1. Home
  2. speakers
  3. 198566-159739

A new measurement specification for speakers?

I am a dedicated low volume listener as my ears have become sensitive with age and past youthful abuse. I have been through a number of speakers with mixed results, trying to find one that sounds good at low volumes. Note that this is not my only criteria for speaker choice, but it ranks high. Frequently it is stated that high sensitivity speakers are good for low volume listening, but I feel there is no connection between the sensitivity rating of a speaker and how it plays at low volumes. 

Speaker sensitivity is typically measured in decibels produced with one watt of input using a 1 kHz tone with a microphone positioned at one meter from the speaker. Roughly useful for determining the kind of amplifier that might be required to drive the speaker, but not an indication of of how it will perform across the audible spectrum with one watt of input. 

For those who have owned or demo'ed a number of speakers, I think we are all aware that each speaker "opens up" at a certain level. By open up, I mean the sound becomes fuller across the spectrum. Good low volume listening speakers obviously open up at lower dB levels. 

What would be a proper means of measuring the level at which a speaker a speaker "opens up"? Technically, in my mind, this would be the dB level at which it delivers roughly linear sound across the audible spectrum, or some reasonable sub-range like 40-10,000 Hz. Call it Spectral Sensitivity. If we send a white noise signal using a select range or spectrum, at what dB level does the sound become linear, i.e. a (roughly) flat line on the graph. 

I am aware of Fletcher-Munson curves, but this does not apply, as we are not talking about human perception, only about the dB level at which the speaker produces the full spectrum of sound presented to it. 

I am not a speaker designer, and there are greater minds in this forum. Does something like this make sense?

zlone
toddalin's avatar
toddalin
659 posts
 

So if the speaker is shown to be "bass heavy" at 1 watt, and this makes it sound "good to the ear" because it compensates for the F-M curve, and it is consistent to its frequency response across its power band, it will be "bass heavy," overpowering to the ear, at loud volumes.

bartsw's avatar
bartsw
233 posts
 

oh ok I see what you mean. I can see many factors likes the room, amp and speakers.

 

unrelated, many if not most speakers and particularly movie content have an intentional peak around 100Hz. for example:

 

https://www.soundstagenetwork.com/index.php?option=com_content&view=article&id=3041:nrc-measurements-arendal-sound-1723-tower-thx-loudspeaker&catid=77:loudspeaker-measurements&Itemid=153

 

https://www.soundstagenetwork.com/index.php?option=com_content&view=article&id=1445:nrc-measurements-focal-sopra-no2-loudspeakers&catid=77:loudspeaker-measurements&Itemid=153

dynacohum
835 posts
 

In an ideal world one would want speakers as uniform in power response as possible, a load invariant amplifier, and a room treated to offer properly placed speakers minimal early reflection. Such a system will “open up” roughly in lockstep with the Equal Loudness curve, aka Fletcher-Munson or ISO.  If you apply corrective equalization in the inverse, you hear the more pleasing balance you seek, in a repeatable scientific manner. Simple in principle but elusive in practice not only because few of our home systems meet the initial criteria but arriving at the corrective EQ curve isn’t a slam dunk either. 

If one is using a preamp with a loudness switch, it is likely to be ineffective because the “inflection point” where the volume and loudness diverge rarely matches the perceived volume in the room with your own speakers. If your amplifier has input gain trims, you might be able to get the loudness in sync with your perceptions, but let’s remember the F-M curves are averages and were arrived at without female test subjects, so YMMV!

Yamaha has a feature on their R-N series called YPAO Volume which uses their room calibration mic to measure the room-speaker response and applies their Variable Loudness with reference to the measured data.  This gets you a lot closer to the promised land, imo. 


Those of us not using the Yamaha hardware can monkey around with various tone controls…I use a Quad 33 and a Schiit Lokius to custom tailor my low volume EQ. 
 

 

jtucker
105 posts
 

What possibly might work would be to plot output level in dB vs input power and look to see if there are any discontinuities in that curve.  Ideally it should be linear.  But I bet it isn't...

kennyc
2,880 posts
 

Wishful thinking.

There is no manufacturer+seller incentive to commit time and resources for both:

  • the complicated task create a standard
  • implementation of the standard

 

More to discover