PC3 Nerd

Tutorials Abound

2011-12-06T12:39:00.000-08:00

Linky -- my channel on YouTube: http://www.youtube.com/poserp

O.k., so it's been awhile since I posted here. Some updates -- unfortunately, the BCRb0x isn't gonna happen. While it's a cool concept, I realized that as a one-man operation trying to sell something like that and support it would be pretty difficult to manage along with all the other things I do in a day (like work and spend time with family). As fun as capitalism can be, there are easier ways...

And so I'm doing video tutorials on YouTube. As much fun as it is to write about nerdy stuff, I feel that there's a paucity of decent video showing people how to do nerdy stuff, especially w.r.t. synthesis on the PC3. The Kurzweil tutorials (see also here, click on the "Video" tab) are a great starting point and mine are not meant to replace those. However, if you've watched them and want something that's more focused on programming (and especially if you want to start discovering new stuff yourself), then my videos will hopefully fill that niche.

I read a lot on the internet about how the UI on the PC3xx series synths is supposedly not intuitive. I personally disagree, in part because I'm a nerd and I like simple UIs, but also because I don't think that's actually true. It is not fancy, but at the same time it's very usable. In fact, I think it's much faster than many other interfaces and has the advantage of being entirely consistent. Wanna do wave sequencing and FM and subtractive synthesis all in the same Program? No problem, and all the UI elements are the same so you don't have to learn a bunch of stuff to use them. Plus, it's darn quick to program with once you've worked through a couple of examples. So my second goal here is to show that the UI is actually very usable on the PC3 and that working with the instrument is generally straightforward and quick.

So enjoy, and check up from time to time on my YouTube Channel as I'll probably be uploading new videos every week for the next couple of months, if not longer. I am in nerd heaven, hopefully you'll join me...

The BCRb0X

2011-01-30T02:19:00.000-08:00

I've posted a demo video on YouTube with a link leading back here (among other places). So, two of you might be wondering what this is all about.

The PC3xx series of instruments are pretty dang cool. One thing that I've always wanted with mine is a step-sequencer sort of thing to use for quickly mocking up beats. It comes with all sorts of fun stuff in the arpeggio department, but I'm more of a knob/button sort of person so I really wanted to have as much control over the sound as possible while creating beats. My workflow is generally a.) make a beat, b.) record it, c.) loop it, d.) add stuff, and so on. I don't record MIDI much so I didn't really need something that would store MIDI. The BCRb0x fills that role; it can be used to quickly make a beat and add some melodic stuff. Or, I can just make some synth loops. Or, I can just space out and do a free-form improv and pretend I'm dancing in the desert.

Anyways, I thought there might be other PC3xx owners out there who could use a "groove assistant" like this when creating songs and/or just enjoying some time making music. So I'ma put this up for sale here in a few weeks.

I've always appreciated quality products, so I'm putting in time to write a decent manual and do some other housekeeping sorts of things that will (hopefully) make this easy and fun to use, whether or not you own an external knob/fader box like the BCR2000 I use in the video. I'll be doing another video in the near future to show how this package can be useful to folks who just own a PC3xx -- they should have some fun too!

So thanks for visiting, I'll have more info soon.

Why the PC3?

2010-12-08T14:31:00.000-08:00

There are lots of different brands of synths out there vying for your dollars (and lots of your dollars at that). So why the heck would someone go with a Kurzweil? I thought it'd be fun to let the marketing materials for the different synths do the talking, with a few comments from me. I'm mostly concerned about programming, so this will focus on that aspect of these synths.

First off, Korg. This is the relevant from the sweetwater site regarding the Korg M3:

Enhanced Definition Synthesis: A new horizon of sound-making potential
The M3 features a newly developed "EDS" sound generating system based on a new chip design that was developed in parallel with the HD-1 engine from the flagship OASYS. It takes full advantage of the M3's 256 Mbytes of PCM source material which is a veritable "best of" from the OASYS.

Using Korg's advanced voicing technology, the 1,028 multisamples, 1,606 drum samples, dual oscillators, four-level velocity switching/crossfading/layering, and stereo sampling allow for an enormous range of high-quality, detailed, and expressive sounds. The filter section, a vital part of the sound-creation process, lets you simultaneously use two resonant filters for each oscillator, and can be configured in series or parallel. Four filter types are provided: low-pass, high-pass, band-pass, and band-reject.

The amp section features a driver circuit that gives the sound a sense of edge and character, along with low boost to help shake the walls when needed. Integrated 3-band EQ allows tonal sculpting without needing to go to the effects section. Detailed envelope generators, abundant modulation routings, plus the all-new AMS Mixers for advanced processing all combine to provide a level of sound shaping far beyond previous possibilities.

Effects
The all-important effect section provides up to five Insert effects, two Master effects, and one Total effect. In addition to great choruses, flangers, phasers, delays and reverbs, the 170 effect types include dynamic processing effects such as compressor and limiter, amp modeling and speaker simulation effects using Korg's proprietary "REMS" modeling technology, and evocative effects like the Grain Shifter and Talking Modulator.

The M3 includes professional features that allow for many cool applications and routings. Dual Common LFOs allow multiple effects to use the same modulation source; an advanced Control Bus provides triggering input from external audio or internal timbres to effects such as compressor/limiters and the vocoder; and the audio inputs can be freely routed to various effect chains and then run to individual outputs allowing the M3 to be used as a 2-in/6-output effects processor.

O.k., so what you're saying to me Korg is that you offer:

1.) 4 layers of velocity switching
2.) Two resonant filters per oscillator
3.) Two oscillators
4.) A whole bunch of samples
5.) Four types of filters (low-pass, band-pass, band-reject, and high-pass)

Wow. I mean, aside from all those samples, how do you get this thing to do stuff beyond some classic synth sound emulations? There are effects, though (up to eight at a time), so there is some potential there... Consider me underwhelmed. That "new horizon" looks a bit like the horizon printed on all those flamingo t-shirts from the 80's:

http://images.google.com/images?hl=en&gbv=2&tbs=isch:1&&sa=X&ei=feP_TJfxHIL6swPw28WrCw&ved=0CDIQBSgA&q=flamingo+sunset+t+shirt&spell=1&biw=1074&bih=646

Next up, Roland. This is from the blurb on Sweetwater for the Fantom G:

Advanced sound engine
Driven by a powerful new audio processor, the Fantom-G provides new sonic depth and versatility. With twice the wave-ROM capacity of its predecessor, it contains a treasure trove of Roland's world-famous SRX-quality sounds, but what really sets this instrument apart is its new ARX expansion, which allows additional SuperNATURAL sound-sets and fully dedicated synth and effects engines to be added to this already powerful live workstation.

...[skip some random stuff about the lcd screen and sequencer]

SuperNATURAL ARX expansion
Fantom-G provides ARX slots that accommodate two SuperNATURAL expansion boards. SuperNATURAL is Roland's unique behavior-modeling technology that enables a new level of organic, emotional expression previously unattainable in synthesizers. Applied to traditional-instrument sounds, SuperNATURAL enables the delicate, organic tonal changes and playing nuances of real musical instruments. It's also a powerful tool for electronic and non-traditional sounds, breathing amazing new life and expression into synthesis.

Effects
Fantom-G is loaded with a state-of-the-art effects engine, with multi-effects available for each Fantom-G sound/part. In 16-part multi-timbre mode, up to 22 effect routings can be used simultaneously. Create complex setups with 16 insert effects per patch, plus global reverbs and choruses. Mastering effects are provided for polishing songs created with the onboard sequencer.

Hmmm... To say "sketchy on the details" is being a bit generous. I have no idea what "superNATURAL" is, aside from it being something about articulation. Perhaps my friend Google can help:

http://www.google.com/#q=superNATURAL+roland

O.k., let's click on the seventh link (a page explaining superNATURAL on the Roland website): http://media.roland.com/en/IT/RD-700GX_IT/SuperNATURAL.html

And... Still nothing, but I think I get it a bit more now. SuperNATURAL is Roland telling you "hey, you want to play a drum, so I'm gonna hardwire a bunch of parameters so your sound will always be like a drum". Sweet, hand-holding! And hardwired signal paths! Hey, at that point you might as well get a prepatched analog synth.

O.k., not so good (too bad also, since you can have up to sixteen insert effects and there could be some fun there). We don't really know what kinds of synthesis are available (at least Korg got that part right), what the heck superNATURAL means aside from fancy user interfaces and tweaking some hardwired parameters, and there isn't even much of an idea about how much sample storage is available (another plus for Korg) without having to read the manual.

O.k., o.k., to be fair the Fantom G is more of a "workstation" than a "synthesizer", so let's look at an actual Roland synthesizer, the V-synth GT. Now we're talking:

Amazing Sounds and Expression Possibilities
The Roland V-Synth GT builds on the original V-Synth, which is renowned by performers and sound designers alike for its expressive synthesis and sound design capabilities. The next-generation V-Synth GT merges the company's innovative Elastic Audio Engine with the Articulative Phrase Synthesis technology to form a new synth with ultra-realistic sounds and amazing expression capabilities. Features such as speed programming, Articulative Phrase Synthesis, and powerful analog modeling ability allow you to break sonic boundaries and create never-before-heard sounds with the Roland V-Synth GT!

...[cut some bullets]

Multiple forms of advanced synthesis
By integrating a new dual core V-Synth engine, the V-Synth GT can combine Elastic Audio Synthesis, Vocal Designer technology, and AP-SYNTHESIS in real time. Roland's proprietary VariPhrase technology is also integrated into the V-Synth GT. Together with its analog-modeling engine, External Audio input, and digital sampling architecture, the V-Synth GT provides an incredibly wide and colorful foundation on which to capture, manipulate, and create sounds.

New Articulative Phrase Synthesis
The V-Synth GT is the first synthesizer in the world equipped with AP (Articulative Phrase) SYNTHESIS. This revolutionary form of modeling recreates the ever-changing behavior, nuance, and sound of an instrument as it's being played. The results are stunningly expressive and realistic, and can be applied to acoustic instrument simulation as well as new, never-before-heard sounds.

Speed programming
Serious sound designers will love the complexity and editing depth the V-Synth GT, but there are many instances where musicians need fresh sounds fast - and that's the beauty of Sound Shaper II. Based on the original Sound Shaper, this enhanced mode lets you chose a sound type, then it brings the most relevant parameters directly to the surface. You'll be creating new sounds in seconds!

Futuristic design
The V-Synth GT is a top-of-the-line instrument. No corners were cut in creating its futuristic yet friendly layout and design. New front-panel features include a color touch-screen with eight universal control knobs, and the addition of dedicated buttons and sliders to make access faster and easier. From top to bottom, front to back, this is an instrument that says "first class."

Whoa, buzzwords... Elastic Audio Engine. Articulative Phrase Synthesis. Vocal Designer technology. Great, time for a new dictionary. Making up your own words is actually a great internet strategy as words are essentially "territory" on the internet; if you can make up your own words, then you can claim your own "territory" in Google. So, at least the marketing minions at Roland are smart.

And hey, some of this stuff sounds kinda interesting. So, it sounds like that superNATURAL thing and this Articulative Phrase Synthesis are somewhat similar, only on the V-Synth you can do custom programming (instead of being stuck with pre-programmed models a-la superNATURAL). That could be fun -- a bowed square wave layered with a hammered saw wave and perhaps we could stack some sine waves that "resonate" like a drum head or something... I'm feelin' it. Er, assuming that's what all of this means. So I'll have to go try out one of these v-synths to get a feel for what all those strange new words mean in practice. A great idea, really, because judging synths based on their marketing materials is a futile exercise anyways.

I really like this idea of layering all these different synthesis types, but I'm a man of numbers so I want to know -- how deep can I layer? So, let's go to the Roland US website and see what they have to say since they can dedicate all the space they want to saying stuff about their products:

Revolutionary Sound Creation
Inside the V-Synth GT Version 2.0 is a dual-core sound engine that powers a staggering number of onboard synthesizer technologies, including Elastic Audio Synthesis, Vocal Designer, AP-SYNTHESIS, COSM analog modeling, and PCM wave sampling and playback. Each core is its own independent synthesizer that can combine any or all of these technologies to create a tone, and two tones can be layered, split, or morphed to create a single patch.

Crap. I can layer two sounds in a single patch and that's it? Maybe the specs can add some more details:

Sound Generator Configuration
Dual-core (2 Tones per Patch: Upper and Lower)
Sections per Tone
Oscillator (envelope x 4 + LFO x 1) x 2, Modulator x 1, COSM (envelope x 2 + LFO x 1) x 2, TVA (envelope x 1 + LFO x 1) x 1, Multi Step Modulator x 1, Tone-FX x 1, AP-Synthesis (Articulative Phrase Synthesis) x 1 (*), Vocal Designer x 1 (*): * Usable on either upper or lower (not both)
OSC1/OSC2 (Oscillator 1, 2)
Analog Modeling: 14 waveforms (SAW, SQUARE, TRIANGLE, SINE, RAMP, JUNO, HQ-SAW, HQ-SQUARE, NOISE, LA-SAW, LASQUARE, SUPER-SAW, FEEDBACK-OSC, XMOD-OSC), PCM/VariPhrase (Preset waveforms + Sampling waveforms), External Input
MOD (Modulator)
4 types (RING, FM, ENV-RING, OSC-SYNC)
COSM
16 types (OD/DS, W-SHAPE, AMP, SPEAKER, RESONATOR, SBF1, SBF2, COMB)
Multi Step Modulator
Provided independently for the Upper Tone and Lower Tone Tracks per Tone: 4, Maximum Number of Steps: 16, Templates: 16, Tempo: 20--250 BPM
AP Synthesis
Source Waveforms: 38 types, Phrase Models: Violin, Erhu, Sax, Flute, Multifade
Vocal Designer
Carrier: Oscillator x 2, Modulator x 1, COSM x 2, TVA x 1, Multi Step Modulator x 1, Vocoder: 4 types (Stereo, Mono, Vintage, Enhance), Formant: 7 types (Maximum)

...

MIDI Parts
1

O.k., so here's the meat. The V-Synth can play back one midi channel at a time. That channel can control a Patch (composed of two Tones). Now, all that cool synth stuff can happen on one of those Tones (but not both at the same time), so I could have some Articulative Phrase Synthesis going on with one Tone while the other uses the Vocal Designer. Per Tone I have an oscillator, a filter, some envelopes, an insert effect... At this point, I think the Korg still wins in the basic synthesis department but not by much; at least it has some velocity switching goin' on. After all, you still get two oscillators and a couple of filters per Patch on both machines. I feel the tides of underwhelment coming in, hopefully those flamingos fly away before they drown...

Hey, we still have Yamaha to contend with! Alright, maybe they can totally rock this and make Korg and Roland look weak (not that Korg or Roland need any help...). Here we go, the blurb on Sweetwater about the Motif XF8:

Pure Inspiration!
Yamaha pulled out the stops with the MOTIF XF8, giving you an unbelievably feature-packed instrument for music performance and music production. This next-generation workstation features an extraordinary collection of voices for you to get creative with. In fact, the MOTIF XF8 features an enormous 741MB of waveforms - that's twice the size of the MOTIF XS and nine times the size of the original! The Yamaha MOTIF XF8 offers everything from acoustic sounds to vintage synths to modern tones for hip-hop production. And with the new OS, complete with tap tempo, automapping, and a customizable user interface, the MOTIF XF8 is more powerful yet easier to use than ever.

Oh boy... Well, the size of ROM is easy (741MB), but as they say size doesn't really matter... If all synthesis could be reduced to sampling, then hey we'd have our winner. After all, the Motif XF8 can have up to 2GB of flash ROM installed, which is quite hefty and rather tempting if you're into sampling. Even the PC3K8 would limp in at second place with its paltry 128MB of flash ROM to compliment the 64MB base ROM. But, alas, synthesis is about much more than sampling, so we're gonna have to try and figure out what sort of synthesis the Motif can actually pull off. Hi-ho, hi-ho, to Yamaha's site we go:

Sound engine that increases the expressiveness of instrument sounds

The tone generator that the MOTIF XF has inherited from the MOTIFXS makes an 8-element-per-voice structure possible. Combined with Yamaha’s XA (Expanded Articulation), this 8-element structure allows you to achieve stunningly accurate acoustic instruments including expressive articulations that up until now have been difficult to re-create in electronic instruments. For example, you can re-create the natural legato of wind instruments, the key release noises of pianos and other instruments, guitar slides and harmonics, and the many different articulations for string instruments (spicato, tremolo and pizzicato, etc.). The AF1 and AF2 buttons on the front panel make it easy to control Voices with multiple articulations.

For synthesizer sounds, the 18 different filter types, advanced 6-stage pitch, filter and amplitude envelopes, unique Wave Cycle feature and assignable modulation Control Sets make the MOTIF capable of reproducing any synth sound from vintage to modern.

The tone generator also supports half damper pedaling through the use of the separately sold FC3 foot pedal. Using half damper pedaling with the damper resonance effect, you can achieve an even more realistic and richer piano sound that simulates the sound board of a grand piano and responds naturally to sophisticated piano pedaling techniques.

Now that's not too shabby! I like this 8-elements-per-Voice thing, although I wonder what, exactly, comprises an "element" and a "Voice". There are 18 different filter types, envelopes to control the pitch, filter, and amplitude of the sound... O.k., now we're cookin' with gas. That's some decent synthesis capability. I mean, with all that ROM we have lots of samples to choose from to use as source material. Plus, the Motif XF allows up for setting up 4 zones on the keyboard at any one time -- that could be fun since we could divide a "program" into four Voices and actually use 32 Elements to create a single sound. Cool.

Unfortunately if we don't want to do any manual-diving, we're kinda stuck and not really clear on what sorts of synthesis can be done on the XF. So, reluctantly, here are the manuals:

http://www.yamahasynth.com/downloads/manuals/synthesizers/motif_xf/english/

Pop open the Owner's Manual and... Looking at the Contents, I don't see a chapter about "creating sounds" or anything similar. Lots of stuff about playing sounds that are built in, and there's a chapter on sampling. Maybe the reference manual will be better... Ah yes, here we go, the "Tone Generator Block". Now, here's what the Motif XF can actually do for us synthesis-wise:

1.) An Element consists of an oscillator, a filter, and an amp.
2.) Eight Elements form a Voice. So, for instance, you could stack eight sawtooth waves in a single Voice.
3.) Pitch, Filter, and Amp all have their own envelopes.
4.) An LFO is available per Element to route to the Pitch, Filter, and/or Amp blocks.

That new horizon thing is a bit more true with the Yamaha; eight oscillators is pretty cool, as are the separate envelopes and filters. You could have eight filters operating at once per Voice, along with 8 LFOs and 8 separate Pitch envelopes.

We have found a contender for the PC3...

-----------------------------------------------------

...or have we? Let's now look at the PC3's page on Sweetwater:

Legendary Triple Strike Grand Piano, plus other keyboards
The PC3 is loaded with the famous Kurzweil Triple Strike Piano, with dozens of variations to accommodate every kind of player. Also included is the Classic Keys ROM, which puts an impossibly large vintage keyboard rig at your fingertips. The recognizable sounds of classic electromechanical instruments, with all of their characteristic imperfections, have been faithfully re-created, along with pedals, amps and speakers. Lush Mellotrons, growling electric pianos, and funky Clavinets are all there at the touch of a button, programmed to sound just like those classic albums. For more vintage firepower, the PC3 is equipped with the most advanced version of Kurzweil's widely lauded KB3 Mode tonewheel organ simulator, which screams, sputters and sizzles just like the original. In addition to newly developed Leslie effects, and a new bank of B3 programs, we've also included sounds from the classic Vox and Farfisa combo organs. The PC3's nine sliders provide real-time drawbar control.

...

Incredibly powerful Dynamic V.A.S.T. synth engine
The PC3's new Dynamic V.A.S.T. architecture is a huge step forward in
synthesis, the fruit of several major strides Kurzweil has made in improving
processing power, flexibility and user-friendliness.

VA-1 Virtual Analog Synthesizer built in
In another huge development, Kurzweil has resurrected the VA-1 Virtual Analog
Synthesizer, having built its synthesis engine right into the PC3. Hear the VA-1's power-shaped, anti-aliasing oscillators combined with a newer, larger set of V.A.S.T. filters and non-linear processors and you'll understand why the internet is buzzing about this feature.

...

Amazing effects power
With up to 16 insert effects plus two auxiliary sends, the PC3 offers more than twice the processing power of Kurzweil's acclaimed KDFX engine. Effects can be chained together in series, or distributed among the PC3's 16 multitimbral channels. And these are simply the best effects in the business - burning distortions, transparent compressors, warm phaser and chorus effects, perfect reverbs - everything from flawless vintage emulations to concert halls which decay into complete silence. All parameters are assignable to any of the PC3's real time controllers - sliders, pedals, wheels, etc.

O.k., so far a bit vague (to be charitable). For a synthesis fiend like me this is what I read:

1.) Dynamic V.A.S.T. synthesis -- a buzzword, to be sure. At least there aren't very many of these (Roland takes the cake in this regard). What is V.A.S.T.? Your guess is as good as mine at this point.

2.) VA-1 Virtual Analog Synthesizer -- another buzzword, but apparently it has to do with a virtual-analog synth that used to exist, but now is part of the PC3. Interesting, but considering all the other marketing-speak we've read today not really all that interesting.

3.) 16 insert effects, and they can be chained together or run in parallel. O.k., this is kind of interesting.

Well, perhaps the Specs listed on the Sweetwater site will give us some clues about what this PC3 thing can do:

# Polyphony: 128 Voice Polyphony, dynamically allocated, lightning fast performance.
# Multitimbral: 16 parts (one per MIDI channel).

...

# User Memory: 1,024 locations for user created preset programs. 1,024 locations for user created MIDI Setups.

Not much there, either. We can play up to 16 parts (which, I believe based on the marketing materials, means Programs) at a time, but still no word on what a Program consists of.

Then, once again it's time to take a trip to the manufacturer's web site:

Dynamic V.A.S.T...

Under the hood of the PC3 is simply the most powerful synthesis engine ever created by Kurzweil. Dynamic V.A.S.T. picks up where the K series left off. Having up to 32 layers per program means that emulations of real instruments can be extremely detailed and accurate, while the synthesized sounds can be extraordinarily large and complex.

Expanding upon our V.A.S.T. architecture, the PC3's new architecture delivers more power and flexibility combined with unprecedented ease of use. Users can now create and store their own DSP algorithms, with nearly limitless possibilities for routing.

In another huge development, we've resurrected the VA-1 Virtual Analog Synthesizer, having built it's synthesis engine right into the new system. Hear the VA-1's power-shaped, anti-aliasing oscillators (including supersaw, triplesaw, and hard sync oscillators) combined with a newer, larger set of V.A.S.T. filters (like the classic-style 4pole low pass w/ resonance) and non-linear processors and you'll understand why the internet is buzzing about this feature.

WTF?? Wait... Did they say 32 Layers per Program? Hey now, the best contender so far was the Yamaha, which offered 8 Elements per Voice. If I'm understanding what this says correctly, then that means the Kurzweil equivalent of the Yamaha Voice is a Program, and a Layer corresponds to an Element (more or less). 32 is a truckload of Elements or Layers or whatever to play with. Assuming, of course, that a Layer has as much stuff in it as one of those Yamaha Elements.

And what's this bit about creating "algorithms" -- "Users can now create and store their own DSP algorithms, with nearly limitless possibilities for routing." Now you've got me thinking... Does this mean I could do something crazy like run an oscillator into a distortion and then into a filter? And then duplicate that 32 times and that's a single Program? I'm interested. Let's keep going:

Cascade Mode...

Every layer in a program can be routed through the DSP of any other layer, in series or blended together, with up to 32 layers. It's the same concept as Triple Mode on our K26 series, without the 3 layer limit. You can run your sound through literally dozens of filters (which can all move independently in real-time) in addition to Kurzweil's signature DSP processors like Shaper, Wrap and Distortion, allowing for some truly dazzling results. This is one more feature which makes the PC3's synth engine so powerful and so different than anything else on the market.

Oh wow... Now I haven't seen anything like this on the other synths we've talked about so far. There are 32 Layers, and I can "cascade" any Layer into any other Layer... This isn't just some random "new horizon", this is new flamingos, new oceans, new sunsets... Me likey. Maybe there's more:

Arpeggiators & Riffs...

The PC3 sports sixteen (yes, 16!) independent, fully programmable arpeggiators featuring 48-step programmable, pattern generators with both pitch and velocity controls - per zone (allowing for up to 16 patterns to playback simultaneously). These are accompanied by a thorough collection of modes, scales and pre-programmed patterns to get your juices flowing.

Wait... Did I read that correctly, that this thing has 16 Zones? Now, if a Zone on a Kurzweil is similar to a Zone on a Yahama, then hot dang I can play up to 16*32 = 512 of these Layer thingies at one time. We've gone way beyond "WTF" and new world territory at this point... This is more like rewriting the laws of physics to create a new universe in comparison to the "new horizons" offered by the Korgs and Yamahas and Rolands.

I mean, check out the numbers:

1.) 32 Layers per Program
2.) 1 Program per Zone
3.) Up to 16 Zones mapped across the keyboard at any given time.

Dude, I gotta figure out what one of these "Layer" things entails. Not much about it on the PC3 page of the Kurzweil site. Maybe there's something in the Knowledge Base:

http://www.kurzweil.com/KnowledgeBase.php?product=181&topic=11&entry=219

O.k., some stuff about assigning Control Sources and... Bingo! About half-way down the page, after all that verbiage about setups, we see a nice screen that says "Layer 1/1" in the upper-right-hand corner. I'ma be bold and presume that this is what a Layer is -- whatever they're looking at on the screen at that point in the post. And joy, they show an "Algorithm" too! Now that looks cool -- you have DSP "block" that you can connect together to form an "Algorithm". That Dynamic V.A.S.T. stuff, then, must mean that you can pick the blocks you want and wire them together however you want. Nice.

I'ma cheat a little at this point since this post is way longer by now than it should be. After some manual-reading you'll figure out that a PC3 "Layer" consists of:

1.) Four DSP "blocks" that can be arranged in any order you want.
2.) A bunch of DSP functions that take up 1 to 4 dsp blocks.
3.) Up to two "Keymaps", or samples, can be loaded into a single Layer.
4.) A dedicated Amplitude Envelope.
5.) Two other Envelopes that can be routed to different destinations.
6.) Two LFOs that can be routed to different destinations
7.) Four FUNs -- mathematical functions -- that can be used to modify any of the Envelopes, LFOs, or whatever else you can load in to them.
8.) A separate PITCH block that can be controlled by envelopes, LFOs, FUN, etc. as well.
9.) A page where you can set Monophonic/polyphonic playback, portamento, and legato settings for the Layer.
10.) A page where you can set up the key and velocity range of the Layer.
...

I could go on, but frankly I'm tired of typing and you're probably tired of reading at this point. Dang, a single Layer in the PC3 is pretty freakin' awesome. For one thing the Layer can do all sorts of different things depending on what DSP blocks you use in it. For another, each Layer can have its own settings for envelopes, LFOs, monophonic/polyphonic playback, portamento, legato, and a whole bunch of other stuff. Remember, you can have up to 32 of these Layers per Program and you can "cascade" the output of a Layer into another Layer.

Now that's what I call a synth, and compared to the competition there really is no competition. Of course the PC3 (and all the other synths mentioned, to be sure) is a lot deeper than what I can write about in a single blog post, but the basic gist of all this is it blows everything else completely out of the water where sound programming is concerned. Seriously, nobody else is even close (at least nothing from the Big 3).

Finally, just for kicks, here are the current prices on the Sweetwater site for the 61 key versions of the instruments I've discussed:

1.) 61 Key Korg M3 -- http://www.sweetwater.com/store/detail/M3-61/ -- $1899
2.) 61 Key Roland Fantom G6 -- http://www.sweetwater.com/store/detail/FantomG6 -- $2299
3.) 61 Key Roland V-Synth GT -- http://www.sweetwater.com/store/detail/VSynthGT -- $3299
4.) 61 Key Yamaha Motif XF6 -- http://www.sweetwater.com/store/detail/MotifXF6 -- $2399
5.) 61 Key Kurzweil PC361 -- http://www.sweetwater.com/store/detail/PC361 -- $1499

I don't know about you, but I think at this point it's gettin' a bit obvious...

Kurzweil Rocks.

PK3 Apparition

2010-09-07T12:36:00.000-07:00

I've mentioned this on forums elsewhere on the 'net so I ought to do so here -- my soundware release will be titled "Apparition". PK3, in case you were wondering, essentially designates what I call my little, uh, company (Poser's Kitchen), and the 3 denotes that this is for the PC3xx line of synths. I'd previously put out a few freeware sounds for the K2xxx using the prefix "PK", so the 3 helps distinguish my PC3xx sounds from my K2xxx sounds. Anyways, my brother snapped a cool pic while we were creature spotting one evening that inspired the name and will appear somewhere in the materials you'll get as part of the soundware package.

This thing is far from done; I've started twice already and gone back to rethink how I generally want this to work. The basic idea has evolved to something like this:

1.) A bunch of "synth" Programs and Setups, with presets, will be the "core" of the package.

2.) Since I want these to play nice in Setup mode, the BCR templates should support stuff like easily assigning a midi channel to a particular "synth" controller template. Also, I want to avoid assignments that coincide with the front-panel slider defaults. I figure that's the only way to be truly safe; if you were to load up four or five "synths" in a setup then moving a slider on the front panel could have unintended consequences if those synths used the sliders' default CCs.

3.) This needs to be easy for anyone with a PC3xx to use/enjoy. I might leave out some stuff because most people would scratch their head and go "uh, how the heck does this work?". Also, in the event that someone doesn't have a BCR, I want the programming to be compatible with other midi controllers (or just the front-panel sliders). There may be some minor feature loss, depending on how a particular "synth" is programmed, but many of the functions (assuming your midi controller has the minimum number of required CCs) will work as advertised.

4.) Each synth should take advantage of performance controls. The first few I programmed mostly used knobs or buttons on the BCR. However, I'd like to add in stuff like velocity/pressure sensitivity, mod wheel functions, and so on. On many synths the player has control over these things; similarly, I want to provide some control logic (for instance, a button to assign aftertouch to modulate pitch and/or frequency cutoff) so it's not an always on/always off sort of affair.

What this means is I'm going back to the drawing board to revamp the stuff I've done and modify plans for some of the stuff I'm working on. Another dark horse is the new OS update; from what I gather it will have some nifty features that I'd like to incorporate into at least a few of my "synths".

And so the labor continues. I've also been getting antsy to actually make music (and not just program), so that might delay things as well.

Ship Shape

2010-07-22T14:08:00.000-07:00

One of the blocks that's eluded me for awhile is Shaper. Not because I couldn't use it, but rather because I wasn't sure about what it was doing internally. I think I have a better idea now, thanks to some perusing I did of old email list postings.

Gordon Reid wrote some enlightening tutorials for Reason users on the Propellerheads website. One in particular on Phase Modulation synthesis is relevant here:

http://www.propellerheads.se/substance/discovering-reason/index.cfm?fuseaction=get_article&article=part25

If I understand correctly (and that's certainly something worth questioning..), I believe that SHAPER essentially does this. If you put in a Sawtooth wave and set Amt to 1.0, you'll get a sine wave as output at the same frequency. In other words, you can think of the sawtooth wave as "picking" which part of the sine wave to play back. Since a (rising) sawtooth wave looks like this:

/|

that means you can consider SHAPER as containing a wave table that outputs a sine wave when you read the table from beginning to end at a steady rate. The shape of the incoming waveform determines the reading rate (a sawtooth increases linearly and then resets, thus the table is read from beginning to end repeatedly), while the Amount parameter determines how to "scale" the input. At 1.0 there is a 1 to 1 correspondence between the incoming waveform and the indicies of the wave table. Less than 1.0 means that one full cycle of the sawtooth waveform will read only partly through the table. More than 1.0 means that a full sawtooth cycle will read the table more than once.

In some ways this is like hard sync, with the sine wave held in the table considered to be the "slave" and the wave present at the input considered to be the "master".

Anyways, that's enough for now. I've got other things to do so I'll have to come back to this later.

graph theory cont.

2010-07-13T12:04:00.000-07:00

I return to this with a bit more wisdom... Here's the layout I've settled on:

op1 -> op 2
|
op3 -> op 4
|
op5 -> op 6
|
op7 -> op 8

This allows things like two independent 4 ops with feedback and up to 5 ops chained in series. The control scheme uses all 51 of the free CCs in the PC3 without a significant loss of function. Ops 1, 3, 5, and 7 can use the amp envelope from their respective pair ops (i.e. 2, 4, 6, and 8) to modulate frequency or gain in conjunction with a variable-shape LFO. Ops 1, 3, 5, and 7 have "fine" pitch controls (i.e., 16384 midi values instead of 128), while the even-numbered ops have a "coarse" pitch control. The even-numbered ops have a "level" control for mixing their output, while the odd-numbered ops have both an op-specific gain control that works in conjunction with the mod sources and the ability to "mix" the ops further down the chain with each other to varying degrees.

That's one design; I have others that I'm working on. There's a cut-down version of this that I'm going to post for free in the next few days.

graph theory

2010-07-06T13:41:00.000-07:00

Today I am exploring FM synthesis and how to set up a flexible "synth" Program for doing FM on the PC3 (to be part of my forthcoming soundware release). A little bird (i.e., something someone said on a forum) tells me that there might be a more direct way to get FM in the to-be-released OS than the usual sine + saw + shaper + amplitude sorta thing that we'd otherwise do on the PC3. That being the case, I'm mostly designing from a level above; i.e. how to lay out various "operators" so that the resulting network has a high degree of allowable signal path complexity. So far, I'm leaning towards this layout (warning -- bad ascii art):

[op1]
|
[op2]->[opA]->[opB]
|
[op3]->[opD]->[opE]
|
[op4]->[opG]->[opH]

The basic idea is this; the signal flows from op1 to op2, where it can take two different paths. It can be mixed with or bypass op2 and proceed to op3 and/or it can be mixed with or bypass op2 and the result be sent off to opA (then B, then C). I currently think that this graph/layout might provide the greatest amount of flexibility relative to the amount of resources required to create it. Assuming one layer per FM op, that's 10 layers (if we assumed 2, we'd still be well under the 32 layer limit with 20). The main problem I've run into is a lack of free midi CCs. I assume one ADR envelope per op (that's 3 midi CCs per op, for 30 total), plus one high-res Frequency knob per op (that's 20 midi CCs since two are required per high-res control). Right now that puts this scheme at 50 midi CCs, and there are 51 currently available for use on the PC3. Thus I cannot implement the routing logic I'd need to make the "network" usable -- 2 CCs for ops 2/3/4 and 1 CC each for the others.

Now, to get back to the title of this post... I wonder if there isn't a more efficient way to implement this? I'd like to keep the "chain" at its current size if possible. At the same time, I'd really like a few CCs around so I could have at least one adjustable LFO per op, and maybe throw in a filter of some sort at the end. So, I will continue to pick my brain and determine if there is a better way. Feel free to suggest such in the comments. Otherwise, I think I'll have to stick with a max chain length of 4 or 5 ops.

A note to prospective commenters; the PC3 currently does not allow "feedback", i.e. feeding the output of a layer into itself.

Soundware is coming...

2010-07-03T01:12:00.000-07:00

It's time to take the gloves off and start gettin' down to business. What does that mean for all you lovely people out there who read this not-too-often-updated blog? Well, it means that you will get the ring-side seat as I work my way through a soundware release. We'll do fun things like explore different DSP routings, talk about the minutae of control sources, and generally get our hands dirty.

So, with that out of the way here's the current plan. Subject to change, but generally speaking this is what I want to put together:

1.) 50 "synth" programs

2.) 24 presets per program, for a total of 1200 presets

3.) A Behringer BCR2000 pre-programmed with all the necessary control templates

4.) Some sort of documentation, and I'll probably make a video demo of each "synth" program to demonstrate how it might be used and how the BCR2000 interacts with it

I don't know what it'll cost yet, but there will probably be several versions available (for instance, you'll be able to buy the collection of Programs without having to buy the BCR if you have your own midi controller; that'll cut down the price quite a bit). Also, I will probably create sequencer files that you can play back at the PC3 to "recall" the presets, so even if you don't have some sort of knob/fader box you would still get 1200 sounds for your PC3.

I've been working on the conceptual stuff quite a bit and I'm about half way through with determining what sorts of "synths" will be included. Certainly a number of relatively faithful emulations of different classic synths will be included -- i.e. your basic Minimoog/Prophet/OB etc. These will be fully-functional emulations with as many features as possible mapped to the BCR. Some will be "mega" synths that you'll load up and play in Setup mode.

O.k., now that I've written all this down I'm committed to the project. At some point in time I'll need to give a name to this mess... Any suggestions?

Introducing: The PK3 VA03

2010-05-19T13:10:00.000-07:00

The PK3 VA03 provides not only squelchy Saw and Square filtered madness, it also includes a built-in 8 step sequencer and it all functions with the stock PC3 -- no external controllers required. Based loosely on the TB-303, the VA03 sports a lowpass filter with up to 48 db of resonance (resonance control is on the Pitch Bend wheel) and includes a low-frequency "bump" to simulate a similar peak in the TB-303's frequency response (see here: TB-303 filter curves). The Program is designed to be used in Setup mode (the included PC3 file will load both a setup and a program) so that SWPRG 1-8 can act as mutes for each sequencer step. Sliders A-H control Pitch (6500 cents), and Slider I controls "tempo". The "tempo" is clocked internally to a sawtooth LFO and is not "locked" to the system tempo, thus it may drift relative to the internal sequencer. With Slider I all the way down the playback rate is approx. 8th notes at 120 BPM. With it all the way up the playback rate is approx 16th notes at 120 BPM. Note that the VA03 Program includes controller assignments on its Info page.

Audio Demo:

VA03 demo (.wav)

PC3 File:

PK3 VA03 .pc3 file

Mega synths on the PC3

2010-05-17T09:33:00.000-07:00

Setup mode is interesting for many reasons, but one of the most intriguing (to me, anyways) is the idea of creating a "mega" synth Setup. One could think of the various synth programs that I'm doing to be just another way to organize some of the PC3's functionality into easier-to-access chunks. This idea is basically along those same lines, but there are a couple of tricks you can do in Setup mode that may (I need to actually program this stuff to see if it works) expand what's already possible in Program mode.

The two bits of Setup mode that I'm currently eying are a.) picking a Zone's effects as the "master" effects for the Setup, and b.) picking a Zone's arpeggiator as the "master" arpeggiator for the Setup. Why would you want to do that? Consider the following scenario:

Let's say you want to model (or do a takeoff on) the Yamaha CS-80. The CS-80 is, essentially, two independent synths stacked together with a master section. One of the limitations (if you can call it that) of the PC3 is there are only 50 free midi CCs for controlling various parameters available for a single MIDI channel. That might sound like a lot, but the CS-80 has a good 70+ controls on its front-panel. This means if you want to emulate the CS-80 as closely as possible on the PC3, you'll have to figure out some way to do it by using more than one Program.

This is where Setup mode comes to the rescue. You can create a single Program for each of the two synth sections of the CS-80 -- both sections are the same. The trick is to then "combine" the Programs using KDFX in Setup mode. This includes, for instance, allowing "voice" A to mix with and/or modulate "voice" B in a Ring Modulator. I won't go into all the details here of how the CS-80 operates; the main thing is to realize that you can do all of that and _still_ have 50 controls left over for KDFX stuff. You won't need all of them (each KDFX chain is limited to 15 FXMods), but you wouldn't have access to them outside of Setup mode because there aren't enough free Midi CCs to control both a single synth voice and to control the effects. Most control surfaces also don't have 70 controllers (the BCR2000 is an exception -- it has 56 knobs (the top row of 8 has 4 pages) and 48 buttons (the top row knobs can also act as push buttons)) available on the front-panel but they can often be "paged". So, a logical layout is Page 1 = synth section 1, Page 2 = synth section 2, and Page 3 = Master section.

And, so, in Setup mode you would have three Programs -- Programs 1 and 2 would be the synth sections, and Program 3 would be the "master" section. This is where Setup mode comes in handy -- since you can specify that a particular Zone's effects be applied globally, you should also be able to load up a "silent" Program in Zone 3 with an Effects Chain attached and apply its effects to all the other Zones in the setup. The Program, in this case, is simply a container for the Effects Chain, and any modifications to that chain via FXMODS should affect all the Zones running through it.

Note that a similar thing can be done with the Arpeggiator. Although the arpeggiator doesn't allow real-time control of the notes for specific steps in a step sequence, it is still quite flexible and you can program in your own step patterns. There are plenty of other parameters that you can control in real-time: note duration, step pattern (both pattern and bank), latch mode, and so on. So for our "mega" synth you would use Zone 4 purely for its Arpgeggiator and control its parameters via another page on your control surface.

Now that's pretty cool -- you get full control over each synth program, control over the effects section, and separate control over the arpeggiator.

But that's not all... (sometimes I feel like an infomercial salesman). You still have 12 zones left over. Since your Arpeggiator and Effects are both global, you could turn your CS-80 into, say, a CS-120, or a CS-160 just by loading the same program you used for Zones 1 and 2 into any (or all...) of the other free Zones. Or, you could load up a different program entirely. This allows for fun what-ifs: what if I had a CS-80 plus a Moog (or PolyMoog) combined to form the Monster Synth of Doom? Or you decided to combine different bits of Yamaha's history into the same "synth" -- mix in a DX-7ish program along with the Moog, or make a CS-Moog or a CS-Prophet... You get the idea. These are Mega-synths, and I think they could be quite cool.

BCR2000 + PC3 fun

2010-04-19T13:29:00.000-07:00

More fun with the bcr2000 and the pc3. One of the cool features of the BCR is its scripting language; you can script a knob to send any number of values between 0 and 16383, or up to 125 bytes of custom sysex (including variables that change as you turn the knob), and do all sorts of other things. One fun thing I did today: I assigned CC#13 to Src1 and CC#87 to Src2 on Pitch for a square wave. Then, I set up a knob on the BCR to split a 16-bit midi value and send the low-order byte to CC#87 and the high-order byte to CC#13. What this does is, basically, create a control with the range of an NRPN. The trick to getting this right on the pc3 is setting the Src2 range correctly. Src1 has a max range of 7200, which will be scaled over the 128 values of the higher-order byte. Doing a bit of math (i.e., 7200/128) we see that each higher-order value increments tuning by almost 44 cents. Setting Src2 to 44 cents, then, gives us the desired result. So, instead of only having 128 values available over 7200 cents, we now have 16384 and thus much better real-time control.

Introducing: The PK3 H8 Oh H8

2010-04-10T07:21:00.000-07:00

Yeah, that's right, Hate Oh Hate. That's what your competition will do when faced with the awesomeness of the PK3 H8 Oh H8. Sixteen well-placed drum and percussion sounds hand-tuned like a fine steak. Mmm, steak. All for the low price of free.

The H8 Oh H8 uses VAST power to recreate/mangle 16 classic drum sounds. Mapped to all-white keys (mmm, white meat) between C4 and D6, the layout is as follows:

C4 - Bass Drum
D4 - Snare
E4 - Rimshot
F4 - Clap
G4 - Low Tom
A4 - Mid Tom
B4 - High Tom
C5 - Closed Hi Hat
D5 - Open Hi Hat
E5 - Crash Cymbal
F5 - Clave
G5 - Shaker
A5 - Low Conga
B5 - Mid Conga
C6 - High Conga
D6 - Cowbell

And it sounds a little something like this:

The Sounds of the H8 Oh H8

This program is "PC3-Friendly"; it uses 5 of the front-panel sliders to control BD Click (Slider A), BD Decay (Slider B), SD Snappy (Slider C), Clap Decay (Slider E), Crash Cymbal Decay (Slider F). The linked audio demo steps through each sound in succession and demonstrates how the sliders control their respective parameters.

You can get the PK3 H8 Oh H8 by clicking below:

H8 Oh H8 .pc3 file

PK3 Mugg available for download

2010-03-24T14:45:00.000-07:00

It was a dark, warm night in Southern California. A vintage synth (small, boxy, somewhat out-of-tune after a long gig) wandered down a dark alley. The old cat was still hip, he thought, even after 40 years. When suddenly....

--BLAM!--

Introducing the PK3 Mugg. This is the first in the somewhat stinky line of synth programs from Poser's Kitchen for the PC3 (hence the PK3 prefix). The Mugg is a good ol' subtractive synth. Three oscillators with continuously-variable waveforms (er, well, almost) -- fade from triangle to saw to square with a single knob (or slider, or piece of midi'd toast, or whatever). The Mugg uses three filters internally to accomplish its magic, with one set of controls for Frequency, Cutoff, and Envelope Amount. A switchable amplitude envelope (turn it off for a much faster attack and no decay) with adjustable Attack, Decay, and Release, can be assigned to give shape to your final sound. The Amplitude and Filter envelopes share the same maximum Attack (0.1 ms), Decay (0.2ms) and Release (0.2 ms) times.

All oscillators have a Range control to shift between one of six octaves. The lower two oscillators have a Fine Tune control for tuning up by as much as a fifth. Each oscillator also has a mix control for adjusting relative levels.

Below are the Midi CC# assignments:

Oscillator 1:

Range: 72
Waveform: 74
Mix: 75

Oscillator 2:

Range: 81
Fine Tune: 82
Waveform: 83
Mix: 84

Oscillator 3:

Range: 88
Fine Tune: 89
Waveform: 90
Mix: 92

Filter:

Frequency: 76
Resonance: 77
Envelope Amount: 78
Env. Attack: 79
Env. Decay: 85
Env. Release: 86

Amp Envelope:

On/Off: 29
Attack: 93
Decay: 94
Release: 95

It is strongly recommended that you also map the following standard MIDI controllers:

Pan: 10
Volume: 11
Portamento Time: 5
Portamento Switch: 65

The PK3 Mugg is available immediately at a fine Internet near you:

PK3 Mugg PC3 File

PK3 Mugg

2010-03-21T15:43:00.000-07:00

Here's a vid:

Here's the lowdown.

This program currently uses 15 layers. A basic "oscillator" uses 5. The anatomy of an oscillator is, essentially:

Triangle Wave Keymap + Saw --> PWM + previous layer --> Mogue Filter + previous layer

These three layers are sent to two AMPENV-secific layers. The first uses an ampenv with no attack time, no decay time, but adjustable release. This is nice for sounds that require a bit of "punch". The second uses an ampenv with 0.1 seconds of attack, 0.2 seconds of decay, and 0.2 seconds of release (at their fastest). The Kurzweil way of doing things is to either multiply or divide the envelope times via a multiplier, as opposed to directly increasing or decreasing the time of a particular envelope stage. So, for instance, an envelope segment of length 0.1 seconds can be lengthened by multiplying it by 50 (the maximum multiplier) to 5 seconds. The switching between envelopes is done using the Enable parameter on the LAYER page. In this case it's set to a midi cc# (assigned to a button in the video) so that one layer is on and the other off when the value is 127, and vice-versa.

The waveform select knob is interesting; it's doing a number of things behind-the-scenes and is, in many respects, the trickiest part of the program. The basic idea is to change the mix between the saw and triangle waves on layer 1 with the first half of the knob's travel, then simultaneously kill layer 1's output and control the width of the PWM block on layer 2 when you move the knob past the halfway point. I'm also slightly adjusting gain (mostly on the triangle wave) to keep the perceived level of the various waveforms more-or-less constant as you move the knob. The switching is mostly done via FUNs. For instance, to switch the PWM layer "on" I use the FUN b > a to determine if the knob is past its half-way point. I am also trying to sweep the knob in such a way that, when you're fading between the saw and triangle waves, you get a full crossfade with only half of the knob's travel. This is done using the FUN b/(1-a), where a = 0.5 and b = the cc# of the knob in question. Interesting things, incidentally, happen when b > 0.5 -- I'll have to post a sound file, but basically the FUN "oscillates" and the rate of oscillation is determined by how much b exceeds 0.5

The coarse tuning knob is implemented using the FUN Quantize B to A. This FUN is cool because you can make anything into a stepped control source. So, B is the cc# in question and A is (generally) a value between 0 and 1. 0.66 is the value I'm using here to give the knob 6 steps. The tuning itself is also somewhat interesting. Since I'm using a keymap for the triangle wave, I have to adjust the TimberShift so that I can bend a sample root through 5 octaves. I'm away from my keyboard at the moment (a happy cat is napping on its cover and I don't want to disturb him), but I believe you shift up (24st). What this does is force a sample that was recorded at a higher sample rate to play back at a much lower sample rate (you can "bend" a sample down in pitch by playing it back at a slower sample rate) so that, on the third click of the tuning knob it's playing at the pitch you'd expect for a particular key. The DSP oscillators don't have that issue, of course.

Anyways, that's the basic gist of what's going on under the hood. I'll post the cc# mapping table and the pc3 file for download in a couple of days after I work out the kinks (among other things, I want to get rid of the bit of saw that you hear when the knob is fully towards the triangle wave part of its travel).

It has arrived!

2010-03-17T19:15:00.000-07:00

And, we reach the turning point. I am now the proud owner of a new PC361. I've been programming and playing off and on today, trying to figure things out and soak up what I can about what this thing sounds like and how it operates in real life. Mostly I've been messing around with some tentative programs; nothing worth posting yet, I'm still acclimating myself to the beast. And what a beast it is soundwise. This is gonna be fun, if a bit tiring at first...

I just.. can't... do it -- not selling my BCR2000

2010-01-20T11:48:00.000-08:00

As the title suggests, I just can't bring myself to sell my BCR2000. The thing is too darn useful. I really like the MPD32, but its controllers aren't quite as flexible as those on the BCR plus it doesn't come with a power adapter (you have to buy that separately) or a power switch. So, I'll have to keep both despite my best efforts to slim down my setup to a single controller.

Why is the BCR2000 so necessary? The biggest feature it has is the ability to send "snapshots" of all its current controller data. So, when I program some crazy CS-80 emulation on the PC3 that requires recalling 80 or so different midi parameter values, I'll have the ability to send (and record) snapshots so I can save my "patches". A real sleeper feature, though is the "learn" function. This nice feature allows you to hit learn, pick any controller, then shoot some midi at it. Then, the next time you operate the controller it will spit back out that midi data. Combine this with the snapshot feature, and you have hardware storage of your "patches". That's right, the BCR will remember the "learn" data through power cycles. Pretty neat, eh?

The Akai is great for doing drums; the rhythms I play are both easier to record and "feel" much better. Plus I can experiment with some of the more interesting rhythms I want to use in my songs that I'm able to play but I can't program. I don't mean weird time signatures, rather, I mean doing stuff like varying the amount of swing or putting notes in the "wrong" places in interesting ways...

There's an instructive moment in "This Is It" (the Michael Jackson performance movie that came out last year) where Michael is trying to get his band leader to play "The Way You Make Me Feel" rather seriously off-beat. The band leader (can't remember his name off the top of my head) tells Michael that he needs Michael to back him up when they rehearse with the full band, because otherwise everyone will think he's playing the part wrong. And, indeed, most people would think that. That is, until Michael starts singing and the rest of the band falls into place -- then it's a monster of a groove. Anyways, to be a rhythmic visionary you have to break out of the straits of quantization, or even "swing" as defined by software, and just play the part.

Kurzweil at NAMM

2010-01-18T10:52:00.000-08:00

I paid a visit to the Kurzweil suite at the Hilton on Saturday after NAMM. For a techie nerd like me this was like walking into an ice cream shop. Anyways, the Kurzweil folks are very cool; I had a good chat with Dave Weiser and co and got to ask some burning questions. For instance; I learned that the PC3 OS runs on a different chip than the DSP, which explains why FUN evaluation times are basically the same. The OS chip is newer, though, so it's doing a lot more than what the equivalent subsystem did on the K2000. The "Mara" chip is dedicated completely to DSP -- a smart move, because then it's not bogged down doing OS duties.

I've thought for awhile that a similar thing would work well with an off-the-shelf processor (i.e., a four or eight core CPU from Intel or AMD) and an off-the-shelf graphics card. Apple, for instance, is pushing OpenCL as a way to use graphics card resources for general computing. A graphics card doesn't work the same as a normal CPU -- both implement a "pipeline", but a graphics card is set up to operate on a matrix of values while a general-purpose CPU operates on one value at a time (or several in sequence, the number of values depends on the length of the pipeline). The OS can handle orchestration, while the graphics card can handle raw calculations for processes that benefit from it's parallel nature.

Anyways, I took a couple of photos of the new PC3K (this is a PC3 blog, after all):

This is the "Blue Ray of Death" v2. The PC3K has both USB host and device ports so you can plug it into your computer for MIDI and plug in a USB flash drive for storage.

A sidelong shot. All the controls on the front panel are the same (or at least they looked the same to me when I glanced over it) as the PC3xx. The keyboard above it on the rack looked to me like the next iteration of the SP2X.

I want to give a big thank you to Dave and the whole Kurzweil crew at NAMM -- y'all rock!

New toys arriving soon...

2010-01-04T11:19:00.000-08:00

It's time to update/optimize my controller arsenal (which is really quite puny, a BCR2000 and a Midiman Radium). I've ordered an Akai MPD32 which should be coming sometime this week. This is to replace both of my previous controllers; I'm not quite sure yet how I feel about that, but the idea (in my mind, anyways) is I want everything to be functional in multiple ways. So, the MPD32 will serve dual-duty as a controller for software and a yet-to-be-purchased PC3xx. In that vein, I'm quite enamored of a little box made by Midi Solutions -- The Event Processor. This little box is cool because, among other things, you can remap one type of midi message to another. So, in the case of the MPD you can't use the pads to send CC data. The pads support polyphonic aftertouch, which may or may not be useful for any particular device. With the Midi Solutions box, you can remap a note and its aftertouch data to a Midi CC # and its value. This essentially turns the pads into pressure "knobs", and could be fun for adding vibrato-ish modulation to all sorts of things. I originally wanted an M-Audio Trigger Finger because its pads can be set up to do this by default. However, the MPD offers many other features that I'm interested in and I got it for a good price, good enough that even with the Event Processor the total still comes out under MSRP for the MPD32. Anyways, for the first while I'll be using the MPD strictly with software, and I plan to continue my Reason/Record tweaking accordingly. More soon, as this is the season of gear acquisition and optimization.

Blech, work

2009-09-23T16:18:00.000-07:00

My day job (computer programmer) is being a bit obtrusive, hence no goodies over the last few days. Somewhere in there I also went on vacation and couldn't get networking operating on the computer I had with me (turns out it just needed a reboot -- good thing I don't do tech support). Anyways, I was watching a video interview with the guy who did the sound design for Wall-E the other day and he did a cool demonstration with a spring suspended from a ladder. If you tap the spring near the bottom and put a pickup on the other end, you get a nice laser-ish sound. The idea is high frequencies travel faster through the spring than low frequencies, hence why it sounds the way it does. It's sort of like spectral delay, only more "smooth". The PC3 has some pretty nifty filters and filter-like effects (especially the Pitcher) which could be useful for such things. I'm thinking that something like four Pitchers spaced throughout the frequency spectrum with adjustable delay times could be fun, on single-cycle or more complex waveforms. More fun would be to implement a delay loop with each of those frequency bands, or in other words doing physical modeling-like things independently on highly-pitched components of an arbitrary sound. Fun, no? The PC3 presents some challenges in creating the delay loops; I'm curious to experiment and figure out if such things can be implemented in the Program side of the synth (as opposed to KDFX).

BCR2000 + Thor = really sweet step sequencer

2009-09-12T10:01:00.000-07:00

I'll have to do a vid of this setup some time, but it blows the previous arrangement I'd be using in Reason 3.0 away. Thor is a pretty cool device, especially for real-time step sequencing. The only gripe I have is not being able to switch patterns (aside from activating/deactivating steps in real-time, then recording those changes and moving the resulting midi clips around the track to get patch changes) on-the-fly. In all other respects though, I think it's a far superior solution. For one thing you get 16 steps. Also, you have control over a bunch of different parameters for each of those steps (pitch, duration, gate time, etc.). Another minor drawback is you don't get the lights on the BCR chasing the current step. That may perhaps be a matter of customizing the .remotemap file. Anyways, this is really slick and it alone is reason enough for me to be happy with my upgrade to 4.0/Record.

The real fun will happen when I get a PC3. I already have ideas in my head for making a tb303 patch in the PC3 that you can sequence using Reason (the key here is the PC3 can do all sorts of fun mapping tricks that you can't do on other synths), along with various custom drum patches that could be sequenced in the same way. I love modern technology.

Nerdiness unleashed -- an idea for sequencing hardware with Reason

2009-09-10T16:57:00.000-07:00

First off, a vid I made a month or so ago:

I post that as a lead-in to the topic at hand; how to use Reason to sequence outboard gear. I'm sure most people think that's not possible, and in some ways they're right. There is no MIDI out in Reason (or Record, as noted below). However that's only partially true.

Reason supports something called "parameter feedback". This allows it to send "feedback" to an interface about what's currently going on in the software so that the interface can do things like light lights, move faders, and so on. This is, for all intents and purposes, sequencing and most midi controllers do parameter feedback via MIDI (i.e., if you send MIDI data through the device that corresponds to mapped controllers, those controllers will respond to the incoming MIDI data). So all you need, then, is some way to pass this MIDI data through to the outside world. Then any control you have "mapped" to Reason will correspond to a control whose data you can send to something else -- an external sequencer, a keyboard, or any other device that responds to MIDI.

Here is where the BCR2000 comes in. There are a number of different "modes" that the BCR2000 can operate in:

BCR2000 manual PDF file (see pg. 10)

The magic happens in USB Mode 4. Do you see what's happening with the "parameter feedback"? It's being sent out MIDI Out A! I don't know how many other MIDI controllers offer that sort of thing, but this is the key to getting Reason's internally-generated MIDI data to the outside world.

This is all mostly theoretical; I've haven't had a chance to try it yet but I'm fairly confident it will work. It might be necessary to route MIDI Out A back into the BCR and then use MIDI OX or something similar to massage it a bit, but my guess is the midi coming out of Out A in this mode would be pretty standard.

Needless to say, this could be very useful for sequencing drums on devices like the PC3 when you don't have access to a nice DAW or you happen to like the Reason way of doing things. The PC3 has all sorts of fun features in its sequencer/arpeggiator, but this could really take it to another level (including, among other things, creating a control-for-control "clone" of the Subtractor in a PC3 patch that you can remotely control from a Subtractor instance within Reason...).

Now, if we could just write our own custom .remotecodec files, you could do some really fun things...

Quick Review: Propellerhead's Record

2009-09-10T08:03:00.000-07:00

Hey, wait, I thought this blog was about Kurzweil stuff! It is, but those things don't exist in a vacuum. My main instruments are still guitar and bass so, when it comes to recording music, they tend to get included in many things. The PC3 includes a number of presets that emulate different real-world instruments, so I'm excited to be able to record stuff that can make use of those sounds along with real guitar and bass tracks. If you're a fan at all of funk, disco, rock, or jazz from the 60's/70's/early 80's, then being able to do that is a very good thing.

I've been using the beta of Propellerhead's new daw-ish software (really, in my mind, an extension of Reason) for about two months now. At first I was somewhat skeptical; I've been using Reaper for some time and I really like it. Among other things, Reaper is cross-platform so I can run it on any computer. It also has some neat and unique features, such as the ability to network computers and use their processing resources to reduce load on your DAW host. Anyways, I don't think Record is a replacement for Reaper (for two reasons, mainly; Record can't load plugins and there is no MIDI out unless you get creative with parameter feedback...) but I still wasn't sure if Record would be worth the money.

That's where Record for Reason Owners comes in. See, I'm still using Reason 3.0. As much as I like Combinators and the Subtractor, there are a number of annoying limitations that prevent one from doing interesting things with them together (I hate the limitation on the number of knobs and control sources/destinations imposed on the Combinator). Reason 4 didn't really solve this problem, but at least Thor provides more in the way of synthesis capabilities. So upgrading wasn't really cost-effective in my view because the new stuff offered didn't justify the price. However, Propellerheads Software is offering both a.) an upgrade for any version of Reason, and b.) Record at a decent price -- $149 -- with the Record for Reason Owners package. Essentially, you get your Reason upgrade and, for $20, you can tack on a daw-ish extension to Reason. That was enough to pique my interest.

Thus when beta testing was announced I decided to give Record a go. And, I'm happy to report, it actually goes quite well.

The things I really like are the main mixer (modeled after an SSL 9000K), the sequencer window, and the Line 6 integration. A note about that; I've read lots of negative stuff about their software and hardware. I agree that it's not an ideal solution (I'd rather mic a cabinet in a nice room), but for my purposes I think the amp sims are pretty good. Record comes with three guitar amps and two bass amps, along with assorted cabinets, for use in the Line 6 devices. This isn't much, but if you've got an iLok then you can get POD Farm for free during September. If you don't have an iLok, you can buy one (I think they're about $39) and essentially get POD Farm for about 50% off. Now that I've played with the included Line 6 stuff, that makes sense as a cheap way to access more amp sims.

I've found myself doing a lot more guitar/bass/drum recording since I've been using Record. Before I'd resigned myself to just doing electronic music because I didn't have a good room where I could set up an amp (or a drum kit for that matter). Record is cool in this respect because, with a bit of tinkering, you can simulate recording stuff in a studio and how instruments actually sound when you mic them. I put together a simple Refill that demonstrates this concept; if any of you have Record feel free to try it and let me know what you think:

Rock Band Refill

And some demos to show how you can get different "sounds":

Demo 1

Demo 2

Sunland

Some nerdy details: What I've done here is model how you'd mic these instruments in the studio. For instance, a microphone doesn't pick up "stereo" reverb, so each instrument runs into an RV7000 (which simulates the room ambience) in mono. For the drums (a Combinator that uses several ID8s to split out the ID8's "acoustic" kit) there are three mics, a mono mic on the kick, a mono mic on the snare, and a stereo mic on the overheads. Of course if you're recording live there's "bleed", so each instrument is routed into the other "mics". When you solo a channel on Record's main mixer, you can hear the "bleed". The nice thing is that, unlike in real life, here it's completely adjustable. The last recording (Sunland) is a bit different from the other two in that I added a stereo "Room" mic to the mix which isn't in the starter song included in the refill. I'll probably update the refill to include it. In fact, I want to put together a whole bunch of starter songs that have different "sounds" to them and put that out as a free refill. Also included are a couple of "tools" for the RV7000; one is for phase correction (it's an RV7000 set up as a simple delay, 100% wet (so you don't hear the original sound) with adjustable delay time) and the other is my version of what a studio might sound like.

And, so, I really like Record because it a.) fits in well with my workflow, b.) stimulates my music-making in a good way, and c.) provides for some interesting patching that can be used to simulate stuff or create weird sonic spaces.

Eh World

2009-09-08T12:17:00.000-07:00

Hello there. This blog will follow my further adventures with the Kurzweil PC3 and related instruments. I've written enough about this stuff in various forums, and I've been blogging for some time, so why not do both together? Yeah, that's what I thought, but I decided to do it anyways. Hopefully I can put together some nice "visual" posts to demonstrate stuff that I'm trying to explain. By "nice" I mean with pictures, videos, sound examples, y'know, all that kind of stuff (maybe flash? Yeah, right, like have time for that sort of wankery...). Also, I'm going to collect as many useful links as I can so Kurzweil users (and some of us are, really. There is no K.A. yet but we might need one in the future) can get their fix. The link section will include links to the various manuals that may prove useful to anyone traversing the, uh, "vast" terrain of Kurzweil's offerings, as well as websites, forums, and all the other nooks of the internet where such stuff can be found. I might also take a shot or two at other synths, most of which are just glorified tape players. Seriously, though, so much of the synth world has stagnated. The Roland Phantom lives up to its name.

I like things shaken, not stirred. There is no comment moderation but you'll have to squint and try to read letters (seriously... I get those things wrong at least every third time). It's annoying, but so is Viagra. Don't do anything illegal, but that changes every day so you'll all be on the honor system. All two of you, that is. I write other blogs for politics, and since you can't gerryrig a vote machine using Kurzweil gear (perhaps the subject of a future tutorial?) then that stuff probably won't fly well here. You can say it but don't expect anyone to care.

And so off we go. This will probably be a somewhat annoying ride in the beginning as I'm still PC3-less. That means I'm only programming in my head. Nuts, did I just say that out loud? The title of the blog is fitting, I guess.