Docendo discimus


  • Calendar

    May 2017
    M T W T F S S
    « Sep    
  • Archives

  • Recent Posts

  • Bling

E-mu Proteus 2000

Posted by brunorc on December 29, 2010

My previous synth – Yamaha CS1x – ended up being used as a sound module. This was the period of my life when I started to write some background/illustrative music. Hardware samplers were already there and it was getting difficult to write a decent music without convincing sounds. One can say a lot of nice things about CS1x – like praising the color, etc. – but saying that its sound are realistic, would be a bold flattery.

The biggest problem with samplers was that you needed samples to feed it. And most of the time good samples were so huge, that you ended up filling the memory with the bass guitar and drums, without even starting to think about guitars. Moreover, samples tended to be expensive, and homebrewing them leaves you with two problems: where to get the sources and how to achieve the proper quality. To paraphrase Jamie Zawinski: some people confronted with the problem of getting the right sound, think “I know, I will use a sampler”. Now they have two problems.

Remember, that was 1999. E-mu was producing entry level ESI samplers with 4MB of memory (not sure about Akai at that time). Professional models could have been beefed up to the astonishing 128MB of memory – usually for the price of a small car. Add to this the requirement for SCSI CD-ROMs and/or hard disks… well, that was an elite sport! Also, editing samples on the tiny screen (compared to 15″ CRT) was a disputable pleasure.

So I thought that sound module would be a better choice. You know, sound module is mostly a sampler – but with all the sampling being already done, so you cannot sample anything anymore. A box of sounds, that removes all the compromises out of your way. Some other wise guys already scratched their heads to baldness, only to stuff all the necessary samples in the best possible quality – to provide you with the set of sounds, that you can only take – or leave. The problem is, that with a sound module (or ROMpler) you are tied to the creator’s definition of “necessary” and “best possible”.

Proteus 2000 - front view

One rack unit wonder

But I was lucky enough to start my mission when E-mu was getting the wind in their sails (and sales). Actually, they mostly had two products they were selling for all those years: Emulator (a sampler) and Proteus (a sound module). Check it up on their page, they are still using the same names, it’s amazing. They never seemed to have (or care about) the marketing department, and while most of their stuff was really created by brilliant engineers (with Dave Rossum being the most prominent one, following the tradition started by Karel Capek), at some point lack of proper market presence brought dark clouds over the company. Anyway…

Display of P2K

That's your letterbox - now decorate the hall.

So, when I was hunting for a sound module, E-mu released their Proteus 2000. It was another generation from their famous ROMpler line: at some point they just gathered some bread-and-butter samples from their Emulator III, packed them into the one-rack-unit box, named “Proteus” and started to sell. Then came Proteus 2 with orchestral sounds, Proteus 3 with ethnic sounds, Proteus FX (with two FX processors added). In the meantime they branched into some more specialized stuff, releasing Vintage Keys – the first sound module armed with their H chip, which brought the power of digital filtering. After that came the Morpheus – the beast from the Z-Plane, offering 197 different types of digital filters up to 14 poles along with complex matrix modulation; and UltraProteus – a selection of the best sounds from the whole Proteus family, with even more Z-Plane filters. Both featured an extension slot on their motherboards, but it was never to be filled. Instead a rich palette of (again) specialized sound modules was spawned: urban Planet Phatt, dancey Orbit and latin Carnaval. Then E-mu released Audity 2000 – named after their analog monster from 70’s – with the intention of compressing the power of modular synths into the one rack unit. It paved the road for the whole “2K” platform, unfortunately suffering from all the child-age diseases.

Proteus 2000 was the first one from the long “2K” line (even although Audity had some features that later came back in OS 2.0, it was never compatible with the generic OSes, nor with any 2K ROMs). It came with 32 MB of sample memory, filled – as usually – with the broad selection of sounds – orchestral, pop, rock, electronica… you name it. But that was only the skeleton: while E-mu was famous of their pristine and selective samples, those were the legendary filters, which gave the muscles to this sound. And while I’m really impressed with their quality as well as diversity, I cannot understand the weirdest decision E-mu has ever made: no realtime control over the resonance of the filter. You read it well – you press the key, the sound starts; from now on you can adjust the filter frequency, but not the resonance. Argh.

Mode buttons

Beware the data dial - the most vulnerable part

But bones and muscles are not enough. To make the sound more vivid, you have to modulate it. Three 6-stage envelopes (two with looping), two LFOs plus 12 assignable MIDI controllers, along with so called “modulation processors” per layer allow you to create some realistic sound effects, as well as complex, evolving synth patches. Yes, synth. Proteus 2000 is a full fledged digital synthesizer, just without the keyboard. But it has its own four knobs, so it is a synth, right? It doesn’t have oscillators – but it has samples, and the way samples are played can be influenced as well: you can retrigger the sample (using footswitch, velocity switch or LFO), as well as change its starting point and delay it, so it won’t play immediately. You can have polyphonic glide (portamento effect). You can introduce variations to your LFOs, so they will sound less predictable – or the other way round: you can sync them to MIDI clock.

But then again, Proteus was only the first pawn to be moved by E-mu. But this time the amount of different sound modules was extraordinary. B3 for organs, Mo’phatt for hip-hop, Virtuoso for orchestral sounds, Xtreme Lead for electronica, Planet Earth for ethnic sounds, Orbit 3 for dance, and so on… E-mu was repackaging their samples like there was no tomorrow, but this time there was still the same platform, only the sound ROMs were different. Later sound modules came with beats and arpeggiators, but that was only the OS, which you could upload thru the MIDI and – presto! – you just have a new, more powerful instrument. And you didn’t have to worry that your rompler couldn’t have sampled anything, since you could just buy new ROMs to enrich the arsenal of your samples. Just in case you really needed some specific sound, which no ROM could provide, E-mu made it possible to create your own ROM (using one of their Ultra samplers) with any samples you liked.

Realtime controls

Realtime control knobs - also useful for editing

So, to summarize: 128 voices of polyphony, 50 different filter types, two FX processors, 23 different ROMs available with four slots on the mainboard. Two MIDI interfaces, three pairs of stereo outputs (that can be used for insert effects, as well as joining the audio signal from external sources) and SPDIF output. You can use up to 32 MIDI channels – and use different arpeggiators on every single one. Every patch can use up to four layers, where every layer has its own filter, three envelopes, two LFOs and modulation processors – and its own settings for all this mess. Just in case it was not enough, one patch can “link” two other patches – an easy way for building layered/splitted sounds out of the existing ones. Complex? Hard to manage on 2×24 chars display? Emigr8 made Proteum, a free P2K editor for Windows. Hating Windows? Use prodatum, another free and crossplatform editor. And you may need it badly, because danger, Will Robinson: data dial encoder is notorious for breaking. After some time your dial may become crazy, then you need a soldering gun iron (up to 25W) and Bourns PEC164215FN0024 encoder for replacement.

How does it sound? In my opinion it sounds great. If you are looking for some realistic replicas of acoustic instruments, guys from E-mu sampled them an awful lot of times, so they know them inside out. If you are into completely synthetic sound, built from scratch – guys from E-mu prepared a swiss-army chainsaw of synthesis (minus realtime resonance). You have to be delicate with your modulations though, otherwise one of those 12-pole filters may just cut your cat into slices. And of course don’t expect the legendary analog warmth: Proteus is purely digital, so you can have difficult times trying to recreate some Moogy sounds in every detail (and don’t try to use anything steeper than 6 poles in such case), but it should be quite easy to get some PPG-like sounds. Below you will find a fraction of 1024 presets that come onboard, and some of 512 user locations (as you may have guessed, there is still a lot of space in mine). Right now I’m coming back to my Proteus 2000 in every free moment (which losely translates to “once or twice a week”), so I hope to get some better sounds soon. If you find those user sounds crappy, it’s rather my fault, so please don’t judge this box on the basis of my sound-designing skills. That would be very unfair. All patches are played by “audition riffs” – another nice feature of 2K platform. In fact, they can be also used as so-called BEATS phrases, turning the sound module into a kind of pattern-based sequencer.

Would I buy it again? Hell yes! Right now you can get a P2K in good condition for about 200€, or even less. Some ROMs are cheaper, some are expensive. Few models of boxes (Proteus 1000, Planet Earth, Xtreme Lead) are thinned out – you will get “only” 64 voices of polyphony, one pair of stereo outputs (no SPDIF) and only two slots on the mobo. Most of the time you should be happy with presets, and even more happy while editing your own – it takes some time, though. Highly recommended gear, go and get it. The only excuse to not to do so is that you already have one. Or two.

Update: You can find more information about this nice box on Yahoo group about Proteus 2000.

Posted in MIDI, Music, Synths | Tagged: , , , , , , | 2 Comments »

Yamaha CS1x

Posted by brunorc on December 3, 2010

In 1990s, Yamaha was busy introducing models from their SY series, successfully merging patented (at this time) FM technology with sample-based AWM synthesis. Later they turned to feature-loaded workstations, like QS300 or W5 and W7, workhorses armed with sequencers; their control surface consisted of display surrounded with function buttons, data wheel… and even more buttons. Users mastered the art of menu-jumping, tweaking parameters one after another. Meanwhile, somewhere in the deep Swedish basement, the completely new shape of synthesizer was being drawn. And after Clavia entered the market with Nord Lead, thus starting the “analog revolution” and bringing knobs back to the public attention, Yamaha knew they would have to jump the bandwagon.

Yamaha CS1x - full frontal

The Blue Meanie

Actually, they hit the sweet spot.

There was a huge problem with QS300 and the W series – they were quite expensive. So Yamaha folks looked at the crowd storming the Clavia stands, took their MU50 sound module and reshaped it into the CS1x synthesizer. They played on people’s sentiment to the well known CS series (with CS80 being the most famous) – old, heavyweight analog synthesizers – but “x” marked the new generation: light, affordable and always in tune. Later, the experience gathered while brewing CS1x was used for creating the AN1x – the one and only Yamaha’s attempt to create a virtual analog. But apart from the chassis (and its color) and similar name, they bear no resemblance to each other.

Having the guts of MU50, CS1x was as far as possible from being a virtual analog. But it proved to be one of its major strengths, to be able to provide multitimbrality of 16 channels. Also, the XG standard emerged just in time to provide the means for the true realtime control over the sound without the need to use bulky SysEx messages. So when it landed in 1996, it was probably the most sought after budget synth.

Display with function and numeric keys

Display with function and numeric keys

It comes with XG-compatible synthesis engine, based on AWM (Advanced Wave Memory) – it means it just plays back samples. With the polyphony of 32 notes it was an ideal companion for a sequencer; it even featured special connector (named “TO HOST”) and a cable that one would plug to the serial port of the PC or Mac, without any need for a MIDI interface. Samples were good enough for the standard of 1996, not convincing though. With some meticulous MIDI programming however, one would be able to get close to the sound of electric guitar, using three different patches, filters, envelopes, FX processors and a ton of MIDI continuous controllers controlling all this mess (not to mention some painstakingly crafted SysEx here and there). I’m afraid that nobody bothered though, maybe with the exception for Yamaha songwriters.


The knobs are back! Notice the Scene buttons above the modulation wheel.

But that was only one face of CS1x – the other one was trying to mimic the red monster from Stockholm. After pressing the Performance button, channels 1-4 were used for building the special, “better” sound. User was able to split & layer up to four so-called material voices (raw samples), as well as use low-pass filters, three envelopes (dedicated to control the filter, pitch and amplitude), LFO and FX processors. Routing wasn’t very flexible, but still some nice results could have been obtained. Moreover, user was able to control the sound in the realtime, using four hardwired (Attack & Release of the amplitude envelope, as well as Cutoff and Resonance of the filter) and two freely assignable knobs – one (Assign 2) could have controller different parameter for each layer, while the other (Assign 1) was common for all layers. Finally, two different settings of those knobs could have been memorized under the Scene settings and recalled instantly, as well as being morphed with the use of modulation wheel. And last but not least, CS1x also had the arpeggiator with some Up, Down, UpDown and Random patterns, plus some more esoteric ones, which would also adjust the velocity and/or filter settings of the sound.

There was also another interesting thing about CS1x: it became popular – as well as the XG standard – and took the advantage of the Internet boom. There were countless web pages (and FTP servers…), from where one could download songs or patches. Right now many of those links are dead, but CS1x definitely took its share of the Internet and used it well.

Panel with edit matrix

One knob to rull them all - edit matrix instead of multiple screen pages.

How does it sound? In XG mode, definitely better than any soundcard of that time. Piano is not-so-real, but not so bad – I mean, the unprepared listener could easily guess, that it was a piano. Electric pianos are rather good (most of them was probably sampled from pinnacles of SY series), organs acceptable; most of the time sounds of acoustic instruments resemble the original ones. XG defines more than one bank of GM-compatible sounds – called variants – so they come in different flavors (like Bright, Dark or Resonant). Yamaha used this to provide 480 basic sounds, with variations mostly focused on fat basses, shimmering pads and soaring leads. This can be useful for some general electronic music stuff, but don’t think it will sound like a real analog synth. If you are looking for this, get AN1x – bear in mind though, that sometimes it goes too far in modeling the vintage gear (you won’t be able to sync your LFOs or delay to the MIDI clock, same for CS1x).

I own this synth from 1998 and at this moment I’m probably a bit biased. I used to work with it a lot, but even when I got other sound modules, I was still using CS1x for scaffolding the songs. Even today I find myself reaching for CS1x and there is always some useful sound I can get out of it. Here you can listen to some samples:

The patch for “Alaska” was created from scratch to resemble the sound used by Eddie Jobson on the first UK album. The rest are slightly modified presets, which demonstrate the already mentioned features: low-pass filter, morphing, arpeggiator and FXs.

Nowadays I wouldn’t recommend buying CS1x, unless you can get it for an outrageously low price – it was always a bit of cheap stuff, but now you can probably get it below 150$. However, if you are looking for something inexpensive, better search for the next of a kin, Yamaha CS2x, which doubles the polyphony and storage memory, has more samples (and of better quality) and drumkits, features hi-pass filter (in addition to the low-pass), more FXs and arpeggiators patterns, has two knobs more – and finally can output arpeggiator patterns through MIDI.

Posted in MIDI, Music, Synths | Tagged: , , , , , | Leave a Comment »

emagic Unitor8 Mk I

Posted by brunorc on November 7, 2010

Yeah, fantastic gear. Bulletproof and reliable. MIDI interface with 8 IN/OUT ports armed with flexible routing matrix (with filters, so you can toss out some transdendental bits of MIDI, like MTC, SysEx or Tune Request) it’s just awesome and makes working with multiple MIDI devices easy.

However, I cannot convince it to forward MIDI Clock signal from one MIDI input to another (and no, I don’t filter it out), which means I cannot use MIDI sync without launching the PC. I thought that maybe the firmware is old, and I started searching on the Internet. Of course, granted emagic got absorbed by Apple, one cannot expect that they would be interested in supporting anything older than four years. Luckily I found it, but it cost cost me a bit, so for the coming generations:

emagic Unitor8 (or Unitor 8, if you prefer) Mk I (serial) firmware update – version 1.3.1

It was definitely worth upgrading for me (had version 1.2.3):

unitor8 firmware revision 1.1.4:
beta version
eeprom cannot be updated! you must replace the eprom with a version
1.20 first.

unitor8 firmware revision 1.2.0 contains the following improvements:
first public release.
eeprom can be updated, but problems when updating via rs-232. you need
unitor8 control 1.0.6 and a special plug. see text file “unitor8 1.20
rs-232 update e”.

unitor8 firmware revision 1.2.2 contains the following improvements:
(not publicly released)

new features:
with amt, midi data can be transmitted inbetween. this feature is used
by logic audio 4.0 and higher.

unitor8 firmware revision 1.2.3 contains the following improvements:
(not publicly released)

new features:
amt realtime nesting (see above) can have up to 4 levels.

unitor8 firmware revision 1.2.4 contains the following improvements:
(not publicly released)

new features:
new command for settings computer unitor8 communication to 230 kbps.

eeprom update via rs-232 now works without the special plug.

unitor8 firmware revision 1.2.5 contains the following improvements:
(not publicly released)

new features:
all 6 frame formats can be striped

unitor8 firmware revision 1.2.6 contains the following improvements:

new features:
can mix video in with burn-in window (video thru)

unitor8 firmware revision 1.2.7 contains the following improvements:
(not publicly released)

new features:
all units (which have at least this version) now communicate with each
other with 230 kbps instead of 115 kbps.

sysex data is now transmitted to / received from pc (rs-232) with
hardward handshaking (rts/cts).

unitor8 firmware revision 1.2.8 contains the following improvements:
(not publicly released)

new features:
new modes for click input: switch to next/previous patch (in patch mode

clock master was not enabled if a device sends only the continue status
byte instead of the start byte.

unitor8 firmware revision 1.2.9 contains the following improvements:

eox (f7) was not transmitted reliably to the rs-232 port.

unitor8 firmware revision 1.3.0 contains the following improvements:

new features:
supports amt level 3 (former amt levels did not help at events at
different, but very close time stamps).

unitor8 firmware revision 1.3.1 contains the following improvements:

new features:
video thru burn in is now possible with ltc read and write.
sysex command for switching back to patch mode with the last selected

1024 byte parts of long sysex messages received by rs-232 port were
transmitted twice in very seldom situations.

Unfortunately, MIDI Clock sent from Novation KS5 still doesn’t get forwarded.

Update: In order to avoid clashes, Unitor accepts the clock signal from the first port, which sends the MTC Song Start command (and resets after receiving MTC Song Stop). KS5 doesn’t send MTC commands, but any decent hardware sequencer does (e.g. Yamaha SY77 or E-mu MP7), and Unitor nicely propagates MIDI Clock received from them.

After all, I recommend it wholeheartedly (I can survive sending MIDI Clock from the PC), however this model can soon end its career. Nowadays you cannot use Unitor8 serial with Mac OS X through a USB – RS232 interface. Also, Apple won’t be releasing any drivers for new Microsoft OSes for those device. To clarify a bit: the Mk I model is “serial only”, while Mk II model has then very modern USB port. And if you are not into syncing to some professional VCR or ADAT tape, you can get AMT8 – it’s just Unitor8 Mk II without the whole syncing stuff. No firmware updates though, but I hope they put the equivalent of 1.3.1 in this.

Posted in MIDI | Tagged: , , , , | 2 Comments »