Speakers driven at low volume vs upping things a bit

[brucew268]

I've noticed for some time that my current speakers tend to not come alive until I've raised the volume to a certain level (suitable to critical/dedicated listening, not to parties). Some have said that issues is true with most speaker/amp combos. Someone else said that is because my 1SC's have a fluid damped tweeter, giving them a bit of inertia to overcome. I also notice that the Stereophile measurements show a fair bit of the midrange in the 5-6 Ohm range, making them a harder load for the amp to drive than say something that is mostly 8 Ohm in the mids but with a shorter dip slightly below.

Any personal experience in identifying patterns for a speaker that is good both at low level detail and a sense of mid-range 'life' at lower volumes?

[Conan]

Usher S520 very similar design to the Proac's. Sound excelent at low volume.

As for pattern, definitely as speaker with an easier load tends to sound livelier.

[awkwardbydesign]

Looking at it from a different angle, I found that going from an active preamp to a TVC improved listening at low levels. Lower noise floor, I believe.

[brucew268]

Richard, I appreciate the benefits of lowering the noise floor and the low level detail that comes through, but it I think I'm talking about something different. Perhaps it was a red herring for me to mention low level detail in the OP. The change in the mids coming to life is as sudden as when playing with speaker position and at one point the stage just snaps into focus. Increasing the volume at some point the voices and other mids just 'snap' into being more real, more energetic, and more dimensional. If it was a noise floor issue, I would think the change would be rather gradual, not to mention that I've done a bit of work on the noise floor, to good effect, but which did not relate to this behaviour. Someone also once mentioned that the phenomena would be once the volume hits a level that it 'charges' the room acoustically. But it would be nice to get closer to identifying whether it is room interaction, speaker impedance, or speaker damping inertia, or something else.

[awkwardbydesign]

Ah, OK. I find that there is usually a volume level for each piece of music, where it comes to life and feels right. Too much is as bad is too little, but every system and every room will have a "right" level. Usually.
Is that closer to what you are talking about?

[brucew268]

Yes. What I'm wondering is if that level can change based on the speaker... Or other?

[walpurgis]

For some reason, proper horn speakers seem to work well at low levels. Maybe something to do with how they acoustically 'couple' with the room.

[awkwardbydesign]

Yes. What I'm wondering is if that level can change based on the speaker... Or other?
Short answer- yes.
Longer answer- the room acoustics will have as big, or even bigger, effect. I have applied a lot of bass trapping and damping to my room, and I can play louder without overloading the room. Which helps with those pieces of music that need to be played louder.

[jaym481]

For some reason, proper horn speakers seem to work well at low levels. Maybe something to do with how they acoustically 'couple' with the room.

I think the bold is the key, regardless of speaker. I've heard the expression "load the room" for speakers, and reckon that's where the magic is. Also, as mentioned above, overloading can be a problem with boomy bass being a common symptom. Treatment of the room should, I think, help both louder and quieter playback.

[SLS]

I canted my speakers recently because of this exact problem. Had Fact8's that don't seem to do quiet at all. Harbeth SHL5+ do.
Doesn't mean Fact8 aren't a very good speaker.

[Barry]

Ah, OK. I find that there is usually a volume level for each piece of music, where it comes to life and feels right. Too much is as bad is too little, but every system and every room will have a "right" level. Usually.
Is that closer to what you are talking about?

Agreed - and it will depend on the system and the room.

In playing horn speakers too loudly, you just hear their colourations more easily.

[synsei]

From my own experience B&W DM2's are incredibly versatile in this respect, sounding fabulously open and transparent at any volume level and they also have the added advantage that they aren't at all fussy about the room they are housed in.

[brucew268]

Long term follow up: I found that part of my problem & solution lay in room treatments. When I was getting the midrange up to a volume that would couple well with the room and give the sense or presence that I desire, the overall sound was a bit too powerful especially off-axis (others using or entering the room). So I put bass traps in three corners of the room and found that helped a fair bit. Interestingly, the choice of traps and their angle on the corner plays a significant role in the detail and soundstage (MF & HF) for good or for ill.

[Reffc]

I'm surprised that no-one has picked up on a few issues related to this topic:

1. F-M: have a look at some F-M curves and you'll see that at low volumes, our hearing isn't very sensitive to bass or treble, hence the "loudness" buttons on older amps. The ear is most sensitive in the 3-4K region due to the resonance of the auditory canal in that frequency band but above and below that, it is less sensitive until volumes are increased

2. Speaker voicing: We've all read about "linear" or "flat" response on axis, right? Well, what you probably wont be told, except by a small handful of manufacturers is at what dBA or dBC SPL that linear curve is tested at. Some speakers are designed for a linear response at fairly high SPLs, many studio monitors included, but this doesn't always make them especially suited to domestic hifi duties if you have to listen at 90dB to get that "correct" balance! Speakers, such as Harbeth, are voiced for a linear response much lower (75 to 80dBA). So how does this work if F-M curves show a loss in sensitivity as volume drops? Simple...the speaker's response is altered to account for F-M effects and you can create a linear response at virtually any fixed level but the trade off is the lower you go with this, the more "Smiley Face" the tuning gets with volume (ie bass and HF are boosted with increased volume) and these designs can sound rather fatiguing with extended listening. For domestic listening, linear response between 75 and 80dB seems about optimal (which is how I've tuned my own designs).

3. Efficiency: Some designs simply take more grunt to get them going but this doesn't directly relate to sensitivity which is measured differently. Sensitivity can vary with frequency, further complicating matters. This is down to both drive unit and cabinet design. Ideally, a loudspeaker should have equal sensitivity across the audible frequency band. The ideal loudspeaker does not exist, so most drive units have a varying sensitivity with frequency, which again is where the crossover comes in. it does not exist just to cross the drive units at a specified frequency and to ensure optimum phase response (many don't achieve that anyway) but also should ensure that a linear or tailored response is achieved. Poorly designed or worse still, without it (unless the drive units are specially designed and made for a linear response with smooth break up), response can vary quite widely and wildly!

I find that I prefer the characteristics of most large, sensitive speakers which tend to sound (to me) more natural and balanced at lower volumes, F-M effects withstanding.

[The Black Adder]

Agreed - and it will depend on the system and the room.

In playing horn speakers too loudly, you just hear their colourations more easily.

+1

[walpurgis]

As this thread has been dragged from the brink of the grave (:D), I'll just say that I've found horn speakers more involving at low volumes than other types (although the LS3/5a does quite well). I think it's probably due to quick transient response, but there may be other reasons too.

[petrat]

Excellent post, Paul :clap:

[spendorman]

Probably the best low listening level speakers that I have are the Quad ESL57's. Might be something to do with the very lightweight diaphragms.

[brucew268]

Interesting to note that most of these discussions tend to focus on the speaker characteristics, when the room can impart as much to the sound characteristics: frequency curve, stage width and depth, perceived tone, perceived dynamics. By use of different acoustic densities and different placement, the perceived voicing can change siginficantly, which may alter what is assumed to be an issue with speakers or other components.

[Arkless Electronics]

I'm surprised that no-one has picked up on a few issues related to this topic:

1. F-M: have a look at some F-M curves and you'll see that at low volumes, our hearing isn't very sensitive to bass or treble, hence the "loudness" buttons on older amps. The ear is most sensitive in the 3-4K region due to the resonance of the auditory canal in that frequency band but above and below that, it is less sensitive until volumes are increased

2. Speaker voicing: We've all read about "linear" or "flat" response on axis, right? Well, what you probably wont be told, except by a small handful of manufacturers is at what dBA or dBC SPL that linear curve is tested at. Some speakers are designed for a linear response at fairly high SPLs, many studio monitors included, but this doesn't always make them especially suited to domestic hifi duties if you have to listen at 90dB to get that "correct" balance! Speakers, such as Harbeth, are voiced for a linear response much lower (75 to 80dBA). So how does this work if F-M curves show a loss in sensitivity as volume drops? Simple...the speaker's response is altered to account for F-M effects and you can create a linear response at virtually any fixed level but the trade off is the lower you go with this, the more "Smiley Face" the tuning gets with volume (ie bass and HF are boosted with increased volume) and these designs can sound rather fatiguing with extended listening. For domestic listening, linear response between 75 and 80dB seems about optimal (which is how I've tuned my own designs).

3. Efficiency: Some designs simply take more grunt to get them going but this doesn't directly relate to sensitivity which is measured differently. Sensitivity can vary with frequency, further complicating matters. This is down to both drive unit and cabinet design. Ideally, a loudspeaker should have equal sensitivity across the audible frequency band. The ideal loudspeaker does not exist, so most drive units have a varying sensitivity with frequency, which again is where the crossover comes in. it does not exist just to cross the drive units at a specified frequency and to ensure optimum phase response (many don't achieve that anyway) but also should ensure that a linear or tailored response is achieved. Poorly designed or worse still, without it (unless the drive units are specially designed and made for a linear response with smooth break up), response can vary quite widely and wildly!

I find that I prefer the characteristics of most large, sensitive speakers which tend to sound (to me) more natural and balanced at lower volumes, F-M effects withstanding.

I was about to make the F-M comment myself but the other aspects you mention are most interesting!

[Reffc]

Just out of interest Jez, is it possible to make an amp with an auto-F-M compensation bias controlled by output voltage? I guess this would be difficult as it pre-supposes connected load and sensitivity are known which they wouldn't be, hence loudness buttons to do it manually.

[Macca]

F-M Compensation device:

http://i903.photobucket.com/albums/ac233/Macca_photos_2009/equalize001.jpg (http://s903.photobucket.com/user/Macca_photos_2009/media/equalize001.jpg.html)