I’m guessing he’s referring to early CDs when both mastering and player DACs had digital roughness. Or maybe he’s talking about digital clipping which was deliberately being introduced in the 1990s to create loudness.
Why wouldn't radio stations just normalize all digital files they're playing to one default volume / loudness? Then "louder" digital songs would only be worse quality / fewer dynamics, but not louder at all.
Sounds like the radio stations are very much complicit. (Also: why are tv or radio ads always so much louder than the rest of their program?)
The loudness war wasn’t about normalizing to 0db vs -3db, but rather about dynamic compression which is a process that raises the volume of softer sounds without affecting the loudest ones as much (or at all). It’s what you use to make the breathiness of a singer’s voice, or a drum kit’s hi hat, more present in a recording.