Cave Paintings, Wabi-Sabi, and Plugins

The other night, Kristin and I saw Werner Herzog’s latest film, “Cave of Forgotten Dreams”:

The movie was a 3D documentary about the paintings found in Chauvet Cave. These paintings are believed to date from up to 32,000 years ago, and are of an artistic standard that is stunning to this day:

The entrance to the cave is believed to have collapsed around 20,000 years ago, essentially sealing in the art until the cave was rediscovered by explorers in 1994.

After the film, we had some deep discussions about art and impermanence. Right now, my main artistic outlet is the development of plugins. And plugins (and software in general) tends to be very ephemeral. A plugin or music software program that runs on a general purpose computer will be lucky if it runs 10 years from now. Windows machines tend to be better in this respect, as an audio app that ran in Windows 98 could conceivably run today. However, the more useful audio applications from the 1990’s tended to run on the original Macintosh OS, and it is difficult to find a computer that can run these applications today. I have several friends that spent years creating work in Supercollider 2, and that work will no longer run on modern Macintoshes. Turbosynth is another “classic” Mac OS application that will no longer run.

The impermanence of digital audio software is obviously not limited to plugins and Mac/Windows apps. Most of the “canonical” computer music pieces were created for systems that no longer exist. In order to get the maximum performance out of limited hardware, the majority of the “classic” computer music languages from the 1960’s through the 1980’s were created using the assembly language of the machine they were run on, and will not function outside of that environment. For example, Music 11 only runs on a PDP-11; Music 360 only ran on an IBM360; Music Kit was designed for the NeXT cubes with built-in Motorola DSPs; the Samson box compositions only ran on the Samson box.

A few of the older computer music languages were written in Fortran (Music 4F and Music V) and still have a fighting chance of being compiled today. Unfortunately, the compositions written for these programs are also hard to get up and running – good luck finding a working punch card reader today. For that matter, many of the optical and magnetic drives from the 1990’s are no longer functional.

In some ways, the impermanence of software brings to mind the Japanese concept of Wabi-Sabi. This is the idea of finding beauty in transience, and that beauty that is “imperfect, impermanent and incomplete.” In the traditional Japanese conception of this idea, wabi-sabi can be found in rustic items such as the simple clay bowls used in the tea ceremony. The cracks and chips of such bowls, and the changing color of the glaze over time as it is chemically altered by hot water, are believed to embody the idea that “nothing lasts, nothing is finished, and nothing is perfect.”

It is tempting to explain away the short lifespan of audio software as wabi-sabi. This perspective would allow me to explain away bugs as “imperfections to reflect upon,” and would carry on the grand tradition of appropriating Eastern religious concepts into a Western philosophical framework without bothering to truly understand those Eastern concepts. However, I’m not able to make the comparison stick.

One could argue that analog musical instruments and equipment decay gracefully with age. The yellowing grillcloths of a Fender amp, the crackle on the lacquer finish of an older guitar, the small tuning instabilities of an older analog synth, the strange electronic howls of a fuzz box as its battery dies…all of these would allow the user to reflect upon the passage of time.

Software does not age gracefully. It works, and then it doesn’t. The life of software is nasty, brutish and short, and death comes quickly via a hard drive crash, or painfully via the bloat of a Windows registry. When it is gone, nothing useful is left behind, except for a box full of toxic metals.

So, how do I, as a software developer and digital artist, respond to the impermanence of software?

For starters, I keep my prices reasonable. Charging several hundred dollars for something that won’t work 20 years from now feels weird. I’ll keep things running as long as I can, but there will come a day when I am no longer alive, and these plugins won’t keep working.

I can also look into embedding my algorithms into hardware. The long term longevity of digital signal processors has yet to be seen, but I have several digital reverb boxes that are 20 to 25 years old, and are still running strong.

Finally, there is source code as a way of survival. My algorithms from my Csound days are still being integrated into open source projects today, and those algorithms are from 1999. I am not working on any open source projects presently, but it something to consider for several years down the road.

Nevertheless, the odds that any of my algorithms will survive 100 years, let alone 32,000 years, are slim at best. At Chauvet Cave, we are seeing the work of a very few artists, and we have no true way of estimating the scope of art in the ancient world, but it is probably save to assume that the vast majority of ancient art has been lost forever to the processes of decay and weather. Life is fleeting, and all turns to dust in the end, so having art that is a reflection of this truth is probably the most honest response.

Comments (22)

  • synthetic

    Good post. I’ve thought about this a lot. Of course we would not have Mozart today if he had backed up to floppy or CD instead of paper, etc.

    I didn’t feel so bad about my PCM96 Surround purchase (even when a plug-in version was released shortly after) because it seems like a long-term investment. Even if the Firewire plug-in feature stops working (and I expect it will in a few years/upgrade cycles), at least I have AES/EBU and analog audio to fall back on.

    And I also avoid high-priced software since it has that a temporary lifespan. The recent Powercore fallout reaffirmed my fear of proprietary DSP systems.

    Perhaps soon you’ll be able to embed your software into hardware, like a Arduino version of a PCM70. Then it will have some shelf life… until the RAM gets corrupted or the caps dry out or the LCD fails or…

    • It seems like hardware that relies upon software for proper functionality is the worst of both worlds.

      I really like Bricasti’s stance on all of this: the M7 has been designed to do the job for the next 20 years. Another approach is something that is instantly obsolete, but has character. The Spin Semi FV-1 has enough juce to run the PCM70 reverb algorithms. Of course, it is unknown how robust the DSP itself is – I thing I may have fried my dev kit’s FV-1 via some accidental static.

  • synthetic

    Eh the world is ending in 140 days anyway, why worry.

    • I thought it was ending on Saturday.

      But you’re right – it is just the Rapture on Saturday. The physical destruction of the world is supposed to happen in the fall.

  • Russell Borogove

    Of the approaches you describe, source code or other documentation of the algorithm is the most appealing to me for long-term survival of your creations. Consider the analogy of living things: none live forever (just like your hardware or software), but they can spread their genetic blueprint, which may live on arbitrarily long, albeit in mixed, diluted, and mutated forms.

  • Hey Sean, thanks for writing this.

    As long as electricity is a constant, there is permanence to software. It’s our desire to continually change the environment in which the software runs that creates the impermanence you speak of. I still have a PowerBook G4 running OS9 just for MacPod, CloudGen, Hyperprism, Steinberg’s D’Cota, and all the other magical OS9 bits of software that never made the leap. Similarly, many Nord Modular G1 users took the “dedicated crappy laptop editor” route to ensure their patch editing capability wouldn’t disappear when Clavia discontinued support.

    Looking at it under the lens of wabi-sabi is interesting though. When you mentioned it in your tweet I expected a different angle. Your plugin UIs have a roughness, a familiar imperfection that makes me much more comfortable using them than a chiseled, beveled, hyper-realistic attempt at depth and physicality. The impermanence of your work is possible, but I agree the comparison doesn’t stick. Wabi-sabi has an intention behind it, an aesthetic born out of rebellion against dehumanized ornamentation and frill, towards the familiar and temporal. The eventual death of software is really in the user’s hands, not the creator’s.

    • In my experience, OS9 isn’t permanent. My feelings about this are totally tied into the early death of my iBook. On the other hand, I got my first Mac right when OSX was released, so my lost work mainly consists of a few SC2 patches. However, it’s true that I could put a new had drive into the iBook and keep OS9 running if I really needed to.

      And I still think of the Nord Modular G1 as the “new” Nord Modular. Man, I’m getting old.

      The imperfection of my UIs stems more from my relative naivete in this area, as opposed to consciously putting any flaws in there. The 2D-ness is entirely intentional, of course, but that stems more from my “studies” (i.e. checking out a few books) of Swiss graphic design from the 1950’s, as opposed to any Zen concepts. Having a UI that is comfortable is definitely a goal of mine, as opposed to attempting to create an intimidating 3D edifice. Look upon my shading, ye mighty, and despair!

  • Nate

    Interesting topic and a reflection point I believe all artists would benefit from spending some time chewing on. I’ll add, there must be more to art than it’s initial manifestation; more to an artist than their artifacts. An idea, a feeling, even a technique, these things reverberate and evolve into much greater things than a piece of canvas or a magnetic footprint on a platter–they live on long after their original manifestations are gone, some with profound impact. In the internet age especially–good ideas will most assuredly proliferate and endure. The most important thing is that we genuinely pursue our passions, speak our ideas, create our art, play our music, and write our code. -tibi ipsi dic vere and to hell with the rest 🙂

  • Jimmi

    MusicKit still compiles & is running on many platforms.
    Really, most of what becomes obsolete are concrete hardware implementations and peripherals. One aspect of a concrete implementation is compiled code targeting a specific set of hardware, of course.
    What to do? Write code in high-level languages that are likely to be supported by future hardware/compilers, and keep hardware dependencies to a minimum or keep them modular. C and variants are pretty safe bets, but library dependancies are gotchas.
    This approach means your source code is the art, not your compiled artifact. Source code is not art on a cave wall, it’s more like a published book. Open source extends the life of your code by encouraging people to copy and improve your work… and fits wabi-sabi well by allowing others to polish, chip, and otherwise augment your work as it enjoys a potentially longer appreciation as art.
    Or… maybe what you make is really a tool? Not that tools are not art in their own right… more like meta-art, maybe.

    • JIMMI! What’s up, man?

      Interesting to hear about MusicKit. How did they get around the 56K code? I wonder if the project to get SynthBuilder running in OSX will ever happen, or if Analog Devices claims ownership of that code base. If it is the latter, it certainly wouldn’t be the first time, or the last time, that ADI sent audio DSP tools into the memory hole (Extended Csound and VisualAudio also come to mind).

      Your source code argument is a good one. Right now, the majority of my code is C++, leaning heavily towards C. I use the Accelerate SIMD libraries from Apple, and roll my own PC SIMD intrinsics, but these are wrapped by my own vector library, which also has an unoptimized set of C functions for debugging. As I continue to work on my plugins, I need to make sure that they are comprehensible by other coders.

      And yeah, my stuff is more tools for artists than art itself. Which is interesting, as far more tools have survived in the archaeological record as opposed to art. Tools tend to be made of durable substances, like rock, bone, or ivory, while art tends to be created in mediums that are easier to work with (wood, fabric, charcoal on cave walls) but are far less permanent. Today, a 12″ single made using computer music tools will probably survive a few decades, whereas the compiled software tools themselves will probably not.

  • Klemperer

    What a great post, with many highly interesting thoughts. I often think about such things from the perspective of a “poetic writer”. Even there, though books “work” for ages and don’t just fall into dust (or can be restored before they would) – things are a tad similar. 99,99% of all writings disappear, many before 100 years are gone, and a vast amount of this is decided by some – at the time – “posh” people. Like now, in our postmodernist times, where a real lot of superb language in works that tell “real stories” are seen as “crap”. And solipsistic descriptions of language in language in language are praised by intellctuals, but not even recognized by people outside those “circles”.
    Then there are people like the sadly enough passed away Tim Conrardy, who really like their Ataris (in his case) – and prolonging the life of software that way.
    It might be a possibility – our times will leave traces everywhere, thanks to technology, films, recordings, billions of blogs, even the usual mix of great information and fragile-ego-hatespeak at KVR^^ – that a future generation, even in the dim and distant foggy future, might want to really, really know how things happened. How this “mp3” (after they read about it in a pdf file, after finding out what pdf was, after …and so on^^) was played…there might be a list of used tools, “hmmm, Valhalla room…” (First some scientists of the future will – here – of course write a bit of crap about Wagner and wild guesses concerning what ancient mythology and Wagner and so on would have to say here – later a technical minded scientist might find something and would, in his evenings, try to rebuild those “ancient computers”….

    One never knows. For me as a budget musician there was always a reason to use Windows (in Germany it was/is a kind of big-city-fragile-ego thing to show off your latest mac in the Berlin café – and nothing against macs, just against those pale faced snobs…). Now, after years, my PC ceases to work with newer software, so I might be glad there is a “jbridge” or something to enable me to use my older plugins who do not get “ported” to 64bit.

    It is nice that things come and go – soon we’ll see 10 000 000 000 people on earth. The need to “write that one novel that wipes all others away” is maybe a capitalist or otherwise ridiculous disease, some fear of dying by pretending one would live “in the book” or “in the software” or “in this painting”…. 99,99999999% of people would get angry if they knew what some nowadays-snob-cultural feuilletonists read into their “ever lasting works”…. Arno Schmidt, a german poet, wrote a nice piece about that. In his “Tina oder über die Unsterblichkeit” ( is it translated?? “Tina or about immortality”, kind of^^) all “famous” people have to live on, and can only die as soon as NOBODY on earth ever remembers a bit of them, of their works…
    Now we don’t have to go knee deep into psychoanalytical studyings to find out what a real poet who sold nearly nothing during his lifetime really wanted to say with that – but the idea is somehow great…

    It is beautiful, beyond words, to find those paintings of the Chauvet caves…so beautiful. But is it important who painted them? Most probably not.
    So if I would live in 20 000 years and someone could try to recreate the ways people made music around 2011 – I guess it would be thrilling, a wonderful experience. (Let’s forget for a while what 10 000 000 000 people will do if 6 did not get it into their heads that luxury and wasting of energy is not what will make a stable fundament for future generations)

    From computer science and ancient paintings to education – seems the more people don’t get told by their parents that they, and only them, are the winning few that are worth more than the rest – narcissm and so on – they’d built nice things AND see the transient nature of them. The rest is unknown – happy accidents, like with the stunning caves. Back to music – that reverb got developers to think about the next step, and then another next one – and that’s all included in the ongoing story of this little part of the technical and musical world…so it’s “in” the reverbs of the future. Kind of^^.

  • Mark

    Thank you for the insightful and interesting writing on this subject.

    Software, be it a plug-in or an MP3 file downloaded off of band camp is infinitely deletable. It’s obsolescence is built in and part of the medium; when you are bored with it, you delete it. Obviously Apple Inc., Microsoft and the rest of the computer manufactures did not become the wealthiest companies in the world, selling a non-essential, luxury product, by building permanence into their OS’s and boxes. If my copy of Vision from the 80’s was still viable, why would I upgrade my PC or buy more software? The midi spec or what you can do with it hasn’t changed in 30 years…

    However, that type of impermanence; planned obsolescence to foster consumption, is not what is being referred to by the concept of Wabi-Sabi. WS refers to the inherent passage of time and it’s effects on the world, the inherent frailty and imperfection in human ventures, the Chaotic nature of Nature and more importantly, the underlying concerns of humanity and the one-ness that unites all things behind our artifice; this is Zen remember… it uses the cognitive dissonance of the concept of universal decay to demonstrate to us what in our nature is eternal. We will try to create beauty even knowing full well that it will be destroyed. According to the principals of Zen however, this does not for one moment let us off of the proverbial hook – on the contrary Zen teaches we must continue to attempt to create beauty and strive for perfection of form all the while acknowledging the frailness of the act. What makes the cave art impressive is not it’s age, but rather it’s lack of it. The soul that drove man to paint those caves 3200 years ago is quite obviously the same one that is still in us. They remind us of the permanence of the Human condition.. extant outside of death and decay and Apple’s constant updates, man still loves and wants and needs the same things now that he did in 1200 BC.

    Anyhow, applying this concept to electronic music or plug-in programing would mean exactly NOT giving in to the concept of planned obsolescence, even full well knowing that your plug-ins or my bug music will most likely not survive 3000 years. Quite to the contrary. Max Matthews isn’t remembered because of a composition or some specific algorithm, but rather for the philosophical contributions he made to the electronic music community (see your reference to open source above). The reason certain pieces of hardware will be coveted as long as people compose or produce music is for the reasons you list in your blog… they die long, graceful deaths. My goal as a musician should be to create something that, like the cave paintings, expresses enough about the human condition, my sense of place (as in terroir), and my acknowledgement of the passing nature of reality to hopefully communicate that to someone else in the future. As a maker of instruments, you can certainly do the same. How long will it be before we forget Matthews, Buchla, Moog? Long time in the future if you ask me…

    Anyhow, sorry for the ramble. It’s a difficult question, an extremely important question (possibly the most important extant – you are basically asking what you as an individual can offer to the posterity of the human race) and unfortunately without a definitive answer. Thank you for having the cajones to ask it… too few of us do.

    • Guy

      Interesting response, but it feels like the word ‘planned’ is something you have added your-self, unless I’ve misread the first post.

  • This article really resonates on me. The philosophical aspects of it have been keeping me awake for long nights — both as part of hardware / software concerns and in the big picture as well. My comment is related to the hardware part of it.

    Tom Owad wrote a short article for Make Magazine vol. 10 (page 184). He stated that antique computers rule the world. People are being forced to have stocks of old computers so they can run the software they programmed / bought / inherited on them these days. The alternative route is to electrically synthesize those old processors and hardware and have an emulated old computer. FPGAs are the core of those emulators.

    Due to my background as an EE and electronics teacher I can clearly see that implementing your reverb algorithms in an FPGA and having a hardware box can greatly expand the lifespan of your work. That can be good or bad but, at least conceptually, is great to know.

  • The reference is here but you have to be a subscriber in order to read it in full length:

  • Sorry for the triple posting but I found a (hopefully) linkable version:

  • the most amazing sounds I have ever heard came from acoustic instruments or life itself, but I dig your FX the most, specially when I’m trying to imitate those sounds just mentioned

  • Hey…this very morning I wake up thinking a hardware version of your amazing Valhalla shimmer would be the answer of all my life’s tribulations… And then I got to read this article… are you going to work on that? Or…do you know someone who is willing to do that?
    Anyway…keep up the awesome work! Cheers from Buenos Aires, Argentina…

  • Peter Drake

    It seems like certain techniques/algorithms are going extinct with their creators. Wolfgang Buchleitner and Keith Barr have passed away without having divulged the critical details of their work, and I read somewhere that Christopher Moore is having profound health issues. At least in these cases their work will survive in hardware for a while, but the algorithms are gone for good.

    There’s something else I can’t quite describe in that these things can’t be reverse engineered by physical observation any more. Even very sophisticated devices from 70 years ago could be worked out by following wires and knowing components, 40 years ago one could clamp on a logic analyzer and trap instructions to (very tediously) work out what a microprocessor was doing. Now there are often several layers between what the chip is doing and what comes out on the wires so it’s really impossible to intercept what it’s doing. Maybe the future holds unforseen techniques for interpolating the mysteries between input and output. For the moment we have these hardware objects that are some kind of reliquary for secret or forgotten techniques.

    • Sean Costello

      First of all, I’m sad about Wolfgang Buchleitner and Keith Barr, and am sad to hear about Moore’s health issues.

      Secondly, older hardware can be reverse engineered. Digital devices from 35 to 40 years ago – man, the Lexicon 224 turns 40 this year! – didn’t use copy protection. They didn’t need to. The ROM code was there for the reading. At the time, the only way you could use that ROM code was to build your own DSP. Today, people can learn from the ROM code.

      In addition, if you have some hints as to how the algorithms work, you can tune the hardware to “degenerate” states and start figuring out how things are working. There’s all sorts of tricks to get hints about what was going on in the hardware. It is harder to get the exact delay lengths this way, but I’m not convinced that those older reverbs had “magic numbers” in there, so much as clever algorithms that were more tolerant of delay length settings.

      Keith Barr was fairly open about the techniques he used. We exchanged a fair number of emails before he died. Most of his techniques are discussed on the Spin Semiconductor website. Keith made heavy use of allpass delays, often in a 2AP + 1 delay “block.” Those blocks could be arranged in parallel (to create diffuse comb filters), or put in series, to create allpass loop delays.

      People have discussed possible Quantec ideas on the Reverb Subculture thread at Gearslutz.

      Christopher Moore published many of his algorithms in patents, which was very cool of him. Undoubtedly he came up with a lot of great ideas after those early patents, as he designed the AKG digital reverb in the 80s, as well as a lot of the Kurzweil reverbs from the 90s.

  • Peter Drake

    Thanks for the Spin Semiconductor recommendation. That development board has been on the back burner as something I might poke at sometime, but now seeing that there’s example code by Barr has me inspired to get it and do some amateur tinkering.

  • Seawind161

    I just stopped by for a copy of VV (because FreqEcho, Room, Shimmer, Space Mod, Supermassive, UberMod and Delay just weren’t enough) and stumbled upon this thread which, in my wildest imagination, I cannot imagine finding on any other software artist’s site.

    If Art is finding more than you hoped for, Valhalla delivers Art. Thank you.


Leave a Reply

Your email address will not be published. Required fields are marked *