I think it must have a lot to do with memory. What strikes me about it is how I would first have read a sentence when I was just learning to read. I’ll use the word voice as the example. At 5 I would have read “voice” like this: Vuh, OY, SUH. After doing that phonetic pattern countless times my memory began to know what it would expect to hear from my voice once I started looking at these symbols I am typing now. So when I see a letter my brain recognizes what the sound is that will be made if I were going to say it aloud. I remember it I think more than I “hear” it. I am remembering the sound. It is like a key on a piano. I read a word and certain keys are depressed which call up a memory.
One way I am proving this to myself is how similar the word is “heard” in my head each time I say it internally. There is no change like I might expect to hear in my own voice. The variation of the sound doesn’t show up.
On second thought, I was able to make the chipmunk voice in my head….