Flat frequency response


I am often surprised by the number of speakers with "gee-whiz bang" technology but can't even get speaker design 101 right. I can see the benefit of avoiding a lot of signal processing but preferences notwithstanding, flat frequency response seems like the logical place to start and then progress from there.

1) Why is it so hard to achieve?

2) Does it matter?

3) Is it reasonable to say when you skip the basics you are only progressing on a flawed foundation.

cdc

Back when I was an avid amateur speaker builder, I borrowed some test equipment from a technician so I could fine-tune a crossover design, my target being the "holy grail" of flat frequency response. As I got closer and closer to "flat", the sound got worse and worse. Still I persevered, firm in my faith that once I reached "the promised land" of flat measured frequency response, everything would fall into place and the angels would sing.

Well, that didn’t happen. When I finally arrived at "flat" response (plus or minus about .75 dB indicated at the microphone location except for the peaks and dips in the bass region), the sound was harsh and bright and imo unlistenable.

After this highly unsettling experience I started spending as much of my spare time as I could in the library of a nearby university, poring though audio industry journals trying to figure out what had gone wrong and what the solutions might be. Here are the conclusions I eventually arrived at:

1. The in-room response at the listening position dominates perceived tonal balance, and the most natural-sounding in-room response has a gently downward-sloping trend as we go up in frequency (the Harman Curve being an example of this).

2. The spectral discrepancy between the first-arrival sound and the in-room reflection field should be minimized.

3. This means that BOTH the first-arrival sound AND the in-room reflections should have approximately the same gentle downward-slope as we go up in frequency.

Therefore, in my opinion, “getting the basics right” includes getting the first-arrival sound right, and getting the reflection field right.

Duke

speaker manufacturer

@cdc

 

The book you need to read is Sound Reproduction: The Acoustics and Psychoacoustics of Loudspeakers and Rooms by Floyd E. Toole now in its third edition. It will answer a lot of your questions.

I honestly prefer the house curve in my room. This is where the the bass region is above the reference level and trending at a downward slope where the up end of the frequency response(20k) is probably at about 6db/oct. In linear scale, theoretically this will give you a flat response but not in FFT.

One other thing to remember is that having a good frequency response is only the tip of the iceberg. How it sound is another thing.

From what I’ve read, and experiences from my attempts at building speakers, it does seem to be important that the overall power output be smooth, with a downward tilt of some sort toward the treble. It really is difficult to get a speaker to simultaneously have a flat, smooth on-axis response while also having a smooth but downward tilted response off axis at all angles. There’s a ton of stuff that goes on between the drivers and cabinet baffle, port, internal enclosure resonances, etc. interacting with each other that make getting the off axis response smooth at all angles a can of worms. If you can’t solve all of that, then some compromises here and there might not be all that audible, while allowing the speaker to perform better in terms of lower distortion, extended response, more headroom, etc. The ideal dispersion pattern for a speaker is not something that has been settled, and may vary depending on tastes, the characteristics of the room being played in, and perhaps the type of music being played. What does seem to hold true though is that whatever the dispersion pattern, narrow or wide, or somewhere in between, or perhaps various combinations of wide up to a certain frequency before narrowing at various rates, the measured response from any angle off the speaker should ideally be smooth across the frequency band, with no stand out peaks or dips. If there are some peaks and dips, ideally they should be random at different angles, and not the same at all angles. However, if they are the same at all angles and you have access to a good EQ, you may be able to improve that speaker quite a bit with EQ.

I think there’s a limit to how smooth and flat the response needs to be, at least for me. I’ve recently compared speakers that measured better in every respect to what I’m using now. I even think they sounded better as a result, but they weren’t really solving anything that my ears couldn’t easily adjust to. My ear is forgiving of certain flaws, and I may even find some subtle colorations charming, so I was not persuaded to move up. My imperfect speakers are close enough, and I think I prefer their slightly wider dispersion in the upper frequencies, even if it isn’t as smooth and even.

Of the various speakers I’ve built, I’ve found that I really liked a minimum baffle design that used a 3/4" tweeter and a 4" woofer. I never tested the response on these, just set the electronic crossover by ear and found them captivating. My girlfriend did too. I’ve also built narrow dispersion horns that could hold a fairly tight pattern down to 600 Hz. I liked those a lot too, played in the same room as the high dispersion speakers. What I think I don’t like are speakers that are narrow up top but don’t hold pattern low enough. If you can’t hold it down to 600 Hz, then keep it wide as high as possible.

Speakers exist optimally and really in room which are tuned by one owner for his own ears...Otherwise it is only a well measured abstract design because not living yet in a room and  not yet working for some ears...😊