How do I best use the recordings to modify my pronunciation?

I am looking for some more direction in how to correct my pronunciation. It never really looks much like what is the visual of the recording.

To be honest, I don't think the voice prints are all that useful. It doesn't really give you that kind of green light/red light feedback to enable you to tweak/improve your pronunciation (like the Duolingo mic test question). You would need to pronounce the sound bite at the same speed and intonation, to get any kind of alignment. But more importantly, just because the graphs don't align, it *doesn't* mean that your pronunciation is necessarily wrong. So in that sense, I think the recording is only useful because it compels you to speak in French, and also because it's passive, you can practice saying the phrase repeatedly. But I don't think it will help you correct pronunciation except for training your ear to the pronunciation.

I try to focus on the peaks in the graphs. Even though my peaks aren't as high as the tutor's, they are helpful for intonation. I have also been able to correct my vowel pronunciation this way. When I mispronounce a vowel, it shows in the graph.

Essentially keep using the listen/record function. From time to time, move your recorded voice print to align it with the model and play both the recordings at the same time - you will be surprised how much your print will come in to line. I have trouble with my Japanese vowels interfering with the French ones! constant and daily use of the listen/record function is ironing this out. I might eventually have French that is only tainted by my Australian accent!