[hybris]

I have a general EQing question that I've been a bit unsure about for a while.

I have several peaks and dips in the area between 100 and 200hz. I've reduced the peaks somewhat, but I'm a bit uncertain if the right thing to do is to push them all the way down to the 75dB reference level or not?

As I also have a lot of dips, I'm thinking that if I push them down to 75dB, the overall / average volume in this area will be too low, since it consists of reference level + dips. While if there are both dips and peaks, the average level will be around 75dB. So is it best to get the _average_ level as close to the reference as possible, or is it best to have as much of the frequency range as possible exactly at the reference level?

I hope I explained that in a understandable way. To illustrate, should the relatively small peaks between 100 and 200hz on this graph be lowered even more so they are at 75dB and as a result removed alltogether - even though I also have dips in the same area?

 

[hybris]

On a related note - moving the mic around it seems like the response is relatively uniform below 100hz, but above the peaks/dips move about quite a bit when I move the mics.. So should I conclude that above 100hz I need to go for room treatment and not eq?

 

[Moonfly]

I'm not sure I fully understand your question but I'll try explain it.

With EQ your trying to get all frequencies play at equal volumes so you can hear them all. When you have a peak at a certain frequency then that frequency will pronounce itself more than the ones that are showing a dip. We really want to be able to fully listen to the entire audio spectrum being played out and large peaks and troughs will have a negative impact on this aim. Its fine to boost your sub bass etc, but the boost output still wants to be as smooth as possible to ensure you get the best, most accurate, and most satisfying playback of material.
When you lower a peak, you reduce the spl difference between the peak and the dip, so even if you cant get your dips up, due to a room mode or similar issue (and dips are harder to fix), reducing a peak will give you a better chance of hearing the frequency within that dip when its played out, giving you this more accurate and desirable playback.

The higher frequencies are affected by your room more than sub bass, they bounce around the room more, and hard surfaces provoke this. The change you see when you move your mic around is a result of proximity to these surfaces. Room treatments will certainly help and I personally feel the room is as important as the system playing in it. Simple things like softer furnishings will help stop reflections and reduce system brightness thus improving the sound. Also by reducing reflections your reducing the number of cancellations occurring where reflections meet the original signal, and that should help reduce dips in response higher up the frequency range. If you look at good dedicated cinema rooms, you'll notice the lashings of soft materials used to improve sound, and eq shouldnt be relied on as an ultimate solution. If you get the room right first, then eq'ing will be much more effective.

At this point I would like to add to a previous comment I made in another thread about smoothing. Over 200 hz you get a lot of frequency dips throughout the range. A lot of them are very precise and are only at a specific frequency. As such I dont get to worried about them. Smoothing helps iron the graph out a bit and remove these which I consider to be distractions so you can better see the overall response, and shape of this response. By all means remove smoothing to take a closer look, but dont get to wound up on very narrow spikes.

 

[hybris]

Thank you for a good explanation. I basically knew most of it already, but from your thorough explanation I deduce that it is a goal to reduce any peak as much as possible regardless of the amount of dips I have in that same frequency range - as this will also reduce the difference between the peak areas and dip areas, and give me an overall smoother response. Correct?

 

[Moonfly]

Thats correct. The smaller the difference between the peaks and dips, the less chance there is that those peaks will drown out the dips causing you to miss material. Using eq it is also possible boost dips a little too further improving accuracy.

In simple terms, the smoother your response, the better it is :T

 

[hybris]

Thats correct. The smaller the difference between the peaks and dips, the less chance there is that those peaks will drown out the dips causing you to miss material. Using eq it is also possible boost dips a little too further improving accuracy.

In simple terms, the smoother your response, the better it is :T

yep. But from 100hz and up the dips/peaks move around a lot when I measure from different spots in the couch, so I'm not sure it will be doing a lot of good. The peak at 120-125hz seem to be fairly consistent from several measurement points, so I guess that may be the only one worth pushing down with eq.

 

[glaufman]

Since your graph stops at 200Hz, I'll assume you're still just looking at bass response, even if it's the upper bass... in this range, room treatments can help a lot, and will do a better job of smoothing out the response at multiple locations than EQ will.
Not to contradict what Moonfly is saying, but while I agree with him for applying smoothing to the plots at higher frequencies, I don't think this is necessary below 200Hz, and can mask and cloud the issues benefits you're testing. You also don't need to smooth every 1-2 dB spike, as at loud volumes it's unlikely you'll hear the difference, just as certain very narrow dips might not be audible.
Another benefit of treatments is that the reduction in reflections will reduce both constructive and destructive intereference (read: raise dips AND reduce spikes) at the same time.

 

[hybris]

Well, I'm mainly focusing on the bass area since the room in question is our living room and I have gotten very limited acceptance for both the cost and visual result of room treatments from my better half I know I could benefit from acoustic treatments for the rest of the frequency range as well. I understand that the right acoustic panels may help the 2-300hz area and upwards. We're experimenting with this at a friends house (he has decay time of 0.6-0.7 secs) so I'll see what he comes up with first.

It is around the 0.4-0.45 area at my place, and I'm generally very pleased with the sound - so there's no immediate rush, but it's a fun hobby.
I've also activated an x-curve dampening the high end slightly.

This is an RT60 measurement from a little while back (haven't got my mic connected (need preamp/xlr cable) yet so it's not full range):

 

[Wayne A. Pflughaupt]

Have you tried adjusting your sub’s phase control? That can have an effect on “raggedness” in crossover region.

To reinforce what Moonfly and Greg already told you, there’s nothing in the 100-200 range worth worrying about, as far as any equalizing is concerned (assuming you have the capability to EQ the mains to begin with?). The dips are too narrow and deep, probably caused by comb filtering from reflections and/or phase between the main and sub. The two peaks are too small to worry about as well. From 100-200 Hz is only one octave, and those peaks are each only a tiny fraction of an octave On top of that, it’s usually an exercise in futility trying to EQ two peaks when there’s a sharp dip between them; ultimately you’ll end up making the dip worse. The peaks would have to be far worse than they are to make it worthwhile.

Regards,
Wayne

 

[hybris]

Have you tried adjusting your sub’s phase control? That can have an effect on “raggedness” in crossover region.

To reinforce what Moonfly and Greg already told you, there’s nothing in the 100-200 range worth worrying about, as far as any equalizing is concerned (assuming you have the capability to EQ the mains to begin with?). The dips are too narrow and deep, probably caused by comb filtering from reflections and/or phase between the main and sub. The two peaks are too small to worry about as well. From 100-200 Hz is only one octave, and those peaks are each only a tiny fraction of an octave On top of that, it’s usually an exercise in futility trying to EQ two peaks when there’s a sharp dip between them; ultimately you’ll end up making the dip worse. The peaks would have to be far worse than they are to make it worthwhile.

Regards,
Wayne

I've tried inverting the phase, which clearly is not a good idea (blue below), but I have a knob on the sub that can move gradually from 0 to 180, I guess I can experiment with somewhere in between if you think that may work?

 

[hybris]

Hehe, funny story - looking at the inverted phase comparison again I was like - hey that looks a lot like the filters I've applied. They did of course increase the effect of the inverted phase a lot.

I now tried inverting the phase and removing all the EQ filters on the subwoofer, which made the difference a lot less pronounced. It now looks more tidy, but it has a relatively wide area that is reduced from 70-90hz. Do you think I could EQ up that, or possible increase the overall volume of the sub and then EQ down the lower end?

 

[hybris]

However, there is definitely most bass when the phase is set where it was originally, so isn't that theoretically the correct phase setting?

 

[hybris]

Played around with the phase to where I got a combination of highest output and lowest dip in the crossover area, and the re-EQed the standing wave control, I guess it looks a bit better between 100-200hz now.. Neverending story.. Not sure what the antimode does in this regard, perhaps I should redo the antimode calibration now that I have moved the phase control knob on the sub?

Purple is the new version, green is the old

 

[glaufman]

However, there is definitely most bass when the phase is set where it was originally, so isn't that theoretically the correct phase setting?

No. The best place for the phase setting is where the bass sounds best, or measures the flattest (or closest to your target curve), which can be quite different from where it produces the "most bass," especially if you're measuring "most bass" with your ears, as oppose to the FR plot. Especially if the FR plot show LOTS of bass, but only at one frequency.

 

[glaufman]

Played around with the phase to where I got a combination of highest output and lowest dip in the crossover area, and the re-EQed the standing wave control, I guess it looks a bit better between 100-200hz now.. Neverending story.. Not sure what the antimode does in this regard, perhaps I should redo the antimode calibration now that I have moved the phase control knob on the sub?

Purple is the new version, green is the old

Sure, go ahead. But that graph looks better at least, so you're getting there.
For the record, I don't have experience with the antimode, but I might go about this a little differently than you seem to be... If it was me I'd shut off all automatic processing, use the manual setting and placements and treatments to the best of my ability, then run the automatic stuff with just a gut check to make sure it's making things better and not worse.

 

[hybris]

No. The best place for the phase setting is where the bass sounds best, or measures the flattest (or closest to your target curve), which can be quite different from where it produces the "most bass," especially if you're measuring "most bass" with your ears, as oppose to the FR plot. Especially if the FR plot show LOTS of bass, but only at one frequency.

I measured with radioshack + REW, and both the peak level and the average level (across the frequency range of the sub) the output was the loudest at the original phase setting.

Both antimode and the Full Band Phase Correction of the Pioneer is activated, so I would assume that they would at least theoretically do a more accurate job of setting the phase correctly than me, but who knows.

I guess it is good advice from glaufman to try changing phase + moving the speakers a little bit (which would basically mean moving them to the left or right a few centimeters - there's no real alternative placement possible) while everything is bypassed, and then rerun the automatic calibrations. The current placement is a result of trial and error when I first moved in, but that was just using my ears, not radioshack+rew.

I also figured since the windows and door (which is a glass door) near the right speakers might influence quite heavily on the dips and stuff from the right speakers, I'll try measuring with different configurations on the blinds / curtains in front of the windows / doors.

Thanks for all the input so far

 

[atledreier]

To hunt down those significant peaks I like to take a few measurements in a small area around my prime listening position and average them. That gives me a better idea of what is consistently peaking in the area, and what is comb filtering.

 

[glaufman]

My only objection to averaging is that you can have a freq that shows up as a peak in one spot, an equal dip in another spot... the averaged curve would show this freq flat, when in actuality you've got a mode causing issues.

 

[cyberbri]

You have variable phase on your sub, so this may not matter as much. But you can also adjust the relative phase with the sub distance setting in your receiver.

In my graph gallery I have a few graphs of differences in frequency response from just changing the sub distance setting.

My sub is literally behind the couch on the right side. About 2~2.5" to listening position (running through BFD eq). The turquoise line here is with sub distance setting at 3'6", and the orange line is at 5'6".

Distances in the receiver ranging from about 3' to 8'. Nothing else changed between sweeps.

BTW, I agree with not worrying about eqing 100-200Hz. You'd have to run an eq before the amp for the mains unless your mains aren't playing that range. Just do your best to get the speakers/sub phase set right for the flattest with the most frequencies at an even level.

 

[hybris]

Thanks for all comments.

A little update, I started from scratch (removing antimode and any EQ) and moved the speakers around a little bit to get the best measurements, as well as testing all possible phase settings.

I ended up with the speakers more or less where they was originally (not a lot of change when I moved them around the space that was available), and the phase at its original position as well.

anyway, long story short - after reapplying antimode and the pioneer - I ended up with a curve that was more or less identical with my last attempt. But it was an interesting journey, at least it verified my phase settings and that I didn't have any better way of placing the speakers than where I had originally positioned them by listening.