EQ Settings for Data

[rxp]

New user here, so go light on me if this is an obvious question that's been answered.

Firstly, WOW! How have I only discovered you today. I've spent years googling speaker measurement data since reading Toole's work referenced on Audioholics. The best I found was Stereophile and the NRC. I have no idea how I missed this. I found my way here due to the recent Denon receiver measurements.

Anyway - for headphones there's a nice project, AutoEQ that lets you target the Harman curve for anything measured. Is there something similar for the speaker data here? I'd love to apply a pre-computed result based on how close I'm sitting to the speaker. Specifically, I sit 1 to 1.5m away from my LS50 and would love any corrections I can try out. I know of course I can do this via room correction but from the research I've read, I only EQ below 150hz. If I had the anechohic data and an EQ made that flat - wouldn't the in-room performance also improve?

 

[Webninja]

I’m still learning how to measure my speakers and how they perform in my room. From what I understand, the room interaction is greater at lower frequencies. There are some that only EQ under 500Hz or 2k Hz.

I suggest trying REW (search for the tutorials on this site), and then depending on your music delivery system, you can apply EQ to help lesson any peaks you measure on your LS50s. I’m using Roon’s EQ for my case.

Oh and welcome aboard, this site is great.

 

[thewas]

Especially above the transition frequency EQ based on anechoic data is a nice approach when the loudspeaker has smooth directivity like your LS50 has. Here I have made such EQ for the LS50, mind you there a two small caveats, the here used NFS measurement setup has minor deviations and also your samples might also have additionally a minor deviation from the here measured one.

 

[Fluffy]

Another AutoEQ aficionado… you couldn't do it for speakers for the same reason you can't really do it on headphones – in headphones, every ear is different and tiny changes in placement and shape makes a big difference in FR. And in speakers, the room is always very dominant, so you can't just create a template that will "fix" every speaker regardless of its acoustic surrounding. That's why for speakers you need to measure the room first and construct a custom correction.

 

[QMuse]

I know of course I can do this via room correction but from the research I've read, I only EQ below 150hz. If I had the anechohic data and an EQ made that flat - wouldn't the in-room performance also improve?

As @thewas_ explained room affects in-room response heavilly up to 300-400Hz so room EQ is always needed if you want to fix the in-room response. You may still have some bumps in the in-room response response in the 400-700Hz region which you can also fix with room EQ but it is advised you do it based on spatially averaged measurements over your LP.

Based on NFS measurement you can EQ speaker anechoic response from 300Hz upwards and you will certainly improve performance. You can also fix resonances in the 80-300 region if NFS measurements shows they are present. If you don't have NFS data available for your speaker you can try to do in-room near field measurement and apply gating, for example you can measure on tweeter exis from 1m distance and apply 5-6ms of gating. Based on that measurement you should be able to EQ your speaker upwards from 1kHz.

What is important here to note is that room EQ and speaker EQ are complementary actions which give best results when applied together.

 

[QMuse]

Another AutoEQ aficionado… you couldn't do it for speakers for the same reason you can't really do it on headphones – in headphones, every ear is different and tiny changes in placement and shape makes a big difference in FR. And in speakers, the room is always very dominant, so you can't just create a template that will "fix" every speaker regardless of its acoustic surrounding. That's why for speakers you need to measure the room first and construct a custom correction.

Based on NFS data you can do speaker EQ which will be valid for every speaker assuming the inter specimen variation is small. Room influence is present in the 20Hz-300Hz (or 400Hz) so room EQ should be done specifically for your room based on in-room spatially averaged measurements.

 

[thewas]

Another AutoEQ aficionado… you couldn't do it for speakers for the same reason you can't really do it on headphones – in headphones, every ear is different and tiny changes in placement and shape makes a big difference in FR.

That reason while valid for headphones as they bypass part of the individual HRTFs isn't valid for loudspeakers as with the same individual HRTFs we hear also the reference that is the real life sounds, so we don't want any compensation for that.

And in speakers, the room is always very dominant, so you can't just create a template that will "fix" every speaker regardless of its acoustic surrounding. That's why for speakers you need to measure the room first and construct a custom correction.

As @QMuse correctly wrote we must separate room and loudspeaker correction, only in the lower bass region room correction dominates but still a loudspeaker that is anechoically linear has better chances to be be preferred when positioned wisely, thus not to near to boundary surfaces. Floyd Toole did research of listeners preferences in different rooms and always the same most neutral loudspeakers were preferred.

 

[Fluffy]

That reason while valid for headphones as they bypass part of the individual HRTFs isn't valid for loudspeakers as with the same individual HRTFs we hear also the reference that is the real life sounds, so we don't want any compensation for that.

It's not only about HRTF. The ear shape and placement of the headphone plays a much more significant role, because what happens inside the cups is basically the acoustic space that the driver needs to contend with. it's like the room for speakers. it's much more dominant in closed headphones, where a slight movement can alter the FR a great deal.

 

[QMuse]

It's not only about HRTF. The ear shape and placement of the headphone plays a much more significant role, because what happens inside the cups is basically the acoustic space that the driver needs to contend with. it's like the room for speakers. it's much more dominant in closed headphones, where a slight movement can alter the FR a great deal.

While a parallel can be drawn I see a problem with headphones with measuring response specific to one's ear and headphone placement on/over them. Contrary to that, with speakers it is relatively easy to obtain good in-room measurement over LP.

 

[thewas]

It's not only about HRTF. The ear shape and placement of the headphone plays a much more significant role, because what happens inside the cups is basically the acoustic space that the driver needs to contend with. it's like the room for speakers. it's much more dominant in closed headphones, where a slight movement can alter the FR a great deal.

But as said that plays absolutely no role with loudspeakers as its the same full transfer path when listening to the live event or a loudspeaker.

 

[Fluffy]

But as said that plays absolutely no role with loudspeakers as its the same full transfer path when listening to the live event or a loudspeaker.

Right. The point is still acoustics and the dependence of the final response on specific acoustic conditions, that makes generalization of solutions unreliable.

 

[QMuse]

Right. The point is still acoustics and the dependence of the final response on specific acoustic conditions, that makes generalization of solutions unreliable.

Not really, as speakers with linear anechoic response have been proved to be preferred. As it applies as a general rule it makes sense to EQ the speaker based on anechoic measurement. Room EQ is still needed, but as every room is specific no general rules apply except that you need to EQ in-room response to the preferred target curve.

 

[thewas]

Right. The point is still acoustics and the dependence of the final response on specific acoustic conditions, that makes generalization of solutions unreliable.

Nope, people prefer on-axis linear loudspeakers with smooth directivity that are room corrected in the bass, its all known and proven research.

 

[Fluffy]

If it was so simple, a solution like the one OP proposed would probably already have existed. Why doesn't it exist, then?

 

[thewas]

If it was so simple, a solution like the one OP proposed would probably already have existed. Why doesn't it exist, then?

Such solutions exist from many companies and hobbyist already for headphones, for loudspeakers the limitation is in the bass region where you need measurements or precise modelling of the room, although for even there now slowly solutions come to the surface like
https://en-de.sennheiser.com/mediaroom/dsp-power-for-all-analog-neumann-monitors-284808
https://docs.linn.co.uk/wiki/index.php/Space_Optimisation_and_Space_Optimisation+
https://www.devialet.com/en-ca/expert-pro-sam-ready-speakers/
where you don't necessarily need microphones.

But again as known by the Harman research an anechoical linear and smooth loudspeaker is preferred indecently of the listening room. By the way more than 10 people by now in this forum have asked my about those filter coefficients for some loudspeakers and none till now did prefer the uncorrected one, although correcting the bass region via measurements is even one step further.

 

[rxp]

Especially above the transition frequency EQ based on anechoic data is a nice approach when the loudspeaker has smooth directivity like your LS50 has. Here I have made such EQ for the LS50, mind you there a two small caveats, the here used NFS measurement setup has minor deviations and also your samples might also have additionally a minor deviation from the here measured one.

That sounds exactly like what I'm after! Amazing - can share the EQ? I'll use EQAPO in Windows to apply. I get Toole's research so I was assuming any correction that flattens anachoic measurements would sound better in a room. So I was hoping there were some corrections based on the excellent data on this site. Another site does exactly that NoAudiopile but there's not many corrections on there.

I've gone down the rabit whole of room EQ before. I have XT32 on my SR6011 and SR6012 (two seperate theaters). While I like what it does for bass, I've never had a big "this is so much better" moment for any room EQ.

Compare to that to AutoEQ sourced oratory1990 (reddit user) measurements (or oratory1990s own corrections) where it's literally night and day. It can turn a horrific sounding headphone like the DT990 into magic. I don't agree with the HRTF issues on headphone EQ btw, we can pick it up on another topic but in my experience it's been fantastic. Even if you did have issues with how circumaural headphones impact HRTF - you can avoid that by using binural mics on your own head. Headphones like the WHXM3's even equalise to a standard target according to wearing condition. I'm a big believer in all this.

In fact the same (genuis!) author of AutoEQ created Impulcifer which uses binural convolution to simulate 7 loudspeakers over headphones. You can then even apply a virtual room target. It's crazy how realistic it is.

 

[thewas]

That sounds exactly like what I'm after! Amazing - can share the EQ? I'll use EQAPO in Windows to apply.

Here you are, you can directly copy it in an EQAPO txt file:

Preamp: -2.7 dB
Filter 1: ON PK Fc 67.20 Hz Gain 3.00 dB Q 2.000
Filter 2: ON PK Fc 170.0 Hz Gain -2.00 dB Q 1.000
Filter 3: ON PK Fc 540.0 Hz Gain 1.60 dB Q 5.000
Filter 4: ON PK Fc 830.0 Hz Gain -2.40 dB Q 2.500
Filter 5: ON PK Fc 1240 Hz Gain -1.20 dB Q 8.000
Filter 6: ON PK Fc 1650 Hz Gain 2.50 dB Q 3.000
Filter 7: ON PK Fc 2640 Hz Gain -2.80 dB Q 2.000
Filter 8: ON PK Fc 4620 Hz Gain -3.40 dB Q 3.000

 

[rxp]

Thanks so much! Going to give it a try now.

PS: for nearfield I found that the Denon/Marantz have Cinema EQ which is just a treble rolloff. I've never used it before but because I was going down the rabit hole of speaker/room correction again I thought I'd try. Really helps tame that treble.

I wish AVR people would just give us simple tone controls. It's ridiculous that we don't have simple Bass/Mids/Highs.

 

[rxp]

I tried the EQ - I did prefer it on. It's very difficult to notice but it's definitely better. I guess the LS50 is a good speaker by default so the EQ is just a slight tweak.

I feel quite surprised now - after never correcting above 150hz when I'm actually trying it and verifying the measurements with REW and a UMIK and XT32 is doing a good job of flatten the response across the range - even across multiple mic placement/speakers. I noticed the app actually lets you define a target curve too, so no need for tone controls. Down the room correction rabit hole I go....

 

[aarons915]

I tried the EQ - I did prefer it on. It's very difficult to notice but it's definitely better. I guess the LS50 is a good speaker by default so the EQ is just a slight tweak.

I feel quite surprised now - after never correcting above 150hz when I'm actually trying it and verifying the measurements with REW and a UMIK and XT32 is doing a good job of flatten the response across the range - even across multiple mic placement/speakers. I noticed the app actually lets you define a target curve too, so no need for tone controls. Down the room correction rabit hole I go....

This is one case that I prefer using Soundstage measurements for EQ since my own in room measurements don't show a 5k resonance but rather show a large 2k resonance. Try these if you want, they are simpler, I only EQ above 600Hz or so because the room can still affect below that. They are a modified listening window based on 2 of the 15 degree curves and the 30 degree response in the Soundstage measurements, I point my LS50 straight ahead so it should approximate my listening window better than their traditional window that is more on-axis.