Hands on/Hands off

One nice thing about being a hobbyist with music creation is that your productivity levels aren't that constrained. Common wisdom states that you should take the time to master a few instruments and focus your time and effort there. When trying to make a living or meet deadlines, good advice!

Since I am doing neither musically, safely ignored! I can instead keep my musical A.D.D. and flit from tool to tool or technique to technique. My most recent set of conflicting interests involve generative music and performance.

On the generative side, I've been writing about Mixtikl and Noatikl which are basically AI engines for music creation. I've been playing with these tools to create algorithmically produced music that sounds "human-like". I intend to do much more here in creating classical and/or jazz pieces that follow an evolving ruleset.

On the other hand, I have also been playing with my new EWI-USB MIDI wind instrument. This is a MIDI device that uses breath and fingerings to produce music. When I first got interested in electronic music, it was the Wave-table synthesis of the Creative Labs AWE-32 (see earlier blogpost) that managed to make some instruments sound almost real. At least on the computer of the mid-90s, they were recognizable if not expressive. Real sounding instruments with MIDI or electronic keyboards has remained somewhat elusive.

This is getting closer and closer to attainable though as technology continues to evolve. With the EWI-USB, there are two types of sound generators I've been using. The more traditional approach is with high quality samples. If you have a multi-sampling instrument such as Logic's ESX24, you can get "real sounding" sounds if you meticulously associate breath control (CC #2 in MIDI speak) with the appropriate amplifiers, envelopes and filters. More interesting is the approach taken by Wollander in their WIVI band offering (a light version of their WIVI orchestra). These sounds are software-modeled which is a fancy term for additive synthesis.

Many sine waves are combined algorithmically to simulate the sonic frequencies emitted by actual analog instruments. I find the results suprisingly good. This sort of instrument also reacts more to the performer. If you are using high quality samples, you tend to sound like the performer who recorded them. The software method, however, reacts to your own nuances of play and breathtone which makes the performance more personal.

I also have used the EWI-USB in non-traditional ways to control analog synthesizers with breath control. In the piece below, I have the WIVI trumpet along with the Arturia MiniMoog. Both are played on the EWI. I also include an electric piano and bass performed on the more traditional keyboard.




Its an early effort for me but I think it has promise. I find the breath control brings much more expressive even to electronic instruments and will be using this quite  a bit more.

In my arsenal for breath-controlled instruments now, I have:

• Garritan's Aria that came with the EWI - mediocre samples but they can be layered up into groups
• Wivi Band - The best quality and realistic sounding instruments I have for the EWI
• Korg M1 patches from Patchmanmusic.com - patches for the Korg legacy synth optimized for EWI
• Moog Modular/MiniMoog - Its easy to create my own patches for this using breath control
• Synth1 - Likewise, I have a set of patches to react to CC #2 breath control
• Madrona Aalto synth - Another soft synth that lets you patch breath control to any filter or effect you wish.

For many of my single voice sounds in the future, I will probably make heavy use of the EWI-USB. 

After several months with my new Akai EWI-USB MIDI wind instrument, I finally got a piece out. I am nowhere near mastering this beast yet, but here is my first effort:



The EWI-USB is a great addition to my gear in that it merges my ancient wind instrument memories (trumpet) with my electronic MIDI studio setup. I am pushing 50 and don't really have the chops anymore to push out trumpet pieces. The EWI lets me use breath control and wind/fingerings to perform any and all software instruments that I have in my collection - literally 100s.

The only catch is that you have to configure the synths or software to respond to breath control (MIDI CC#2) and/or to aftertouch. The device itself, as I have already blogged is round $240 street price but as always, that is sort of the tip of the iceberg.

The EWI comes with a set of Garriton samples that are mediocre at best and allow you to play samples that to paraphrase Douglas Adams, sound almost, but not entirely unlike the real instruments they portray.

After mucking for awhile with the EWI, I decided that I needed a better trumpet sound and went with Wollander's WIVI band (as I also blogged earlier). This product is a set of software-modeled instruments which is a fancy way of saying, it sounds REAL and doesn't use samples! A very difficult feat!

The piece above, however, is using 1980s sample technology in the Korg M1 legacy synth. In this piece I am using Korg patches that have been created specifically for wind instruments by Patchman Music. These are modifications that use aftertouch (since the Korg does not directly use breath control) to affect the volume and timbre of the sounds. The sounds aren't entirely real, but they do blend together well and support breath control extremely well.

This is where the EWI leads to more purchases. The patches for the M1 were another investment but well worthwhile. I now have several very interesting ways to use the wind controller:

• Wivi Band - this gives me very realistic sounding trumpets, trombones, tenor saxes, oboes, tubas etc.
• My old ESX24 sampler in Logic - if you set up the CC#2 (breath control) in the modulation matrix to control relative volume, sample selection and cutoff, you can make sampled instruments in Logic 9 sound very realistic.
• Aalto synth - this is a Buchla inspired synth that you can "wire up" to use breath control which can be routed to anything. Sounds "synthy" but its great to use with the EWI.
• Synth1 - The ubiquitous freeware synth has some nice patches set up for breath control as well.

The EWI so far is fantastic. I can play trumpet parts once again as well as sax or any other wind based interments and it also adds expressive possibilities to my software synths that aren't possible with just a keyboard.

Is that art???

Every once in awhile, I find myself in a classical mood and one thing I notice is that I start thinking of the composer I want to listen to. This is very different than anything with jazz or popular music. With minor exceptions, classical music is always associated with its composer rather than the performer. Rarely am I thinking, I want to hear the Philadelphia symphony or London philharmonic - I'm thinking Mozart or Bach.

There are a few exceptions such as a particular singer or performer such as Yo yo Ma or other virtuoso, but usually its much more about the composition than the artist's interpretation. When it comes to other forms of music, it is mostly about the performer rather than the song writer. Interesting stuff. Few, however would question the artistry of a great composer nor of a performer.

So, with electronic music... There are many ways to create compositions. There are keyboard based synths that are "played". There are trackers or sequencers that can be edited manually and there are other ways to create musical compositions. Usually, the process involves playing midi type controllers into a DAW of some sort that records the events in individual tracks for more editing.

In the case of trackers (Sunvox, iSequence, Renoise etc), the events can also be manually entered via QUERTY keyboards or pecked out a note at a time with a midi keyboard. So, is this still art if it is more an engineering process rather than performing? I think most would argue yes since it is a bit more in the vein of composing - not to compare my electronic "noise" directly with Beethoven, but you get the point.

Taken one step further, there is generative music. This style of composition abstracts the process one step further and has the composer giving various voices and parts a set of musical "rules" to follow in a piece. Rather than banging in exact notes and meters, the composer creates a set of rules and guidelines to put into an algorithm that in turn generates the sounds. This is analogous to a jazz musician sketching  out a piece, telling the musicians which keys to noodle around in and then letting them improvise.

I've written before a bit about Intermorphic's Mixtikl program and recently I've purchased the "big brother" version called Noatikl. I haven't found any generative music software quite so deep or approachable, though to be sure, there is a steep learning curve!

Mixtikl was the first product I used - initially on the iPad but I also bought the desktop version. Mixtikl emphasizes the "mixing" process and allows you to combine standard waves or loops along with generative elements that evolve their sounds according to the rules provides. There are many tiklpaks available with pre-built generative (and non-generative) sounds but you can also create your own.

The internal sounds generated by Mixtikl use a built-in modular synth called Partikl which is very powerful though somewhat thin sounding. I have made many of my pieces on SoundCloud with pieces of mixtikl or even entirely in mixtikl. The limiting factor, in my opinion, is that the internal synth isn't all that rich sounding. The features are great, but the sounds themselves and filters, are sort of meh - not surprising considering that they have to run on less powerful CPUS.

Enter Noatikl - This product can also create sounds and run them through the built-in Partikl synth. More impressive though, you can just pump out MIDI on channels and put the MIDI stream directly into your DAW and use any synth or software instrument you have! When paired with Logic 9, this allows me to generate music in any instrument I have on my desktop (I have hundreds if not more!).

The creative process here starts with a Logic 9 template with 16 tracks - one for each MIDI channel. Using Noatikl, you can wire any of its voices to MIDI channels - even having multiple voices per channel if you want polyphony. Within the Noatikl environment, I set up the "rules" for the piece including:

• What key and scale type should the piece play in
• For each voice - should it create single notes or chords?
• For the chords - should it use keyed harmonies or offsets to pitches?
• For rhythmic voices - fast, moderate or slow - dotted notes or only regular notes - what percentage/probability for each note duration?
• For patterns - what rhythms should be used? What note velocities or pitches - what is the probability of each one

The list goes on from there. You can set rules for MIDI CC events and almost anything you can think of. Noatikl and Mixtikl are basically artificial intelligence agents for creating sounds. 

Once the rules are in place, I hit play in my Logic DAW and listen to the tracks. When I find a combination that works, hit record and all of those MIDI events are recorded in Logic 9. At this point, I disconnect the piece from Noatikl and pick up work in the DAW. I might want to remix, add effects or mute/unmute certain sections. 

Once completed, you end up with a generated piece that you can then mix/master or mangle as you see fit. This too is only the surface of these powerful programs. They also offer a scripting language, can act as "hyperinstruments" (they listen to what you play into them and then harmonize with what you are doing) etc. I will write more as I learn more but I am using Noatikl and Mixtikl extensively in my modest works.