HydrogenAudio

Hello everyone!
I'm trying to get my feet wet with the room acoustic correction. Hoping you guys can point me in the right direction.
First thing I did, I installed REW program and even though I didn't have a calibrated mic, I measured my room using a regular USB microphone (not calibrated).
The results are shown in the picture below:


1) Can someone help me understand what I see here?
2) Now, let's assume I get a calibrated mic and repeat the measurement and get a correction EQ curve. What do I feed this curve to in order to correct my room?

Thanks in advance!

the picture is probably a measured frequency response.
btw the easiest way to do this is with mathaudio room eq plugin for foobar2000. which is also freeware and has a good tutorial.

also if you use bad microphone, it's possible to get results that are worse than no correction at all, so it really only makes sense with a good linear-FR microphone.

Maybe this link is of use: https://www.avsforum.com/forum/91-audio-theory-setup-chat/1449924-simplified-rew-setup-use-usb-mic-hdmi-connection-including-measurement-techniques-how-interpret-graphs.html

Thank you both!

Don't try to boost the dips  (standing wave nodes (https://en.wikipedia.org/wiki/Standing_wave)).    A 10dB boost to "fix" a 10dB dip requires 10 times the power, and the nodes & antinodes are different in different parts of the room so extreme boost will just cause more problems.

Generally, standing waves are corrected/minimized with bass traps and other acoustic treatment.    You are "trapping" or absorbing the reflected  waves and that helps to smooth both the nodes and antinodes.    

The microphone doesn't have to be perfectly-flat but a calibrated mic will come with a calibration curve.   As far as I know you can't load the calibration curve into REW so you'll probably have to compensate manually.

1) Can someone help me understand what I see here?
2) Now, let's assume I get a calibrated mic and repeat the measurement and get a correction EQ curve. What do I feed this curve to in order to correct my room?

I would need to know how your system is set up to be more helpful.  As I recall, you had new Focal speakers.  Are you using an AVR with room correction enabled or something else?

For #1, you are seeing the room response of your system at one specific point.  The portion of the graph below 200Hz is dominated by the frequency response of your speakers (and the mic, since it isn't calibrated) and room modes.  Room modes can create peaks and nulls that result in uneven LF response.  These can vary widely with location, especially if you are too close to the speakers.  You have quite early LF roll off which may be the mic (how does your bass sound overall?) and three nulls at this point--30Hz, 48Hz and about 120Hz.  The 120Hz looks sort of what Audyssey or similar products do to combat peaks.  Is this with or without any room correction already engaged?   The portion of the graph to the right of 200Hz  represents the part where the room response is determined by the interactions between direct and reflected sounds.  Look up "Schroeder frequency" for more on this.  The response is fairly flat to 10KHz except for the numerous sharp dips caused by nulls between direct/reflected sounds or between L/R speaker outputs.  This is called comb filtering and it is hard for room EQ to do anything about it.  Comb filtering is also very common--everyone has it to some degree--and those flat frequency responses you see for speakers always involves smoothing.   With 1/3 octave smoothing, your curve would look really good except for the early roll-offs. Your response also dives after 10KHz, this could be the speakers, the mic, room materials, etc. 

For #2, there are many different ways.  An AVR or amp with Dirac or Audyssey does this process with it's own mic and system.  If you get a UMIK-1 from MiniDSP, they make products that just do the room correction (hence their name) using the curve from the microphone.  I'm not sure I see a whole lot in  your graph that would benefit from Room EQ, except perhaps the bass roll-off, the HF roll-off and maybe that 120Hz dip.  You can't take out dips easily, and then only if they are rather broad. 

So, first you need accurate information, and this means a calibrated mic and readings taken from as many positions as possible.  If you listen in a single chair, you can still take 6-8 readings in a small area.  Then you need to consider careful speaker placement and testing--small changes in position can sometimes cause large changes.  After that, perhaps some minor room treatments.  Then Room EQ.

the picture is probably a measured frequency response.
btw the easiest way to do this is with mathaudio room eq plugin for foobar2000. which is also freeware and has a good tutorial.

also if you use bad microphone, it's possible to get results that are worse than no correction at all, so it really only makes sense with a good linear-FR microphone.

I installed the plugin and added it to the Foobar2000 but couldn't figure out how to run it within the Foobar. Do you remember how to do it?

Edit: nevermind, I figured it out :-)

Don't try to boost the dips  (standing wave nodes (https://en.wikipedia.org/wiki/Standing_wave)).    A 10dB boost to "fix" a 10dB dip requires 10 times the power, and the nodes & antinodes are different in different parts of the room so extreme boost will just cause more problems.

Thanks for pointing it out! I thought those narrow dips might be those but wasn't sure...

I would need to know how your system is set up to be more helpful.  As I recall, you had new Focal speakers.  Are you using an AVR with room correction enabled or something else?

Thank you for your comment as well!
Currently, I'm using a Pioneer AVR without any room correction or EQ enabled.