Best approach to REW measurements + EQ (EqualizerAPO)

[flaviowolff]

Hi, all!
After moving my miniDSP DDRC-24 to the living room, I'm trying to EQ the bass region of my stereo monitors using REW + UMIK-1 and EqualizerAPO.
There are lots of differente tutorials over the internet, and apparently people use very different methods. What brought my attention the most is the fact that some people measure the channels separately, thus creating independent filters for each channel, while some measure both together and just average that and create a single filter.

An example of this last approach is this video, which is highly regarded on youtube:

What do you guys think is the best approach for EQing with REW + EqualizerApo?

 

[Vasr]

I have found individual measurements (but with crossovers enabled to a sub if using one) using RTA rather than multi-position the best balance between results and expediency. Yes, I once did for a 5.1 situation, a 5-position measurement then averaging it in REW, etc. I wouldn't wish that procedure on my enemies.

The reality is 90%+ of the benefit occur simply by knocking off the room resonance peaks below 500 hz even with a single position measurement using an uncalibrated mic. The rest is all looking at charts and overthinking it for dubious benefit.

 

[Kouioui]

@flaviowolff are you no longer using the miniDSP DDRC-24 with its DiracLive room EQ solution in the system with the stereo monitors?

 

[EchoChamber]

I have found individual measurements (but with crossovers enabled to a sub if using one) using RTA rather than multi-position the best balance between results and expediency. Yes, I once did for a 5.1 situation, a 5-position measurement then averaging it in REW, etc. I wouldn't wish that procedure on my enemies.

The reality is 90%+ of the benefit occur simply by knocking off the room resonance peaks below 500 hz even with a single position measurement using an uncalibrated mic. The rest is all looking at charts and overthinking it for dubious benefit.

Agree, I’m just starting to get a grip on REW. I spent about an hour yesterday measuring left and right speaker. The biggest gain in sonics was to tame a bump at 82 Hz by a few dB. I also added a sharp high pass filter at 25 Hz for good measure.

 

[flaviowolff]

@flaviowolff are you no longer using the miniDSP DDRC-24 with its DiracLive room EQ solution in the system with the stereo monitors?

I moved it to the living room setup. Unfortunately i can't purchase another unit as of now.

 

[flaviowolff]

The reality is 90%+ of the benefit occur simply by knocking off the room resonance peaks below 500 hz even with a single position measurement using an uncalibrated mic. The rest is all looking at charts and overthinking it for dubious benefit.

Glad to read that. Do you think that measuring both channels together is enough for that task? Like what the video suggests. Also, do you think it's worth it to fiddle with RePhase?

 

[Kouioui]

As you own a Dirac setup it would be interesting to compare the results with listening tests to both methods. I think the biggest takeaway from Julian's video is the use of the Harman house curve with REW and the way he goes thru his process step-by-step, explaining the why and how of it all. I happen to like the results with my setup he outlines but there are several other approaches that have some additional advantages.

The most thorough technical results driven one is what mitchco uses in his series of articles on Audiophile Style but he does charge for his personal services, which he should. I think for someone whose audio content is PC-based, a UMIK-1/REW/APO method like Julian's gives the biggest return on investment given the time it takes and has made recording, mixing, and general listening more enjoyable for myself using my monitoring setup.

 

[Vasr]

Glad to read that. Do you think that measuring both channels together is enough for that task? Like what the video suggests. Also, do you think it's worth it to fiddle with RePhase?

Doing both together makes no sense to me unless both speakers were in perfectly symmetrical situation and suffered from the same room modes. Most rooms aren't that way at all. Even in this case, doing them separately would also give the same result.

Doing them individually corrects them for their individual issues. The issues from their constructive and destructive interference with each other is of dubious benefit IMO (and only for the same content coming through both at equal volumes which is hardly a good model of reality). Just phase/distance align them correctly. The two ears on our head aren't modeled by a single mic to correct for the interactions between the two.

 

[Wayne A. Pflughaupt]

The reality is 90%+ of the benefit occur simply by knocking off the room resonance peaks below 500 hz even with a single position measurement using an uncalibrated mic. The rest is all looking at charts and overthinking it for dubious benefit.

I sure wish more people could wrap their heads around this. Many seem to think that because they have 30 filters available, they can’t let any go to waste. I’ve seen people plaster 7-8 filters in the space of one octave!

Regards,
Wayne A. Pflughaupt

 

[Wayne A. Pflughaupt]

Glad to read that. Do you think that measuring both channels together is enough for that task? Like what the video suggests. Also, do you think it's worth it to fiddle with RePhase?

As Vasr noted, individual measurement is best, unless you have a perfectly symmetrical room and positioning. In my situation, the left speaker is on a corner and the right is near a door opening. Naturally, the left one had a nasty peak at about 80 Hz caused by the corner, that the right did not.

Personally I don’t have a problem EQing the upper frequencies if needed. For example, the Paradigm Atoms in my bedroom system have a broad hump that runs between about 2-6 kHz. Taming that was beneficial. Just keep in mind that any EQ of the upper frequencies requires matching filters for left and right. Mismatched filters will whack the imaging. In my experience 3-400 Hz seems to be the point that determines if separate EQ filtering can be used, or if matching filters are required.

Regards,
Wayne A. Pflughaupt

 

[Kouioui]

Doing both together makes no sense to me unless both speakers were in perfectly symmetrical situation and suffered from the same room modes. Most rooms aren't that way at all. Even in this case, doing them separately would also give the same result.

I think anyone attempting room EQ should follow the proper placement guidelines one would normally use for any listening setup and what works best for the room dimensions before starting the process, such as middle of the shortest wall of a rectangular room. Proper toe-in, listening distance forming an equilateral triangle, and symmetrical placement verified by physically measuring the distances from the front wall to the back of the speaker cabinets makes for a better monitoring setup regardless of any room correction. Most of the better monitors only need help under 500Hz as many have mentioned but measurements will let you see if there are any issues to be aware of above that.

For those who are forced into an asymmetrical setup, it makes sense to perform REW measurements of individual speakers or if you suspect any large response variations between the two, which can be easily verified using outdoor or middle of the room close-up sweeps of each speaker. If you can't locate the speakers far enough from the front wall to minimize the increase in bass from room boundary effects, then you can either adjust using the trims on the speakers themselves for actives or let your room EQ do the job itself.

 

[dasdoing]

there are no perfectly symetrical rooms I have ever seen. It would need the door in the perfect middle of the backwall (or frontwall of course); never seen a room like this. and perfectly symetrical windows.
measuring seperatly is important for the stereo image. your phantom center will move left and right in diferent frequencies if L and R FRs differ too much

 

[Kouioui]

there are no perfectly symetrical rooms I have ever seen.

Sure there are, mine is a 24'x14'x8' cement block construction with a 10ft sliding door made of double-paned glass centered on the back wall that I cover with drapes and no other doors or windows. Not everyone can afford to build dedicated mix or listening rooms but it's not hard to find a contractor in the US that can accomodate special requests such as mine when building a home.

 

[whazzup]

This was my recent sequence of events after getting the xonar u7 and finally moving the sub to its own channel:

01.
In Peace eq / equalizer apo, upmix stereo to all channels so sub has sound (running 2.1 setup, yours may be different).
Plus change Peace's max db value to 60, because I noticed even at -30db preamp reduction, some sounds can still be heard.
Windows volume set to 97 (Windows has some audio artifacts at max volume, check the other threads).

02.
Turn subwoofer x-over to as high as possible (150-200hz), because I'm using peace to handle x-over.
Measure sub alone (Vol ~75% to max so I can reduce gain to taste later + use Windows volume slider instead of having to go to the sub).

03.
Auto eq and import the values into Peace. Measure again and now based on results, manual tweak the filters.

Plus because I want to use peace to do a steeper x-over, I added a Butterworth low pass at 100hz with Q of 10. Note that I did this with prior knowledge of my main speaker's roll off.

Listed as one step, but the trial and error took more than an hour.

04.
Do the same for left and right channels each, alone (I drop the sub channel preamp to -60db).
I also added a Butterworth high pass at 95hz, Q 20, as mine are satellite speakers. No point asking babies to produce sub 100 anything. And my left speaker is against the side wall, so it was pretty important to eq them separately.
Remember to also use Rew to run some single tones to verify that the filters / crossovers are working as intended.

05.
Now I turn sub back on and do another round with Left channel + sub, Right channel + sub (mainly to see how the x-overs are working), and finally all 3 speakers together. Already touched on above, but tweaking the x-over of the speakers and sub took some time. Just make sure to make bigger changes at the start, like 4-6db and more, to see the response before refining.

06.
I also found it useful to generate a single tone in Rew, then use the arrows to increase / decrease the frequency values (without stopping the playing). You can then REALLY hear how certain frequencies get softer or louder in real-time and correspond that with the measured graphs that you are seeing.

07.
Do a RTA of the background noise in the room. Will be useful to identify any unnatural peaks.



Took me more than half a day, with the measurements and re-measurements. Be ready, it's time consuming, if you're new to it like me. Right now my subwoofer is -8db preamp reduction. If I really need to I can bump it up to 0db, crazy loud for my small room.

I tried RTA for a while, but after I realised the results are not too far from single measurements, I only use it as sanity check against the single measurements. And to measure background noise.

 

[Severian]

I roughly follow the method in the OP and it is giving me good results.

 

[Taralius]

Julian on OP video is averaging all the individual speakers pair measurements before creating the filters and importing them to REW.
I would like to do a bunch of Left and Right separate measurements around main listening position.
At the end of the process I will find my self with 7X2=14 measurements.
Do I just need to average them and create the filters as I would have done with 7 speakers pair measurements?

 

[Taralius]

What I would like to know is:

If I have a left single measurement and a right single measurement, do I have to:
a) average the measurements, create the filters for the combined curve and include one correction file on eq app?
or
b) create a set of filters for left speaker, create another set of filters for right speaker, and then include two correction files on eq apo?