Mains peak, how to address?

[Orange55]

I have successfully EQ'd may mains and are loving the sound, after a few trips down the wrong path which seemed to kill the soundstage.

The questions I have is about the peak at around 600hz and how if at all I should address this?

Here are my graphs before EQ, Mains averaged:


After EQ, Mains averaged.


Here are the filters I am using at present:

Left Main
Freq Q DB
67 5.1 -10.2
98 3.0 -5.0
161 10.0 6.0
280 1.2 1.5
900 3.0 2.0


Right Main
Freq Q DB
69 4.6 -9.6
159 7.0 -5.5
186 10.0 6.0
900 3.0 2.0




As stated I am, very happy with the sound, however wonder if it may be improved by tackling the 600hz peak but looking for advice on if and how?

[Wayne A. Pflughaupt]

Welcome to the Forum, Orange!

IMO, you’re over-equalizing. For starters, there’s nothing in the range from about 150-450 Hz that needs any attention. Likewise, your 900 Hz filter was most likely unnecessary.

Your main problems are the 600 Hz and 67 Hz peaks. If your equalizer isn’t one of those listed in REW’s EQ drop-down selection, I’d use the RTA function. That way you dial in a filter in real time that will bring down both peaks with minimal effect above or below the peak. The goal isn’t necessarily to totally flatten them, but to get them down in the same general range as the rest of the graph.

With the 67 Hz peak being asymmetrical, it might be helpful to use a broader filter than 10Q, to drag down the area between 67 Hz and ~125-150 Hz with it. It may be beneficial to center the filter below 67 Hz. The idea is to bring it all down to something relatively linear w/o dragging down frequencies above ~150 Hz.

At that point the only area that might need addressing is everything above 1 kHz, as it looks like it will all be exaggerated. It’s hard to recommend a filter, since response above ~6 kHz is drooping. I suggest that you point the measurement mic directly at the speakers, not the ceiling; that way you can see what they’re doing out to 20 kHz. At that point you can decide what to do, but it looks like it may need a very broad filter centered somewhere between 4-6 kHz.

Regards,
Wayne

[Orange55]

Hi Wayne

Many thanks, I have implemented your suggestions and I think things are looking better. Now I just need to have a longer listen to ensure it all sounds ok.

Mains averaged, with no filters and the Mic facing directly at the mains.


Mains averaged, with the EQ in place.



The filters I now have are:

Left Main
Freq Q DB
67 5.8 -9.0
99 3.0 -5.0
606 7.0 -4.0
5000 1.0 -3.0

Right Main
Freq Q DB
68 4.9 -8.0
571 7.0 -3.0
5000 1.0 -3.0

Am I correct in assuming I am seeing nearly a +/- 3db response across the mains?



PS: I know the target level has changed, but I did go through and ensure everything was correctly set-up prior to the readings.

[Andysu]

Welcome Orange

Have you used the prediction setting that overlaps the frequencies this makes it easier to see which ones need tending too?

I’m also confused as to which graph is left and right? I normally type in some text and paste it onto the graph.

[Wayne A. Pflughaupt]

I think the graph is both left and right combined, Andy.

Orange, not sure what happened, but with your new before-EQ looks, there is a big dip just below 200 Hz that wasn’t there in the first before EQ graph?

The 600 Hz peak probably needs more than 4 dB. Also, in that frequency range, above about 400 Hz, it’s best to use the matching filters for both channels.

I’m surprised that the highs don’t look any better. What mic are you using?

Regards,
Wayne

[Orange55]

Hi Wayne

Yes I am not sure either, used the same position as last time.

Here are the traces of the two mains before EQ on the first measurement I posted.


And here are the two traces for the second measurements I posted, before EQ.


To me the dips look fairly simular, but on the second reading in the red trace there seems to be a boost between 250hz & 350hz, now that is very strange.


I am using the Behringer ECM8000 mic, the reduction in the 5K frequency is defiantly noticeable from a listening point of view, has helped relax the music.

Regarding the amount of roll off in the high end, I have seen this thread, do I need to me aiming for the kind of roll of I see in these graphs?

http://www.hometheatershack.com/foru...alisation.html

[Wayne A. Pflughaupt]

To me the dips look fairly simular, but on the second reading in the red trace there seems to be a boost between 250hz & 350hz, now that is very strange.

Yeah, pretty strange. You can expect to see some deviation from subsequent measurements, but nothing like that. I would expect that some other factor, like a nearby door being opened or closed, would effect a greater frequency range than that. It's so specific it looks like there was an EQ filter in place.

Anyway - I suggest identical L/R filters for the 600 Hz peak. Below that I'd EQ separately. If you want us to review your EQing, it would be better to see the left and right graphs with the filters you used, rather than the combined graph like before.

Regarding the amount of roll off in the high end, I have seen this thread, do I need to me aiming for the kind of roll of I see in these graphs?

Not necessarily, your graphs just looks like others I've seen when people (based on bad advice) had the mic pointing towards the ceiling. Either you're not getting a good reading up that high (perhaps from not having a custom calibration file?), or maybe that's what your speakers really are doing. You may just have to trust your ears: Try to flatten it out some more and see what it sounds like. If it sound better, more "air" and detail, great. If things sound "hissy," strident, etc. then keep it as it is with no EQ.

Regards,
Wayne

[Andysu]

Ornage

I could do with an orange and mango juice right now. Your making me feel thirsty.

That’s fairly close. That sharp 80Hz and 160Hz 180Hz I think?

Does your room have any odd shaped walls around the front?

If you placed the speakers side by side in the middle of the room and run the sweep again (independently) they might be very, very close. Then space them out one foot per speaker and keep repeating the sweep till one of them starts to show irregularly peaks dips.

Hmmm, saying that I might try the same experiment, myself.

[Orange55]

Thanks for the advice.

One of my speakers is closer to a side wall than the other, which is why they may differ slightly, but not able to address there positions.

I will work on some more measurements over the next few days and post the left and right as seperate graphs.

I definantly have the mic pointing at the speakers, but only have the calibration file for the mic from this site.