main screen. Grab one or more samples from your drive, drop them into the Distributed Waves window, right-click, and select Build Gig from Loaded Waves, and you have a new patch.

In the Instrument Editor (a standalone application that runs without loading the main GigaStudio interface), you can modify many patch parameters. You'll find screens to configure key and Velocity zones; set and enable loop points; and add amplitude, filter, and pitch envelopes. Some parameter values have limits — for example, the maximum attack time for a pitch envelope is 10 seconds — and the amplitude envelopes are ADSR only. Additional flexibility in those controls would be nice.

MAKE IT QUICK

Numerous features simplify your work flow. The QuickSound feature, for instance, scans your hard drive at startup and identifies any new sample files. The program can automatically build keymaps for samples QuickSound locates, which makes creating large collections of sound effects easy. Also, it is simple to set up links to the other applications — a sequencer and audio editor, for instance — that you will no doubt use in a GigaStudio session.

GigaStudio's small number of effects algorithms is one of the program's few drawbacks. If you're doing acoustic-instrument emulations, though, that shouldn't be an issue. The program ships with only three effects, and you receive a fourth, free, when you register. (NemeSys reports that other third-party effects are in development.) The effects sound great and are easy to apply and work with; the reverb quality is especially good. Creating effects automation is simple and lets you design interesting, time-varying sounds.

GigaStudio is not the best option for creating avant-garde or alien soundscapes, though it is great for playing back samples you created elsewhere. It's an excellent choice for anyone experienced with hardware samplers and for novices as well. The manufacturer made arrangements with Hollywood Edge to create QuickSound-compatible versions of many of its best sound effects, and that will certainly make it one of the most efficient sound-effects and Foley platforms around. Rumor has it many Hollywood composers have dumped their hardware and added GigaStudio to their racks — not surprising, given the power and intuitiveness of the software.

REAKTOR 3

Native Instruments recently released version 3 of Reaktor, its flagship modular sound-design software, and processing and playing samples is an important part of the package. (Reaktor integrates two formerly separate NI products: Generator, a soft-synth builder; and Transformator, a dedicated sampling app.) Reaktor ships with several dozen example sample players that use the program's numerous sound-manipulation tools. Right out of the box, you will find time stretchers, sample granulators, as well as sample players that have been optimized for loops and drum sounds. You can substitute your own audio files for any files used by the existing players, mixing and matching components to create unique sound-design tools.

Reaktor runs as both a VST and a DirectX plug-in, and it can also be used as a standalone. More than just a sampler or sample player, the program is a toolbox full of sound-generating and processing modules that you can combine in nearly any combination. Winner of an EM Editors' Choice award, Reaktor offers everything from common synth elements (such as oscillators and envelopes) to low-level mathematical functions. The software is a tweaker's delight, yet if you stick with the hundreds of examples that come with the software, no assembly is required.

When you open an example, you'll likely see a massive array of knobs and sliders with abbreviated identifying labels. For example, News Cool Meat Chopper presents a multitude of parameters (see Fig. 3). Some divulge their names and functions when you hold the cursor over them, and others don't. (If no text is displayed, the patch developer hasn't included any.) If you leave the Properties dialog open, you can make changes to the displayed parameters, which are reflected immediately in the open device.

Building a multitimbral sampler in Reaktor is a straightforward process. Beginning in the Structure window, select 1 or more of the 24 basic sample Instruments and paste it as many as 15 (if it's mono) or 7 (if it's stereo) times. Then, connect the outputs of each Instrument to the Audio Out. All of the sample Instruments include preloaded example audio files, so you should be able to hear audio as soon as you connect them.

Next, set the MIDI receive channel in the Instrument Properties panel for each instrument. If you have a multichannel sound card that supports ASIO, you can direct each sampler to a different physical output; if you don't, mix all the signals to one stereo output. Substitute your samples for the default ones, and you're set.

The number of processes you can add to your samplers is astounding. If you're interested in waveshaping, drag a WaveWarp module and insert it between the sampler and the Audio Out. How about a morphing vowel filter? That's also available to process samples, as are other filters and vocoders, dynamics processors, pitch and time shifters, chorus and delay effects, and more. Additionally, you can add, delete, or modify modules, even as a sample plays back. Reaktor automatically incorporates any adjustments you make in real time.

ANALYZE THIS

Many of Reaktor's sample-transformation features are based on analysis and resynthesis techniques. Those features are among the program's most impressive and are unlike anything short of Symbolic Sound's hardware-assisted Kyma system. When you load a new file into a resynthesizing sampler, Reaktor performs a quick spectral analysis of it, which it then uses for controlling the sample's playback rate and pitch. You can save the analysis data, which makes it load more quickly next time, or you can discard it when you close the work session.

Among the modules that use the resynthesis process is the Sample Pitch Former. This versatile module lets you control a sample's pitch and formants independently. Like other Reaktor samplers, you can incorporate a map of as many as 127 sample files in the Sample Pitch Former and then configure the sampler to move through them automatically as it plays back. You can also slow down the sample-position pointer until it reaches a dead stop and freezes on a single sample. For manual control, assign the position pointer to a slider or fader, or use an LFO or even a random function to automate it.

One reason Reaktor has such a large and loyal user base is that Native Instruments has made it quite easy for users to share patches. Hundreds of user-contributed example files are at the company's Web site. Moreover, the thorough documentation — complete with tutorials and a reference entry for the more than 200 modules — makes getting started with the initially austere-looking program much easier.

Reaktor doesn't have as many ready-made templates as some other programs. Unlike GigaStudio, for example, Reaktor doesn't let you drag 16 samples to different channels on the interface to build a multitimbral configuration. Because Reaktor's architecture is so open and flexible, however, with a little effort, you can create nearly any sampling process imaginable. Reaktor 3 is as close to the “endless sampler” as any you'll find.

SOFT SAMPLECELL 3.O

Digidesign's Soft SampleCell is the latest incarnation of a computer-based instrument that, until recently, required dedicated digital signal processing (DSP) hardware in one of your computer's expansion slots. Introduced more than a decade ago, SampleCell is a popular sampling platform for Macs and PCs. The first SampleCell, a joint project between Digidesign and Opcode, was a Nubus card with editing software only for the Mac. A PCI version was later developed for both platforms.

The latest edition, Soft SampleCell 3.0, works just like previous versions, but one element is missing: the expansion card. Consequently, Soft SampleCell costs $950 less than its PCI-based counterpart yet offers the same functionality and more, including a 4-pole resonant filter. Instead of dedicated DSP hardware, Soft SampleCell harnesses the power of your computer's processor. Now almost anyone with a modern Power Mac, a compatible floppy drive (for the authorization disk), and sufficient RAM to run OS 9.0.4 can run SampleCell, even on a computer without PCI slots.

Soft SampleCell is a Mac-only application that requires Open Music System (OMS) for MIDI communication. If you're a Mark of the Unicorn Digital Performer user who refuses to give up FreeMIDI in favor of OMS, Soft SampleCell won't suit your needs. If you have Digidesign hardware, you have a choice of DirectConnect or Direct I/O for multichannel audio output; otherwise, you have to use Sound Manager, which limits you to stereo output. If you invested in expensive ASIO hardware with multiple outputs, you're out of luck.

Soft SampleCell can open AIFF, WAV, and Sound Designer I or II files, but it has no facility for directly recording samples. It doesn't import non-native sampler formats such as Roland or Akai; however, translation programs can convert sampler discs to SampleCell's native format. SampleCell is such a popular format that several other virtual samplers support it.

TAKE IT TO THE BANK

The main window in Soft SampleCell displays a bank of Instruments, each containing a sample or a multisample in mono or stereo (see Fig. 4). The number of Instruments in a bank determines the window's width. Each Instrument appears in what looks like a channel strip showing its name, MIDI channel, key range, audio outputs, panning, and level.

A button at the top of the Bank window opens the Sample Map window, in which you import samples, preview them with an onscreen keyboard, and define their Key Groups and Velocity Zones. Another button opens Misc Parameters, which contains controls for pitch-wheel range, tuning, and Velocity response. Also in the Misc Parameters window, you can specify the lowpass filter's frequency, resonance, and slope (either 6 or 24 dB per octave). The Resampling Quality pop-up, at its highest setting, helps to minimize audio artifacts when pitch is transposed more than seven semitones.

Other buttons let you assign three EGs, two flexible LFOs, matrix modulation, and parameter response curves. In addition to the usual attack, decay, and release times, the EGs provide both a sustain level and a sustain decay time. Matrix modulation offers 18 sources and 24 destinations, adding flexibility that's missing from many samplers. You can use Velocity to control the sample start time, for example, or control filter resonance with Aftertouch.

A button in the Sample Map opens the Loop window, an editor for defining loop points and trimming audio. That integral loop editor is the only way you can edit the contents of sample files in Soft SampleCell, whether they're located in open Instruments or on disk. For each file, you can specify a single sustain/release loop or one loop for sustain and another for release. The program can create crossfades and automatically locate potential loop points. A Waveform Display shows the entire sample, and a Loop Display zooms in on the loop points. An especially welcome feature is the ability to adjust loop points as the loop plays.

It is disappointing that Soft SampleCell has no effects processor, but you can create auto-panning and simulate many effects in the modulation matrix. To add reverb or compression, you can assign plug-ins to SampleCell's audio track in Pro Tools or your sequencer application.

Soft SampleCell's 100-page manual comes in paperbound and bookmarked PDF versions, giving the advantage of holding a book in your lap while searching its contents in Adobe Acrobat.

Soft SampleCell appears to be optimized for use with a Pro Tools system; however, it should be useful with any sequencing program that supports OMS. Multichannel audio requires Digidesign hardware, so those users who do not use Pro Tools are limited to stereo output. Although the program doesn't support third-party audio hardware and Digidesign feels no need for multiple instances, Soft SampleCell's well-conceived feature set makes it a flexible sample player.

SAMPLERCHAN 1.3

Nick Whitehurst's SamplerChan is the newest soft sampler on the market, having come out of beta just days before this writing. It operates as a VST plug-in but doesn't run using FXpansion's VST — DX Adapter within Cakewalk's Sonar. The program is 16-part multitimbral and supports samples as high as 32-bit, 44.1 kHz resolution. It provides special features for triggering samples and can automatically generate rhythmic and strumming note patterns.

SamplerChan's interface is split into three main work areas: Mixer, Editor, and Wave (see Fig. 5). Each has features and provides access to options related to different stages of your work flow. In the Mixer view, you set the levels for each of 16 MIDI channels; assign a Program Change and, if necessary, a Bank Change number to trigger the sound on the current channel; adjust the two effects-send levels; and perform various file-maintenance operations. The large graphic Mixer interface is nicely laid out, but you can't resize it.

Building a multitimbral setup in the Mixer is a breeze. First, drag a file from the Explorer-style window on the left of the screen directly into any of the 16 channel slots; then repeat as needed as many as 15 times, and you're set. If desired, you can specify a maximum polyphony (at most, 64 notes) for each channel. You can also enable the in-line limiter and route each channel to any of the eight possible outs on a per-channel basis. To patch a channel into the plug-ins provided by your host software, assign it to either of the two effects sends.

SENIOR EDITOR

SamplerChan, like many of the other programs, lets you work at various levels of its architecture, but some of the Editor's performance parameters are unique. For example, like the other samplers in this roundup, SamplerChan's patches consist of samples assigned to parts containing various splits and layers. Parts are combined into instruments, which are stored in banks. In addition to the familiar tuning and volume controls that you can assign to a sample, the program offers numerous ways to add nondestructive rhythmic components that affect the way a sample plays back.

Using the tools under the Rhythm tab in the Editor, for example, you can build long sequences of note events that trigger your sample each time a MIDI Note On is received from the host. We grabbed a single bell tone from the hard drive, loaded it into SamplerChan, and then created a 24-note sequence that altered the sample’s pitch and Velocity. When Steinberg’s Cubase played a middle C, SamplerChan played the bell with the 24-note sequence we had created. We continued to add layers to the patch by loading other drum samples and building unique sequences for each of them. Before long, we had some elaborate, evolving polyrhythmic textures playing back in sync with the tempo of the Cubase sequence.

The Editor screen also provides features that are especially suited for working with drum sounds. Switch the display to Drum mode and drag any number of drum samples from the file area at the screen’s left onto the Editor. SamplerChan automatically assigns each sample to a different key, and you have an instant multisampled kit.

SamplerChan offers robust matrix-modulation features. In addition to traditional modulation sources, such as three multiwave LFOs and three five-stage EGs, it has an envelope follower that tracks the sample’s amplitude envelope and allows you to use it as a mod source. You can control the pitch of a sample with the amplitude-curve data, for example; as the source sample grows louder, the destination sample’s pitch gets higher. A less-familiar modulation destination is the effects-send level; modulating it with an LFO, for instance, creates an effect that fades in and out during a recurring time frame.

In the Loop

Access to SamplerChan’s extensive looping features is obtained in the Wave screen. You have options to automatically fade loop points in or out, fine-tune a loop’s pitch and length, and set the loop direction (forward, backward, or bidirectional). Like Propellerhead Software’s ReCycle, SamplerChan offers tools to slice a loop and manipulate the slices, letting you slow down or speed up the loop without changing its pitch. You can even assign each slice to a different key, providing options to completely modify or mangle the original loop or simply add a little swing to it.

By pressing the Extract Groove button, you can store a copy of the time markers associated with a sliced loop and reuse them elsewhere. Those and other loop-manipulation features remove SamplerChan from the standard soft-sampler world and add considerably to its value as a composition tool.

Perhaps more than most soft samplers, SamplerChan is a live-performance application. One useful performance feature is the aptly named Performer. The Performer’s virtual joystick (and many other controls) lets you move among four quadrants. Like the physical joystick on certain Korg synths, you can map the joystick movements to numerous performance parameters, such as filter cutoff frequency, LFO depth, and the length of an envelope segment. You can even map the movements to the levels of different parts so that samples fade in and out as you move the stick around. You can record all joystick movements in your host software to re-create or automate the performance of your sounds. The Performer is a great feature that, by itself, is worth the sampler’s price.

SamplerChan’s Options screen offers various ways to customize the software. You can link to as many as three external applications directly within the program, optimize various aspects of its performance, and specify as many as 99 levels of undo. The documentation (in PDF only) does an especially good job explaining an audio signal’s flow through the program, and it includes clear graphic flowcharts. Lucid explanations help you understand how best to exploit SamplerChan’s resources.

SamplerChan is the most unusual program in our roundup. It’s an excellent choice if you’re interested in working with loops and grooves but don’t yet have dedicated tools for the job. It also makes sense for anyone who wants a more traditional soft-sampler experience. If the developer adds DirectX Instrument support and works out the bugs with VST—DX Adapter, non-VST users will have a chance to use SamplerChan.

SampleTank 1.0

A sample player in the form of a VST Instrument, IK Multimedia’s SampleTank comes in DJ ($199), L ($249), and XL ($499) versions; the main differences are the size and scope of the included sample libraries. SampleTank DJ and L are unable to open instruments found in only the XL library; you can upgrade by paying the difference in retail price. XL also includes ST Converter, a utility that converts Akai S1000 and S3000 files to SampleTank format.

SampleTank’s installer disc has 289 MB of mostly multisampled instruments, in addition to 182 MB of third-party demo sounds. Installing SampleTank DJ from a second disc has an additional 323 MB of loops, electronic drums, and synth samples. Move up to the L level for a wider palette ranging from electric and acoustic guitars to orchestral and vocal sounds totaling 593 MB together with the basic and DJ instruments. SampleTank XL supplies another 613 MB and increases the variety of sounds even more, adding up to a grand total of 2 GB. The quality of the included sounds varies considerably, but most sounds we tried out are pretty good. All instruments and samples are in a proprietary format, so you can only open them in SampleTank.

Because it’s a VST Instrument, SampleTank requires a VST 2.0—compliant sequencer program, though it loaded with no problem in Sonar using the VST—DirectX Adapter. Depending on CPU speed, its maximum polyphony is 128 notes. With enough processing power, you can run as many as 8 copies simultaneously in Cubase VST and 16 copies in Logic Audio. Each copy responds on 16 MIDI channels, with a different instrument on each channel.

SampleTank appears onscreen as a plug-in window with many knobs, buttons, and a display area (see Fig. 6). Choose specific instruments from a category list that appears at the center of the display area. Select one instrument for each MIDI channel, which you specify with a row of 16 buttons, and stipulate the instrument’s volume and pan position. Clicking on an Info button reveals supplementary information for each instrument, such as multisample size and the copyright holder’s name.

Written in Stone

Instrument editing is minimal: because SampleTank is a sample player in the strictest sense, reassigning samples or key ranges isn’t possible. Those capabilities would be a welcome addition to a future revision, along with user-defined envelopes and filters.

In any instrument, however, you can control as many as four real-time parameters using onscreen knobs or MIDI Control Change (CC) messages. Parameter names appear above the knobs and in greater detail on the display’s left side, just below the current instrument name. If you click on the parameter value that appears just below each knob, SampleTank displays its MIDI CC number and its value. Because you can change parameters with MIDI messages, you can automate changes by recording them into your sequencer. You also can save any changes you make to an instrument’s default parameter settings.

With pull-down menus containing 27 effects algorithms, you can assign as many as four effects to each instrument. Four slots for effects each have an On toggle button and an indicator LED. The first slot always provides compression and four bands of EQ; the other three are user-programmable and range from envelope filter and rotary speaker to Slicer bpm, an effect that transforms a sustained sound into a rhythmic phrase. When you click on a slot, four or five adjacent knobs are labeled with appropriate parameters. You can control all parameters with MIDI CCs and sync some of the effects to MIDI tempo.

SampleTank’s Search function allows you to enter keywords like "Indian" or "180 bpm" to quickly find samples appropriate for your needs. You can specify the number of notes each instrument can play, thus controlling its CPU usage, and indicate whether an instrument plays polyphonically or monophonically.

SampleTank comes with a small stapled, 40-page manual that’s also provided as an Acrobat file. Each SampleTank version includes a list of instruments in paper and PDF formats.

Stella9000 2.5

Stella9000 is part of Koblo’s Studio9000 suite of software instruments for Macintosh. Based on Koblo’s Tokyo engine, Studio9000 includes three soft synths; a virtual drum machine; and Stella9000, an 8-note polyphonic sample player. Stella9000 works as a VST plug-in instrument, as a standalone program, or with any software that supports OMS or FreeMIDI for MIDI input and output. For audio I/O, Tokyo supports ReWire, Sound Manager, DirectConnect, Direct I/O, and MAS in addition to VST 2.0.

Stella9000’s highlights include freestanding operation, an amazing multimode filter, three invertable ADSR generators, a sophisticated arpeggiator, and a modulation matrix with 16 sources and 28 destinations. On the downside, Stella9000 is 8-note polyphonic, and each instance runs on a single MIDI channel. Surprisingly, it does not have an effects processor or facilities for editing samples or rearranging multisample keymaps, though you can control a sample’s start point.

Stella9000 opens SampleCell instrument keymaps (as long as they’re assigned to a single MIDI channel), and it reads AIFF and Sound Designer II audio files. (A WAV translator is available from Koblo as a free download.) You can select from more than 250 factory programs in the Presets menu; most are effects or other sounds oriented toward dance and techno music, and many sound a little out of date. It’s been almost two years since Studio9000’s last release, and it’s due for an update to compete with stiff competition.

Sound-shaping Tools

Stella9000’s architecture is synthlike; it’s almost identical to Vibra9000, Studio9000’s most refined virtual synth. However, in place of the synth module’s oscillator waveforms are samples and multisamples. All of Stella9000’s controls transmit MIDI, and you can control every parameter from your sequencer or any external MIDI source.

One of Stella9000’s especially welcome aspects is that it has only one window, which lets you see what’s going on at all times (see Fig. 7). The onscreen front panel contains 68 knobs, 18 buttons, and many stylized pop-up parameter displays. Flashing LEDs provide plenty of visual feedback when data is sent or received. The panel is divided into sections labeled Sample, Filter, Env, LFO, Arp, Mod, and Global. The Sample section has controls for sample-playback offset, tuning, Velocity modulation, and pitch-bend range, as well as amplitude attack and release. Buttons turn on looping and reverse playback.

Stella9000 has the same Filter section as Vibra6000 and -9000, and it’s one of the coolest filters ever. In addition to balancing the output of simultaneously available lowpass, bandpass, and highpass filters, you can change their stereo positioning with the Spread knob. A pop-up menu reveals eight types of filters, including 2-pole, 4-pole, 8-pole, and a notch filter with two 4-pole filters. Two comb filters produce resonances at multiple frequencies, and there are combinations of two or four 2-pole filters. Additional knobs control cutoff, resonance, key tracking, Velocity sensitivity, and distortion.

Stella9000’s three LFOs have separate unipolar and bipolar outputs and a choice of six waveforms. Each LFO has its own attack-decay envelope, and dedicated lowpass filters can reshape the LFO waveforms. You can sync LFO rate to sequencer tempo.

The modulation matrix provides eight source-to-destination routings, each with its own depth control. Modulators include envelopes, LFOs, note number, pitch bend up or down, and real-time controllers. Destinations can be nearly any front-panel parameter, including modulation depth.

Studio9000’s minimum system requirements are more modest than some: you need 40 MB of free RAM, Mac OS 8.0, and 16 MB of disk space for installation. You can select which instruments to install–Stella9000; Gamma9000; or Vibra1000, -6000, and -9000–depending on your resources and requirements.

Unity DS-1 2.1.3

BitHeadz’s Unity DS-1 was the first full-function virtual sampler available for Macintosh, and it’s still one of the few virtual samplers that records samples. With comprehensive sample editing, extensive modulation routing, and the ability to work alone or with all popular sequencers, Unity goes beyond emulating hardware and embraces the advantages of software-based samplers. On the Mac, nothing else comes so close to reflecting a hardware-based sampler. A scaled-down Unity Player ($199) is also available from BitHeadz.

Unity is optimized for G4 and Pentium III processors, but it will also run on a 200 MHz PowerPC 604 or a Pentium II. On Macs with dual processors, Unity processes most voices on the second processor and leaves the main processor available for your multitrack audio-recording program. With sufficient CPU power, Unity plays as many as 256 phase-locked stereo voices. You can open as many as eight copies of Unity at the same time, providing 128 independent MIDI channels.

To make efficient use of memory, Unity DS-1 is a suite of applications divided into Unity DS-1 Editor, MIDI Processor, Keyboard, and Mixer. At first, so many applications appear to be a hassle, but to play Unity as a standalone program, you only need to run one. When you use a sequencer, the Unity engine runs in the background, so you don’t have to open any additional programs. Also included are a number of utilities for organizing sample libraries, playing Standard MIDI Files (SMFs), and accomplishing other tasks.

In the Unity DS-1 Control Panel, you can specify parameters such as the maximum RAM (as much as 1 GB), maximum polyphony, sample rate, Velocity curve, MIDI and audio I/O devices, controller routings, and other details of Unity’s operating environment. The Control Panel gives you the most direct control of how the Unity engine functions.

For live performance, Unity DS-1 MIDI Processor lets you speedily select and play sounds, set up splits and layers, and use Unity’s sophisticated arpeggiator. You can modify tuning, panning, effects sends, and other real-time parameters as well.

The Mixer displays 16 MIDI channels, each as a different channel strip, with a master section for controlling overall volume and global-effects depth. You can select sounds, mute and solo channels, and control the same real-time parameters that you can with the MIDI Processor.

Sample Editing

Whether you’re tweaking programs for a particular project, creating samples from scratch, or retooling a sample library created for another sampler, the heart of Unity is the Unity DS-1 Editor. The Editor provides tools to manipulate samples, multisamples, and programs.

The Editor’s Samples page is a full-fledged sample editor in which you can view and edit every waveform in the current bank (see Fig. 8). You can define sampling rate (as high as 96 kHz), bit rate (from 8 to 24), root note, fine-tuning, tempo, loop points, and other details. Converting mono samples to stereo (and vice versa) is easy using a single pop-up selection.

The Samples page’s Munge menu provides powerful audio-editing functions, including normalize, pitch shift, time stretch, change length, split, invert phase, and crossfade loop. Although there’s no dedicated loop-editing window, an auto-loop feature lets you scroll through potential loop points. You can even play the sample continuously as you adjust loop points.

The Record button on the Samples page opens the Record window, where you record samples in stereo from your computer’s audio inputs. Although you start sampling by hitting the Record button, you can’t set a threshold that triggers Record as you can with most hardware-based samplers.

The Multisamples page is where you insert samples into the different zones in a multisample. Zones can be displayed in an onscreen keyboard or a list. You can designate each sample’s key and Velocity range, start point, tuning, volume, panning, and so on. You can even define crossfades between zones, mixing a sample from one zone with another at the split points at which one zone ends and another begins. That helps to minimize transitions between zones.

The Programs page is divided into Configuration, Main, Modulation, Effects, and Global sections, each of which is accessed by clicking on tabs. On any Programs page, you can open, add, and delete programs from a bank. An overview of the program’s algorithm appears on the Configuration page, where you can graphically route oscillators, filters, and effects to the output.

The Main page displays parameter controls for two oscillators and two filters. There are 16 filter types, including lowpass, highpass, bandpass, and allpass with varying characteristics, slope, threshold, and comb.

You can route 12 sources to 60 destinations on the Modulation page as well as program flexible LFOs and five-stage EGs. The number of modulation routings, LFOs, and envelope generators is practically unlimited, depending on your computer’s horsepower. LFOs can sync to your sequencer’s tempo.

The Effects page is divided into four quadrants containing controls for two insert effects and two global effects. Pop-up menus provide lists of effects types, with different sets of parameter sliders for each effect. Insert effects include everything from flanger and multitap chorus to parametric EQ and distortion. Global effects are limited to reverb, reflection, and delay.

The Unity DS-1 Manual is a 274-page PDF file that’s clearly written and well indexed, though bookmarks would be nice. Unity comes with a well-rounded, 300 MB sample library that includes a 12 MB General MIDI (GM) bank. An abundance of Unity-format discs, ranging from classic keyboards to percussion hits and breakbeats, is available from BitHeadz and other manufacturers.

Unity DS-1 is a mature program with a lot of function and flexibility. If other samplers replicate its breadth of capability, many people will find replacing samplers with computer software much easier to accept.

Virtual Sampler 2.56

Speedsoft’s Virtual Sampler runs as either a VST or a DirectX Instrument and also functions as a standalone application. It’s a surprisingly deep program considering its price and offers advanced features such as the ability to draw filter curves and LFO shapes by hand and employ any of 15 Velocity curves. As an added bonus, it can load VST Instruments and pass their output through its processing stages.

Virtual Sampler is available in two versions. The standard version ($40) has all the features of the professional version ($75) except ASIO support and the ability to import Akai samples. In addition, rather than functioning as a DirectX or VST Instrument, the standard version works only with DirectX.

Virtual Sampler provides as much as 64-note polyphony and 16 discrete outputs. Its main screen toggles among 12 work areas, only one of which can be open at a time (see Fig. 9). An onscreen keyboard appears at the bottom of the display, regardless of which work area you’re in. Various icons, such as those for file management, are always accessible at the top of the display.

The program has the ability to manage enormous amounts of data. At the bottom layer of its architecture are Splits (keymaps) that define the range over which individual samples are triggered. Splits include various performance parameters, such as an amplitude envelope, two filters, two effects, two LFOs, and a setting for both pan and detuning. A single Virtual Sampler voice can have a maximum of 65,535 Splits, though it’s unclear what you would do with so many.

A Preset is the next higher level and includes a MIDI Channel assignment, a discrete output route, and a location in a Virtual Sampler Bank. A Bank holds as many as 128 Presets, and each is assigned a Program Change number according to its location within the Program map. You can load a limit of 64 Banks during any work session. Like some hardware devices, Virtual Sampler uses the term Multi for a specific setup of 16 Presets, one per MIDI channel. You can store as many as 16 Multis and access them with a combination Bank/Program Change command. Fortunately, Virtual Sampler’s Bank Manager feature keeps track of the files that you’ll accumulate.

Tweaky Clean

A list of the currently loaded samples appears on the Sample page, with controls to set the sample start and end points, adjust loop points, and perform various DSP functions. A large graphic display of the selected sample appears, and though you can zoom in to a high level of magnification, no increments show you precisely where you are in your sample. One handy feature lets you see exactly where any given sample is being used; just click on the Used column for a list of the Presets employing the sample, complete with the sample’s Bank and Split locations. You can also replace that sample with a new one and retain all the settings and performance parameters of the original.

Assigning Splits to key zones is simple using the dedicated Zones screen. On that screen, you can either use the Hi or Lo Key scroll controls to set a range, or drag the beginning or end of the horizontal bar that represents each Split in the Split window. By toggling the View menu to Velocity, you can adjust the Velocity zones for each Split without leaving the work area.

Virtual Sampler has manifold effects and can use VST plug-in effects even in the standalone version. Its envelopes are particularly easy to use. The single five-stage amplitude envelope can have logarithmic or linear segments on a segment-by-segment basis. If you designed an envelope but want to change the time it takes to evolve, click on the EG-multiplier button and scale the envelope from 1 to 50 times in single-unit increments (15, 25, 35, and so on). You can save as many as three envelopes and reuse them during the same work session too.

Many conveniences make your work flow smoother. One is the ability to preview a Preset by triggering it with a single note or with one of several chords. If you want to hear what the preset sounds like played as a major seventh chord, just set the Chord dialog to that chord; whenever you play a single note on the virtual keyboard, you’ll trigger the full chord. One minor nuisance is that when you hold down an arrow key to change a parameter value, the values do not scroll, so you must continually reclick the mouse to increase or decrease the value. As a work-around, you can just double-click on the parameter field and type in the desired value.

Virtual Sampler can import WAV and AIFF samples, SoundFonts, and even Steinberg LM4 drum kits. It also reads AKAI S-1000 and S-3000 data and retains presets (including Splits) and samples (including loop points, panning, and volume). We have hardly touched on the range of customization options, from altering the program’s response to controller messages to scaling key tracking, which rival those of some hardware samplers. The program includes well-written and thorough documentation and was stable throughout the testing period. Virtual Sampler is a real bargain and well worth a look.

Volkszämpler 1.0

Noted for its sampling hardware, CreamWare recently released Volkszämpler, a software-only application that runs as a VST plug-in. The program ships on a single CD-ROM and includes a modest (by our standards) 70 MB of samples from diverse commercial collections. It offers as many as 32 stereo notes of polyphony, and you can run multiple instances if needed. A Mac version should be available by press time.

Volkszämpler loads a startup screen with four images appearing across its center. Clicking on an image opens one of the four separate work areas, which is a nice arrangement because it lets you view only the area you want to work in (see Fig. 10). However, with multiple windows open, the screen can easily become cluttered, and you can’t arrange or tile the open panels. Nor is there "flyover help" or even any text describing what each screen is for. After you become familiar with the software, that won’t be a problem. You can resize only some of the windows, and none use the standard Windows conventions to indicate that they have that capability. Once you get past the nonstandard interface, however, you’ll discover useful options.

From the Top

You must create a new Program (Volkszämpler’s basic patch) before you can audition or load samples. Access samples from the Sample Browser window, which toggles between a view of your samples and Program files (a maximum of 999 Programs can be loaded in one work session). To load a sample, drag it from the Browser to the Keygroup List. If the Sample Editor window is open, a waveform display appears.

In the Keygroup window, you choose a key range by setting the High and Low key points. Then open the Options subwindow to access additional performance parameters. The Options window provides access to four screens: Global, Zones, Filter, and Envelopes 1 and 2. In the Global screen, you set Tuning, Volume, Pitch-modulation sources and amounts, and LFO levels for all the Zones in a program. If you move to the Zones screen, you can work in even finer detail on individual Zones, setting a tuning offset and Playback mode (one-shot, looping, and so forth), for example. The Zones area splits into six additional subwindows, and it’s difficult to remember each setting’s location. To illustrate, Loudness and Pan settings for each Zone are in the Main Zone window, but Tuning and Filter controls are in the Additional window.

You can tweak your Program using a single filter, but you’ll find only resonant 2-pole low-, high-, and bandpass types. Modulating the filter resonance and frequency is easy, but a small glitch in the current version makes it difficult to select all the sources using the mouse (the manufacturer says that will be fixed in the next upgrade). Also at your disposal are two ADSR envelopes that are a cinch to manipulate by entering numerical values or by dragging on the envelope display itself.

You can make adjustments to parameter values by dragging with the mouse; in most cases, moving vertically produces coarse changes, and moving horizontally provides finer control. Further, you can type values directly into most parameter fields. Sometimes, the range of parameter values isn’t intuitive; filter frequency ranges from 0 to 99, for example, and there’s no indication of how that corresponds to hertz values.

In the Sample Editor screen, you can perform standard edits such as cut, copy, and paste as well as reverse and extract (crop). You can also access the two built-in DSP functions, Normalize and DC Offset. Zooming is easy: just highlight a range while holding down the mouse button; when you release, the display fills with the highlighted range. You can use the + and — keys on your keyboard (but not on the number pad) to zoom within fixed levels. Recording (threshold with a prerecord option, manual, or MIDI-triggered) occurs in the Sample Editor window as well. Before recording, you must assign Volkszämpler to an insert in your host software and create a new Program.

Volkszämpler packs enough features to make it a useful tool for adding sampling capabilities to your rig. Many features, though, are buried in sub- or even sub-subwindows, and keeping track of feature locations is often hard. Moreover, we ran into occasional display glitches, even with an update from the manufacturer.

Still, the program supports files with resolution as high as 32-bit, 96 kHz. It is also among the more efficient samplers in the group. We layered ten 16-bit, 44.1 kHz mono samples in a single Program using extensive modulation, and Cubase’s VST Performance meter bounced comfortably between 10 and 20 percent. With a bit of added development, perhaps including a little design rearrangement, Volkszämpler could be a good choice for your first soft sampler.

VSamp 3.2

Malcolm Haylock’s VSamp costs a fraction of what other virtual samplers cost, yet it probably performs most sampler functions that you need, and it’s easy to set up and use. In addition to a standalone version that works with OMS or FreeMIDI, you also get VSamp VST, a plug-in instrument for VST 2.0 hosts. The 20-page PDF manual clearly explains every window, command, and function. VSamp runs on any Power Mac, Centris, or Quadra.

VSamp offers sample playback on 16 MIDI channels simultaneously, with as many as 64 notes and 128 samples per channel. On a fast computer, more than a thousand notes of polyphony are possible. Despite its benefits, VSamp doesn’t record samples, and it has no effects processing. The program offers stereo output through the Sound Manager, and the plug-in version provides four assignable stereo outputs.

Samples are assigned to Instruments; Instruments are arranged in Banks. Samples must be in the same folder as their associated Instrument and Bank. VSamp provides windows for editing Instruments and Banks but has no sample waveform editor; setting loops requires a separate audio-editing program.

VSamp’s appearance is utilitarian and not very exciting. Its Instrument Editor looks like a spreadsheet with all samples listed in a column and essential parameters displayed in rows (see Fig. 11). Each row shows a sample’s name, file type, lowest and highest pitch, original pitch, pan position, and Velocity range. Click on the Sample field to insert a sample; you enter most parameter values by typing them into their appropriate fields, but you can enter pitch data by clicking on the field and playing a note on your MIDI controller. You can also copy and paste data between fields.

At the top of the Instrument Editor is a checkbox for Track, which allows all the Instrument’s samples to change pitch in response to MIDI notes; uncheck it for unpitched drum samples. Amplitude, Filter, Tuning, and LFO buttons open another window each. The Tuning window lets you specify pitch offsets and pitch-bend range. The LFO window supplies four knobs for setting frequency, mod-wheel depth, and the amount of pitch and filter modulation.

Open Windows

An AHDSR envelope is displayed graphically in the Amplitude window; view or enter values for each stage in five corresponding fields. Two knobs determine how the envelope responds to Velocity and Expression (CC 11). You can assign an amplitude envelope to either an entire Instrument or one sample at a time.

The lowpass filter has six knobs: Frequency, Resonance, Pitch Mod, Velocity Mod, Breath Mod, and Gain Trim. Click and drag to change their values or type in the corresponding fields. Values entered in the Filter window are applied to the whole Instrument. Regrettably, the filter has no envelope control.

The Bank Editor window contains a list of a total of 16 Instruments that you assign to MIDI channels and audio outputs. (Unless you’re using VSamp VST, your only audio output is Output 1, the Sound Manager.) You can assign as many as four Instruments to the same MIDI channel. Add Instruments to a Bank by clicking on an Instrument field and selecting from a list in the file dialog.

When the Keyboard window is open, you can play notes with your mouse and play a range of 14 notes from your computer keyboard. Unlike onscreen keyboards in other virtual samplers, VSamp’s keyboard can’t be used to specify key ranges while you’re editing Instruments.

You can download trial copies of VSamp and VSamp VST at www.kagi.com/smaug/vsamp/. After paying a registration fee and obtaining an authorization number, you’ll receive access to the VSamp FTP site, which has more than 160 MB of downloadable VSamp instruments and samples.

EM associate editors Dennis Miller and Geary Yelton have been sampling everything in sight for the past three months and will be giving their computers a much-needed vacation soon.

We welcome your feedback. E-mail us at emeditorial@primediabusiness.com.