Nardone et al. did some work back in 1989 looking at motor recruitment patterns in a low threshold slow twitch and high threshold fast twitch motor unit from the gastrocnemius muscle during concentric, isometric and eccentric contractions typical of a set of calf work.
What they found was that the slow twitch MU was recruited during the concentric and isometric part of the movement, and the FT MU during the eccentric phase, and that the selective activation of the FT MU was MOST pronounced when the eccentric phase was done relatively quickly.
From what I've read and from talking to a hard-core muscle physiologist that has competed in bodybuilding nationally, the eccentric component should be performed relatively quickly but controlled for maximal MU recruitment, thus I would agree with a faster eccentric movement.
The reason the slow movement is so painful is because the muscle is ischemic (no blood flow) and metabolites such as lactate, H+, Pi, and K+ are accumulating. However, that doesn't necessarily mean more muscle growth. I'd opt for more motor unit recruitment, especially Type II's since they are responsible for activating Type II fibers that are the largest and contribute to mass and power vs more pain unless you’re into pain.
For size, both fast explosive concentric movements (speed sets) and slower (but not super slow) movements would mix up recruitment patterns on the concentric end, but I'd keep the eccentric part about the same, quick and controlled. For maximal mass, you'd want all the fibers to respond. Sometimes too much speed work misses some of the slow twitch units and while they are smaller, you'd still want to develop them to add to overall mass.
W6