Hacker News new | ask | show | jobs
by digitailor 1206 days ago
This is exactly what happened in digital audio signal processing and recording, where word size represents amplitude. 12-bit audio was the first word size that provided a pretty good noise floor by the late 1980s, a real improvement over 8-bit. And by the mid-80s the CD format was already providing 16-bits for playback, which really is good enough for most playback scenarios. The 16-bit DSP era was just a few years longer than the 12-bit and quickly gave way to 24-bit, which provides a noise floor good enough for almost anything audio processing and recording related and is still the standard after more than 20 years. I have gear that defaults to 32-bit now, obviously just for power-of-2 convenience in software dev, which is annoying because the file sizes are bigger for basically no reason.

(The master buses in DAWs and digital hardware use even larger word lengths these days, but it's not really the same thing, that's a summing and calculation process)