Lee Barry Composition (And Photo) Diary

 Seed from found sound:         Curio No. 49 by meta4s AI Versions:

My first AI jazz album using digitized cassette demos from the late 80s and early 90s. Most of it was never performed. or if it was, the performances were mediocre and not recorded well. The "restorations" are akin to the restorations of old films or recordings on 78s.  The model (v4.5) sounds less than optimal (and better than v6), and interestingly has a warbly cassette sound, so it inherits the imperfections of the medium. The corollary with vinyl would be that the restorations would include crackles and pops. The lo-fi quality actually gives it a faux-retro sound, or the sound of a live recording where the tape had degraded. To do restorations in a more musical way, you would let Suno create provisional versions from digitized tape, then transcribe them in some form to a MIDI file, and then use the audio file from that to create the final.  Be Bach : The original recording was me playing the bass melody, but the horns play it on the AI gen. Suno filled out the arrangement...

 

  Untitled

  Became a soft ambient song. AI shortens the time from initial spark to mastered recording. 

 Seed:    

There was an interesting study that was done a few years ago on the complexity of lyrics. But there’s a difference between those who write scientific papers about lyrics and those who actually write and sing them, which has an interesting connection with AI-generation of songs where the “singers” are just LLMs shuffling patterns to match riffs. When I work on a lyric, I’m focusing on how the quality of words to be used musically, especially at the rhythmic level. How syllables are mapped over note values is crucial. I think lyrics should be as economical as possible, which is the issue with translations from English because you’re “deracinating” the natural speech rhythms. With the Esperanto lyrics, the music sounds fine on its own with the vocals just being another instrument. If you don’t understand the words, then it’s essentially an instrumental. Most people don't tune into lyrics. The music is most important on first hearing, but lyrics can soak in over time.

Why am I so attracted to AI for generating material? I had been involved in generative music in the ‘90s creating music from seed structures, and I’m still interested in that, but sometimes I find myself wasting too much time going to websites, logging in, thinking of text prompts to generate music–when I can generate music from an idea or simply pick up a guitar or a bass and I’ll have an idea within seconds. Like this morning: There was a Bill Clinton quote in my diary on 9/22/2010: “They always like you better when you’re gone.” I heard a 12/8 minor blues in it. For the past 4 or 5 months I have been generating music from words and phrases that are in my diary and I jot them down in my Songday notebook. I wouldn’t necessarily want to generate music algorithmically as a first choice. It would be interesting to have my own large language model that would generate music based on the inputs that I give it, such as a rhythm or melody and see what happens. A lot of creativity is “Let’s tr...

  September Songs  

  [View Gallery]   The first images are photos of the TV that I took as I was watching various commemorations and footage of the event. When I got downtown, I started taking pictures of buildings that serve as a snapshot memory.  On 911, they had closed the office at about 10:30 and I started to walk around to take photos. I had gotten my first camera about a month before, and I wasn't completely familiar with it, so most of the photos were overexposed. Later on I inverted them, which is a look that I liked. So I thought that I would take overexposed pictures and then invert them in the same fashion.