Paul Hales on room equalization on Soundstage! Network

[Few]

Long-time speaker designer Paul Hales was interviewed on Soundstage! Network. He has interesting arguments supporting his position on room equalization. “Would you equalize a piano to compensate for the resonances in a music room?” (Paraphrased). I think there are a couple of unaddressed gaps in his logic, but it’s thought-provoking and well thought out. I had to cringe through the interviewer’s interjections (how did they choose this guy?), but it was time well spent overall. The link is for part 1, but there is a part 2 as well.

I’d be interested in others’ impressions.

 

[Curvature]

“Would you equalize a piano to compensate for the resonances in a music room?”

Yes, and this is done all the time in performance spaces, either physically using architectural means ("treatment") or electronically, since almost all concert halls have reinforcement.

EQ filters, speakers, rooms and ears can all be represented in mass-spring-damper models. They seem more different than they are.

Edit: Typo. I would also add that changing speaker and listening position can be interpreted as a kind of EQ.

 

[Few]

This doesn't really address Hales's point. If I understand him correctly (and it would be better for others to draw their own conclusions from his statements), he asserts that by equalizing to compensate for room resonances, you detrimentally affect the direct sound which, by definition, arrives at the ears without having been affected by those resonances. If the room has a mode that causes a peak at 60 Hz and you implement a notch filter at 60 Hz to compensate, the direct sound arrives with a big notch at 60 Hz, not with flat response. He claims that one result is a significant loss of impact in the bass.

A resonance in a speaker, on the other hand, affects both direct and indirect sound, so using filters to eliminate the associated peaks and dips also improves the speaker's time domain behavior (again, that's my take on his statements, and I may be misrepresenting them). Hales also ties this to a need for the off-axis sound to match the on-axis sound so that corrections for one angle do not cause problems at another.

 

[kemmler3D]

A resonance in a speaker, on the other hand, affects both direct and indirect sound, so using filters to eliminate the associated peaks and dips also improves the speaker's time domain behavior

This is also true of room EQ, though.

I see his point as you've explained it, but I think it only applies in large rooms where your ear can readily differentiate direct and indirect bass... I'm a bit out of my technical depth but I think this is not applicable to small rooms, i.e. most of our homes.

 

[Coltr@ne]

@Few

PSI Audio AVAA C20 electronic bass trap | Stereophile.com

"Bass—the final frontier," declared Captain James T. Kirk. I have no doubt: The biggest problem in nearly every listening room is getting the bass to sound right. Today, we voyage to the frontier of bass response. A Brief History of Time

www.stereophile.com

 

[Few]

To be clear, I'm not championing Hales's cause. I hoped to encourage others to listen to the interview and share their own impressions. I may very well be distorting his argument as it passes through my own biases. Hales (unlike the interviewer) is an articulate guy, so whether you agree or disagree, I think he does a good job of making his argument.

Hales also argues for wide directivity speakers because (he argues) most natural sound sources are not very directional. He mentions some of his own speakers that radiate quite uniformly into 180 degrees, but he doesn't say why it wouldn't make more sense to go for a full 360 degrees, according to his logic. Some of his speakers are very shallow and intended to mounted very close to a wall, so maybe that's the reason.

Part of his pitch is, I think, that humans acclimate to their listening environment so things sound wrong if we place a sound source into that environment but remove the environment's influence. I can see that logic holding together for a close-mic'd recording where someone's goal is "they are here" reproduction--you want the saxophone to sound like it's being played in your listening room. I didn't catch him clarifying how the ambience from a recording in a large venue works into this. Do we want our system to sound like someone packed Symphony Hall into our living room?

Again, these are just my impressions so others might come away with a different sense of what Hales is saying. And then there's the question of whether what he is saying holds up.

 

[gnarly]

I strongly agree with Paul Hales. "Room-correction" with DSP is a misnomer. The only way to correct a room is acoustically.
And agree that the cost of trying to obtain the smoothest in-room response possible below Schroeder, is a loss of transient impact.

 

[Timcognito]

A piano is point source part of what DSP does is correct imbalances for the stereo image at the listening position. Few listen to a piano that way and there is almost no way to correct a whole large room. Also, IMO only because I'm out of my league of technical expertise.

 

[amirm]

If all you listened to was piano, then you could live without EQ of your system as well. The times bass modes get hit are small and brief...

 

[Few]

Here’s a specific case in round numbers: If a reflected sound travels 4 meters farther than the direct version of the same sound, then its arrival is delayed by almost 12 ms, which is the period of an 85 Hz wave. So the first cycle of the 85 Hz component of the sound can’t be influenced by the reflection which hasn’t arrived yet. If the listener places a notch in the response at 85 Hz because the reflection contributes a peak to the response, the first cycle will be notched out even though it could not have been influenced by the room resonance. Hales argues that the attempt to address the room resonance distorts that first cycle and audibly detracts from the speakers ability to deliver natural impact at low frequencies.

That seems like a different case than the slow build-up of energy associated with a resonance with a high Q (quality factor). That slow build-up isn’t from any propagation delay. It’s just inherent in narrow peaks in the spectrum—Fourier says narrow peaks build up and also decay slowly, meaning over many cycles.

Hales acknowledges that his recommendation to not use dsp to flatten room resonances is inconsistent with the usual advice, and a rare case in which he disagrees with Toole et al. I have not done my own experiments so I’m left unable to verify or refute the various claims. It’s an interesting topic, though.

 

[janbth]

Here’s a specific case in round numbers: If a reflected sound travels 4 meters farther than the direct version of the same sound, then its arrival is delayed by almost 12 ms, which is the period of an 85 Hz wave. So the first cycle of the 85 Hz component of the sound can’t be influenced by the reflection which hasn’t arrived yet. If the listener places a notch in the response at 85 Hz because the reflection contributes a peak to the response, the first cycle will be notched out even though it could not have been influenced by the room resonance. Hales argues that the attempt to address the room resonance distorts that first cycle and audibly detracts from the speakers ability to deliver natural impact at low frequencies.

My take: This the sort of argument that is made from the lack of understanding of how room modes work. Room modes are a phenomenon that appears below the transition frequency, when wavelengths are on the order of the dimensions of the room. In particular, the traveling wave analogy breaks down. Room modes, standing waves, resonances - are collective phenomena. One or two cycles of sound is irrelevant. A peak in the frequency response is not created by «a reflection». It is built up by several repeated cycles.

That seems like a different case than the slow build-up of energy associated with a resonance with a high Q (quality factor). That slow build-up isn’t from any propagation delay. It’s just inherent in narrow peaks in the spectrum—Fourier says narrow peaks build up and also decay slowly, meaning over many cycles.

What is being said here?

 

[kemmler3D]

To be clear, I'm not championing Hales's cause. I hoped to encourage others to listen to the interview and share their own impressions. I may very well be distorting his argument as it passes through my own biases. Hales (unlike the interviewer) is an articulate guy, so whether you agree or disagree, I think he does a good job of making his argument.

Hales also argues for wide directivity speakers because (he argues) most natural sound sources are not very directional. He mentions some of his own speakers that radiate quite uniformly into 180 degrees, but he doesn't say why it wouldn't make more sense to go for a full 360 degrees, according to his logic. Some of his speakers are very shallow and intended to mounted very close to a wall, so maybe that's the reason.

Part of his pitch is, I think, that humans acclimate to their listening environment so things sound wrong if we place a sound source into that environment but remove the environment's influence. I can see that logic holding together for a close-mic'd recording where someone's goal is "they are here" reproduction--you want the saxophone to sound like it's being played in your listening room. I didn't catch him clarifying how the ambience from a recording in a large venue works into this. Do we want our system to sound like someone packed Symphony Hall into our living room?

Again, these are just my impressions so others might come away with a different sense of what Hales is saying. And then there's the question of whether what he is saying holds up.

I think I follow, it sounds like this is connecting a subjective experience with a specific concrete observation. Seems ripe for an experiment.

If we're concerned about losing impact by notching the direct sound, it at least sounds kind of reasonable to me, although I could be persuaded there's a mistake here.

I would point out that in your example, 12ms is maybe long enough to hear direct vs. reflected, but you won't find many phenomena in music that are so short. Even a pretty tight bass drum sample lasts 10x longer than this. The initial attack of a drum happens on a scale this long, generally a bit longer. An entire drum hit, or a single short bass note is many times longer than this. So the benefit of EQ becomes beneficial within a short window relative to musical events.

I do think more generally, if you're concerned about this stuff, phase distortion from bass EQ is worth thinking about, and can be corrected somewhat with FIR filters.

But I think it's definitely debatable whether the 'impact' from direct sound / initial attack is worth the overall crappy sound you get from uncontrolled bass modes. It's not like masking from excess decay time slopping all over the place doesn't hurt impact, too.

I would also say yes, your ears adjust, but even listening to live music in a venue, uncontrolled bass modes sound like crap and can wreck the experience. It's not like your brain can totally filter out +10dB at 120hz and -10dB at 80hz. The notes go up and down in volume like a merry-go-round.

 

[MattHooper]

I’m a big fan of Paul Hales’ Transcendence speaker line from the late 80s. I owned the T5 floor standing speakers, and I still use Hales T1 and TCenter speakers for my home theatre. Absolutely perfect for my purposes.

I remember talking to him when I owned the T5s and mention that they were one of the most beautiful sounding speakers I had heard, but if I could ask for anything more, it would be greater dynamics. He totally agreed with me and apparently that was the stage in which he had already decided to pursue higher dynamics (with low distortion).
The next prototype he showed at an audio show was the Alexandra speaker that had high sensitivity, with excellent dispersion characteristics. Apparently it absolutely wowed show goers. But his company folded shortly after. And then he moved into the pro audio and Home installation field, where he pursued his desire for very high dynamic range and high spl capabilities, along with neutrality and well controlled dispersion (he’s a fan of Floyd Toole). I wish I could hear some of his speakers from his current Theory company.

Anyway…. In terms of what he was saying in that sound stage podcast, I’ve heard him a little bit before on this.

My own experience introducing subwoofers and room correction to my floor standing speakers in my 15’ x 13’ listening room (which also has a large room opening into a hallway), is that I preferred the sound without the subwoofers and room correction. Although the measurements showed that the room response, especially in the bass had been flattened out much better, I found the subs/room correction took a bit of the punch and life out of the sound. Could’ve been of course just my having become accustomed to certain room nodes or whatever. My son felt the same way when he did a comparison.
But at least my experience did somewhat align with Paul’s descriptions. (not that this anecdote establishes he is right)

 

[Redacted]

EQ filters, speakers, rooms and ears can all be represented in mass-spring-damper models. They seem more different than they are.

I had a PhD guy working for me who previously designed car suspension system. He was a good software engineer, as was a guy I had that did astrophysics, but I digress. Point is he first introduced me to "inerters" precisely because when I talked to him about the sums for filter design, he pointed out how the sums don't quite map to real mass/spring/damper systems. There are some YT videos that explain better than I can about inerters.

 

[Few]

Nobody is arguing that crappy rooms sound good. The question is whether using frequency-selective filters to modify the output of speakers can fix low frequency room effects. Hales argues that such filters do more harm than good because they alter the direct sound, which plays a major role in the initial impact of something like a bass drum. He may place more weight on that aspect of loudspeaker performance than someone else does (he plays drums). A pipe organ aficionado may not care as much.

It is widely accepted that speaker corrections at mid and high frequencies should address only the direct sound, and ignore room reflections. One way to reframe the original question is whether the common practice of doing the opposite at low frequencies is well founded in what is known about sound perception. Okay, not exactly the opposite. But some comments here do seem to favor the exact opposite (completely ignore the first cycle or two).

And let’s not lose track of the fact that standing waves are built from counterpropagating traveling waves, so the notion that traveling waves are irrelevant when standing waves are present is at odds with the underlying physics.

 

[kyuu]

Hales argues that such filters do more harm than good because they alter the direct sound

That's irrelevant. The sound you hear is already irrevocably altered by the room mode. Room modes are undesirable equalization, and their corruption of not just the bass but of a significant portion of the audible frequency range (due to masking) is undeniable. They can't be ignored, and they way they are often dealt with without room correction is by having bass deficient systems to avoid putting energy into them (i.e. the "musical" sub/speaker). There are not good acoustic treatment options for them. Even if you turned your room into an anechoic chamber, you still wouldn't be rid of room modes.

Oh, and as far as "you wouldn't room EQ a piano", beyond the already good responses as to why you actually might, one other major point is that playback systems in a domestic space are not instruments. What goes for an instrument does not necessarily apply to a speaker and vice versa.

 

[Few]

Please explain how a room mode affects a sound that has not interacted with the room. Or are you proposing that the sound “knows” the walls are there even though it hasn’t had time to reach them?

 

[kyuu]

The direct sound from the speaker is not perceptually dominant under Schroeder, the room is. No one is proposing any such thing as the room mode affecting the direct sound. The direct sound simply isn't relevant.

 

[Curvature]

Regarding direct sound. The ear has a fusion interval (~time averaging) which increases as frequency decreases. Direct sound as a concept, hearing the first arriving wavefront separately from other reflections, makes sense only above 1.5kHz or so. Below that you hear a gradually widening perceptually-processed average of direct and reflected sound that maps well onto steady-state acoustic response as measured by a microphone. Therefore correcting an acoustic dip or peak at 300Hz means correcting the perceptual defect as well.

 

[amirm]

Here’s a specific case in round numbers: If a reflected sound travels 4 meters farther than the direct version of the same sound, then its arrival is delayed by almost 12 ms, which is the period of an 85 Hz wave. So the first cycle of the 85 Hz component of the sound can’t be influenced by the reflection which hasn’t arrived yet. If the listener places a notch in the response at 85 Hz because the reflection contributes a peak to the response, the first cycle will be notched out even though it could not have been influenced by the room resonance.

You are not hearing a single cycle. After the initial cycle, you have follow ups that mix with the direct sound, creating those standing waves. There will be many such cycles, making the modes clearly pronounced.

Hales argues that the attempt to address the room resonance distorts that first cycle and audibly detracts from the speakers ability to deliver natural impact at low frequencies.

I will have to watch the video but this seems like a layman argument, devoid of reality. You are not listening to the sound waves in slow motion with your eyes! The ears don't work that way. Indeed, try to get a transient response at low frequencies and you find that such a thing is an oxymoron -- it doesn't exist. Bandwidth limited 80 Hz tone cannot show any kind of transient by anyone's layman intuition.

Measure your room and use a PEQ to just pull down a single peak. The difference is not only large in frequency domain (less boominess) but also time domain in how tight the bass becomes. This is simple stuff and a speaker design must know it.