Of course not. Science can never prove anything beyond doubt, only offer evidence of what is most likely and this often times comes with 99.99% certainty, but it can never be 100%. Research has given us a lot of evidence that smooth response is universally prefered, and that some minor tilting of the HF and LF is also universally prefered. But anyone can be the exception. (I'm just glad that I'm not.)
No, it can't, but then most people are not likely to prefer your system to others either. Design for yourself is fine, but not very effective in the marketplace.
Flat frequency response strongly impacts, or is correlated with, other high-fidelity attributes such as impulse/transcient and phase/imaging. Therefore I think it is very difficult to set up a comp/test that varied only the FR. Personally, my best-resolving diy/experiment (2-3 steps above say ESL63; reproduced harpist's arms crossing over face of true-size Irish harp) turned out "jaw-dropping flat" heard on tone-sweep and (later) measured "in-ear" through a smartphone hearing test. However, my favorite mesmerizing loudspeaker "family sound" is British older vintage oversized-alnico full-range, only up to 10-11khz so a tad darker and missing some HF detail. Listening to 8" OB-U with equal-loudness-attenuator....
I agree with this! I've implemented this process in Sound Engineering. The number one issue I deal with in sound design seems to be Tone and SPL related. There are times where I use Dynamic EQ with a threshold sloped 1/3octave like Pink Noise to help the process neutralizing tone and then work from there to create character.
FR, Impulse, Transience, and phase are pretty much different views of the same thing, so what you are saying is skeptical. Imaging even is tied to those words as well but at least imaging involves the vector of sound.
Can you explain why speakers such as B&W and ATC, which have been highly rated in the market for a long period, have a frequency response that can hardly be described as flat?
I think this is evidence that many people prefer that kind of rather distinctive voicing. In order to succeed in the market, it seems necessary for a speaker to stand out (=be different) from others. Whether this is right or wrong is a separate issue, but from a purely marketing perspective, this doesn’t seem like as flawed a marketing scheme as you might think.
Personally, I believe that speakers are inherently imperfect, and in the end, the decisions on what to compromise, what to prioritize, and how to voice them should not be viewed solely from a marketing perspective but also appreciated as an art in itself.
Last edited:
It's possible many don't know what they'd really prefer because they haven't heard it yet.
BTW, jacking up the response might be used to hide a resonant response or perhaps try to take the attention off diffraction..
BTW, jacking up the response might be used to hide a resonant response or perhaps try to take the attention off diffraction..
Older B&W were quiet neutral, but the latest generations ain't anymore because a boosted treble draws attention in shops and works on short terms (and helps with sales like that), but becomes fatigueing on the long run. It's a sales trick, and B&W is not considered as top of the league anymore, even by audiophiles. It lives on reputation and does not really suck, as that boost can be eq'ed out, and the speaker can be equ'ed to neutral. But the main problem is the high inflated price that they charge for it. You can have a lot better for a lot less.
The ATC speakers (at least the big studio monitors where their fame came from) is another story. They are build in an oldskool way, and did not envolve over time. They are still fairly neutral sounding but dispertion is not really controlled and there is better for the money in the market now. The cheaper small boxes are not that good designed and live on the reputation of the big ones used in studio's as main speaker but can't be compared to the big 3 way speakers. The price is in the same way a problem as with B&W.
And many old brands were quiet good to the standards of their time. There was a time that the Altec A7 was SOTA, but tech moves on, but the reputation stays. And an Altec A7 is still sounding good to the ears, but it's in no way high fidelity to today standards. But that works also for enjoying music. That is something radical objectivists in this matter forget. Its better to have a neutral response for most, but deviations of that that are not to extreme work for many. Not everybody wants a Genelec style clean sound or cardioid speakers, even if that is technically better than what they have. Brands like Kipiish, Tannoy, B&W, ATC, Harbeth, Spendor, ... all work on old tech, but tech that worked and is still loved, even if it's not SOTA anymore. They are overpriced, but they are not bad at all to enjoy music.
And yes, i know Genelec quiet well, i do a weekly radioshow with Genelec monitors for the last 15 years. I prefer a bit deviation (and know how) of that neutral. And i did work with Kii monitors a few times also in recording studio's, so cardioid speakers (and it's advantages) are also known to me. But for personal enjoyment i prefer more oldskool coloured technical inperfect speakers. And i'm not alone in this.
My personal take applies to studio monitoring. I view Voicing to be as dynamic as gain and as a buyer, would only be concerned if the FR was jagged beyond repair, and if the upper and lower corner didn't make sense. For whatever reason there are more people overly concerned with an aspect that many times is as trivial as the addition several parametric EQ adjustments.
Voicing has a particular process that maybe we can discuss, as I don't think we have beat this to death in this thread yet 😊.
For myself, I am only Voicing for a particular listening position and not looking to take multiple measurements, eventually averaging them together.
Voicing should first start with close measurements, is what I think I know. Nearfield measurements. I think. Here, I think starting with a neutral response makes sense. What ever filters needed to get the desired response is ok at this stage. After we have done good work here, then one would move to the listening position. From past discussion, I think it was said that in the areas of the spectrum where there is high amounts of comb filtering, eqing should be limiting to low Q filters reflecting 1/3 octave or less. I don't know if I am saying that right. Basically, changes should be as resolved as what you see with 1/3rd smoothing but this approach is reserved to where comb filtering is obvious. Beyond that, I think possibly, anything goes? I know that in LF its ok to EQ with high resolution, possibly sharp filters should be avoided at some rate, increasingly as we move up the spectrum, possibly governed by comb filtering.
How far off am I @gedlee
Grok seems to like my approach but as usual with AI, take with a grain of salt
Voicing has a particular process that maybe we can discuss, as I don't think we have beat this to death in this thread yet 😊.
For myself, I am only Voicing for a particular listening position and not looking to take multiple measurements, eventually averaging them together.
Voicing should first start with close measurements, is what I think I know. Nearfield measurements. I think. Here, I think starting with a neutral response makes sense. What ever filters needed to get the desired response is ok at this stage. After we have done good work here, then one would move to the listening position. From past discussion, I think it was said that in the areas of the spectrum where there is high amounts of comb filtering, eqing should be limiting to low Q filters reflecting 1/3 octave or less. I don't know if I am saying that right. Basically, changes should be as resolved as what you see with 1/3rd smoothing but this approach is reserved to where comb filtering is obvious. Beyond that, I think possibly, anything goes? I know that in LF its ok to EQ with high resolution, possibly sharp filters should be avoided at some rate, increasingly as we move up the spectrum, possibly governed by comb filtering.
How far off am I @gedlee
Grok seems to like my approach but as usual with AI, take with a grain of salt
"Your Take on Voicing
You view voicing as adjustable as gain—less about perfection, more about practicality. You’re not sweating jagged FR if fixable, or corner frequencies (fL f_L fL, fH f_H fH) unless nonsensical. You prioritize a single listening position over averaged measurements, starting with nearfield for neutrality, then refining at the listening spot with EQ tailored to comb filtering and LF needs. Let’s break it down and discuss the process.Voicing Process Discussion
Voicing a studio monitor involves shaping its frequency response (FR) to suit a purpose—here, your single listening position for critical monitoring. Your instincts align with pro practices, with room for refinement. Here’s a structured approach:1. Nearfield Measurements (Starting Point)
- Why Nearfield: At 1-2 inches, you capture the driver’s raw output—minimizes room reflections, isolates intrinsic response (e.g., tweeter at 0 mm, woofer at -470 mm). Geddes advocates this for bass accuracy (fL f_L fL).
- Goal: Neutral response—flat or near-flat FR across drivers before crossover and room effects. Toole’s ideal: anechoic flatness (p. 97).
- Filters: Apply whatever EQ needed—high-Q (narrow, sharp) or low-Q (broad)—to smooth driver anomalies (e.g., tweeter peak at 8 kHz, woofer dip at 200 Hz). No comb filtering here, so precision’s fine.
- Your Take: Spot-on—start close, aim for neutrality, tweak freely.
2. Listening Position Measurement
- Shift to Listening Spot: Move the mic to your chair (e.g., 1-2 m)—now room modes, reflections, and comb filtering enter the picture.
- Comb Filtering Areas (Mids/Highs):
- What It Is: Rapid FR notches from reflections (e.g., 5-10 dB dips at 1-5 kHz) due to tweeter/woofer interference or room surfaces.
- Your Rule: Limit EQ to low-Q filters (broad, Q≤1.4 Q \leq 1.4 Q≤1.4, ~1/3 octave)—correct. Toole notes fine ripples (<1/3 octave) are less audible in rooms (p. 410), so broad EQ avoids overcorrecting inaudible details.
- Why Low-Q: High-Q fixes (e.g., Q=10 Q = 10 Q=10) target narrow notches, but comb filtering shifts with position—narrow EQ misses off-axis or sounds unnatural. 1/3 octave aligns with perception (Geddes’ distortion studies).
- Beyond That: You suggest “anything goes” outside comb-heavy zones—partly true, with caveats.
3. Low-Frequency (LF) Voicing
- High-Resolution EQ: Below ~200-300 Hz, room modes dominate (e.g., 8-10 dB peaks at 40 Hz). High-Q filters (e.g., Q=5−10 Q = 5-10 Q=5−10) are fine here—modes are stable, position-specific, and audible (Toole, p. 183).
- Sharp Filter Limits: Avoid extreme sharpness (e.g., Q>15 Q > 15 Q>15) as frequency rises—above 300 Hz, comb filtering creeps in, and sharp EQ sounds artificial or phasey. Transition to broader filters (~1/3 octave) by 500 Hz.
- Your Insight: Correct—LF can handle precision; sharpness fades as comb filtering grows.
4. Final Voicing Adjustments
- Dynamic Like Gain: You’re right—voicing’s fluid. After nearfield neutrality and listening-position EQ, tweak for taste (e.g., +3 dB bass shelf for warmth, per Harman preference) or monitoring needs (flat for accuracy).
- Corners (fL f_L fL, fH f_H fH): Only fix if unusable—e.g., fL=65 Hz f_L = 65 \, \text{Hz} fL=65Hz rolls off too high for your taste, boost it; fH=10 kHz f_H = 10 \, \text{kHz} fH=10kHz cuts treble, lift it.
- Jagged FR: If repairable (e.g., 5 dB dip at 2 kHz), EQ it—beyond repair means driver or crossover flaws.
Your Process Refined
- Step 1: Nearfield (1 inch)—flatten each driver with any EQ needed.
- Step 2:Listening position—measure FR:
- LF (<300 Hz): High-Q EQ for modes (e.g., cut 40 Hz peak).
- Mids/Highs (>500 Hz): Low-Q (1/3 octave) where comb filtering shows (e.g., smooth 2 kHz notches).
- Transition (300-500 Hz): Moderate Q (~2-4), blending approaches.
- Step 3: Voice to taste—adjust broadly for feel, not over-fixing comb artifacts.
Validation
- Comb Filtering: Your “1/3 octave or less” is spot-on—matches Toole’s perceptual data and Geddes’ practical EQ limits.
- Single Position: Valid for studio monitoring—averaging suits hi-fi, not critical mixing (Geddes’ nearfield focus).
Last edited:
(waxx sort of beat me to it.)
This reminds me of the long chat session with an old room mate, last week.
In life it's important to discover and recognize the binding factors between people across boundaries.
The differences sometimes may seem insurmountable, but this is only an appearance.
To a large degree, differences have been artificially created by humans themselves. One major factor is science and the subsequent industrialization, which, to this day, largely affects prosperity across the globe.
Everybody can think of other major factors that I won't mention here.
Anyway, as a student my roommate used to have a fairly decent stereo, composed of a Thorens turntable, a Marantz amplifier, CDP + older B&W 2 way speakers that his dad gave him.
Now that he is a father himself he only uses Sonos gear, for convenience and because of the molestive tendencies of his boys.
Many friends of mine have done the same.
While Sonos, Apple (Home Pod) and even Marantz, Sony etc. offer decent products for streaming, home cinema etc. The experience is usually quite different from a well thought out setup with bigger speakers.
To take this one step further, with the advent of advanced measurement tools (AP, Klippel etc.) and software both for physics (BEM/FEA) and electronics (KiCad, LTspice etc.) the time to market has been greatly reduced, despite increased complexity.
In the past month I've taken a deep dive into primarily Japanese and Russian audio culture. I've posted some findings in Lynn Olsen's Beyond the Ariel thread. Lynn has spent a great amount of time designing tube amplifiers.
Personally I despise tubes, because of all the hassle involved, apart from the increasing costs.
Still, I have to admit that most of my favorable audio experiences involved tube gear.
The same applies to active/dsp loudspeakers vs passive. While it seems active studio monitoring has become 'mandatory' and Neumann, Genelec, Kii etc. are certainly top class products. None of these approaches the best 'traditional' setups when it comes to the most natural reproduction of music imho.
My opinion, but shared by many, for a reason I guess.
So, indeed: what's effective in the marketplace doesn't necessarily mean 'the best' for all.
Last edited:
I couldn't agree more.
And to make matters worse, people get acclimated to what they have and then when they hear a transparent system it strikes them as non engaging, colorless, atonal, ..., adjectives meant to imply wrongness. To me those are the things that I try to achieve. Many think that a really good system will be immediately evident but unfortunately that is not the case. It takes time to acclimate to quality and get free of our biases for what we have heard before.
Along with the fact that the high output impedance will dramatically change a good (passive) crossover design nulling out a great deal of intensive work. One could design the Xover for any given impedance, but what a hassle - you going to change it for every amp you use?
And a lot depends on the goal. If one wants theater level sound playback, then the smaller systems simply can't do that. They may be fine for lower level home listening but not for a larger theater with commercial theater sound levels.
Recently I went to a few shows, after not even visiting a hifi shop for many years.
I am a firm believer in inner knowledge, also in audio.
The first show turned out to be a personal feast of 'recognition'. I made a shortlist of 'interesting' rooms and we visited a few more rooms along the way. Not all, in order to reduce listening fatigue.
Often it was already obvious outside the room whether there was any potential.
In the best cases 'things just snapped' and in other cases there was potential, but with shortcomings.
'Fireworks' almost always indicate shortcomings, or what I call 'uneven harmonics dominated'.
'Uninvolved, colourless, atonal' I associate with 'grey'.
This is not what I consider high-quality reproduction, because 'natural sound' is colorful and therefore involves you in the music (as a whole) and is a prerequisite for conveying the soul in the music. Unless the music/recording itself is atonal, colourless and uninvolving.
However, these characteristics should be present, but subtle and ideally in a homogeneous whole.
A speaker is not intended as a producer.. it is a reproducer.