Birdkids - Crow Intelligence System Modules


#1

Guys, I think BirdKids just made their version of the Audulus Module for eurorack.


#2

I don’t know, I get a lame vibe from this presentation, no offense to Birdkids. I can see how you can pitch the idea “anything is possible,” but as you all know from years of developing Audulus, bringing what is possible into actual use takes a long time to develop. I get that the idea of having knobs and information on a control surface while working is attractive, as I was discussing on the other thread, but for myself, I am not looking for another eurorack module as a controller, nor am I interested in forking out all that dough for it. But everyone has all their unique needs and interests so I am sure there will be people who will want this.

But it also brings up the notion that heading toward Audulus working as an AUv3 on an iPad might be a way to futureproof the software. Very soon, I think that money rather than size will be the limit of processing power (iPhone X), so we will probably see the iPad start to become a preferred interface for music production/performance.

I think the one big question could be, “what is stopping the best coders in eurosynthland from designing modules in for Audulus,” which is probably on par with, “what could possibly be done so that the CPU load for patching in Audulus becomes better than anything else out there by a significant margin?”

Makes me think of attracting some serious mathematicians. Could you partner with MIT and pose hard questions about processing limitations?


#3

The key differences between ipad, audulus, ES-8 combo and Crow Intelligence more than just knobs, its also the sheer number of connectors as well as the opportunity to optimize performance in a way that’s impractical to do while apple is playing software intermediary. There is a big difference in price point and distribution. Audulus is priced so that anyone can afford it and primarily distributed through the app store, where the Crow intelligence patcher software is bespoke to it’s hardware and made with an eye to interact with the outside world. While @taylor is constantly under the gun to patch issues created by iOS updates and maintaining connectivity with systems like au3, IAA and audio bus for many different use cases of the Audulus user base, Bird Kids just has to stay USB compliant, since all the processing is hosted by the module and you are just editting with a keyboard and mouse. There is definitely something to be said for being able to detach the usb cord then continue to use the patch one put together with a crazy generative sequence inside it.

Candidly I think there is an opportunity to parter with bird kids here since Audulus 3 is obviously more developed than the Crow Intelligence patcher software, and has a huge module library. They could pay @taylor a licensing fee and make nest Audulus inside the Crow Intelligence or maybe the user groups could work together for creating tools to create consistent standards for node-based patchers in a eurorack environment.


#4

Yeah that’s one angle. I am not so taken with the eurorack lust myself, as I wasn’t too taken with Elektron’s stuff or the OP-1. It would be nice to own all of this, but I kind of like the idea of being a bit minimal, but also the idea that other people aren’t minimal, while I am and that makes us different which is good.

I get it in a way, but all of these considerations are aiming at a very saturated demographic at the moment. That’s not to say that one shouldn’t consider what you are suggesting.

I am used to the complaints about apple and I understand. I think that some of this is quite mistaken on very serious level though. Apple is juggling a million things and succeeding in a big way. I was around for the apple II computer, just as I was around when I had my 286 running my bbs on my 14.4 k modem. I watched DOS turn into windows and I watched windows fail, over and over and over…Apple was/is (always has been) such a breath of fresh air. I imagine it is difficult to juggle the snapchat crowd with the “every time something new comes out I max it and haunt the forums with future feature requests instead of making music” crowd.

Myself, I don’t want to think too hard about my OS and I believe that putting the end user first often puts the devs in difficult positions. But what about the efficiency of math in modelling circuitry? I think that there is a serious subject for inquiry there, one that I wish I was involved in. My hunch is that some heavy innovations in logic are at our doorstep, but that the majority of computer science students have jumped on the raft too far downstream – that everything is actually not so simple as it seems and there are alternative base architectures.

All of this touches on a really fascinating debate. Can you patent math? You can claim ideas, but in the VCV Rack, Euroland world optimization techniques might be trade secrets but I don’t think you can stop eavesdroppers from using algorithms and I don’t think you should be able to. Which then almost leads to asking, “if I side with an open source ethos, how can I balance forwarding the love of all of this creativity while generating the appropriate amount of income to support it?” Then you might get to thinking, if I make successful affordable things that sell a lot of units in an app store, isn’t it smart not to neglect this opportunity and avoid pouring too much of my energy trying to market mainly to Eurorack customers?

But then going back to your intelligent point, Audulus could be imagined as a Eurorack OS.


#5

I think they take turns being frustrating and inept, but I have way more apple products than pc products, which tells you ultimately which way I lean. But I want there to be competitive platforms out there, and right now the Surface Pro range are just way better than iOS for almost everything except …Audulus.

:rofl: I couldn’t put it any better.

From what I can tell, FPGA chips are really where it’s at with the component modeling. The way the technology works, the sample rate is in the Mhz and is already being used in products like Roland Boutique range and Novation Peak. They even started making their way over to eurorack with modules like the Intellijel Shapeshifter.

It sure would make life easier to have a lingua franca !


#6

Ultimately the choice between running a piece of software on a dedicated device or general purpose computer is simply whether the hardware is capable of running the software with an adequate level of performance. Dedicated hardware often has the edge in terms of raw performance but can also be expensive, particularly when produced in small numbers. Audulus could certainly benefit from faster hardware, and I’m sure there’s room for improvement in the code itself. Higher sample rates would go a long way toward eliminating some issues and better MIDI interconnectivity, others. Wavetable synthesis became popular because it was a way to calculate waveforms that didn’t require real time processing. FPGA based devices like the Peak simply take this concept to the next level. While this is an effective technique for creating a digital oscillator, other parts of the typical synthesis chain don’t lend themselves so easily to digital recreation. Filters, delays and other non-linear circuits are computationally difficult and usually require significant simplification in order to be processed in real time. The underlying mathematics is fairly well understood, however solving non-linear equations numerically typically requires multiple iterations and is difficult to accomplish with high precision in real time with reasonably priced hardware. For these cases analog circuitry still has a place. I think its interesting that Novation chose a mix of digital and analog circuitry for the Peak and stress this in their marketing material. As more and more Eurorack modules become digital in nature, one has to wonder what the ultimate end result will be. At this point, module to module interconnections remain analog but it’s not particularly efficient to take a digital bitstream, convert to analog, then immediately convert it back to digital for the next unit in the chain. Far simpler to have software modules running on a common platform. I think it likely that ultimately we will see a mix of general purpose computing hardware (iPad, Mac, PC etc.) with high speed I/O for capturing and outputting sounds and connection to the remaining analog elements. The economies of scale make single purpose computer systems prohibitively expensive. It’s instructive to note that most professional music production is carried out using general purpose hardware, and ultimately the vast majority or recordings are digital bitstreams. I think a more interesting area for exploration are the mechanisms used to interact with musical devices. Historically we have been forced to adapt to the peculiarities dictated by the physics of the instrument. Fretboards, keyboards, rows of holes, etc. have been the interface used to perform music. Now we have pad arrays, knobs, touch surfaces, proximity sensors etc. As we decouple the human interface from the music making machine, what new methods of expression will we create? Watching someone perform with an Ableton Push and a PC running Live, one can’t help but think that we’re seeing a significant change in the way music is produced, and as a result, a change in the nature of the music itself. Self playing pactches take this concept even further. They are certainly musical, but the creative process is rather far removed from the end result. These are the areas that I find most engaging.


#7

A super intersting perspective here. I had an Ableton Push 2, but sold it months later at a $350 loss. I now have the KM BopPad which I am more into because I am whacking a surface with a stick that is radial and velocity mapped.

Although all-in-one setups may be more efficient in some ways, I sure notice how unique modular setups (I don’t mean modular synthesis in particular, but just having maybe 4 or five separate interfaces/instruments) get more positive feedback from people around you. It seems like if you have a giant machine that “does it all” people tend to act like and treat you as though “the machine is doing it.”

I have also been following these little articles lately that discuss chip architecture. Now I think that the emphasis on “AI” is, to me, slightly misguided and a bit irritating. But the notion of neural network chip tech is interesting

https://www.wired.com/2017/04/race-make-ai-chips-everything-heating-fast/

There was another interesting article in Make magazine in the past year that had to do with the absurdity of chip architectural patents, and how the designs are so dated and built for the wrong purpose even though they are foundational in the industry. I wish I could track it down…

Anyway, I think we are in for some serious changes with computational efficiencies, but I also think that some smart Math grads could be really looking into DSP, and other modelling techniques as a practical testing ground to work on pure theoretic math.


#8

Actually computational methods used in solving systems of non-linear equations is a pretty hot topic for research at the moment. Most natural systems are non-linear and improvements in technique always bring immediate benefits. Weather forecasting and climate modeling is a prime example. At it’s heart, its just simple physics, but the system is so complex and involves so many non-linear feedback loops, that modeling it with any degree of accuracy is hellishly challenging. Makes a diode ladder filter look like child’s play. Faster hardware always helps but can only go so far. I think it’s safe to assume that there are already plenty of grad students working on the problem.


#9

@stschoen made some interesting points about how digital input versus analog with strings, holes, or keyboards are areas ripe for further development and exploration.

I think being able to run patches on Linux running on devices like the Raspberry Pi will enable people to develop their own custom hardware interface modules/controllers.

The strength of Audulus in these situations would be it’s ease of programming versus something like Pd. Since these setups would be programmed on other devices, you would have minimal resources for Audulus being used for graphics/animation.

Such distributed computing solutions involving traditional MIDI, ES-8 Eurorack interfaces, computers, iPads, iPhones, and microcomputers like the Raspberry Pi all oriented around an Audulus programming environment would provide maximum flexibility and minimize development efforts.

As he also pointed out, you could incorporate analog modules for tasks that are challenging to recreate digitally.

More niche uses like music based upon non-western scales could be more easily be constructed and supported in such an environment as individuals or small companies would have fewer barriers to entry too.

Such an environment could also support the development of low budget approaches to Eurorack as well which would also translate if someone should decide/afford to go with Eurorack.


#10

That’s was basically the plan when I started my case. While I have plenty of vactrols and funky analog circuits, my collection of digital modules slowly grows. I think this owes mostly to them being so fun and secondarily for them being ultra predictable.


#11

If you’re looking to build you own interface, rather than a processing unit, here’s like a good way to go. I’ve wanted to build a MIDI hammer dulcimer controller for quite some time and this was the best piece of hardware I found. Most of the low level programming is already done, so you can focus on the unique elements of your design.

http://lividinstruments.com/products/builder/


#12

I built a multi-temperament and microtonal quantizer a while back if you’re interested. The first supports up to 16 variable intervals per octave and the second up to 39.


#13

I’d be interested in seeing a picture of that…


#14

If it’s an Audulus patch, I’d be very interested in seeing it.


#15

I added the 24 scale and microtonal quantizers to the Modules/Utility Category


#16

Yes, you can. Usually in the form of proprietary algorithms that are complex enough to require a computer to generate them.