Alvin Lucier Auto-Transcription MP3

Alvin Lucier’s composition “I Am Sitting in a Room” is so ripe for adoption, it’s surprising that the work isn’t revisited more often. In the 1970 original, Lucier recorded himself stating something plainly and then recorded that recording being played, and so on and so on. The intelligibility of his spoken statement (which begins: “I am sitting in a room, different from the one you are in now. I am recording the sound of my speaking voice …”) dissolved in direct proportion to how the piece’s sonic intrigue accumulated.

Late last year, I wrote in the Disquiet Downstream about Japanese artist Kanta Horio’s version of Lucier’s “I Am Sitting in a Room” (, which involved not someone speaking but, instead, the sound of flat panel speakers left to their own devices. Thus Horio emphasized the parallel between Lucier’s work and John Cage’s investigation of the silence of contained spaces, notably anechoic chambers.

And now, musician C. Reider has updated “I Am Sitting in a Room” using audio-transcription software. Reider (aka Vuzh) is a longtime and frequent poster on, and that blogging-community service earlier this year started providing the software developed at so users could post to their journals from their phones. Employing what it terms “Voice-to-Screenâ„¢ Messaging,” SpinVox takes a user’s recorded message, converts it to text and posts the text to the user’s LiveJournal webpage. (SpinVox is not alone in this realm. A similar service is provided by And a company called posts the recordings as audio snippets, rather than as transcribed speech.)

A few days ago, Reider decided to try out the new service on Lucier’s composition. He did so over the course of 11 increasingly — often humorously — contorted iterations. Reider explained his method to me in an email message:

In this version, instead of the text being replayed back into a room until it’s rendered into ambient mush, the text is read into the telephone, transcribed by SpinVox and the resulting transcription was re-read back into the telephone. The result was that the original text was rendered into unintelligible garble by the end of the eleven generation cycle. Actually, it was already unintelligible by the second generation.
The idea for the rendition seems natural enough. After all, Lucier’s “I Am Sitting in a Room” always resembled a conceptual-art game of telephone. The results of Reider’s experiment are listenable to at his webpage. Or if you want a shortcut, here are the first file (MP3), the fifth (MP3) and the final/eleventh (MP3).

Reider’s initial text, like Lucier’s, doubles as performance and instruction:

I am sitting in a room different from the one you are in now. I am recording the sound of my speaking voice and I am going to recite it back into the telephone again and again, until the auto transcriber of LiveJournal’s voicepost feature reinforces itself so that any semblance of this original text is destroyed. What you will read, then, are the natural limitations of auto-transcription technology, compounded by the low fidelity of the little microphone in my telephone’s handset, poor phone-line quality and probably also by my own poor enunciation. I regard this activity not so much as a demonstration of physical fact, but more as a way of entertaining myself because I’m bored tonight.
It then, during the course of the feedback cycle, becomes something along the lines of:
Hi, I’m sitting in the room from the one you’re sitting and now I’ve recorded the voicemail by speaking for us and I’m going to reside it back to the telephone again and till I had it deprived till delivered to you. No, for his posed to be reinforces is so bad and the reason month. Have the original text destroyed by you how…
… and so on. Of course, note that for Reider’s version of “I Am Sitting in a Room,” he had to read his take on what SpinVox typed. So his own processing of the material — as he puts it, his enunciation — is as much a part of the algorithm as is SpinVox’s technology.

It’s also interesting to contrast the nature of the sonic transformations in Lucier’s original and in Reider’s version. In the Lucier, the sound as a whole decayed with each generation of recording. In the Reider, each recording is no less clear, in sonic terms, than its predecessor; only meaning decays.

More on Reider at and more on Lucier at his Wesleyan University webpage,

Leave a Reply

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