EQ and house curve help

[patate91]

I'm working on my house curve using REW. I'm using the MMM technique at my listen place. I want to EQ only for this spot.

First which smoothing is best for HF eq : 1/3, 1/6, 1/12?

Second with REW's auto target caculator I have a 1,5db difference between speaker L and R (44,1db And 45,6 db) (speakers are at the same distance) Should I use 44,1db, (the lowest) target for both speaker? Intuitively I would say yes.

Third the speakers crossovers are 600hz (miniDSP), I plan to EQ between 20 and 300hz. Auto HF eq using REW is mostly from 2000 to 20khz. What about 300 to 1000hz? REW suggest one or two EQ correction in this region, mostly lowering peaks (Q >4). Should I leave it untouched or trust REW's recommendation? I tested with EQ in this region and it sound good and got consistent measurements too.

I'll be able to post mdat later.

 

[RayDunzl]

Second with REW's auto target caculator I have a 1,5db difference between speaker L and R (44,1db And 45,6 db) (speakers are at the same distance) Should I use 44,1db, (the lowest) target for both speaker? Intuitively I would say yes.

That seems like a pretty low level, if it is an accurate SPL reading...

Have you tried running the sweep at 75dB or so?

 

[patate91]

That seems like a pretty low level, if it is an accurate SPL reading...

Have you tried running the sweep at 75dB or so?

With SPL I have between 75 -80db C. The 44,1db comes from RTA. I think it's just a configuration issue.

 

[patate91]

That seems like a pretty low level, if it is an accurate SPL reading...

Have you tried running the sweep at 75dB or so?

Note that I dis measurement yesterday at 83db Z, and got the same behavior.

 

[RayDunzl]

Looking at some old calculated target levels here for Left, Right, and both speakers:

MartinLogan
Before Correction 69.9 and 70.3 and 76.7
After correction 73.0 and 72.8 and 78.8

JBL LSR 308
Before Correction 73.5 and 72.3 and 78.1
After correction 78.5 and 78.5 and 83.6

The levels correlate well with the level of the sweep tone. Wondering why yours is so low.

---

As for which to use, higher or lower, I don't use REW to create filters, so, "don't know". The above are results using automated AcourateDRC for corrections.

Experiment and see.

 

[patate91]

I did use the lower db level for my curve EQ, and got a more solid center image (It's not surprising though that level match speakers will give a better image). Note that mid and bass level seems equal for both speakers. I used the 1/12 smoothing.

At low level listening I prefer a more flat HF response. Boost in the bass region.

Changing music genre sometime I prefer >1000hz roll off. How tired am I seems To affect what I prefer too.

 

[Hipper]

I use no smoothing for the 0-500Hz region and 1/12 smoothing for the rest. This is to enable me to see what's going on but also not to get too bogged down in the detail higher up the frequencies. It's pointless trying to equalise narrow dips and peaks at this level as we apparently don't hear them. Hence the use of coarser smoothing.

My general measurement results tended to be around 50-80dB but after I added lots of room treatment ended up mostly around 60dB. This was using a 75dB signal on REW. You need to get at least to 60dB to reduce the effect of any ambient noise.

Regarding the left and right speaker disparities, is it just in the mid to high frequencies where the problem is? That will certainly effect the placement of your image. There are test discs that have signals on them that will tell you if your speakers are aligned properly for a strong centre image. In any case you could first check to see if it is not a speaker issue by swapping the speakers round. You can also swap cables round. If the differences remain then it is something asymmetrical in your room, perhaps where the sound reflects off the side walls. You can then either EQ for it (like using a balance knob in the old days!) or move one speaker in or out a bit, or toe in one of them differently, or use some absorbent material equally distant from both speakers to prevent sidewall reflections.

When I use filters for the 0-200Hz I take them exactly from REW. I apply them through a Behringer DEQ2496 using its Parametric Equaliser (PEQ). This has only ten filters so it covers me up to around 180Hz. After that I have to use the Behringer's Graphic Equaliser (GEQ). As I move up the frequency range I employ wider filters. For example, the PEQ filters have widths of between 1/7 to 3/2 octaves. For 315Hz I used 1/3 octave but at 8kHz I used 13/3 octave. At my advanced years (67) I don't hear so well at 8kHz so I boost this considerably to get some more percussion (hi-hats for example). Of course the width of these filters depends on my interpretation of the frequency graph and the limitations of what I can do on the Behringer's GEQ. The MiniDSP, if that is what you are using, may be more flexible and offer more exact filters.

I EQ for flat throughout the frequency then add to my taste.

These days all my measurements through out the frequency range are done with both speakers at the same time. In listening I couldn't tell any difference with doing them separately.

 

[ernestcarl]

I'm working on my house curve using REW. I'm using the MMM technique at my listen place. I want to EQ only for this spot.

First which smoothing is best for HF eq : 1/3, 1/6, 1/12?

Second with REW's auto target caculator I have a 1,5db difference between speaker L and R (44,1db And 45,6 db) (speakers are at the same distance) Should I use 44,1db, (the lowest) target for both speaker? Intuitively I would say yes.

Third the speakers crossovers are 600hz (miniDSP), I plan to EQ between 20 and 300hz. Auto HF eq using REW is mostly from 2000 to 20khz. What about 300 to 1000hz? REW suggest one or two EQ correction in this region, mostly lowering peaks (Q >4). Should I leave it untouched or trust REW's recommendation? I tested with EQ in this region and it sound good and got consistent measurements too.

I'll be able to post mdat later.

I just use the shortcut keys to quickly switch around whichever smoothing method I want to use.

I notice that 1/48 with RTA in MMM resemble 1/12 smoothing so sometimes minimal smoothing using 1/48 or 1/24 afterwards suffices.

Remember that the high frequencies esp. using MMM isn't going to look the same due to impossibility of replicating the movements you make and the way you orient the microphone each and every time.

Generally, I'd try to EQ one region first (e.g. 20-300Hz) and re-meaure and adjust... then listen afterwards with and without EQ, and make further adjustments by ear -- what looks nice on a graph may sound too boomy or anemic, uneven etc. in reality.

There's nothing wrong in trying out REW's suggestions based on your own target and filter task settings, but always do an A/B listening test afterwards and/or perform sweeps & MMM re-measurements to verify.

The high frequencies (usually just applying a 'house curve') is something that can be done or refined much later.

Sometimes, a simple HF shelving as a 'tone curve' is enough in that region if the speakers are already well-behaved on- and off-axis.

It can also be useful to look at the overlays SPL/phase prediction window to get a bigger preview instead of just confining your view inside the constricted EQ Window.

example:

 

[ernestcarl]

If the differences remain then it is something asymmetrical in your room...

In my couch setup, there's a lot more absorption on the left side which causes a difference in listening even though single point measurements and pink noise show same SPL. So I have deliberately increased the volume on that side by about a +1dB to improve the balance in imaging.

 

[patate91]

My room is treated on both side (ceilling, back walls and corners). I guess the asymetry is very plausible. Moving the speakers around is tough since they are huge, horns can move on top though. Maybe I'll fiddle a little bit with them.

For MMM, I'm getting very consistent measurements even with HF. I guess room treatments, but horn, listening distance is pretty short certainly help.

I'll perform more listening tests in the coming days with different genre. Thanks guys.