When I visualize a car in my imagination, the experience is obviously different from seeing a car. It's not even close. Similarly, if I imagine a sound, the experience is obviously different from hearing it. In part this is due to shortcomings of my imagination. But I suspect it's not just that. Rather, imagined experiences are qualitatively different from actual experiences. This isn't the difference disjunctivists get at between hallucinations and veridical experiences. I am willing to concede that a hallucination and a veridical experience could be phenomenally the same, but then both would be different from imagined experiences. There are, of course, structural analogies. Imagining a red triangle is related to imagining a blue square much as seeing a red triangle is related to imagining a blue square. And there may be some resemblance between imagining a red triangle and seeing a red triangle.
Here's a hypothesis about dreams: The "visual" experiences in dreams are phenomenally more like the ones in visual imagination than like seeing, but one's ability to tell the difference is suppressed.