Music, when digitised, has two dimensions. It has amplitude, ie loudness, and it has frequency.
When being recorded it is sampled at a precise rate. If the data isn’t presented to the DAC at precisely the same timing of the original sampling when it was recorded it will sound poor. So, the data, which represents the amplitude of the music must be correct, and error correction techniques ensure this, but crucially the timing must be consistently spot-on too. That’s why different streamers etc sound different.
Noise is electrical mush which interferes with the timing of the data to the DAC, not audible hiss.