On a regular ST it isn't great. It's mono and the output doesn't have an anti-alias filter, so high frequencies sound very "tinny". The regular ST doesn't have any DA converters, they use the CPU to modulate the frequency and volume of the ST's Yamaha soundchip at very high rates to output a sample. This uses up a lot of CPU time and is one of the reasons why most ST games don't use samples for music.
The STE is definitely on par with the Amiga, it has dual 8-bit DA converters setup as stereo with up to 50KHz sampling rate playback. It also has an anti-aliasing filter clean up the sound. Playing a mod on an Amiga and STE will sound pretty much identical.
