So on a random day, I encountered this video and as someone who averages 15,000 minutes of music a month, I thought, "I should be able to do this." But then I tried it, and quickly realized those 15,000 minutes come from repeatedly listening to a pretty narrow rotation of artists, recognizing a random song by a random artist from a 0.1 second clip just wasn't happening. That's when the idea hit: what if I built the same concept, but customizable where I pick my own artist(s), game mode, difficulty, and all that? So here it is, I miserably lost at someone else's version, so I built one that I could actually be good at.

Tech: HTML (page structure/markup), CSS (theme and layouts), JS (game logic: artist search, fetching songs from iTunes, round flow, scoring, results screen). Difficulty tiers use Spotify popularity to rank tracks, with the top-ranked tracks treated as "hits" (iTunes still provides search and 30s previews). Songs are dealt from a shuffled "bag" per difficulty (pickOneRound()) so the same track can't repeat until every other song in that pool has been used. Playback uses Web Audio (AudioBufferSourceNode.start(when, offset, duration)) for sample-accurate timing; artist avatars upgrade via YouTube Data v3 (- Topic channels, 100/day, Map-cached, referrer-restricted). Theme toggle reuses the same data-theme + localStorage pattern as the rest of this site.

What I learned: a plain <audio> element controlled with currentTime + setTimeout isn't precise enough for a true 0.1 second clip, it's fine on desktop but unreliable on mobile Safari, sometimes producing no audible sound at all. Switching to a decoded AudioBuffer with start(when, offset, duration) fixed it, since the browser's audio engine handles the exact timing instead of JavaScript timers trying to. Also learned the hard way that Deezer's API blocks direct browser requests and Spotify's requires a secret that can't safely live in public client-side code, so iTunes's Search API handles search and 30s previews while Spotify is now used separately just for popularity ranking.

Next: Share results feature, update way to filter "hits"