J
Story
Genesis and Sega CD – Audio Filtering
> We’ve obviously lost some detail here, but the samples do track the general shape of the sine wave. A smart audio driver or audio device will be able to mostly figure out what the original wave looked like.

You haven't lost any information. There is nothing to "figure out" as the analog waveform will be exactly reproduced after the final filter stage.

Only if the frequency being recorded is half the sample rate (a.k.a. the Nyquist frequency) or less. That's the case for that nice pure sine wave in the example, but in more complex cases you'd absolutely be losing information.

Whether the lost information matters of course depends on the frequency range you can actually hear; in the case of CD-quality (44100Hz) audio you probably won't notice any issues unless you're a dog (or unless you're taking a higher-sample-rate recording and downsampling without any kind of anti-aliasing), but the article discusses a PCM chip that natively supports 32552Hz and is receiving a signal that was naïvely upsampled from 16276Hz. Especially in the latter case, you absolutely will hear the information loss.

They should watch the Monty video on YouTube
it's a fair watch:

D/A and A/D | Digital Show and Tell (Monty Montgomery @ xiph.org) 3/2/2013 https://youtube.com/watch?v=cIQ9IXSUzuM

The low-pass-filtered versions all sound noticeably worse to me IMO - like the audio's being smothered under a pillow. Unsurprising, since smothering the treble is usually the point of a low-pass filter.