I’ve been far away from academia for a long time, but last time I read papers on voice processing it went something like this:

Abstract: we’ve achieved [very good results] using this one simple trick…

Body: actually, we will maybe not tell what was the corpus we used to measure how good we are. We’re also going to omit several important steps where they can be omitted nonchalantly, so that reproducing what seems to be a thorough description will be a pain

So, I don’t know if it ever is all nice and simple, but man could it be better if things were always done in good faith and professionally