It's a common misunderstanding that samples are like pixels in a digital image. The notion that because samples are quantized, the signal can't take on any intermediate values. This intuitive understanding is incorrect; as I pointed out earlier, quantization adds noise. The original signal is still captured and is always present regardless of bit depth. Only the sampling rate determines the harmonic content ("detail") of the signal. Watch Monty's "Digital Show and Tell" around the 9:20 mark, where he lowers the bit depth, and the noise floor goes up but the sine wave he's sampling is still there unchanged.

Bit depth is about noise floor, and that's all. The output signal of 24-bit has the same "resolution" as 16 bits, just less noise. See Production Advice's video "The truth about bit-depth and digital audio resolution" for a practical demonstration!

@@nicksterj sorry ... this is just not true. Bit depth is indeed about resolution as well as noise floor. Funny how this is so misunderstood.

@@51bpm You didn't watch the video I recommended, did you? The harmonic content of the signals nulls out perfectly, leaving only noise - meaning that the musical "detail" in the signals is equivalent. They sound exactly the same except for noise.

Also watch Nigel Redmon's "Audio Dither Explained," where he shows how dithering preserves ALL the detail in the signal regardless of bit depth. There is effectively no resolution limit imposed. So no audible difference between a 24-bit signal and a dithered 16-bit signal.