Blog haze: it’s the dimmed state of consciousness that comes from reading too many blog entries about such and such a remix creative commons resampled installation art piece about intellectual prop . . . there, see I’ve already fallen asleep. But wait, this is really, truly, incredibly cool. If you don’t believe me, just watch the video.

The bizarrely-capitalized sCrAmBlEd?HaCkZ! takes an audio sample, slices it up into rhythmic slices and categorizes it in a database by spectral content, and then (here’s the good part) retriggers the audio based on an input. In other words, you can beatbox into a microphone and get a totally mashed-up version of MC Hammer in response. Check out the entertaining video for a sense of what’s going on.

The software itself will be unleashed on the masses, open source; it was created with c++, Python and PureData. (Okay, so you have to know not one but three geeky things. So we’ll probably just watch the video and leave it at that.) As near as I can tell from the software page, Pd does the audio/video playback, C++ does the analysis, and Python is the database. There are no effects: this is really just a (very clever) playback mechanism.

For the world beyond cool online videos, I do think vocal input has been wildly underutilized in new music tech. I could see an interface like this completely reconceptualized for a wide variety of material (think melodic or timbral re-slicing) . . . and who’s to say you have to use 80s music videos?

Via Music thing.

Recycle TV: Remixed TV Beats

DIYers, take note — this previous TV remixer project also released source code, in Max/MSP (no mucking about with C++ and Python code, if that scares you!)

  • m15a

    very interesting! agree about vocal input. (almost) anyone can talk. been meaning to do something with that myself . . might just steal something from this idea. 🙂

  • admin

    Well, that raises an interesting question — how else to implement ideas like this. C++ and Python and Pd are a bit much for most musicians, and developing in a single app could mean more ways to customize the software for different musical applications.

    I could see doing some of this sort of analysis in Max/MSP, or even Reaktor. Tricky bit is doing something with the spectral analysis — using that to trigger something else is most definitely the hard step.

    This is certainly a far cry from a vocoder, but the idea of abstracting input from output is related.

    Curious to hear ideas.

  • tom

    Sven's cool video is now on YouTube:

  • dawhead

    admin's comment drives me nuts. i don't hear people saying "its too hard to play the piano, there must be an easier way to realize this idea". how many "musicians" are going to come up with the idea of an N-dimensional space to represent spectral similarity? and when they do, is python likely to intimidate them? finally, if you want to mess inside the tools, you've got to learn the tool builders skills. if you want to build or tune a piano or a guitar or even a MIDI stick, its an entirely different set of skills than playing such a beast. likewise for this tool

  • admin

    I think you misunderstood me, if I'm following you correctly. I agree with you completely. I'm not suggesting an easier way of doing exactly the same thing. I'm suggesting possible *different* ways of doing *different* things . . . given this fellow has already built his own, excellent performance act. I admire both the technical and artistic acheivement of what he's done (and this is a rare case in which each is served up in good measure).

    Anyway, to someone skilled at C++, Reaktor can be foreign and challenging.

    And yes, to many musicians, the thought of C++, Python, and even Pd can be frightening, hence my comment in the story. I'm guessing some readers here will want to sit back and watch the video. I also know for a fact many readers here are skilled coders — by hobby or profession — so your point is well taken; I'm the last person to disagree (or I wouldn't be noting what programming tools are used in the first place).


  • m15a

    i consider myself a tech person before a musician (although maybe regretably). but i still wouldn't want to deal with 3 languages if i didn't have to. easier solutions would likely lead to more (tech/software) "hobbyists" trying out something like this than "musicians". of course those two groups overlap, as shown by this website (cdm, that is).

    also, i might think that there are some people who think that the piano could be improved upon for their own creative purposes. (for one, microtonalists, um, xenoharmonicianista . . . nm) the difference is that "piano" is well established and harder (more expensive) to fiddle around with. and then there's john cage, who was a musician who did mess with the piano technology very successfully.