Some people don’t care about the quality of the voice. As long as the words aren’t choppy and the audio doesn’t buzz away what needs to be understood, a more metallic voice isn’t a problem. By cutting out wider ranges of empty airspace, VoIP codecs can achieve smaller file sizes. With smaller files, your audio data can travel through the network easier and have fewer issues with network congestion. The concept is simple; the bigger the file, the more chances it could fail. A big file needs more network capacity or bandwidth to reach its destination intact, and the file can be delayed or chopped more often by random spikes or bad performance on the network. Smaller files can get through faster.
Unlike physical objects, you’re not dealing with big targets that can take a hit versus small targets that will be demolished; the files have the same virtual durability, and smaller files can slip through problems with less delay risk. Ideally, you’ll want a mixture of both smaller file sizes and better quality. The art and science of VOIP codecs aren’t in the extremes, but in trying to get better quality while compressing or cutting the audio in better ways.
Video and music streaming is no different. The files have different intents and can get away with different levels of performance, but figuring out how to make any file size smaller without affecting quality is the main goal.