Is Full Range EQ a no go?

[Orange55]

As many have said before me the knowledge on this web site is fantastic and I have read with interest and put into practice lots of the techniques I have seen online, house curve, min number of filters etc ...

However, there is one area which appears not be to addressed and that is Full Range EQ, everything appears to focus around the 20hz to 200 hz. I have worked on getting the full range from 20hz to 20Khz smooth and with great effect I believe from what I hear on my 5.1 system using the ECM8000 mic and XENYX preamp.

But am I doing something which should be avoided?

[brucek]

am I doing something which should be avoided?

A couple of issues here.

The first is the quality of the EQ you're using. If it's a BFD or the like, then there are issues of noise floor introduced along with concerns over the quality of the full conversion through the ADC and DAC chain.

Second is the tendency of many to try and use other than very low Q filters. Trying to adjust peaks and dips at higher frequencies is problematic unless you placed your head in a vise while listening. Only very broad filters should be used.

brucek

[Wayne A. Pflughaupt]

However, there is one area which appears not be to addressed and that is Full Range EQ...

I keep meaning to do a write-up on the subject...

In the mean time, you can find some good info on the topic at this thread.

I have worked on getting the full range from 20hz to 20Khz smooth and with great effect I believe from what I hear on my 5.1 system using the ECM8000 mic and XENYX preamp.

What are you equalizing with?

Regards,
Wayne

[Orange55]

Hi Guys, thanks for the guidance, very helpful and I have now made changes based on your advise.

Firstly I am using a Tag Mclaren AV32R to do the EQ, with Linn Ninka mains and a B&W PV1 sub. My Centre and Surround are the matching set to the Linn fronts.

I spend 80% of my time listening to 2 channel music, hence my desire to eq the mains, I don't use the sub with music as the mains go down to 50Hz. Any assistance would be great.

Here is my fronts before and after with the eq used, I removed some of the 300hz plus at the weekend with no adverse effect to the actual sound.


Front Left - No EQ


Front Left - with EQ


Front Left - Settings:
71Hz -3.0 10.0
105Hz -10.5 3.8
147Hz 5.4 2.3
191Hz -7.5 5.8
4700Hz -3.0 1.1



Front Right - No EQ


Front Right - with EQ


Front Right - Settings:
72Hz -3.0 1.3
131Hz 3.7 5.1
175Hz -5.5 5.1
222Hz 5.5 3.4
290Hz -3.0 5.1
4700Hz -3.0 1.1




Finally my sub has rather two drastic drops that I have been unable to do anything about, from moving he sub changing phase and EQ etc; Overall I think the sub graph looks good, but wondering if there are any secret tricks. Plus there are far to many filters, but I can't seem to reduce them and still get the desired response.


Sub - no EQ


Sub - with EQ



Sub - Settings:
21Hz 2.0 4.3
33Hz -3.1 3.1
41Hz 5.0 3.7
57Hz 4.0 5.0
61Hz 4.0 3.4
72Hz -8.6 3.9
97Hz -14.5 2.2


Thanks in advance.

[weverb]

Wayne is definitely your man to help guide you through eq'ing your mains. He has been a great mentor to me through my journey of eq'ing my entire 5.1 set-up. As brucek stated... be sure to use only wide/broad filters.

Looks good so far!

[Wayne A. Pflughaupt]

Thanks for the kind words, weverb!

Orange55 wrote:

I spend 80% of my time listening to 2 channel music, hence my desire to eq the mains, I don't use the sub with music as the mains go down to 50Hz. Any assistance would be great.

Does the Tag use FIR filters? It’s curious that the left and right graphs look so different, both before and after EQ.

What’s with the peculiar Target curve? I’d want response out to 20 kHz, if the speakers are capable of generating it. Is the Tag causing that roll out?

The Target curve for the sub, I’d suggest shelving at ~30 Hz rather than 50 Hz, if you have enough headroom. Also, why did the sub’s Target change from 72 dB to 80 dB?

It looks like most of the sub filters are indeed needed. It’s a shame you can’t get better before-EQ response from the sub. If there are some other placement options you haven’t explored yet, I’d surely look at them, especially a corner, if there is one available. A sub with 8” drivers usually needs all the help it can get.

Regards,
Wayne

[Roger Dressler]

Wayne A. Pflughaupt wrote:

Does the Tag use FIR filters?

2nd order IIR biquads, per John Mulcahy.

[Orange55]

Wayne A. Pflughaupt wrote:

Thanks for the kind words, weverb!

Does the Tag use FIR filters? It’s curious that the left and right graphs look so different, both before and after EQ.

What’s with the peculiar Target curve? I’d want response out to 20 kHz, if the speakers are capable of generating it. Is the Tag causing that roll out?

The Target curve for the sub, I’d suggest shelving at ~30 Hz rather than 50 Hz, if you have enough headroom. Also, why did the sub’s Target change from 72 dB to 80 dB?

It looks like most of the sub filters are indeed needed. It’s a shame you can’t get better before-EQ response from the sub. If there are some other placement options you haven’t explored yet, I’d surely look at them, especially a corner, if there is one available. A sub with 8” drivers usually needs all the help it can get.

Regards,
Wayne

Hi Wayne

Thanks for your feedback, it appears I have made a couple of school boy errors.

Firstly I have applied a house curve to my mains, whereas they should be flat and the house curve onto the sub.

As for the overall response from the mains I wonder if it is the position of the speakers that could be causing that.

The sub levels are different as I uploaded an old graph for the non-eq.

So next steps are to apply all of the learning from above and from the other thread and will post my new measurements for review and a picture of the actual room as that may help with any odd readings.

Thanks again

[Orange55]

Ok lets try again, staring with the sub.

Here are the pre and post eq graphs reflecting the new roll-off. Sadly again using 8 filters.


Sub Without EQ


Sub With EQ



Sub Settings:
20Hz 4.0 1.8
34Hz -3.1 6.0
41Hz 5.3 5.0
44Hz 5.0 3.2
60Hz 6.0 6.3
72Hz -13.0 5.0
90Hz -8.9 3.0
103Hz -7.6 3.0



Here is the room in which the speakers sit, to see if this explains why the fronts are so different pre and post eq.

System



And here are the graphs for the fronts:

Front Left Without EQ


Front Left With EQ


Front Left Settings:
45Hz 4.5 4.6
46Hz 4.5 6.3
54Hz 5.0 1.8
70Hz -12.2 6.9
104Hz -11.6 2.1
131Hz 5.0 3.9
156Hz 5.2 7.8
186Hz -2.2 6.0




Front Right Without EQ


Front Right With EQ


Front Right Settings:
50Hz 5.0 5.0
53Hz 5.0 5.0
55Hz 5.0 5.0
70Hz -10.9 4.0
89Hz -3.6 2.0
126Hz 4.6 5.7
172Hz -2.6 5.6
214Hz 5.0 10.0


Any help and comments are more than welcome, the main issue I see if far to many filters and I hope some person with more experience may be able to guide me in the right direction.

Thanks

[weverb]

On your mains, couldn't the first three filters on each main be done with one filter with a larger bandwidth?

Also, looking at your picture, you are dealing with the same issue as me, no room symmetry around the mains. Therefore, they are going to respond/perform differently.

[brucek]

graphs reflecting the new roll-off

I'm as baffled as Wayne by your sub target.

It's as if you've baked in a peak. I've circled the drop off that you've programmed in. Can you discuss why you do this. It results in a peak at 30Hz........ It's exactly what we're attempting to avoid.


Below is a more conventional house curve.
It's an log interpolated 8dB rise from 70Hz to 30hz.


brucek

[Orange55]

Hi brucek

I can imagine your thought when you saw the graph .

I have clearly miss understood what I had read on the house curve. I see my mistake now. What great service you get here.

I think I can see why I made that mistake and I am probally very well positioned to write a dummies guide to eq given the basic errors I have made.

[Orange55]

weverb wrote:

On your mains, couldn't the first three filters on each main be done with one filter with a larger bandwidth?

Also, looking at your picture, you are dealing with the same issue as me, no room symmetry around the mains. Therefore, they are going to respond/perform differently.

I'll see if I can do that, thanks.

[brucek]

miss understood what I had read on the house curve

Yeah, the file of the jpg I posted is a simple:

30 8.0
70 0.0

You're just trying to boost the bottom end a bit, but usually you'll stop at 30Hz. It's generally a good idea to tick the log interpolate feature checkbox which draws a straighter line between your points, otherwise it will hump up between points and sound a bit bloated.

The pic below shows the difference between a normal 80Hz crossover target and then compares it to a 5dB house curve. You can see the boost in the target. The new target line will be used by REW to recommend its filters.



brucek

[Wayne A. Pflughaupt]

Orange55 wrote:

As for the overall response from the mains I wonder if it is the position of the speakers that could be causing that.

Judging from your picture, I’d say so:

The left speaker, it’s the corner that’s causing the boost in the 70-100 Hz range to be worse than on the right.

Good looking system, by the way – I’m drooling!

Since the processor is using traditional IIR filters (thanks for the info Roger), I think it’s over-equalizing. First, I’d loose all the boosted EQ settings below the crossover frequency for both channels. That's the subwoofer's territory.

For the left channel, the narrow 131, 156 and 186 Hz filters are all “fighting” the overly broad filter at 104 Hz. “Equalizing the equalizer,” in other words. Rather than using two closely-spaced filters at 104 and 70 Hz, I’d try a single, broader filter at 70 Hz or a bit lower. That would pull down the 70-100 Hz peaks without “digging in” to the frequencies above that point. A side benefit would be that the same filter would also reduce the 40 Hz peak.

For the right channel, the same thing. A single broader filter south of 70 Hz will bring down that peak with minimal effect on upstream frequencies. The 172 Hz filter isn’t needed, but something in the 125 Hz range might be after applying the <70 Hz filter. The 214 Hz filter is way too narrow and boosted too much – I’d lose it.

That said, it wouldn't hurt to save the processor's settings so you can compare them to my recommendations. Use the ones that sound better to you.

For your sub, I’d say it’s probably your placement that’s killing you. Will it fit in that corner behind the left speaker? Corner placement often works wonders in asymmetrical rooms like yours. It should boost your output and get you better extension, and any response anomalies will probably be easily equalizable (unlike with your current placement).

Regards,
Wayne

[weverb]

Told ya Wayne was "The Man" when it comes to eq'ing the mains.

[Orange55]

Wayne A. Pflughaupt wrote:

Since the processor is using traditional IIR filters (thanks for the info Roger), I think it’s over-equalizing. First, I’d loose all the boosted EQ settings below the crossover frequency for both channels. That's the subwoofer's territory.

Thanks Wayne, I kind of did suspect there were to many filters and now you have explained your recommendatons I can see the error of my ways. I'll give it a bash over the weekend and see how it sounds.

Wayne A. Pflughaupt wrote:

For your sub, I’d say it’s probably your placement that’s killing you. Will it fit in that corner behind the left speaker? Corner placement often works wonders in asymmetrical rooms like yours. It should boost your output and get you better extension, and any response anomalies will probably be easily equalizable (unlike with your current placement).

Yes I know the placement is not great, I have tried moving it a little but now into that corner due to space issues. I will have a look and see what can be done, at least now I have ditched the house curve I know what it should look like.

Here's to a weekend of fun!

[brucek]

I’d say it’s probably your placement that’s killing you

I would get those mains out from being behind the TV and so close to the walls. At least make them equal or slightly in front of the rack and test... you'll be surprised in the new soundstage that you getting.

Yeah, put the sub in the corner (since you're going to move the speaker out, there will be room).

brucek

[spreston]

Hi Orange,

I tried to EQ full range with a Behringer DEQ2496. With one set of speakers, the effect was rather interesting and it was fun to fix the EQ work done by recording engineers.

However, I found the output of the Behringer to be less than transparent (maybe the DAC, maybe other issues in the box). As a stand-alone DAC it works but I have heard better. In the end, I decided to run my mains directly from the poweramps. The sub is a different beast entirely.

You might have a different experience running your mains with your EQ and using different DACs.

[Rudy81]

I am following your post with interest since I have just recently made my first measurements. I am wondering why the roll off beyond 10kHz? I have a very similar plot and can't figure out why the tweeters roll off that way. Am I setting something incorrectly?

Thanks,
Rudy

[sfdoddsy]

Most in-room measurements from the listening position will show that kind of treble roll-off. I wouldn't worry about it.

[Wayne A. Pflughaupt]

Rudy81 wrote:

I am following your post with interest since I have just recently made my first measurements. I am wondering why the roll off beyond 10kHz? I have a very similar plot and can't figure out why the tweeters roll off that way. Am I setting something incorrectly?

Most often, it’s the result of poor measurement technique – specifically, pointing the mic at the ceiling. The mic’s pick-up pattern is omnidirectional for most of the frequency range, but at the highest frequencies it becomes directional. So naturally, if you use a vertical orientation the only information it’s getting at the extreme frequencies is what’s bouncing off the ceiling – if that.

Presumably, if you’re taking full-range measurements, you’re interested in seeing what response at the listening position is. You want a graph that shows both what your speaker are generating and what you’re hearing at that position. The best way to accomplish this is to set up the mic at the listening position and aim it at the speaker.

Regards,
Wayne

[brucek]

The best way to accomplish this is to set up the mic at the listening position and aim it at the speaker.

The generic file that we provide from Cross Spectrum is frontal incidence. Certainly for near-field measurements the mic should be oriented horizontally and pointed at the speaker for accurate high frequency readings, but for listening position measurements with the downloadable cal file, to take advantage of direct and reflective signal, Cross Spectrum (and our download page advice) recommends to point the mic vertical with a 10-20 degree tilt toward the front of the room.

brucek

[weverb]

brucek wrote:

The generic file that we provide from Cross Spectrum is frontal incidence. Certainly for near-field measurements the mic should be oriented horizontally and pointed at the speaker for accurate high frequency readings, but for listening position measurements with the downloadable cal file, to take advantage of direct and reflective signal, Cross Spectrum (and our download page advice) recommends to point the mic vertical with a 10-20 degree tilt toward the front of the room.

brucek

This is the technique I use, but I also had Cross Spectrum provide me calibration files for both orientations.



Oh, and thanks to Wayne's great mentoring, I EQ full range.

[Wayne A. Pflughaupt]

brucek wrote:

The generic file that we provide from Cross Spectrum is [calibrated for] frontal incidence. Certainly... for listening position measurements with the downloadable cal file, to take advantage of direct and reflective signal, Cross Spectrum (and our download page advice) recommends to point the mic vertical with a 10-20 degree tilt toward the front of the room.

Yes, I know vertical is what Cross Spectrum recommends, but it’s poor technique in general and counter-intuitive in particular for a mic that’s calibrated for horizontal orientation. Not to mention, it’s contrary to long-accepted industry standards for in-room, full-range audio spectrum analysis: I’ve never seen any professional system calibration done utilizing a vertical mic, and I haven’t seen any manufacturer of hardware RTAs, nor any software program (such as TrueRTA), recommend it either (although admittedly I probably don’t know all of them that are available).

Vertical orientation doesn’t “take advantage” of the direct signal at the extreme high frequencies - it minimizes them, due to the aforementioned high-freq directionality issue with the mic.

As far as reflective signals, for the life of me I don’t see why anyone would want to insure that they they’re captured for a full-range measurement, since by all accounts they degrade the accuracy of a listening-position measurement. But below 10 kHz or so they’re all going to be there anyway with horizontal orientation, so nothing is lost, reflections-wise, by aiming the mic at the speaker, yet you gain the advantage of seeing what your speakers are delivering above 10 kHz at the listening position.

weverb has the right idea – if vertical orientation is the desired method of measurement, the mic should be calibrated for it, to compensate for the natural high-end loss.

Regards,
Wayne