Mystery about SPL measuring

[Toe in]

Hi, I have wondered about one thing quite some time. When you are doing an SPL analysis with REW ( and other software) with a microphone, the resulting graph always looks like a mess, and you will have to smoother it to watch it. Now, why is this happening? The curve seems to be made up of a lot of spikes that has an amplitude of around 20db or so. This must be some glitch - perhaps a result of the way it's measured. I mean, there can't be such enormous variations in amplitude. We would have heard that.

Does anyone have an explanation?

 

[John Mulcahy]

Measurements in rooms have large response variations, at low frequencies they are predominantly modal resonances and at higher frequencies they are predominantly the result of reflections adding to the direct sound with a phase difference that depends on their path length compared to the direct path. A reflection as strong as the direct sound on a path that corresponds to 180 degrees of phase shift at a particular frequency will completely cancel the direct sound at that frequency. Reflections are usually lower than the direct sound and only yield partial cancellations. The variations are real and always present in rooms, our hearing system is able to largely ignore them and give us the perception of something pretty close to the direct sound.

 

[Toe in]

Very interesting. So all these hundreds of spikes up and down are due to reflections in my room?! I haven't thought of that.
This strengthens my belief that the human hearing is a mystery. For example, compared to all these speaker- and room-related distortions, an amplifier or DAC is almost perfect. Even the cheapest of electronic equipment is thousands times more distortion free than a high end speaker! And still, we hear differences between various amplifiers, etc.