Why is Weak Bass recommended here?

[-Matt-]

(Sorry for the clickbait title)!

I've been looking back over the reviews of some of the most recommended speakers and one thing they seem to have in common is really weak bass.

Perhaps I'm making an incorrect assumption, but I'd expect that in the ideal world we would aim for the in-room response to have a gentle downward slope which carries on all the way to 20Hz (dashed blue line below). Furthermore, when it rolls-off, it is better if it drops sharply so that it doesn't put bumps into the region that potentially overlaps with the sub.

As a reminder, the manufacturer's specification is -6dB at 38 Hz which (if you are generous) could be achieved by that weird shelf in the bass response.
Preference rating for this is 6.5 or 8.1 with a sub.

Obviously small bookshelf speakers (like the Kef R3) will struggle to generate deep bass, so let's look at a tower type speaker (the Revel F35).

Again, the bass starts rolling off at about 100 Hz.
Manufacturer's spec is 55Hz, 46Hz, 35Hz (-3 dB, -6 dB, -10 dB).
Preference rating for this is 4.9 or 7.4 with a sub.


Now let's look at the Revel with the highest preference score currently listed in the review index (the F328Be):

Again, the bass starts rolling off at about 100 Hz.
Manufacturer's spec is 24Hz (-10dB); 26Hz (-6dB); 35Hz (-3dB).
Preference rating for this is 6.5 or 7.8 with a sub.


At this point I'm starting to wonder, is the Klippel actually able to measure the low bass frequencies properly?

But now let's look at this not recommended speaker (the B&W 805S):

Despite being a small bookshelf type speaker, the bass manages to extend down to about 60Hz before starting to roll off.
Manufacturer's spec is -6dB at 42Hz, -3dB at 49Hz.
Preference rating for this is 4.5 or 6.3 with a sub.

The low preference score and non recommendation for the 805S are due to the bumps in the frequency response. Such bumps could be caused, for example, by port resonances which possibly could have been reduced if the speaker wasn't tuned for such a deep response. So manufacturer's likely have to trade-off bass extension for smoothness of response.


Maybe it is an unrealistic hope, but I really don't want the bass to start rolling off until well below 80Hz (especially in floorstanding speakers) in order to have a flat response on which to impose the THX recommended 80Hz crossover.

The "with sub" preference scores are strongly biased toward speakers that don't have a deep bass response. (They assume that a perfect subwoofer will handle any low frequencies that the speaker can't manage). But if the bass roll-off of a speaker is higher, then it means the high frequency extension of the sub needs to reach further up - which is not necessarily the case and could be difficult to integrate.

Particularly in cases where people are planning to listen without subwoofers, I think a lot more emphasis should be given to bass extension.

 

[fpitas]

One practical reason might be that Amir doesn't get many full-size speakers to test. He has yet to test his Salon2s, which I bet will be more like your desired performance.

 

[abdo123]

One practical reason might be that Amir doesn't get many full-size speakers to test. He has yet to test his Salon2s, which I bet will be more like your desired performance.

Not really.

 

[abdo123]

Anyway I would start reading here: https://www.audiosciencereview.com/forum/index.php?threads/analytical-analysis-room-gain.23211/

TLDR; -3dB on 35Hz and -10dB on 20Hz would measure flat in-room. Without considering any boundary reinforcement.

and from 100Hz down to your primary room mode there is a gentle 10dB increase

 

[fpitas]

Not really.

I wonder what room gain, and gain from reflections does to the low bass.

Edit: I see you already answered.

 

[-Matt-]

But all the graphs I posted were already "estimated in-room responses" (as such, shouldn't they already include room gain).

And by the way - for what sized room - and for what distance between the speakers and the walls?

 

[abdo123]

But all the graphs I posted were estimated in room responses (shouldn't they already include such room gain).

No, they don't. Like you mention every room is different. What they do is they try to sum up the reflections to find a "speaker sound signature" that is more or less consistent across multiple rooms.

there has been some target curve research that looked at the in-room measurements alone (instead of anechoic measurements alone) and decided on a 1.2 dB/oct slope as the best sounding slope (as opposed to the 1 dB/oct slope suggested by the anechoic measurement model). but the difference is not worth talking about and I encourage everyone to find their own slopes anyway. that's the fun part about measurements in general.

And by the way - for what sized room - and for what distance between the speakers and the walls?

Check out the thread. the details about the room are laid out in detail. you can also use REW's room simulator.

 

[FrankW]

But all the graphs I posted were already "estimated in-room responses" (as such, shouldn't they already include room gain).

Not really.

Estimated ^^^
Not estimated https://www.stereophile.com/content/revel-ultima-salon2-loudspeaker-measurements

 

[-Matt-]

So the ideal response on the "estimated in-room response" graphs should have some sort of rolled-off bass?

Maybe it would be a good idea to sketch some approximation of this ideal response onto the graphs as an aid to the eye?

 

[abdo123]

So the ideal response on the "estimated in-room response" graphs should have some sort of rolled-off bass?

Maybe it would be a good idea to sketch some approximation of this ideal response onto the graphs as an aid to the eye?

I would say it's more nuanced than that, the in-room measurement above shows 100Hz to be ''loudest'' frequency in the range. Which is a choice.

 

[Zapper]

Amir documented the problem of measuring the low bass response in towers with multiple bass drivers and ports here: (smaller speakers with single woofers are unaffected)

Almost since I started measuring speakers using Klippel NFS, we have been having trouble with tower speakers and ports. The more drivers, the more ports, the worse the results there. A few months ago I experimented and realized that increasing the microphone distance from the speaker helps fair bit. The low frequency sound field is quite complex due to multiple radiating surfaces. The closer you sample to the speaker, the harder the prediction for the far-field is. Increasing distance helped with that but seemingly it was not enough.

So I reached out to Klippel with carefully scenario showing the improvements one gets from increasing distance asking if there is a better solution. I am at the limit of how wide I can go and so is the fixture. To my pleasant surprise, Christian who is the architect of NFS responded that there was an alternative undocumented way to make even better progress. The issue currently is that I set one reference point for measurements and that is set at tweeter axis. High-frequency is where complexity is very high and we need that center to be more optimized. With these tower speakers, the acoustic center for the bass is way lower causing that to be an added problem. Numerical precision is lost quickly and as such, NFS software limits the expansion order in bass. This in turn causes too little energy to be summed to represent the total bass output. With me so far? It is fine if you are not. Let me show the results.

Turns out it is possible to set the reference point for different frequency ranges. That allowed me to set a much more optimal point for bass frequencies, fixing the issue we had.

All the measurements you show were done before he learned how to make corrections for the problems of deep bass measurement.

Revel's own measurements, done with a different method, show much better bass response. However, there is concern that there are different issues with Revel's bass calibrations. Here are some of Revel's plots for comparison (source: spinorama.org)

 

[DJBonoBobo]

 

[Koeitje]

But all the graphs I posted were already "estimated in-room responses" (as such, shouldn't they already include room gain).

And by the way - for what sized room - and for what distance between the speakers and the walls?

The in-room response estimation is only accurate above the room's Schroeder frequency.

 

[-Matt-]

View attachment 278805
View attachment 278808

These are the sort of responses (or even deeper) that I'd hope to see from floorstanders.

Amir documented the problem of measuring the low bass response in towers with multiple bass drivers and ports here: (smaller speakers with single woofers are unaffected)


All the measurements you show were done before he learned how to make corrections for the problems of deep bass measurement.

This is exactly the sort of problem with the Klippel measurements that I feared. Makes sense because the woofers get further from the measurement axis as the speakers get bigger.

 

[-Matt-]

So the ideal response on the "estimated in-room response" graphs should have some sort of rolled-off bass?

Maybe it would be a good idea to sketch some approximation of this ideal response onto the graphs as an aid to the eye?

I would say it's more nuanced than that, the in-room measurement above shows 100Hz to be ''loudest'' frequency in the range. Which is a choice.

Maybe the sketched curve could show the theoretically ideal rolled off bass response which would be needed to match the THX 80Hz crossover? Anything that extends below this would just be gravy!

 

[delta76]

I've been looking back over the reviews of some of the most recommended speakers and one thing they seem to have in common is really weak bass.

weak bass is a relatively easy fix (read, subwoofers). other things are harder, or even impossible to fix.
you can't expect a small bookshelf to have strong bass, that's the law of physics. that's why we have pref. score for the measurement alone and with a (theoretically perfect) sub.

 

[Zapper]

These are the sort of responses (or even deeper) that I'd hope to see from floorstanders.

As you can see, the Klippel works fine for smallish monitors with good bass response.

This is exactly the sort of problem with the Klippel measurements that I feared. Makes sense because the woofers get further from the measurement axis as the speakers get bigger.

True. It's a complicated system and I suspect even the Klippel engineers are still learning how to take optimum measurements. This is not the first measurement issue Amir has encountered - there was variability in bass due to temperature, and an issue with high frequency irregularities due to reflections from a contact detector apparatus on the microphone.

The most foolproof bass measurements are the ones done outdoors on top of a >25 meter radio tower, where there are no nearby surfaces to interact with. Even "anechoic" chambers aren't really anechoic at low frequencies.

 

[-Matt-]

weak bass is a relatively easy fix (read, subwoofers). other things are harder, or even impossible to fix.
you can't expect a small bookshelf to have strong bass, that's the law of physics. that's why we have pref. score for the measurement alone and with a (theoretically perfect) sub.

Of course. I even made a comment to this effect in the op.

The point is... If I made a speaker that only had a tweeter (no woofer at all) and somehow it had a nice smooth response but only above 3kHz. Then in theory this could get a very high "with sub" preference score. But good luck finding an ideal sub that you can use with a crossover frequency of 3kHz!

"With sub" preference score completely forgives vast failings in bass whilst strongly punishing small ripples at higher frequencies. Or perhaps I'm wrong. Does it specify an upper limit for the frequencies that can be contributed by the ideal sub?

 

[fpitas]

Of course. I even made a comment to this effect in the op.

The point is... If I made a speaker that only had a tweeter (no woofer at all) and somehow it had a nice smooth response but only above 3kHz. Then in theory this could get a very high "with sub" preference score. But good luck finding an ideal sub that you can use with a crossover frequency of 3kHz!

"With sub" preference score completely forgives vast failings in bass whilst strongly punishing small ripples at higher frequencies. Or perhaps I'm wrong. Does it specify an upper limit for the frequencies that can be contributed by the ideal sub?

I think Harman recommends 120Hz max.

 

[-Matt-]

I think Harman recommends 120Hz max.

But I doubt that is part of the equation for calculating "with sub" preference scores. Or is it?