Review – MOTU’s Mach Five
Mark of the Unicorn’s Mach Five sampler has been around for a number of years now, but it was only last year it received an update, See what’s on offer in this close up look.
by Warren Burt, Nov. 2013
Mach Five Three – the All-in-one, with Elegance
Mark of the Unicorn’s Mach Five sampler (Mac and PC) has been around for a number of years now, but it was only last year it received an update to version three. The update was more of a complete makeover, and the dozens of new features that have been added catapult it into the position of the top three or four software samplers available today. It’s a strong contender for the number one slot, but in the software world, things change so rapidly, I don’t want to make predictions like that. Suffice it to say that the software is a highly efficient, pretty easy to learn BEAST, and it really can lay a claim to be just about the only sampler/synth that you would need. It’s absolutely serious competition for Kontakt5, EXS24, or any of the other flagship samplers out there.
First of all in the features list is – ta dah! – a printed manual. One can actually work on screen with the sampler and have a paper document next to you to refer to things. No more switching from one window to another! And you can actually learn about the software without a computer present! This kind of innovation in documentation is one that more manufacturers should consider. (grin) The manual is clearly written, and if there is a bit too much “see chapter 3-5, p. 46” for my taste, the software is so complex that I don’t know how else a manual for it could be written.
Last year I was considering getting a “real sampler,” as opposed to a sample player, such as Garritan’s Aria, for which I was writing my own sfz files. I wanted a sampler with more features which offered greater ease of programming, more options for sound treatment, and of course, the ability to micro-tonally tune it and have possibilities for scripting. The promo for Mach Five Three promised all this and more. I then ordered it early in the year, and after the obligatory four month wait (Australia is at the end of a very long and leaky supply line for many things technological – I just waited three months for a windscreen (!) for my Zoom sound recorder!), the software finally arrived. It required an iLok, and sure enough, a week before the software arrived, I lost my iLok. So there was more delay while I got a new iLok and renewed all the licenses on it.
Once all the problems were solved, it was very easy to install, and because of the iLok, I tried installing it on a number of computers to see how it would fare: A Toshiba i3 laptop with XP, a Dell Centrino Duo with XP, an ASUS 101 netbook with Atom processor, running XP; and an ASUS Vivo Tab running Windows 8. The software worked fine on all of these. CPU percentages varied from computer to computer. I made a patch using the “Organ” oscillator, the most CPU intensive module in the software, and ran a random composing program with Algorithmic Arts’ ArtWonk that generated many overlapping notes per second – a real torture test for the software. The average CPU percentages for this were Toshiba: 20%; Dell: 32%; ASUS XP netbook: 55%; ASUS Vivo Tab: 48%. In standalone mode, using each computer’s internal sound card, sound on the Toshiba and ASUS XP netbook was flawless. There were some breakups on the Dell, and more breakups on the ASUS VIvo Tab. Running as a plug-in under AudioMulch, and using an external sound card (Roland UA-4FX) sound was flawless on the Toshiba, Dell and ASUS XP netbook. There were still a few breakups on the ASUS Vivo Tab, but adjusting the latency to 4096 and using an ASIO driver for the UA-4FX eliminated those. The conclusion here is that the more CPU power you can throw at Mach Five, the better, however, if you have an older and slower machine, it will still work just fine with lower polyphony.
The main purpose of the program is to sample, and there are several types of sampling available. In the program these are called Sample, Stretch, Slice, Ircam Granular and Ircam Stretch. Sample is the traditional sampler, and you can easily do bulk loading and mapping of a set of samples very easily, although the method is a bit different than is described in the manual. All kinds of layering, looping, crossfading, etc. are available here. Stretch uses granular techniques to play the sample at the same preset tempo at all pitch levels. If you get the granular settings right, it can be very effective. Slice is a beat-slicer, and you can set the slices automatically, or you can let it do the slicing for you. There are a number of controls to control the fineness of the slicing and once you have a slicing you like, you can do a one-click mapping of the slices to a MIDI keyboard. It can also save a MIDI-sequence of notes which reflect how the slices can be played to reflect the rhythm of the original. Ircam Granular is licensed from the French Ircam music research institute. It uses a much more sophisticated kind of granulation to stretch sounds. It can seamlessly time-stretch just about any sound, although you do have to fiddle with the settings a bit. The most critical control, I’ve found, is the “jitter” control, for “realistic” time stretching (“realistic” in the realm of virtual sonic fictions, that is!). Ircam Stretch uses a phase vocoder FFT algorithm to stretch sound. It’s extremely CPU intensive, but with the right settings, sounds absolutely delicious. And I recently made a program with it where I stretched a piano sound, then exceeded the polyphony limits my Toshiba XP i3 machine could deliver, and the glitches made were absolutely gorgeous. So even driven into malfunction, the Ircam Stretch produces gorgeous (though maybe unintended) sound. Any and every control on all the sampler modules can be MIDI-controlled, and once you drop a sample into the program, you can freely alternate modules, hearing how it can be treated in each one of these.
In addition to sampling, there are several kinds of oscillators: Analog, Analog Stack, Noise, FM, Wavetable, Drum and Organ. Analog is an analog oscillator emulation with the usual suspects: sine, square, triangle, saw, pulse and noise. Using this and the full range of LFO controls, VCAs, Filters etc. (more on these below), one could use Mach Five Three as a fully featured analog synthesizer emulation. The Analog Stack is eight of these oscillators ganged together, for some work with additive techniques. Noise has eleven different kinds of noise, some with quite effective controls for varying timbre. FM is a four-oscillator FM synth with twelve algorithms for connecting the four oscillators. This one sounds great. Wavetable provides dozens of preset wavetables (short waveforms) to play with, and you can drop and drag your own into them. Drum is a combination of a gated tone and a gate noise. At first, when I read about it, it sounded very uninteresting, but I was delightfully surprised to find that it made a wide variety of very attractive percussive sounds. Organ is a drawbar organ emulation. Very good sounding, and very CPU intensive. You won’t be using this one to play Jimmy Smith licks live on your netbook with (believe me, I tried!), but with a powerful enough computer, it shines.
The program uses a hierarchical organization – Oscillators (whether sampled or waveform based) are grouped into Keygroups, which are grouped into Programs, which are grouped into Multis, which travel to a Mixer, which has lots of Aux sends and receives. The Keygroups go through a selectable patching of VCA, LFOs, Envelope Generators (several types, including user drawable multipoint envelopes), and Filters (dozens of types!). There can be an unlimited number of Programs in any multi, and Mach Five Three can accept up to 64 MIDI channel inputs at any one time. For a recent improvisation, I programmed 16 Programs into a Multi. Each Program had between 10 and 128 samples in it. Each program was on a different MIDI channel. I played this with my Roland PCR800 keyboard. The program was on my Dell XP Centrino Duo machine. The playing of the sampler, and the sound quality was absolutely flawless. Latency was pretty near close to zero – playing didn’t feel mushy at all.
The hierarchical organization means that an effect, for example, can be applied at any level of the hierarchy. You can have effects applied to Oscillators, Keygroups, Programs, Multis, or as sends and receives in the Mixer. Routing can get as complex as you want.
Many people buy a sampler for its content. I’m not one of them, being a roll-your-own kind of guy. The program though, comes with 45GB of content. (I just went out and bought a portable 1TB drive for these, and then proceeded to load the rest of my 30 years of samples onto the hard drive as well – my whole sonic history on one handy drive! Which of course I then immediately backed up.) Some of these are very sample intensive single instruments, such as Piano and Guitar. The scripting on these is very good, although you’ll need a hefty CPU to handle some of these such as the piano. The Universal Loops and Instruments is a very fine collection of General MIDI and bread-and-butter timbres, and the MachFive Biosphere set is a whole library of wonderfully glowing textures. Two DVDs are devoted to the UVI Xtreme FX set, which it looks like it will take months to get through. The effects seem quite well recorded and I’m sure will prove useful. As well there are excellent Piano, Bass, Drums and Guitar sample sets. I’m sure some people, more commercially oriented than I, will find these sample sets a real reason to buy the sampler.
In terms of importing other companies sample libraries, the program is very versatile. It will import EXS24 (including Garage Band instruments), GigaSampler, Kontakt 1-4 samples (not Kontakt 5 though), Sample Cell, SF2, and VSampler2 files with ease. Although not listed, I also found it would import many kinds of .sfz files as well. When I got the sampler, I scoured the internet for free samples in many of the above formats. I found that EXS24, Giga, SF2 samples imported with ease. Kontakt 1-4 usually imported well. .sfz files imported the samples, but usually some programming was necessary. And UVI .dat and .ufs files are the native format for it, so those will of course import with all programming and scripting intact. Importing a variety of sample types with all the programming is quite a feat, and if the importing is not perfect across all platforms, it’s still very good and will allow you to at the worst import the samples and then make a Mach Five native program with them. For the EXS and Giga sample sets I found, however, no additional programming was necessary. I have a large library of VSampler3 samples made over 15 years. For these, I found that VSampler would save them as Soundfont .sf2 files, which easily imported into Mach Five. The 16 Program Multi mentioned above was quickly and easily mostly made with .sf2 files imported from my VSampler3 library.
Scripting is supported, as is a very powerful arpeggiator. This is an area I haven’t gotten into yet, but on the basis of the features available in the F Grand 278 piano sample set, and the Telematic guitar set, it looks like it is deep, comprehensive and very powerful. The scripting is done in the Lua language, which I’ve been meaning to learn for the past couple of years. Looks like it’s time to do that.
Microtonality is also supported. It is extremely easy to implement, and each program can have its own tuning. The ease of tuning was the reason that I chose Mach Five Three over Kontakt 5. Having written tuning routines for Reaktor, I know how convoluted that process is. With Mach Five Three, one simple loads a Scala .scl file for a 12 note scale, or a Scala .kbm file to put in the mapping for a scale of more than 12 notes, followed by the Scala .scl scale for the desired tuning. And when a program is saved, the tuning data is saved with it. So my requirement that a sampler be able to give me easy, multi-timbral, multi-tuning resources which are recallable instantly is more than met by Mach Five. For a micro-tonalist, I think Mach Five is the only choice.
As mentioned above, there are about 47 different kinds of effects (more were added in the latest 3.2.0 update), which can exist in any kind of chaining at any level of the hierarchy. Each effect has a number of presets and each allows the user to make and store their own presets as well. And of course every control is capable of being controlled by MIDI.
As you can see, the program is very deep, and a LOT of thought has gone into it. I wouldn’t hesitate to recommend it to anyone, with the proviso that for best performance you’ll want a hefty computer with a lot of CPU and a big fast hard drive to keep samples on. I have a LOT of software synthesizers (I especially love my family of LinPlug machines), and I’m certainly not giving up on those, but I can see that a lot of my work in the future is going to be done with this very sophisticated and marvellous piece of software.