2. speakers
  3. 1278114262

The pursuit of bass...

Most people would probably say that the most important thing for a speaker to get right is the midrange and I'd have to go along with this myself. If the midrange is too shouty or too recessed or just tonally inaccurate, the other qualities of the speakers are pretty much pointless regardless of how great those qualities might be.

However, with that said, I do not think that the midrange is the most challenging part of a speaker to get right. In my opinion, that award goes to the bass frequencies. Sometimes I listen to a speaker and it is indeed the midrange or treble that is holding back the sound quality, but far more often, for me, it's the low frequencies. I find that I’m more forgiving of midrange flaws. Midrange flaws don’t impact my musical enjoyment as much as bass flaws.

Now, of course, the bass is also the most vulnerable to room acoustics which just exacerbates the problem. But, mostly I blame the speakers themselves. Many fine speaker manufacturers simply ignore the problem by rolling off the bass early. I won't name any specific brands to avoid a flame war, but this is very common. Often I will see specs for frequency response that indicates -6db at 50hz. This is typically very unsatisfying bass. Also, many speakers are not balanced properly across the frequency spectrum so that while they might dig down to 35hz or 40hz, they don't sound like it because the bass is always underwhelming compared to the volume level of the midrange and high frequencies.

The speakers that do attempt to reproduce good solid 40hz bass, often still sound quite bad when reproducing those frequencies. And I'm not talking about pipe organ bass here, I'm talking about the 35hz to 55hz range. Unfortunately, really good minimontors that are amazing with 60hz up, really are missing out on a lot of the presence and atmosphere created by those low frequencies. Emotionally connecting to the music and suspending disbelief is easier with quality bass reproduction. For some reason many audiophiles are willing to live without it. I can understand this since pursing quality bass can be a frustrating endeavor. Also there is cost to consider. Good bass typically comes from bigger speakers and is therefore considerably more costly. But even the best mini-monitors usually commit the sin of omission in the low frequencies.
jaxwired
audiokinesis
2,750 posts
 
Putting my designer's hat on, bass is largely a matter of juggling tradeoffs. It's not that difficult to get good bass if you start out with a big enough box. If box size is severely constrained, we must make some hard choices between bass extension and efficiency (and it's not quite so simple as it looks at first glance).

On the other hand, the midrange and lower treble are much more difficult to get right. That's where I may spend hundreds of hours during the design phase.

A set of equal-loudness curves will graphically illustrate what the ear's sensitivity is at different frequencies across the spectrum. This is a good roadmap for telling us where the ear is especially critical and where it is relatively forgiving.

equal loudness curves

There is a great deal of useful information in these curves.

Now the curves still don't tell us how much each portion the spectrum contributes to overall subjective enjoyment. In my opinion bass in particular needs to be balanced relative to the rest of the spectrum. This means different things for different speakers, and the behavior we want from a given speaker may change depending on the room acoustic situation.

In my opinion if one places high priority on really good bass reproduction, the mini-monitor format is too inherently compromised to really deliver. Make the aesthetic tradeoff to a floorstander, taking into account your savings on speaker stands, and you will be well ahead of the game. Or add a good subwoofer system to those mini-monitors.

Duke
dealer/manufacturer
plato
1,543 posts
 
Hi Duke,

Thanks for posting the "Equal Loudness Curves". This must give the speaker designer something to think about in that if you build a speaker with measured flat response between 3kz and 4kHz it will be perceived as "bright" by the average person.

I would also say that because of anticipated room gain you would not need as much boost in the low bass as the curves would imply but correct me if I'm wrong on that.

So what this means is that a speaker that measures perfectly flat will sound bright to the average person. Also anemic in the bass if there is inadequate room gain.

This makes sense to me as I use an EQ in my large-room system to boost low bass slightly, cut the response a dB or two at 2k, and 3 - 4dB at 4k, and I add couple dB's of boost at 16kHz.
tim916
52 posts
 
Building full-range speakers is pretty much a waste of woofers and wood due to the realities of small room acoustics. A speaker should be designed to play down to 70-80hz with little distortion at high volumes. The equalized subwoofer(s) take over below this point. Even massive speakers like the Wilson X-2 and Rockport Arrakis will benefit from this configuration.
audiokinesis
2,750 posts
 
Plato, the equal loudness curves illustrate that the ear itself is not "flat"; I wouldn't consider them to be any sort of "target curve" for loudspeaker frequency response, but they do show where a designer needs to be careful; in my opinion you correctly identified 3-4 kHz as a particularly critical region. A dip in this region does make a speaker more forgiving, but it also results in the overtones of many instruments being under-represented, which detracts from a realistic presentation.

As to why a "flat" speaker usually sounds bright, I think that has to do with microphone placement during the recording process and the acoustic differences between our listening rooms and concert halls. Some of it may have to do with loudspeaker radiation patterns as well (many speakers have an off-axis flare in the 3-4 kHz region because the tweeter's pattern is quite wide there).

Tim916, the effects of small room acoustics in the bass region is indeed a significant issue. Since the room-interaction peak-and-dip pattern can change dramatically with a small change in listener location, the improvements from equalization are usually limited to a small area, and the response may actually be a lot worse in other locations because of applied EQ. Of course, this may not matter much if listening is confined to the sweet spot. Now I noticed you said "subwoofer(s)", leaving the door open for more than one. A distributed multisub system will reduce the amplitude variance (smoothe the frequency response) as well as the spatial variance (less change in frequency response from one location to another). As a result, not only is equalization less likely to be needed, but if so it is more likely to be an improvement over a wide area.

It is possible to spread out the in-room bass sources somewhat even in a conventional pair of speaker boxes (no sub), by using multiple woofers spread apart or locating the port in a different plane from the woofer. The improvement is not as dramatic as with scattered multiple subs, but it is still worthwhile in my opinion.

Duke
dealer/manufacturer
jax2
3,031 posts
 
Hey Duke - Can one correct (or compensate to any degree) for a room-generated suckout? My room has a consistent suckout with every speaker I've measured, at around 80hz , from the listening position. I understand that a low-end suckout is very common with most rooms and varies in terms of where it occurs within the range - correct me if I'm off base. Alas, I must listen nearfield (7-8 feet ears to speakers) so am locked into a rather narrow latitude of movement of both listener and speakers. There is some, but it is rather small. I'm currently using some treatments but wondered what, if anything, might be most effective with dealing with a room suckout like that? I'm guessing it may be limited to changing the listener position or the structure of the room itself. Current treatments include bass traps, diffusion at first reflection, and absorption on back wall, and between the speakers.

A brief follow-up question - I've been using pink noise to gauge such things. What is the difference between pink noise and white noise (noticed the recommendation for white noise in the Geddes reference)?
More to discover