Quote Originally Posted by Cary B. Cornett View Post
I had to think about that one. With headphones, you have only the direct sound getting into your ears from the transducer. With speakers, you hear not only the direct sound, but multiple reflections from different surfaces, all with different delay times. The result is multiple different comb filters with different spectra that mask each other. This is also why the idea of using a little reverb to mask the latency is workable.
Its not about direct or bounced sound. Put you fingers in your ears and talk. Do you hear your voice differently?

Put headphones on and talk, you will hear similar changes to the sound. Your voice will be louder with much of that occurring due to bone conduction.

Bone conduction latency is very short, the difference in latency between the low latency bone conduction and headphones causes comb filtering.

Remove the covering over your ears and relative level of sound from bone conduction drops significantly reducing the relative level of the comb filtering effect.

Adding reverb to the signal doesn't remove the comb filtering, but does help to mask mask the perceived effect.