EQ targets for wide dispersion speakers

[Piranesi]

I recently assembled a pair of Linkwitz Pluto - inspired 3-way speakers. Like the original Plutos, they are omnidirectional up to about 3khz. Out of the box, the frequency response was pretty funky, so I've been using REW to design a parametric EQ for them.

With a typical (Harman-like target) roll off of about 2db/octave starting from 2khz, they sound pretty dark in comparison. I compared them side by side with typical sealed box speakers- definitely subjectively much darker. With a flat or even slightly up-tilted response from 2khz onwards, they start to sound more natural.

Has anyone else experienced this? Do the dispersion characteristics of a speaker affect the ideal frequency response target?

 

[thewas]

Has anyone else experienced this? Do the dispersion characteristics of a speaker affect the ideal frequency response target?

Yes, as above the modal region we mainly hear the tonality of the direct sound, see more about it here:
https://www.audiosciencereview.com/...ut-room-curve-targets-room-eq-and-more.10950/

 

[Piranesi]

Thanks for the answer. It took a while but I've read your original post in its entirety.
Is there therefore a way to estimate what an appropriate target curve should look like, for a loudspeaker of a given dispersion pattern and listened to at a given angle?

 

[TimVG]

Thanks for the answer. It took a while but I've read your original post in its entirety.
Is there therefore a way to estimate what an appropriate target curve should look like, for a loudspeaker of a given dispersion pattern and listened to at a given angle?

Yes, but ideally you need the spinorama measurement for this. I don't know the technical details of the loudspeaker, but if the on-axis sound has a neutral regression, and the listening window is flattish - I would limit correction to about one octave above the Schröder frequentie of the room.

Can you post the room response without EQ?

 

[dasdoing]

the problem is that what we percieve as reverb is percieved added to the direct sound in a 500ms FR window.
unfortunatly there doesn't seam to be a secure way to window a room meassurement to what we hear yet......once we have a windowing like that we can all EQ to flat

 

[Piranesi]

Here's the raw measurements with psychoacoustic smoothing- certainly not pretty, though I'm not sure how much of that is due to my crappy measuring setup. These are measured from a central listening position- I haven't done moving mic measurements yet.

 

[Piranesi]

Here's the impulse reponse in case it helps.

I have absolutely no idea what this means.

 

[Absolute]

There's no "target" to aim for that will guarantee a better sound. The whole research more or less concluded that whatever the curve looks like in a room with a flat, smooth measuring speaker is the correct curve. And they go on to recommend a broadband tonal EQ (tilt function) if there's a need to adjust tonality according to preferences/recording/room interaction.

Once you start to "room correct" the speaker you introduce all sorts of crazy to the speaker. To combat the room you're best off trying to minimize SBIR related dips and peaks and try to avoid strong specular reflections early in time.
If there's EQ involved above a few hundred hertz, it should be speaker correction based on anechoic measurements. If not possible to better the speaker that way, buy better speakers.

This is a rather short and sweet version of what the research has taught us.

Too simple for audiophiles, I guess

 

[dasdoing]

you can try a frequency depend window of 5 cycles in REW to filter direct sound only.
but you have to create a minimum phase version of the meassurement first so that all frequencies start at the same time

 

[ernestcarl]

Here's the impulse reponse in case it helps.

View attachment 115827

I have absolutely no idea what this means.

With all traces check-marked, and the time scale view zoomed-out so wide one can't really say much about anything here.

I'd start with figuring out how to display the frequency response graphs first using more usable scaling e.g. 50dB and/or with 5dB horizontal divisions. Static, single-point measurements are almost never going to look as pretty esp. at a distance.

If the time (e.g. phase) information is of interest, then you should use a timing reference or adjust thereafter as dasdoing mentioned.

*It's quite obvious that you are missing a whole lot of bass and low-mid energy, though.

 

[jlo]

Has anyone else experienced this? Do the dispersion characteristics of a speaker affect the ideal frequency response target?

You can check here : http://www.ohl.to/calculators/targetcurve.php
If you set DI for high frequencies to a lower value, as for an omnidirectional loudspeaker (ie DI at F2=2), you see that the high frequency target is much higher (corresponding to a lower slope).

 

[dasdoing]

If the time (e.g. phase) information is of interest, then you should use a timing reference or adjust thereafter as dasdoing mentioned.

the minimum phase version thing only works for the FR though (it will remove wrong dips that are simply caused be delayed frequencies falling out of the window). the phase information will be wrong

 

[dasdoing]

You can check here : http://www.ohl.to/calculators/targetcurve.php
If you set DI for high frequencies to a lower value, as for an omnidirectional loudspeaker (ie DI at F2=2), you see that the high frequency target is much higher (corresponding to a lower slope).

it seams impossible to use this in small rooms since there is no RT60 possible.
also the DI is estimated by the user?

 

[ernestcarl]

the minimum phase version thing only works for the FR though (it will remove wrong dips that are simply caused be delayed frequencies falling out of the window). the phase information will be wrong

Finally starting to understand now (slowly) after experimenting some.

 

[dasdoing]

Finally starting to understand now (slowly) after experimenting some.

Do you have anechoic chamber graphs for your speakers? would be intresting to compare the result of this method to those*. Unfortunatly there are no graphs for my speakers.


*though (even) the (perfect plot of the) direct sound probably isn't the same as the anechoic response (diferent distance, production diferences, some very short reflections adding to direct sound) there should be similarities

 

[ernestcarl]

@dasdoing

Only "pseudo-anechoic" directivity measurements taken inside my cramped room.

The exported minimum phase version of the LR averaged FR in the listening couch looks rather different, though:

All traces are 1/12 and with no windowing -- except for the cyan & red trace (LR vector avg FDW 15 cycles).

Listening position is ~15-20 degrees off-axis from the use of toe-in to avoid immediate side-wall reflections.

Due to the positioning and extreme angling of my speakers -- which are very close to the sidewalls, BTW, I still find myself wanting to apply a small amount of post EQ above Schroeder -- not much, and this is mainly to match channels.

Only notable strange thing that stands out to me is the 5.65 kHz hump in the anechoic FR which isn't present at the MLP. They all look similar enough, at least.

Not sure these Sceptre S8 monitors are even considered "wide" dispersion speakers -- compared to others like Revels, no, not really.

 

[dasdoing]

@dasdoing

Only "pseudo-anechoic" directivity measurements taken inside my cramped room.

The exported minimum phase version of the LR averaged FR in the listening couch looks rather different, though:

View attachment 115868

All traces are 1/12 and with no windowing -- except for the cyan & red trace (LR vector avg FDW 15 cycles).

View attachment 115865
Listening position is ~15-20 degrees off-axis from the use of toe-in to avoid immediate side-wall reflections.

Due to the positioning and extreme angling of my speakers -- which are very close to the sidewalls, BTW, I still find myself wanting to apply a small amount of post EQ above Schroeder -- not much, and this is mainly to match channels.

Only notable strange thing that stands out to me is the 5.65 kHz hump in the anechoic FR which isn't present at the MLP. They all look similar enough, at least.

Not sure these Sceptre S8 monitors are even considered "wide" dispersion speakers -- compared to others like Revels, no, not really.

not sure I understand your graphs here.

what I meant is comparing the anechoic messurement with the LP one in minimum phase version and 5 cycles of frequencie dependent windowing

 

[jlo]

it seams impossible to use this in small rooms since there is no RT60 possible.
also the DI is estimated by the user?

You're right, there is no "true" RT60 in small rooms (especially because the soundfield is not really diffuse) but estimation of 60dB decay based on RT20 or RT30 is fine enough (ie use REW).
The DI can be estimated by the user (or from manufacturer datas). For standard boxes, it is generally quite easy. It may be a bit more complicated for a horn depending on its geometry.
But to calculate the target slope, you don't need very exact values.

 

[ernestcarl]

not sure I understand your graphs here.

what I meant is comparing the anechoic messurement with the LP one in minimum phase version and 5 cycles of frequencie dependent windowing

Oks. Let me check... BTW, I need to mention that I also just noticed some discrepancy in the way REW exports a wave file IR of the Min phase version using the file menu and the one found in the 'measurement actions tool' button -- probably has to do with my export settings or whatever.

Min phase version or the FR is nearly identical with the original LR Vector avg it was derived from -- both with FDW 15 applied.

 

[ernestcarl]

anechoic messurement with the LP one in minimum phase version and 5 cycles of frequencie dependent windowing

I thought you meant anechoic measurement taken in the nearfield (orange trace). So what you mean is at 2m distance at my MLP using FDW 5 like this?