[Freo-1]

Found a interesting website that allows one to estimate how much power is required to reproduce music:
http://www.crownaudio.com/elect-pwr-req.htm
 
After runnig some calculations for various listening levels, this brought some insight as to why my 88db/w speakers sound better at normal listening levels with a 150 watt integrated unit as opposed to a preamp/power amp combo where the power amp was 30 wpc Class A (the class A has 6db of headroom).  At low level listening, the class A did have that magic.  However, with moderate listening levels on symphonic music, the extra power makes a big difference.
 
I can see why many audiophiles go for the higher power amps.  For Low Wattage audiophiles, one REALLY NEEDS high efficency speakers. 

[DS-21]

The numbers from the Crown calculator are only valid outdoors, or in an anechoic chamber.

Try it. Open an an SPL meter app on your phone, play a tone, and walk from one of the speakers to your listening position. You will not see anything like the falloff with distance one finds outdoors or in an anechoic chamber.

In fact, if the tone is broadband (pink noise, white noise, etc.) and the system does not employ multiple subwoofers to randomize room mode excitation and minimize mean spatial variance in the modal region, the overall SPL at the listening position may well be louder than at a meter from a speaker. Depends on the room modes.

Still, unless the room is very small and quiet,  30W on 88dB/W/m speakers is a marginal combination, to be sure. 3dBW more (i.e. 60W) would likely be fine even for DSCH 10, though.

[Letitroll98]

Thanks Freo, an interesting toy to play with if nothing else.  But I gotta ask, how loud do you guys listen?  My regular speakers are far from efficient, but the calculator says I need 20 watts for 90db, which is really loud in my room, even for peaks.  Of course for 110 db I'd need 2005 watts, but I kinda have grown attached to being able to hear, so I won't be peaking out at that level anytime soon.

[poseidonsvoice]

http://daedalusaudio.com/images/loudspeaker%20sensitivity.pdf

Best,
Anand.

[opnly bafld]

My main system speakers are @ 88 dB 1w/1m and I use one of two amps, a 15 watt stereo tube or 540 watt mono ss.
The only time I can tell a difference is in demo mode, at my normal listening levels they both sound very good to me.

YMMV,
Lin

[Freo-1]

http://daedalusaudio.com/images/loudspeaker%20sensitivity.pdf

Best,
Anand.

Thanks much for the link!  This makes the point about needing sufficient power even more relevant. 
 
The dynamic peaks when listening to a full on symphony are quite astounding.  Its not so much about how loud one wants the music, its more about how CLEAN of a signal one needs to accurately play back the dynamic peaks (especially with digital sources).

[AJinFLA]

http://www.cordellaudio.com/he2007/show_report.shtml
"The Peak Power Demands of Well-recorded Music"

[opnly bafld]

See post #33 (by Tom Danley) for some interesting comments about power needs:

http://www.whatsbestforum.com/showthread.php?3771-2.4KW-amp-anyone&p=60551#post60551
 

[Rclark]

Found a interesting website that allows one to estimate how much power is required to reproduce music:
http://www.crownaudio.com/elect-pwr-req.htm
 
After runnig some calculations for various listening levels, this brought some insight as to why my 88db/w speakers sound better at normal listening levels with a 150 watt integrated unit as opposed to a preamp/power amp combo where the power amp was 30 wpc Class A (the class A has 6db of headroom).  At low level listening, the class A did have that magic.  However, with moderate listening levels on symphonic music, the extra power makes a big difference.
 
I can see why many audiophiles go for the higher power amps.  For Low Wattage audiophiles, one REALLY NEEDS high efficency speakers.

This

[Freo-1]

Excellent links.  Thanks for those.  Interesting reading.

The more one looks into this, the more higher power amps look like a good option (assuming it still sounds good at lower volumes, not always a easy trick).   

The current Pass Labs amps all employ single ended class A at low power, and shift accordingly as power demands go up.

[poseidonsvoice]

Excellent links.  Thanks for those.  Interesting reading.

The more one looks into this, the more higher power amps look like a good option (assuming it still sounds good at lower volumes, not always a easy trick).   

The current Pass Labs amps all employ single ended class A at low power, and shift accordingly as power demands go up.

I think the time has finally come to have your cake and eat it too, however, the diyer in me always asks why be picky when you have a buffet line? Build an F5, then build a 120 Push Pull OTL, then build an Ncore 400, then a 20 watt SET using GM70's and so on. Some amps you will like for nominal levels, others you can rock out with. Some are incisive and others are buttery smooth. That's what Mike Galusha does and I am sure he has fun. And that's what I do as well. If I didn't have a day job, I would have 1 project per week!

Thanks for starting the thread - there are important points discussed here that are sometimes assumed!

Of course I have desirable advantage - 95 dB sensitive speakers with multiple subs. I can play with pretty much anything...

Best,
Anand.

[DS-21]

http://daedalusaudio.com/images/loudspeaker%20sensitivity.pdf

The part converting Watts to the IMO more appropriate measure of dBW is good, but otherwise that paper makes the same basic mistake the Crown calculator does - assuming that all the sound is the direct field response from one loudspeaker. In small rooms, there is significant SPL contribution from reflected sound.

Though there is a material difference: the Crown calculator is targeted towards people who may actually need to know how much power is needed for a given SPL from a loudspeaker of a given sensitivity outdoors, but here it is simply misapplied to small rooms by someone who doesn't know any better.

The authors of that paper, however, should know better.

Since "the factual basis" (as the paper snottily puts it) of SPL falloff with distance from the source is completely wrong in the context of a small room, Claimed "Scientific facts" (again, to use the paper's rather schmucky language) are "scientific false," because of the authors' blithe disregard for context.

But hopefully nobody's going to take my word for it, and instead people who are properly skeptical of my assertions will do a simple test. Walk from the speaker to the listening position while something's playing, and observe the SPL falloff as you move away from the speakers on an SPL meter.  JL Audio offers a free SPL meter for iPhones that is adequate to this task. Perhaps there is a similar app for other ecosystems. No calibration or anything is required, as the absolute SPL value is unimportant.

[JohnR]

... assuming that all the sound is the direct field response from one loudspeaker.

Well, it does provide a correction factor for two speakers.

[Davey]

The part converting Watts to the IMO more appropriate measure of dBW is good, but otherwise that paper makes the same basic mistake the Crown calculator does - assuming that all the sound is the direct field response from one loudspeaker. In small rooms, there is significant SPL contribution from reflected sound.

Though there is a material difference: the Crown calculator is targeted towards people who may actually need to know how much power is needed for a given SPL from a loudspeaker of a given sensitivity outdoors, but here it is simply misapplied to small rooms by someone who doesn't know any better.

It's not a mistake, it's just a factor not addressed.  Rating speaker sensitivity with a standardized measuring distance and in anechoic conditions is the only way to provide a meaningful "baseline" rating that would allow (reasonable) comparisons between speaker A and speaker B.

Once the speakers are installed in a reverberant environment (listening room) it gets much more complicated, and I don't think it's unreasonable for a speaker manufacturer to not make estimations for that variable.

The 6db SPL falloff with distance is completely right if anechoic conditions are the measuring environment. 

I thought 'everyone' knew that a doubling of power only provides a 3db SPL increase.  So, the dbW "conversion table" doesn't provide any helpful information regarding amplifier requirements.

Cheers,

Dave.

[JLM]

I tried for years to like SET's but gave up due to their lack of a "commanding grip" of any speaker that didn't have significant colorations (seems to me that for the most part colorations increase with efficiency). 

As elluded to above, high output and high fidelity at low output are mutually exclusive (for solid state) but high output is very expensive with tubes.

Based on my mellow taste I've decided that 93 dB/w/m efficiency is the best compromise all around.

[DS-21]

It's not a mistake, it's just a factor not addressed.

Problem is, it's a material and relevant factor, and failing to address it makes the analysis worthless except outdoors or in an anechoic chamber. Which would be OK, but such a glaring error combined with the holier-than-thou tone of the article is really off-putting. 

Rating speaker sensitivity with a standardized measuring distance and in anechoic conditions is the only way to provide a meaningful "baseline" rating that would allow (reasonable) comparisons between speaker A and speaker B.

Agreed, mostly. "Mostly" because I would add "and a standard drive voltage."

But "how to rate loudspeaker sensitivity" is not the question at hand. The question presented in the relevant part of the article is "How do you put [sensitivity, amplifier power, and dynamic range requirements] together?" In the context of sound reproduction in a small room, the article abjectly failed to put those things together in a meaningful manner, because it wished away one of the most important variables in that analysis.

Once the speakers are installed in a reverberant environment (listening room) it gets much more complicated, and I don't think it's unreasonable for a speaker manufacturer to not make estimations for that variable.

Wishing away the existence of the actual environment in which the device is intended to be used is a sort of "estimation" for that variable, no?

If they cannot provide some correction for a very relevant and material variable - and perhaps, to be fair, there is no reasonable "fudge factor" for it - they ought not to make "how much power do they 'need'" recommendations at all! Rather, they should stick to providing sensitivity (anechoic conditions, standard distance and drive voltage) and thermal/mechanical power handling limits of their devices. Data on power compression would be nice, too. The paper in question, interestingly, is utterly silent on the thermal or mechanical power handling of the firm's devices.

I thought 'everyone' knew that a doubling of power only provides a 3db SPL increase.  So, the dbW "conversion table" doesn't provide any helpful information regarding amplifier requirements.

I would hope they do know that. I added that part to my post mostly because I wanted to say something nice about the posted article where it was technically sound rather than so hopelessly flawed as to be misleading, before I critiqued the so hopelessly flawed as to be misleading section.

That said, thinking in terms of dBW rather than Watts is helpful to break free of the various marketing talking points about amps. It helps one avoid the "well, amp A benched at 180W/channel but amp B only benched at 145W/channel, so amp A is a lot more powerful" mistake. Or the, "I can't believe my '500W' sub amp only benched at 480W! What an overrated POS!" mistake. One often sees less educated "audiophiles" taking insignificant differences in wattage (among other things...) far too seriously.

At any rate, again, I hope people don't take my word for it, but rather investigate for themselves what kind of SPL falloff with distance they get in their in their own rooms with their loudspeakers in their current locations. It only requires free tools and a couple minutes.

[a.wayne]

I tried for years to like SET's but gave up due to their lack of a "commanding grip" of any speaker that didn't have significant colorations (seems to me that for the most part colorations increase with efficiency). 

As elluded to above, high output and high fidelity at low output are mutually exclusive (for solid state) but high output is very expensive with tubes.

Based on my mellow taste I've decided that 93 dB/w/m efficiency is the best compromise all around.

I'm in agreement , high efficiency speakers tend to have unbalanced bandwidth and sound (coloration) to get a full balance usually cost you sensitivity unless large speakers with multiple drivers and then other issues .

[a.wayne]

It's not a mistake, it's just a factor not addressed.  Rating speaker sensitivity with a standardized measuring distance and in anechoic conditions is the only way to provide a meaningful "baseline" rating that would allow (reasonable) comparisons between speaker A and speaker B.

Once the speakers are installed in a reverberant environment (listening room) it gets much more complicated, and I don't think it's unreasonable for a speaker manufacturer to not make estimations for that variable.

The 6db SPL falloff with distance is completely right if anechoic conditions are the measuring environment. 

I thought 'everyone' knew that a doubling of power only provides a 3db SPL increase.  So, the dbW "conversion table" doesn't provide any helpful information regarding amplifier requirements.

Cheers,

Dave.

It gets pretty close if you have your speakers are away from the walls, if not the bass will not fall at the same rate , but it will in the mids/highs ...

regards,

[Rclark]

As alluded to above, high output and high fidelity at low output are mutually exclusive (for solid state)

not anymore.  with a bullet.

[avahifi]

I posted this a while back in a different thread, but the info might be useful here too.

The real problem is that the audiophile for the most part has no idea the power is an exponential function as a factor of voltage swing.  Exponential functions are definitely not intuitive.

In any event, if you have two speakers who's input sensitivity is 3 dB apart, the less sensitive one needs double the amplifier power for the same acoustic output.  If the speakers are 6 dB apart, the less sensitive one needs four times the power, 9 dB apart needs 8 times the power, and so on. A 3dB difference in sound levels is pretty much established as a difference anyone can easily hear, but not a large difference. It also requires twice the voltage swing to generate 3dB more output.

Power equals RMS voltage SQUARED divided by load resistance.  However apparent loudness is much better defined by RMS voltage rather than power. Peak V x .707 quantity squared divided by 8 in general.

For example our 200W/Ch Fet Valve 400R will swing 60+V peak, our 300W/Ch Fet Valve 600R will swing 70+V peak (both into an 8 ohm load both channels driven in phase).  That is a 10V voltage difference.  However that is about a 100 Watt power difference.  The apparent sonic loudness capability is much more in line with the voltage output difference (about 10 percent more) not power difference (50 percent more).

Note that measuring power into 4 ohms is a cheap way of doubling your power ratings (do the math) as long as the amplifier under question can supply the necessary twice the current needed for the low impedance load.

As an aside, in the real world, "gas milage" is also a non intuitive based exponential number.  Your car's gas economy is much better served by considering "gallons per one hundred miles" than gas mileage.  How important is it to improve your gas mileage by 10 miles per gallon?  Not much if you are going from 40 MPG to 50 MPG, you are only saving one-half gallon per one hundred miles.  But from 10 MPG to 20 MPG, that saves 5 gallons per hundred miles.  Do the math!  Makes me wonder why truckers drive so fast when they would save hundred of dollars with of gas on every long distance trip if they just stuck to 50 mph or so and the reduced drag with save them more than they would realize.  Again, do the math, its just simple algebra.

Frank Van Alstine

PS Yes I know that watts by themselves are not an end-all measurement, the quality of these watts (or preferred, volts) are probably even more important.  But at least the data above should help clear your head a bit.