Scene 1 is my choice.
Honestly, and I'm not doing this to be funny or anything, I read Scene 2's first sentence and my attention waned at the disassociated POV describing how Harry looked. I read Scene 1 perfectly fine and envisioned it, and already had no issues imaging a placeholder for Harry's appearance, but when Scene 2 came up, I had to focus on slotting pieces of color and texture onto Harry first and then get into the scene.
That's a gear shift. Am I making sense why I chose the first over the second?