As a linguist and practiced language learner AND lifelong classroom outlier, I have a couple thoughts which may or may not be informative, and which are most certainly unprofessional.
The challenge with assessing which of those methods would "work better" in the classroom (the ordinary vs Yvain's) is that teaching, wich children especially, depends on acheiving two different sets of results: success in catching and holding children's interest and motivating them to perform more extensive internal elaboration on the content of their lessons (assessing this is not a far cry from assessing what sort of TV ads will prompt audiences to elaborate on the content and become convinced), hence the vapid games and tricking them into thinking vocab is interesting---and success in teaching the material in a way that takes unknown information and makes it not only known to but understood by the children.
IOW: vocabulary has to be memorized and not systematized because the link between sound and meaning is, in all cases except onomatopoeia, arbitrary. But language as a whole is not necessarily taught best by brute force memorization, as context and understanding of the context in which vocabulary words are used and the history of those words can prompt students to elaborate on the culturally-loaded passle of meaning encoded by that series of sounds, thereby strengthening the associations to the memory of the related mental sound-file, ie the vocabulary word, improving memory of that word, and increasing potential for accurate usage by the language learner.
I am quite good at "filling in the gaps" and the main reason is that some teachers made off-hand comments about why and how to do this when I was young. I caught and heard those comments, and all through high school I worked at learning how to do what they said. I would never have figured it out without those teachers' comments, so if teachers more often and more directly taught these kinds of learning skills and "test-taking skills," I suppose many more people would be better at this process. My entire K-12 education came from curricula which strongly emphasized teaching students "how to learn" in addition to teaching information and life skills. My younger sister did not have this education for as many years. She is in college now and often calls me and asks me how to write a certain sort of paper. As I flip back through the direct lessons I received on such things in school, I explain to her how to teach and guide herself and why those broad learning methods will help her. For the first time in her life, she is a confident student not cowed by the thought of having to write a paper. That is some anecdotal evidence for the importance of teaching people principles and meta-skills. And if our teachers learned in their education schooling the information that is already out there about teaching meta-skills, then they would not find it so very difficult to teach. Unfortunately, many of the teacher certification courses out there do not provide this information.
"As soon as I notice it, it will go away."
Wow, you are blessed. When I hear sounds in my head, whether remembered or imagined, I feel as though I literally hear them. They are not merely background noise... on some days all of the music in my head gets so loud I just can't think straight and I have to find a way to silence my inner world. When I hear a melody, even in isolation, I hear full harmonization in my mind, which is why if I start singing along with a friend I have to work at sticking to the melody and not expressing the accompanying harmonies I hear in my mind. Because hearing them so vividly while knowing the sensory sells in my choclea are not vibrating accordingly is sometimes frustrating, thus by creating phsyical expressions of the sounds I hear in my mind, I reconcile my external reality and my internal reality. All this, too, is anecdotal evidence, and evidence of perhaps nothing more than my own strangeness.