# Equalize Speaker using impulse response

So I have this school project where my aim is to improve a loudspeaker in general, my idea is to mesure the impulse response of the room using a mic, then use this mesure to conceive a loudspeaker that can compensate the effects of the room. My question is how could I exploit the room impulse response in conceiving that filter let's, and what kind of filters should I use? would that even be possible?

PS: sorry if the question sounds dumb, I have zero experience and knowledge with this kind of stuff.

• Welcome! Your question doesn't sound dumb at all. What you're looking for is an equalizer :) – Marcus Müller Jun 7 '18 at 13:02
• Thanks Marcus! Yes I need an equalizer, but what I am looking for is how this equalizer function, I mean what should I compare the room response too in order to correct it, and is it possible using Matlab :). – Mohamed Mahrach Jun 7 '18 at 13:32
• I'd recommend reading up on the zero-forcing equalizer, as I think that's the one easiest to understand without a lot of estimation theory :) – Marcus Müller Jun 7 '18 at 13:53
• Pretty much everything is answered here. – jojek Jun 7 '18 at 21:36

This depends a bit on what you want to get out of this and how much effort/work you are willing to put in.

Doing a room EQ that actually works and makes it sound consistently better is quite complicated. There are commercial systems available but they tend to be complicated and expensive or tied to a specific product.

A really good freeware option to play around with is Room EQ Wizard https://www.roomeqwizard.com/

If you want to do this yourself, I suggest you break it down in a few separate steps that can be tackled individually

1. Figure out how to make a good measurement in a room (noise, reverb, microphone, ADC, SNR estimation, etc.)
2. Figure out what to measure (using the tools from step 1) that's representative enough of the entire room/speaker/listener situation
3. Extract perceptually relevant parameters from these measurements
4. Turn these parameters into some sort of correction that would move the parameters towards some perceptual target
5. Design a filter and/or algorithm based on this correction
6. Design the process/HW/system to actually implement and apply the filter or correction

All of these steps have been studied in detail and there is a fair bit of literature available on any of them.

You should understand that a room doesn’t have a single impulse response. The locations of sound sources and listeners matters.

While the term “impulse” response strongly suggests firing a shotgun (with blanks) a wide band continuous noise may be more suitable.

An equalizer that is typically used in communications or radar/sonar is trying to improve SNR and interference rejection. People are not matched filters.

Sound quality is something different.