Chris Randall: Musician, Writer, User Interface Designer, Inventor, Photographer, Complainer. Not necessarily in that order.
June 2, 2018
by Chris Randall
Yeah, it's been a minute since an AI video, but we're gonna get back to that now. Readers may remember a series of experiments I did back in 2011/2012 with touchscreen-based control paradigms (here, here, here, and here, with some absolutely stellar discussions about usability in the comments.) Those were admittedly somewhat early days for the entire concept; the iPad had only been out a couple months when I started those experiments, and the idea of an app-based control paradigm was a fairly new thing.
Fast forward to 2018, and shit has progressed a bit, and people are generally used to using touchscreens for control. The reason for the video above isn't really about experimenting with the control paradigm, since that's pretty well-trod territory by now. I'm coming at things from a different angle. I've used a monome for years now, and I have a Max4Live step sequencer for that platform I've written that is pretty much only useful for me, and that I'm very happy with. However, I was using it last weekend, and I got to thinking that it would be dope if I could record control gestures along with the beats. Obviously, the monome itself is kind of shit for that sort of thing, so I first "ported" the control logic for the monome to a JUCE app, so I could run it full screen on a touchscreen monitor. When I did this, I was able to break out all the unlabeled control buttons to dedicated buttons, and improve the pattern memory and such.
After that, I gave each lane a four-bar gesture recorder; there are three gestures in all, and the X, Y, and Z planes can be assigned to any parameter in Live. (In the quick demo above, I generally have them going to effects sends and suchlike.) The sequence memory and control is hosted in the M4L patch, but the gesture recording and playback is hosted in the JUCE app. Note this is running on a separate computer entirely from the one hosting the Live session. (It is, in point of fact, that little Intel NUC, stuck to the back of the monitor with double-stick tape. It is communicating with M4L on the host computer via OSC over my home wi-fi network.)
There are actually 10 lanes of gesture recording; in the video above, if you look closely, you can see them labeled D1 - D6 (the drum lanes), Bass, and S1 - S3. I don't actually use the non-drum ones in the video, but they're there and working. There are 8 banks of 8 patterns, and each pattern has its own gesture memory.
I could easily add more buttons to where I could control the session entirely from the app, and not have a Push2 there, but there's no sense re-inventing the wheel. The purpose of the experiment was to proof-of-concept fast, intuitive real-time control of a Live session from a separate computer's touchscreen, and I'm pretty happy with things so far. The next step is to try to put together a whole song (or several songs) to perform; this isn't great for writing as it requires a lot of prior preparation. But for performing, I think there's a lot of potential to be explored.
Side note: the two synth pads I play towards the end are both Quanta. That fairly major undertaking is reaching its final stages, and the synth is perfectly usable in a session now. So yay for that!
May 24, 2018
by Chris Randall
Between going to Superbooth (which, taking prep and recovering from jetlag in to account, sucks up almost 2 months of my time) and working on Quanta and our complete catalog rewrite, I haven't had a whole lot of time to work on music lately, which puts my regimen of one LP or two EPs a year at risk. I am of the firm belief that I can't really offer opinions on how things should be done if I'm not doing them myself, and music tech is no exception to that.
My general habit when I'm working on AD stuff (either Research or Development) is that if I come across something interesting that sparks my imagination, I stop whatever it is I was doing for an hour or two and try to develop it in to a framework. Then, when I have 10-15 of these half-songs stored up, I go through them and pick five that have some sort of sonic relationship to one another, and then finish and mix them all in one sitting. This particular EP departs from that, as the first track, getProcessor, I actually finished some months ago (January, I think) for a podcast, and the last track (strike_332) I recorded and mixed in one sitting yesterday. The other three fit that general description, though.
Unfortunately, this method of working precludes making process videos, for the most part, because the tracks are written and recorded over a fairly long span of time, and pop in to existence at unplanned moments. So, in lieu of a whole slew of process content, I'll just write a bit about each track, based on what I remember.
As I noted above, I finished this track a few months ago as an exclusive for the Headroom podcast. It starts with my usual noise bed (construction of which is detailed in several AI videos), and some synth burbles courtesy of a Reaktor ensemble. The drums are, in a rarity for me, based around a sampled loop, the beginning of 23 Skidoo's 1983 track "Coup". I re-architected it heavily, of course. The rest of the drums are various Reaktor ensembles (I believe the kick is a Blocks patch), working on drum samples I created through my usual methods specifically for the track. The synth "lead" is Basic; I actually played it in via the Linnstrument, then reorganized it with Replicant and Eventide H3000 factory.
This has the usual noise bed. The 3-note bass motif is from Live's Wavetable synth; the chord pad in the beginning is Phosphor, and the granular synth line is the first recorded appearance of Quanta in its early stages. (It couldn't even load samples back then; this is the grain engine working on the sidecar oscillator.) The chord line that doubles the main pad towards the end is also Quanta, in a somewhat later stage; the source sample is an RS09. The burbly glitchy synths you hear throughout are Basic, run through Automaton and Replicant. The drums are Axon for the kick and hat and a reworked field recording run through a whole raft of plugs for the other stuff.
This is a fun one; all the rhythmic elements are from a Max4Live patch that runs in real-time, all synthesized on the fly and played algorithmically. (Thus, it's essentially different every time I run it.) I made controls for density, and the ability to turn individual sources on and off, and that's it. It was quite challenging to put melodic stuff over the top of this. Since the source is all on-the-fly, I was able to tune the harmonic elements of the percussion to match the root, which really makes the melodic part boxy. The melodic elements come from two main synth hits; both of these are Quanta run through various effects (Grind, Eos, Filterstation, etc.). I played them in manually with the Linnstrument. This track is far more abstract than my usual stuff, but it was fun to make, and I'll probably do more like it in the future.
In this track, I was experimenting with controlling feel in weird ways; the song is in straight time, but I gave the bass a dotted 8 LFO, and the snare has a triplet feel done with the predelay in ADverb2 (the new version has a much bigger early reflection situation than the original, so you can time it appropriately.) This results in a track that can either feel straight, swung, or in trips depending on which rhythmic element you latch on to. The foot is just a resonating filter; I don't remember how I made the other rhythmic elements, to be honest, but it sounds like the usual group of glitch plugs (Replicant, Automaton, H3000 Factory, etc.). The pad is Quanta with a heavy sidechain compression, as is the synth hit, which is driven with follow actions in Live, randomly choosing a MIDI note.
This track I wrote and recorded in its entirety just yesterday; I installed Form, a Reaktor Player synth that is part of Komplete, and was experimenting with sending it real-time controller info from a Roli Lightpad Block. I got that interesting drone, and the rest of the track just fell in to place around it. Usual noise bed (you're hearing a ambience around a pyramid in Mexico underneath everything), and the initial drum loop is, in a super extra rarity for me, lifted straight from another song. In this case it is from Chris Carter's 1980 cassette-only release "Electrodub 2." (I had parted it out for a demo video, and I accidentally landed on the loop while sample surfing, and it laid right in there.) I believe it is a CR78 through an early digital reverb. The kick is, like most of my tonal kicks, just a filter self-resonating (two different flavors in this song, natch) and the other rhythmic elements are created much as above, with glitch plugs working on field recordings for the most part. The rolling snare that appears halfway through is a 606 snare through Unfiltered Audio's excellent Fault plug. The granular synth thing that goes throughout is Quanta on a sample of my CS5, and the pad is also Quanta, working on an MKS80 sample, which I played in real time.
May 6, 2018
by Chris Randall
Superbooth 18 ended yesterday, and tomorrow I'm off to TXL to begin my journey home. So today is (theoretically) a time for quiet reflection. Largely because everything in this damn country closes on Sunday, despite their professed atheism.
Obvs we had no Eurorack to show, as we're not making that format any more. We had already paid for the booth before we came to that decision, and there were no refunds available, so we decided to just go ahead and come anyhow, and show our new software. I've attended many trade shows, but never shown software, so I wasn't quite sure how to go about it. I decided on a small Windows machine running Bitwig to best show off Quanta in all its hi-resolution MPE-capable glory. (In retrospect, we should have nutted up and bought a Surface Studio like the Bitwig guys used; live and learn.) There was a Roli Seaboard Rise 25 attached to this. In addition, we had a 12.9" iPad Pro with my Linnstrument attached to it. This configuration turned out to be pretty good, since the controllers attracted as many people as the software.
One interesting thing to note: this is my third time showing at Superbooth, and at Superbooths '16 and '17, we had quite a few people come up and see the Eurorack-only display and note their sadness. At this show, we only had one or two people that saw the software-only display and be demonstrably sad. This leads me to believe we made the right decision with respect to deprecating the hardware.
In any case, Simon-Claudius did the vast majority of the customer-facing demonstrations, allowing me to mostly interface with other companies and do the various press interviews. This worked out very well, as I'm no longer really physically able to stand at the booth for an entire 10-hour session by myself. I mean, I can do it, but the people towards the end of the day would get some salty comments. "Yeah. It's a fucking plugin. Yay. Go away."
But all's well that ends well. We made some new friends, got to hang out with some old ones, and talked about Quanta a lot. All in all, a successful show. I will be very, very glad to be home with my wife and cats and garden and pool and driving in a city where bicyclists aren't on a suicide jag, though. Once I have my wits about me, expect some high-quality video demonstrations re: Quanta. I also did one of Sonic State's excellent "Meet The Makers" videos with Nick. Until then, here's a fun video from Bedroom Producers Blog. For extra credit, read the comments, then come back here and make fun of them.
April 19, 2018
by Chris Randall
Carl Mikael's Cabinet of Curiosities is one of my favorite YouTube channels; his videos often get me thinking, and this one from last week was no exception. One thing I have always found intensely irritating in my time in this business is religious zealots, who use their confirmation bias as a bludgeon. Carl-Mikael is, in a round-about way, addressing that in the above video, which is, like most of his content, worth a watch. "Pragmatism beats dogmatism every hour of the day..." as he says, a viewpoint I wholeheartedly endorse. Give his channel a sub, as well. He puts up quality content.
April 18, 2018
by Chris Randall
A month or so ago, I got turned on to the concept of Dante, which is an audio-over-Ethernet protocol used primarily by big production houses and live sound reinforcement. (I imagine it was invented by someone that had to run a 48-channel snake from the stage to the front-of-house every night, and got sick of being covered in spilled-beer-and-shoe-dirt slime.) In the simplest terms, think of a network, but instead of files and web sites, it serves digital audio. You use normal IT shit like switches and CAT 6 cable, but your goal is shunting massive channel counts of digital audio at ludicrously low latencies instead.
I had known about it for years (you can't help but see the references if you're looking at high-end AD/DA converters, since the usual suspects in that world all have Dante capability). I didn't really think about the ramifications until a friend beat me over the head with the concept. My ultimate goal in my home studio/office is simplicity. The fewer cables I have run, and the fewer conversions I have to deal with, the happier I am. It is especially attractive to me because I run multiple computers of different flavors, and having their I/O talk to each other in more-or-less real time would be excessively handy.
Once I was tipped to the potential, the full OCD Experience kicked in, and I started thinking about replacing my current rat's nest of I/O and monitor controlling. The main attraction to me is ultra-low-latency computer-to-computer connections. I've always thought it is dumb to convert to AES or SPDIF to do a digital computer-to-computer audio interaction, and so do the Dante people. Nominally, a computer-to-computer connection in Dante would be the same as any old-fashioned way, with expensive converter boxes in the way. But Audinate (the company that invented Dante and is the Keeper Of The Holy Scriptures regarding the format) got that sorted in a big way, with three pieces of software that make Dante in a home studio an attractive option.
The first is Dante Controller, which is essentially a virtual patchbay that lets you connect Dante sources and destinations. Dante gear all has a Gigabit Ethernet port, and you basically just run everything to a normal Gigabit Ethernet switch in a star fashion. Dante Controller sees everything on the network, and lets you set clock masters and routings and shit. Controller is free.
The second piece of software is the Dante Virtual Soundcard. This is an ASIO and CoreAudio/WDM driver that works like any other sound card driver; it has 64 ins and outs (which is mildly comical in something like Live. Did you know the I/O panel scrolls? Neither did I) and ludicrously low latency. Any computer running the Virtual Soundcard and connected to the Ethernet rig throws its I/O to the network, and shows up in the Dante Controller patchbay. The driver is US$29 per computer.
The third piece of software is the one that seems like magic to me, and which is useful whether or not you have a Dante system. I only assume it isn't more well-known is because Audinate's business model doesn't lend itself to marketing to hobbyists and home studio folks. It is called Dante Via, and basically lets you route _any_ audio source in your computer to any other. Think of it as Soundflower or Audio Hijack, but on pharmacutical-grade methamphetamines. You can run either this or the Virtual Soundcard. It shows up as an ASIO destination in software that supports that, or a WDM/CoreAudio destination elsewhere. Instead of 64 I/O, you get 8 channels, but otherwise, it is more or less the same, as far as how the DAW works. You just drag-and-drop your sources to your destinations, and you can mix-and-match anything as you see fit and turn any piece of software or I/O in to a sender/receiver. Dante Via is US$49 per computer (which is incredibly cheap considering what you get) or US$59 for a combo of the Virtual Soundcard and Via. There is a 15-day demo of Via on the site.
I purchased a Focusrite RedNet X2P to be my main monitor controller. It comes with a pair of Focusrite's high-end mic pres (their good stuff, not the prosumer Scarlett series). This can be powered off a POE Ethernet switch, so it's just one CAT6 cable to the switch and that's that. It is built like a god damn main battle tank, and is one of those Just Plug It In And Go Because It's Really Well Made For People That Don't Want To Dick Around kind of things. My Adam nearfields go from that, and my desk situation is sorted. I connected the Skull Canyon NUC and my main computer to the switch via their Gigabit Ethernet ports, and with the former running the Virtual Soundcard and the latter running Via (I also use the computer for games, and would like to hear YouTube videos and shit), I have VCV Rack running on the NUC, with multi-channel audio at no noticeable latency running in to Live, with Live Link providing wireless sync over my home wi-fi network. (Note that the X2P comes with a license for the Virtual Sound Card, as well as some Focusrite plugins I'll probably never use, but never say never, right?)
It is early days for Dante in a small studio right now, and many of the I/O solutions that are promised aren't quite here yet. I can easily just plug in a high-end convertor or mic pre rack to the system and it'll just show up. Audinate is releasing a 2-channel class-compliant USB dongle so you can plug in iOS or visiting laptops to the network with ease, and there are several 2-channel AES, SPDIF, and analog solutions to bring legacy gear in to the fold. ProCo and Radial also have similar small-and-cheap solutions either already released or in the works. I've only had this rig working for a day, so I can't really speak to its robustness, but it performed flawlessly with the above VCV-on-another-computer situation, as well as an hour or so with World Of Warcraft and a brief writing session in Live. I gave the mic pres on the X2P a quick check to make sure they worked, but I can't really speak for their all-around applications at this point. I'll put up another post with further thoughts after a month or so with this system.