What design parameter(s) enhance low volume listening?

[MikePM]

Although I don't post often or at all, I have been reading this forum for some time and completed my first 6.5" 2-way DIY design about 6 months ago. They sound great at average or higher volumes, but are a little compressed when listened to at low levels. Unfortunately, I rarely have the chance to listen at loud or even average volumes, and would like to know your thoughts on what design decisions go into making a speaker sound good when played at moderate to low levels.

At one time I noticed a thread where Jon Marsh alluded to designing a speaker that sounded good at lower volumes. Unfortunately, I could not find the thread and don't recall it going into detail about the choices he was making to get this result.

A speaker dealer I spoke with thought that designing to achieve a relatively flat impedance curve would achieve the result. I have a sufficiently powerful Rotel amp so I don't believe speaker efficiency is the key. Maybe a three way would do it, or possibly using a sub-woofer?

Any thoughts would be appreciated.

[joecarrow]

For a speaker that's only used at low volumes, I'd recommend equalization that takes your ear's perception of volume as a function of frequency. Check this out:

http://en.wikipedia.org/wiki/Equal_loudness_curve

A lot of old integrated amps and receivers had a button on them labeled, "Loudness". They didn't make them louder, they just added some EQ and made them sound like they were playing louder.

The speaker dealer was full of hot air when he brought up the impedance curve. I really think that you'd benefit from some signal processing. What do you use as your signal source? If you can connect to a computer sound card, you might be able to experiment with this without spending any money- just see if it gives you the effect you're looking for before building anything new. You know they're capable of making the sound you like, now you just need to find a a way to make them do it quietly.

[Gir]

A lot of old integrated amps and receivers had a button on them labeled, "Loudness". They didn't make them louder, they just added some EQ and made them sound like they were playing louder.

So THAT'S what that button did! You have no clue how confusing it was for me as a little kid... : )

[chasw98]

Here is an article that describes the effect joe is talking about and how it relates to modern electronics. It is only one page long and describes what you are hearing.

Chuck

[cjd]

I think it's a lot of things. Certainly, some drivers don't operate as well at low input levels, and our ears are funky as well.

I will say, I often listen to music at ambient SPL levels (56-58dB, maybe up to 60). Yeah, it's missing something on some types of music (big orchestral stuff bothers me the most I think) but that's JUST an SPL issue. Everything is there. Some EQ probably could help tweak that perception and make it less "missing" but...

C

[cameronthorne]

Also, what are you using as a volume control? Could a digital volume control (throwing away bits to make it quieter instead of attenuating the signal) result in a compressed sound at very low volume levels?

[Amphiprion]

I wouldn't really think so, I think it would just increase the noise floor, which would/should still be above the ambient noise in a room by a decent margin. I think loudness controls (or a more advanced form of some such equalization) would be the first thing to try. Someone correct me if I am wrong though.

[cjd]

Also, what are you using as a volume control? Could a digital volume control (throwing away bits to make it quieter instead of attenuating the signal) result in a compressed sound at very low volume levels?

I believe that this would be the case, not absolutely sure. I know if you're using a Squeezebox the way they mess with things you get a little extra range (~80% to 100%) without losing data bits, and there is definitely a change in signal/noise ratio...

There's two sides to this question - one is whether you hear it all, the other is where it sounds "right".

C

[MikePM]

Thanks for the input. What I can say is my 20 year old Allison Acoustics three ways didn't seem to lose as much as the volume was turned down. From this discussion, it may be that using a sub that can fill in the lower octaves as you turn down the volume, would work to make the sound "seem" more normal.

As an alternative, would using a design that moves more air, i.e., larger or multiple woofers, with less excursion, as opposed to relying on a smaller woofer work more consistently as the volume is changed.

What I'm trying to suggest is that drivers have certain characteristics at a particular input level, and that some are more consistent as the volume changes. Is this something that is tested or measured?

[cjd]

It's probably not entirely a simple thing to measure, but looking at various parameters may work. Off the top of my head, Bl is going to play a role - anything else that measures linearity will, and you can certainly also do measurements at varying output levels to see if things change.

I think sensitivity will play a role here, as some drivers may literally take a certain amount of current to get moving adequately. Or maybe it's an amplifier problem at that point. The concept that a system can "wake up" at a certain volume is not unusual though.

What design did you use, if you're willing to share?

C

[cotdt]

For low-volume listening, a smiley-face response curve (U-shaped frequency response) is best, because at low volume our ears are not as sensitive to the frequency extremes as compared to the midrange. But at higher output the smiley-face response curve can become fatiguing fast and would no longer be considered an accurate rendition of the recording.

[Mazeroth]

For low-volume listening, a smiley-face response curve (U-shaped frequency response) is best, because at low volume our ears are not as sensitive to the frequency extremes as compared to the midrange. But at higher output the smiley-face response curve can become fatiguing fast and would no longer be considered an accurate rendition of the recording.

Exactly what I was going to say. Boost the lows and the highs. :T

[MikePM]

What design did you use, if you're willing to share?

I built the Zaph L18, and like I said they are otherwise excellent. They just seem to hide somewhat when the wife or kids are in bed and I have to keep the listening level down. I was a little reluctant to mention the particular design since I didn' want to sound as if I were at all critical.

It is just apparent that my old three way, while less detailed and neutral overall, did, nevertheless sound more open when they were played at lower levels.

[Chris7]

The comments about the ear's equal loudness curve are most important, but there is also the factor of the speaker's efficiency. For some reason (perhaps related in some way to power compression), high efficiency speakers tend to sound more alive at low volume levels. This may be another reason why the L18 design needs higher SPLs to liven up.

[jdybnis]

I've read that most dynamic cone drivers suffer from hysteresis at low levels as a property of the surround material. I have no way of confirming this but it seems plausible. Electrostatics and ribbons are also said to work better at low-levels than dynamic cone drivers. The B&G planar-magnetic drivers might be worth investigating, and Magnepan too.

[Dennis H]

[contrarian mode on]
I've had a bunch of gear with 'loudness' EQ controls. Never used them, hated the sound. SL's latest experiments with rear tweeters on his dipoles imply, when you get the power response right, the volume control acts like a 'distance' control, like you were walking farther away from the band. You don't get a smiley curve when you walk away from the band and you shouldn't use one when you turn down the volume. You just need speakers whose sound replicates the walking away thing when you turn down the volume.
[contrarian mode off]

[joecarrow]

A note on "loudness controls", and why you may not like them-

First, they level of correction is dependent on the actual SPL produced. Some receivers dynamically changed the correction with the position of the volume knob, and some did not. If your speakers' efficiency was far from the efficiency assumed by designers, then it wouldn't scale very well. Also, some did not even really scale with volume- it was either on or off.

Second, the old loudness filters suffered from the same problem that was the reason that for a long time audiophiles were staunchly against external equalizers ("give me a straight wire with gain!" was the battlecry), the electronics of the day simply weren't that good. Opamps were noisy and had problems with distortion and bandwidth. It was a tossup whether any additional components between the signal source and the amp were more of a benefit or detriment. These days it is possible to do your equalization with vanishingly low noise and distortion.

Also, Mr Contrarian, do you really want it to sound like you're walking away from the band? I'm not sure if I do. Anyway, thanks for the counterpoint. This stuff is definitely worth experimenting with, just to know for sure how it suits you.

[Dennis H]

A note on "loudness controls", and why you may not like them-

First, they level of correction is dependent on the actual SPL produced. Some receivers dynamically changed the correction with the position of the volume knob, and some did not. If your speakers' efficiency was far from the efficiency assumed by designers, then it wouldn't scale very well. Also, some did not even really scale with volume- it was either on or off.

Second, the old loudness filters suffered from the same problem that was the reason that for a long time audiophiles were staunchly against external equalizers ("give me a straight wire with gain!" was the battlecry), the electronics of the day simply weren't that good. Opamps were noisy and had problems with distortion and bandwidth. It was a tossup whether any additional components between the signal source and the amp were more of a benefit or detriment. These days it is possible to do your equalization with vanishingly low noise and distortion.

Also, Mr Contrarian, do you really want it to sound like you're walking away from the band? I'm not sure if I do. Anyway, thanks for the counterpoint. This stuff is definitely worth experimenting with, just to know for sure how it suits you.

Nah, nothing to do with noise or bad electronics. I'm not that picky. I just didn't like what the loudness control did to the frequency response. I had an old Yamaha preamp that had the loudness control on a pot rather than a switch so you could dial in however much you wanted. I always left it turned all the way down. It just sounded better with the loudness control off.

[Chris7]

Another problem is that there wasn't consensus for the longest time on what equal loudness curves to use. Some preamps (including Yamaha) switched from Fletcher-Munson to Robinson-Dadson in the 1990s, but now the evidence seems to be pointing back to something similar to Fletcher-Munson being the best approximation. Consumer gear also usually leaves out the middle hump in the equal loudness curve.