I think "speaking" isn't actually one skill in itself, but rather the combination of a few skills. From experience, I can say that lots of audio input (as long as you're actually paying attention to it, and ideally memorizing it) does indeed help with certain aspects of speaking, i.e.:
- accent
- the "melody" of the language
- confidence in the sounds (this, to me, is something slightly different from a good accent; others may disagree)
One thing that lots of audio has not really helped me with is fluid production, i.e. stringing words together naturally and speaking at a normal pace.
A lot of people say that television is more helpful than radio, because then you have physical cues (topic, picture, facial expressions, etc.) to help you make the correct associations with the words. Personally, I don't have the patience for news broadcasts or the time to watch a lot of movies in my target language, so what works the best for me is cheesy soap-opera type TV shows. Plus, it seems to me that TV shows often have more dialogue than films do.
When you see the circumstances in which a particular character says certain words – and then the reaction they get to those words – you can learn very effectively, because you're bypassing translation and just going directly from oral expression to understanding meaning.
