Emu Systems Emax II
What do you get if you cross an Emax with a Proteus? An Emaximus perhaps, or maybe even a Protemax? Kendall Wrightson test drives Emu's new 16-bit sampler to find out.
If the Emax was an integrated version of the Emulator II, it would be logical to assume that the Emax II is an integrated version of the Emulator II. This is not the case. In reality, the Emax II is a mixture of Proteus hardware and Emax software. Aesthetically it looks exactly like the original Emax, except that it has a limousine black livery rather than battleship grey.
The Emax II comes in both Standard and Turbo HD versions, and both are available in either keyboard or rack formats. The rackmount Turbo HD, as supplied for review, boasts an internal 40 megabyte hard disk and four megabytes of RAM, allowing 53.7 seconds of mono sampling at the maximum sampling rate of 39kHz; 13.4 seconds maximum is possible on the Standard version (1 Mb of RAM, no hard disk). Two megabyte memory upgrade kits will shortly be available, allowing a Standard Emax II to be upgraded to seven megabytes in full, and the Turbo HD to eight megabytes (giving 107.4 secs!) Although the current versions of the Emax II can play back samples in stereo, they can only sample in mono. However, Emu report that a stereo sampling capability will be available as an upgrade. With the stereo playback hardware already in place, an upgrade would presumably consist of a new sample input stage and some new software.
The internal floppy drive on both Standard and Turbo HD versions is a conventional DS/DD 800K type, rather than a high density DS/HD 1.4 megabyte drive as fitted on the S1000. Consequently, you will need a lot of disks to store your samples if you rely solely on floppies, and loading sounds from disk can be a slow business - two minutes for a 1Mb bank. By comparison, the Turbo HD's internal SCSI hard drive performs the same function in only 12 seconds (10 times faster). To open the way to interfacing with powerful external data storage devices, both Standard and Turbo HD versions of the Emax II have SCSI (Small Computer Systems Interface) fitted as standard, as well as some very professional hard disk utilities, which make backing up data and copying from one hard disk to another much easier.
The Emax II's maximum sampling rate is 39kHz - the same as Proteus - rather than the customary CD-standard 44.1 kHz. This is a shame, because for the sake of 5kHz or so, the Emax II bandwidth is 20Hz-19kHz, not the holy 20Hz-20kHz. More importantly, it denies the Emax II a possible direct digital input/output update for direct digital sampling and for backup onto DAT recorders etc.
In a specification-led market, this may damage the Emax II's perceived attraction. However, my experience of samplers has proved that specifications tell you very little about what a machine will be like to use in practice. There are many other very important considerations - like how it actually sounds, how easy it is to use, what kind of third-party support is available, how many sounds are available, and so on - and in these areas the Emax II scores very highly, as we shall see.
Operationally the Emax II is identical to the original Emax. Commands are divided into 'modules' - groups of functions - that are accessed through the use of the numeric keys.
You can layer two Voices together — Primary and Secondary - to form a Preset. When editing a Voice, the Emax II asks if you wish to edit the Primary or Secondary component (or both). Up to four Presets can be stacked, but only if they are arranged sequentially. A fixed 28K of RAM is set aside for Preset data, regardless of total memory size, which means that you can actually run out of Preset memory if your Presets are complicated.
The Emax II also features a built-in sequencer, and some memory is obviously required to hold sequence data. Up to 99 sequences can be held in memory, and the sequence memory is shared with sample memory. Sequence data doesn't take up anything like as much memory as sample data, so in practice you don't have to worry about running out. A collection of Voices, Presets and Sequences is called a Bank - only one Bank can be held in memory at a time, and Bank size depends obviously on the amount of memory installed in your machine.
Presets and Sequences can be loaded up from other Banks saved on disk, but Voices cannot. This is quite a problem, because it means that it is necessary to load in an entire Preset just to get at one Voice. Often the Preset will be too large to load into memory, so (take a deep breath) it's then necessary to save the Bank you're working on, load in the Bank containing the Preset with the Voice you require, create a new Preset containing only that Voice, save this edited Bank, load back the original Bank you were working on, and then load in the single Voice Preset from the edited Bank!
To help you find your way around, all of the Emax II's functions are printed on the front panel. When a module is accessed, its LED illuminates and functions can be selected by keying in the function number. This user interface is much quicker than the usual system of scrolling through endless functions until the right one appears, which one tends to find on comparable samplers, and encourages the user to explore the Emax II's full potential. The only problem with the user interface is the lack of a large graphic display, which would make trimming samples a little easier.
The Emax II organises itself as follows. Once you've sampled a sound, it can be assigned across the keyboard as a Primary or Secondary Voice. 99 Voices can be held in memory at once. A Voice can be transposed up four octaves and down as far as you like. It can then be digitally processed - looped, truncated and so on, then dynamically processed, ie. enveloped, filtered, chorused etc.
Original Emax disks will load into the Emax II and the sounds can be replayed with enhanced fidelity and dynamic range. However, a totally full Emax disk will not load, because more memory is used in the 'expanding' process. It's also possible to save an Emax II Bank of less than one megabyte as a 'compressed' Bank, which can then be loaded into the original Emax.
Although the Emax II cannot sample in stereo (yet), it can play back stereo samples. Two mono Voices can be converted into a stereo Voice by layering them on top of each other and activating a Preset Management function called, not surprisingly, Stereo Voice. The two Voices to be combined in this way must share the same sample rate and original key and be mapped over the same keyboard range. (The Digital Processing functions, Change Pitch and Change Sample Rate, can be used to convert samples to conform to these requirements.) If these criteria are fulfilled, the two Voices immediately adopt extreme right and left pan positions.
One of Proteus's most heralded features is its 32-voice polyphony, although it is only 16-part multitimbral (ie. it can replay 16 different sounds at the same time). In comparison, the Emax II is also 32-voice polyphonic but has 32-part multitimbrality to take advantage of the extra voices - 32 different samples can be played simultaneously.
The interesting feature of this polyphony is that the Voices are configured as 16 mono or 16 stereo Voices. This means that the Emax II is capable of 32-voice polyphony, but not as 32 mono Voices, only as 16 stereo. In other words, the total number of individual mono or stereo Voice combinations cannot exceed 16.
Audio output is via four stereo pairs of jacks - a Stereo Mix Output plus three extra stereo pairs (A, B and C). Each Voice can be assigned to any pair of outputs. As with Proteus, the extra (ie. non-mix) outputs are via stereo jacks, and if a stereo plug is inserted, the ring serves as a signal return which sums into the main outputs.
A Mono Mix Out is also provided, as well as a Stereo Headphone jack. There are also sockets for two footswitches and one foot pedal (volume type). Clock In/Out jacks allow the Emax II's internal Sequencer/Arpeggiator to drive (or be driven by) TTL signals, such as those provided by pre-MIDI equipment.
SCSI and RS422 ports are also present. The RS422 port is provided on the Emax II so that OMI's excellent CD ROM player (and therefore the existing Emax library CDs) can be used without modification. This is a very important consideration, because it gives Emax II owners instant access to a huge and varied sound library. And some people would say that Emu have the best sound library going.
The Emax II's sample input is via a rear panel ¼" jack. I would have preferred a balanced XLR-connection, and perhaps Emu might consider XLRs on the stereo sampling upgrade. Prior to sampling, you must select a sample time, sample rate, input level/gain and a sampling threshold. With a blank memory, a new sample will automatically be assigned to Preset 00. However, when adding to a Bank, it is necessary to create a new Preset or overwrite an existing one. You can monitor the sample input at all times via the main outputs, even when not in sample mode.
The default keyboard range for a new sample is Original G1/Low C1/High B1. If a second sample is taken, its default range will be Original G2/Low C2/High B2, and so on up to G5/C5/B5 for the fifth sample. This is intended as a default for quickly sampling another synth, but it would be a good idea to have several sample range defaults available - say for drums, keyboard instruments, and so forth. This would mean that you could instruct the samples to be assigned to every other key, every fifth key etc, without having to programme it each time. Individual Voices can be tuned by +/— 45 cents, which lets you correct slight tuning inaccuracies in the original sample.
The sample quality is all you would expect of a 16-bit device: very clean and well defined, with an excellent dynamic range (presumably due to the 18-bit sample playback stage). The ability to stretch a sample up four octaves and down as far as you like provides lots of scope for creating weird sounds. It also means that one sample can be played across the entire keyboard range, and being 16-bit, samples played very far away from their original pitch don't become noisy. In fact, with synthesized sound sources there is no need for multisampling, one sample can sometimes be enough to adequately cover a whole five-octave keyboard.
Sample editing is performed by entering the Digital Processing module. The first functions that you will probably use here are Truncate and Loop. The Emax II supports one sustain and one release loop, and if you have trouble finding a loop point manually, you will find that the auto-loop function works brilliantly. For trickier samples still, the crossfade loop will always produce a usable result, though you should copy the sound first, as crossfading changes the sample irrevocably.
Samples can be spliced together (with splice crossfade), or combined (digitally added together) if they have the same sample rate. Samples can also be reversed without losing the loop points. The Gain/Attenuation function allows you to optimise the sample level to achieve maximum signal-to-noise ratio on playback - ie. to 'normalise' the sound. This very useful feature allows low level or badly sampled sounds to be drastically improved, and is not presently found on competing samplers.
Another very useful function is Taper, which digitally fades a sample out from a specified point, to remove unwanted reverb or clicks at the end of a sample without creating a sharp cut-off.
The Change Pitch function makes it possible to alter the pitch of a Voice in order to splice or combine it with another sample of a different pitch. The Change Sample Rate function, not surprisingly, changes the sample rate of a Voice, which again may be necessary for combining Voices in various ways.
Interestingly enough, a sample can be converted upwards to 44.1kHz, so the Emax II's Proteus-based hardware is obviously capable of playing back samples at this rate - which begs the question of why it wasn't chosen as the maximum rate. [Probably so as not to harm Emulator III sales - Ed.] Both the Change Pitch and Change Sample Rate functions take one minute to execute on a one second duration sample.
The Emax II's Dynamic Processing section contains all the facilities for real-time control and modification of samples, including an analogue VCA and VCF. Both the VCA and VCF envelopes have Attack, Hold, Decay, Sustain and Release (AHDSR) segments. The filter also has variable resonance, envelope amount (allowing you to create positive and negative filter sweeps) and keyboard tracking. Subjectively, it is very warm sounding compared to a digital filter. Its cutoff frequency can be modulated by Mod Wheels, Foot Pedal, or any MIDI Continuous Controller.
Keyboard velocity can be routed to level, filter cutoff, filter resonance, pitch, filter attack, and pan. I found that using all these options together could create some very interesting percussion effects. A Control Enable function allows you to exempt certain Voices from certain kinds of controller data - very useful for filtering out Pitch Bend on only one Voice in a split keyboard arrangement.
The Dynamic Processing module also contains a Chorus function, which is very effective at fleshing out mono Voices. Chorus clones the Voice, sending one to the extreme left and the other to the extreme right. The two Voices can then be detuned by up to 20 cents, which produces extremely rich and warm sounds. A stereo Voice can also be chorused, which produces very rich textures indeed.
Modulation and Pitch Bend data, Aftertouch, Foot Pedal, and MIDI Controllers A and B (assignable) can all be used to modulate filter cutoff, LFO depth, pitch, volume, attack rate, panning and crossfade of a Voice, to provide some real-time expression facilities. The two footswitches can be set to control sequencer and arpeggiator start/stop, sustain, release cross-switch and advance Preset.
The Emax II includes an unusually powerful arpeggiator, which offers six assign modes, up to eight key repeats, 15 extensions and three harmonies. It can synchronise to external MIDI Clock (or a 24, 48, 96 ppqn click) and transmit arpeggiated notes over MIDI. A glissando effect plays every chromatic step between the notes played (including extensions) and the 'Cruz' control doubles the speed of the arpeggiator each time a new note is added.
The tempo range is 40 to 240 bpm and the arpeggiator note value can be from 1/2 to 1/96. A keyboard arpeggiation range can be selected so that the keyboard can effectively be split. But the best thing of all about the arpeggiator is that different settings can be stored with each Preset.
The on-board sequencer of the Emax II could perhaps be more accurately described as a 16-track MIDI recorder/player, because it offers absolutely no editing commands - no quantise facilities, nothing; not even a metronome.
One simple way around this latter omission is to sample a click sound and assign it to an arpeggiated Preset running at the same tempo as the sequencer, and then record it. Like the arpeggiator, the sequence will then synchronise to external MIDI Clock or a click, and you have a metronome to play along to. (This metronome sequence can then be saved and used whenever you start a new song).
These limitations apart, the sequencer is great as a musical notepad, offering 96 ppqn accuracy and faithfully recording all MIDI Controllers. However, the sequencer's main value is for live use - replaying sequences produced on an external sequencer, to save you dragging a computer along to gigs. To make this easier, Supermode allows the simultaneous transfer over MIDI of up to 16 tracks into the internal sequencer. Each track is automatically assigned its own Preset and MIDI channel. Talking of which...
Every Emax II Preset has its own settings for MIDI Receive Channel, and you can set global MIDI parameters such as Receive Patch Changes On/Off, Local Control On/Off, Controller Transmit/Receive numbers, etc. A MIDI Overflow facility is available to pass extra notes (ie. above the Emax II's polyphony) to a second unit.
Perhaps the most interesting MIDI function, however, is the ability to load a Bank from a hard disk with a Program Change command. This is done by turning the HD Bank function on, after which the Emax II will interpret the next Program Change number it receives as the number of the Bank to be loaded from hard disk.
In fact, these days, the official MIDI spec is somewhat out of date in respect of Program Changes, as many devices offer well over 128 patches - and in an application like this, there is the added problem of specifying which drive to load data from. When they get around to updating the MIDI spec, no doubt a 'Media Select' command will be incorporated. In the meantime, Emu provide a method that doesn't break the rules, but provides the facility.
The Emax II provides two means of actually creating original sounds as an alternative to sampling: Transform Multiplication and Spectrum Synthesis.
Transform Multiplication combines two samples in such a way that their common frequencies are accentuated, while uncommon frequencies are discarded. For example, if speech is TM'd with a sustaining sound, the sound appears to 'talk' in a very similar way to a vocoder. The results of the process are difficult to predict, but always interesting. There is a serious drawback, however. It takes 39 minutes to Transform Multiply two one second long Voices - enough time to write a quick song!
Spectrum Synthesis offers facilities similar to additive synthesis (as used on the Kawai K5). To construct a sound, you begin by creating a Spectrum - a waveform generated by adding together up to 24 sine waves. The pitch and amplitude of each sine wave can be independently programmed, so the resultant waveforms can be very complex indeed.
The Voice to be created (which you can set to any length) is divided up into 24 'time slices', called Partials, and a Spectrum can be allocated to each of these Partials. The Emax II can interpolate between Partials, to save you having to create 24 different spectra, and the interpolation can be smooth or stepped, giving you a choice between a gently evolving sound or a much harsher effect.
As with TM sounds, Spectrum Synthesis sounds take a long time to calculate - about six seconds per time slice for a one second sound - and it's a shame that the Emax II can't perform the TM/SE calculations as a background task.
Nonetheless, the results are fascinating, and some of the SE sounds provided with the Emax II are unlike anything I've ever heard before - very dramatic, and perfect for film music.
A disk of useful Spectra is included, along with a disk of short instrument attacks and loops (like those found on Roland's LA synths). Perhaps Emu intend making a dedicated SE/TM based synth at some point in the near future, with the calculations performed in real time by DSP chips rather than by software - it would certainly be an interesting machine.
Listening to Gerry Basserman's demo sequences, the Emax II's dynamic range and clarity are well demonstrated. One of the most striking things about the Emax II is its ability to produce unusually warm sounds for a digital machine. The combination of digital and analogue technology produces a hybrid instrument, equally able to deliver accurate, sharp samples and lush synth-like pads.
In its rackmount form, the Emax II's main competitor is undoubtedly the Akai S1000. The S1000 has become a studio standard by virtue of its appeal as a production tool - its 16-bit 44.1 kHz stereo capability and quoted 20Hz-20kHz frequency response has made it very popular with studios.
The Emax II appears to be aimed directly at the musician, though, as it is first and foremost a musical instrument; to compete with the S1000, the Emax II needs to appeal to both musicians and engineers alike. To this end, Emu would do well to offer a few more 'production tool' features, such as audio/SMPTE triggering and 'Timestretch', and doubtless these could be implemented with a software update. Also, XLR connections would certainly inspire greater confidence in a professional environment.
The most important upgrade, however, will be stereo sampling - should true phase-coherent stereo sampling be provided, then the advantage of being able to replay 16 stereo Voices simultaneously would increase the Emax II's appeal immeasurably.
Another shortcoming (from the studio point of view) of the Emax II is the choice of top sample rate. In today's climate 44.1 kHz sampling is simply a must, because it is a standard, and digital domain transfer is now an everyday experience in top recording studios.
In its current form the Emax II will appeal firstly to Emax owners wanting to upgrade to a 16-bit, 16-voice machine. For them, the mark II is the obvious choice as all the library sounds (including those on CD ROM) are compatible. Also, there would be no need to learn how to use yet another machine as the Emax II is functionally almost identical to the original Emax.
As far as non-Emax owners are concerned - and that's most of us - with its sequencer, output routing facilities and excellent Controller provision, the Emax II is unrivalled in an application which calls for high quality sampling combined with live sequencing. How the Emax II fairs in the world of the recording studio now depends on its popularity with musicians, for it is they rather than engineers that could make it successful in a market presently dominated by the S1000.
Many thanks to Stephen McLaughlin for his invaluable comments.
Emax II (rack and keyboard) with 1 Mb RAM: £2850 inc VAT.
Emax II Turbo (rack and keyboard) with 40Mb Hard Drive, 4Mb RAM: £5290 inc VAT.
Emu Systems, (Contact Details).
Review by Kendall Wrightson