animated 50 days graphic by Prisca Choe

animated 50 days graphic by Prisca Choe

50 days of coding karaoke covers

Today is the first day of a new 50 days of making for working on a project for COVERS BY MP3: music, migration, and memory of the Korean diaspora. Some goals are to use the 50 days of making format to iterate on this project before it’s first release in July, to flex my documentation skills and make it easier to collaborate and communicate by sharing while building!

what are we making?

Yesterday, Prisca and I met for Day 0, reviewing what we have done so far for a part of the project these next 50 days will focus on which is an installation where we’ll allow performers to create new memories by choosing a karaoke song to perform in front of a green screen as they are projected into a new karaoke composite with song lyrics in Korean/English or both.

some of what we’ve done so far

  1. Made a quick prototype using p5.js and ml5.js and sing king on youtube. Tested with background collected on a pinterest mood board aligning to four different background theme categories.

Screen Shot 2025-08-15 at 12.35.33 PM.png

Screen Shot 2025-08-15 at 12.53.02 PM.png

  1. Started a new prototype using OBS Studio, testing background removal and, video capture, creating scenes, compositing sources and adding filters

    Screenshot 2025-11-14 at 5.21.00 PM.png

    Screenshot 2025-11-14 at 5.07.32 PM.png

Screenshot 2026-01-14 at 8.13.07 PM.png

Screenshot 2026-01-05 at 9.38.52 AM.png

Screenshot 2026-01-05 at 9.45.11 AM.png

Screenshot 2025-11-19 at 2.49.02 PM.png

  1. The green screen has arrived! first tests, clearer and more responsive to motion(dancing) background removal.

Screenshot 2026-03-18 at 7.11.59 PM.png

Screenshot 2026-03-18 at 7.54.56 PM.png

Screenshot 2026-03-18 at 7.36.09 PM.png

  1. Testing songs with Korean lyrics, using TJ Karaoke as the Youtube channel to search for songs

Screenshot 2026-03-18 at 8.31.53 PM.png

Screenshot 2026-01-08 at 11.24.10 AM.png

  1. quick testing of using OCR text recognition plugin to extract text and redisplay it on the screen, currently leaning against this approach for a first iteration due to added complexity

Screenshot 2026-03-27 at 8.04.18 PM.png