Mp3 converter are a sensitive topic.
This is where the sound
Or should I say, this is where no sound should be made. The problem
with converter is, that you don't take notice or think about them, unless
you suddenly hear some unwanted artifacts in your music or speech
and more complicated part of converter lies on the encoding side
while decoding (playing) is much simpler.
How to decode mp3
is exactly specified in the ISO standard.
It only provides sample
implementations (one simple and one complex) for encoding. That means
that the encoding quality differs much more from codec to codec,
than the decoding part.
You can test the decoding quality in a very scientific way by comparing
bits. A decoder cannot sound "good" or "bad"
but just decode correctly or make mistakes.
The limits for encoding
regarding "good" or "bad" sound are set much wider
and the door for opinions on this is wide open. It mainly depends how
smart the developer implements the basic concept and how good they can
handle the trade off between speed and quality.
A good encoder
doesn't necessarily have to be slow, but it tends to be that way because
the process is actually very complex.