How to eq sub only and intergrate with mains

[Joffieb]

Hi

I am running Harbeth P3ESR's and a B&W ASW675 sub. I want to integrate the sub and want to interfere with the Harbeths as little as possible. I am running a MiniDSP 2x4 into the sub so can do time alignment and 5 band eq as well as setting a crossover. I'm looking for advice how to integrate the sub only eq'ing the sub by using measurements. So far, if I measure the sub alone and apply the generated filters, once the full system is measured, the low end where the speakers and sub play together flops.

Any advice please.

Thank

 

[sm52]

I am running a MiniDSP 2x4 into the sub so can do time alignment

To coordinate the time of subs and mains, you need to set a delay for the main ones. Not for the sub.

 

[Wayne A. Pflughaupt]

The best way to keep subs from interfering with the main speakers is to high pass the latter, thereby allowing the subwoofer to handle the bass range alone.

Lacking that, your best bet would be to run REW on your Harbeths to determine where they roll off at the low end. Then, either low pass the sub to operate below them, or if intentionally overlapping the two, only EQ the sub below the Harbeth’s operating range.

Regards,
Wayne

 

[Joffieb]

The best way to keep subs from interfering with the main speakers is to high pass the latter, thereby allowing the subwoofer to handle the bass range alone.

Lacking that, your best bet would be to run REW on your Harbeths to determine where they roll off at the low end. Then, either low pass the sub to operate below them, or if intentionally overlapping the two, only EQ the sub below the Harbeth’s operating range.

Regards,
Wayne

Thanks,

I have managed to calculate a delay to add to the sub using a MiniDsp that has removed one wide null between 115Hz and 155Hz.

I can't eq or HP crossover my mains separately. Also, not sure how to spot the roll off as there's a dip and then the lower peak may be a room mode.

So my problem is that I can't work out filters to only apply to the MiniDsp on the sub because of the overlap with the mains. If I just measure the sub and create filters, the mains at those frequencies are ignored for the filters and if I measure the full system but only apply the filters to the sub, the mains at that frequency are not filtered.

I've attached my measurements with my progress so far.

 

[Wayne A. Pflughaupt]

So my problem is that I can't work out filters to only apply to the MiniDsp on the sub because of the overlap with the mains. If I just measure the sub and create filters, the mains at those frequencies are ignored for the filters and if I measure the full system but only apply the filters to the sub, the mains at that frequency are not filtered.

Yes, I understood the problem from what you described in your first post.

Give us a full-range measurement of each of your main speakers separately. I'll take a look and see if I can give you an idea of where they roll out.

Regards,
Wayne

 

[ddude003]

The specs for the Harbeth P3ESR are: A Two-way, sealed-cabinet, stand-mounted loudspeaker. Drive-units: 0.75" (19mm) magnetically shielded, aluminum-dome tweeter; 5" (110mm) magnetically shielded, Radial2-cone woofer. Frequency response: 75Hz–20kHz, ±3dB. Nominal impedance: 6 ohms ("easy to drive"). Sensitivity: 83.5dB/2.83V/m. Recommended amplification: >15W. Power handling: 50W program...

As this is published information I would still suggest you take your own measurements as Wayne suggests...

 

[Joffieb]

Thanks guys,

Here are the speaker only measurements

 

[Wayne A. Pflughaupt]

Judging from your mains-only measurement, it appears you’ll have to limit subwoofer EQ to 40 Hz and below.

From the readings in your Post #4 file, #3 looks good, with perhaps turning the sub down a bit.

Alternately, #9, 10, and 11 look identical for all practical purposes (meaning, you’re not likely to actually hear a difference between them). They look good, except perhaps needing a slight level increase from the sub. The depression between 40 and 70 Hz will be virtually impossible to fully eliminate since the “corners” are so sharp.

Hopefully one of these two will sound acceptable (#3 or #9, 10, 11), with or without a bit of adjustment with the sub’s level.

Readings are merely a starting point. Ultimately, go with what sounds best, even if the graph it delivers isn’t the picture of perfection.

Good luck!

Regards,
Wayne

 

[Keith_W]

From this graph it is apparent that your left and right speakers are not symmetrically positioned in the room. There are huge room modes with the right (red) speaker, and the left looks much better behaved.

There is also something odd going on with that dip between 2-3kHz and the plummeting treble above 10kHz. I looked for some published measurements of your speakers and I found the Stereophile review. That plummeting treble >10kHz could be explained by (1) wrong mic orientation (2) wrong cal file (3) measuring speaker off-axis. I don't have an explanation for the dip between 2-3kHz, but I suspect this would be the XO region. Some speakers have narrower directivity at XO, which would manifest as a dip when measured off-axis, but since there are no published CEA2034 of your speaker I don't know.

Blue is the sum of your left and right speaker. The reason we sum the response is because it makes it easier to look at what is happening at low frequencies. I am sure you can immediately see the problem - there is a huge null at 65Hz. Study of individual measurements tells you that the main culprit is the right speaker.

I would suggest an 80Hz low-pass for your sub to try to fill in that huge 65Hz null. But before you even do that, ask yourself if there is any possibility for finding a better position for your L/R speakers. The principle is: position your speakers and listening position until you have as few nulls as possible in your freq response whilst ignoring the peaks. Then you chop all the peaks off with DSP. DSP can easily fix peaks. But nulls are another story - the best way to get rid of them is to reposition equipment or buy more subwoofers. You can try to fix nulls with DSP but it's not easy.

As an aside, I can see a lot of very loud and very early reflections in your Energy-Time Curve. Are these speakers on stands, or are they sitting on a table?