Scratching, captured. Photo (CC-BY-SA) karl sinfield / sindesign.

Add this to the Internet of Things: imagine data recording scratching and scratch performances.

Technologists Jamie Wilkinson, Michael Auger, and Kyle McDonald propose a new way of storing scratch moves as data. They’re not just working in traditional ways, either: they’re hacking turntables and optical mice and cameras, and imagine not only recording performances, but having machines recreate scratching. (Robots!) And they want your help. Kyle writes:

i’m going to be leading a group at art hack day ( brooklyn, january 26th-28th ) about scratch markup
language, a tool for recording performances from turntablists.

this describes the general idea and who we’re
looking for. we need everyone from web designers/developers, to
hardware hackers, coders and musicians. if you’re interested, or know someone who is interested, contact me or join the google group

(We cover the awesomeness that is Graffiti Markup Language on Motion, which goes further to explaining why this sort of data storage can be powerful and enabling.)

All is described – rather bizarrely – in an image. (Can we have plain text, please? It is, at least, a pretty picture!)

Proposal/poster image (CC-BY) Kyle McDonald.
  • Out of interest, what HAS GML done for Graffitti Culture? 

    It is interesting to ponder how taking a physical 'traditional' and creating ways to document and replicate it digitally changes the 'core' of the tradition.

    Similar to the path that oral storytelling has taken through the written word and I guess, the motion picture. The group experience of watching a movie is far removed from communal storytelling – and debatably, not for the better.

  • Fantastic. It's about time.

  • q-bert of ISP helped to develop a notation system for turntablism a while back..

  • These guys have been looking at this since before timecode vinyl

    • Lachie Challis

      Yes, this whole thing is nothing new. In fact the 'old' scratch notation method was featured on at least one hip-hop documentary, I think the title is "Scratch".

  • salamanderanagram

    i have a reaktor ensemble i rigged up that records incoming freqs off a traktor record and uses it to scratch real time audio. i suppose it would be pretty simple to record the the actual playback speed to an audio table, is that what this is on about?

  • Mutis Mayfield

    I will contact them.


  • peterkirn

    There are graphical notation systems for scratch notation. What is different here is, they're looking to make a digital, reproducible format. That's as different as saying you don't need MIDI or wave files because you have music notation – the digital format is a different animal. Now, that said, I'd love to see the two related.

  • Freeman

    If you know C++, you can hack Mixxx pretty easily to record timecode vinyl motion. The pitch and position information is all there in the code, you'd just need to dump it out to a file and go from there. (It would take beginner-level C++ knowledge to write it to a text file, then you could open that up in Excel and plot it or go from there.)

  • Mutis Mayfield
  • we've found that mixxx is really unstable, so we're not planning on hacking it directly. though we have used the open source xwax library (which mixxx uses internally) to decode traktor+serato vinyl with openFrameworks.

    there have been a bunch of projects in this area before, especially people hacking with ms pinky. the problem we see is that these projects are old, abandoned, or closed source. has a lot of potential but it's still being developed in a closed source way — and it's focused on a single method of recording. we want to have a format that is general enough that people can use all different techniques to generate the SML data.

    i used a pic cause i wanted to include the TTM notation example 🙂 but i just posted the text to flickr too

    if you look closely you can also see a first draft of SML in the background…

  • @Peter: one of the plans is to use ScratchML to transcribe TTM notation. Having the "MIDI" data of the performance opes up a lot of doors. Thanks for the writeup

  • chandrabose21

    yes, please go ahead and use ScratchML, you will know the difference , Thanks @bose
    Proposal Forms

  • I proposed a method for getting the crossfader signal in a study. Was very happy with it. Have a look here:

    The method is described in paper 3. I'll send it to those who want to read. Might be other things of interest there too (it's a doctoral thesis presented 2010).

    Also check out Alexander Sonnenfeld's S-notation, the most thorough system out there.

  • Bas