If you look at the way analog audio is converted to digital audio, you would be forgiven in thinking that the conversion back to analog would lead to a stair stepping effect. Analog audio is sampled at precise moments, and looking at those samples, it looks like the result steps from one sample to the next. As the video below shows, that isn’t the case.
The video is from Monty Montgomery at xiph.org and it’s a solid watch all the way though. Don’t worry if they lose you at some point. They get you back soon enough. This also has the best explanation of dither (adding predictable digital noise to actually reduce the noise you hear) that I have ever seen.
This may be an engineers explanation, but it does a good job of showing you what the conversion looks like, from an analog perspective and a digital one.
Keep in mind, this isn’t the end all, be all of digital audio. It doesn’t take into account, or at least add enough into the discussion, the values of higher bit depth (16 bit vs 24 bit) on the quality of sampled audio. I understand they can’t be completely comprehensive, but the benefits of a higher bit rate aren’t expressed, and that is too bad. It’s a valuable part of what they are talking about.
Hat tip to the excellent Create Digital Music for pointing this out. Even if you aren’t a musician, if you deal with digital audio, you should subscribe to their site. And head over there for some more discussion of the topics this video brings up.