Power requirements - does more power matter that much?

sbo6

VIP/Donor
May 18, 2014
1,661
594
480
Round Rock, TX
I spent a bit of time running the numbers to see about how much amplifier power matters as part of a FB post. I thought some folks might find it interesting and may be able to correct me if I'm incorrect in my quick assessment. I appreciate any feedback and comments.


Here are my thoughts on power and what I call speaker efficiency falsehood:

Speaker efficiency is commonly measured by applying 2.83V (assumes 1 W with an 8 ohm load) @ 1KHz @1 meter measured on axis. The current works out to be .35 Amps btw.
For a speaker with a nominal 8 ohm impedance and an efficiency rating of 88DB:
- +3DB requires 2x power, +10DB = 10x power. So let’s say we want to achieve an average (not peak) of 94DB (pretty loud). You would need 4x the power = 4 watts. For 104DB you’d need 10x = 40 watts. And for 107DB you’d need 2x = 80 watts.

- Now take into account the sweet spot distance to the speakers. Using the SPL to DB calculation and using our 88DB efficient speaker (measured @ 1 meter) let’s say our listening chair is 4 meters away = -12DB reduction. So now 80 watts yields 95DB (107DB – 12 DB for distance).

- There are other variables like efficiency gain via 2 speakers = +3DB depending on proximity, decrease in efficiency via room treatment (especially absorption) , speaker phase swings and voice coil impedance changes over high power / high temps, and other factors.

However here’s the biggest flaw I see in speaker efficiency and required power estimates: When is the last time you played a song that is a 1KHz tone? Music is incredibly complex spanning ~20 – ~20KHz simultaneously and often in bursts with dynamic swings. So all the calculations and resulting power estimates above are mostly meaningless because they are predicated on the signal equating to one frequency (and a fairly high frequency at that). Woofers are notorious for drawing significant current and are primarily excluded at 1KHz, so IMO you need much more than 80 watts to achieve 95DB @ 4 meters. Based on impedance and phase graphs for many speakers I’d take a swag and say you need 2x the average power to avoid compression, clipping, etc. For peaks I'm sure significantly more is required. And that, IMO is why power matters.
 

DonH50

Member Sponsor & WBF Technical Expert
Jun 22, 2010
3,947
306
1,670
Monument, CO
Peak-to-average in music was measured ages ago at around 17 dB, a power factor of 50, so to hear a symphony from ppp to fff you may need a lot more power. Movies have been documented with ~30 dB.

IME people greatly overestimate their average power needs and underestimate their peak power needs.
 

sbo6

VIP/Donor
May 18, 2014
1,661
594
480
Round Rock, TX
Ultimately it very much depends on speaker efficiency and impedance, room size and your expectations in terms of loudness. It all matters. Also, if as Don says, Peak-to-average in music was measured ages ago at around 17 dB, by my calculations above that equates to ~40x (I know Don has it at PF = 50). So for the scenario above

- If 10W is required for 86DB average power then for a 17 DB average to peak (86DB -> 103DB) ~400 watts is required.
- If 20W is required for 89DB average power then for a 17 DB average to peak (89DB -> 106DB) ~800 watts is required.

That's a lot of power for a fairly typical speaker type and room size.
 

Ron Resnick

Site Co-Owner, Administrator
Jan 24, 2015
16,017
13,347
2,665
Beverly Hills, CA
. . . people greatly overestimate their average power needs and underestimate their peak power needs.

I think this is a technical explanation in support of the subjective view held by me, MikeL, Jeffrey_T, JimFord, PeterA and many others that (if sonic quality remains constant as power rises) seemingly excess or "unnecessary" power or "headroom" is good because it allows for the sound to remain effortless on dynamic peaks.

Once a speaker is sufficiently insensitive that moderate power is required (say, 50 to 100 watts; once the power requirement exceeds SET or PSET output levels) then, I, personally, believe that throwing a lot of power at it (say, 250 watts, 400 watts) is a good thing. (I fully appreciate that current matters as much as watts, depending on the impedance nature of the loudspeaker load.)
 
  • Like
Reactions: sbo6

microstrip

VIP/Donor
May 30, 2010
20,806
4,698
2,790
Portugal
Power is the product of voltage times current. When we are addressing just the power we are looking at a very incomplete picture of the electrical situation. Some modern amplifiers such as the DartZeel NHB458 have displays that compute the instantaneous power in real time - people would be very surprised looking at the very different indications we get at similar sound levels with different speakers and different types of music.

BTW the power meter in most (almost all ) amplifiers is just for cosmetic purposes and can be very misleading.
 
  • Like
Reactions: sbo6 and Alrainbow

Alrainbow

Well-Known Member
Dec 11, 2013
3,189
1,387
450
From many systems I have heard it’s the opposite lol. more has no inherent flaws i think. And I don’t see any info on instantaneous peaks here. this is many times the peak. And add on recovery time of a given amp.
 

sbo6

VIP/Donor
May 18, 2014
1,661
594
480
Round Rock, TX
I think you can consider WRT peak over average as instantaneous or longer durations.
 

DonH50

Member Sponsor & WBF Technical Expert
Jun 22, 2010
3,947
306
1,670
Monument, CO
17 dBW = 10log10(Wratio) ==> Wratio = 10^(17/10) = 50.1187... So not exactly 50 times but pretty close...

Using the above example, with no room gain and a pair of 86 dB/W/m speakers 8' from the listener, 1 W yields about 81.3 dB. That's pretty loud to me. If you have a 100 W amp you can achieve about 101.3 dB (20 dB increase), again pretty durn loud. 50 W/speaker gets you about 98.3 dB. Drop to 83 dB/W/m speaker and you need twice the power. Move twice the distance away (16') and you lose 6 dB, meaning you need four times the power to make up the difference. You can drive yourself crazy figuring out how much power you need. For me, it's a short drive. :)

Online calculator: http://myhometheater.homestead.com/splcalculator.html -- All of these have various issues, but if you want a rough ballpark figure it at least gives you an idea, and allows you to play some "what-if" games.

For full-range speakers without subs, you have to consider equal loudness as well. It can take 20 dB or more for deep, deep bass to sound (feel) as loud as midrange frequencies. 10 W may be all you ever need in the midrange and yet need 1000 W in the bass.

https://en.wikipedia.org/wiki/Equal-loudness_contour



HTH - Don
 
Last edited:

BlueFox

Member Sponsor
Nov 8, 2013
1,709
406
405
Being retired and living on Social Security and my savings, I have to say ‘Power is like money. The more, the better’.

Of course it can’t be counterfeit, but good, and clean power.
 

Alrainbow

Well-Known Member
Dec 11, 2013
3,189
1,387
450
I spent a bit of time running the numbers to see about how much amplifier power matters as part of a FB post. I thought some folks might find it interesting and may be able to correct me if I'm incorrect in my quick assessment. I appreciate any feedback and comments.


Here are my thoughts on power and what I call speaker efficiency falsehood:

Speaker efficiency is commonly measured by applying 2.83V (assumes 1 W with an 8 ohm load) @ 1KHz @1 meter measured on axis. The current works out to be .35 Amps btw.
For a speaker with a nominal 8 ohm impedance and an efficiency rating of 88DB:
- +3DB requires 2x power, +10DB = 10x power. So let’s say we want to achieve an average (not peak) of 94DB (pretty loud). You would need 4x the power = 4 watts. For 104DB you’d need 10x = 40 watts. And for 107DB you’d need 2x = 80 watts.

- Now take into account the sweet spot distance to the speakers. Using the SPL to DB calculation and using our 88DB efficient speaker (measured @ 1 meter) let’s say our listening chair is 4 meters away = -12DB reduction. So now 80 watts yields 95DB (107DB – 12 DB for distance).

- There are other variables like efficiency gain via 2 speakers = +3DB depending on proximity, decrease in efficiency via room treatment (especially absorption) , speaker phase swings and voice coil impedance changes over high power / high temps, and other factors.

However here’s the biggest flaw I see in speaker efficiency and required power estimates: When is the last time you played a song that is a 1KHz tone? Music is incredibly complex spanning ~20 – ~20KHz simultaneously and often in bursts with dynamic swings. So all the calculations and resulting power estimates above are mostly meaningless because they are predicated on the signal equating to one frequency (and a fairly high frequency at that). Woofers are notorious for drawing significant current and are primarily excluded at 1KHz, so IMO you need much more than 80 watts to achieve 95DB @ 4 meters. Based on impedance and phase graphs for many speakers I’d take a swag and say you need 2x the average power to avoid compression, clipping, etc. For peaks I'm sure significantly more is required. And that, IMO is why power matters.


Btw

Its sensitivity your discussing not efficiency and i read an amplifier handbook once that stated typically what determines power necessary is your listening level in relation to 1 watt of usage . In one test with 1 watt being used at 89 db levels , 200watts plus was necessary for unclipped dynamics ..

So 100 watt if that level is 86 db on said speakers for eg..
 

Mark Seaton

WBF Technical Expert (Speaker & Acoustics)
May 21, 2010
381
141
390
47
Chicago, IL
www.seatonsound.net
When we see specific numbers on a spec sheet, it's only natural to want to believe they mean the same thing on one spec sheet to another. The reality is that those numbers are but one snapshot of how that amplifier behaves in real operation. Demands from different speakers and to a lesser degree, different music, can represent dramatically different experiences of subjective power and useful loudness in a room on real content.

We rarely get a good look at what an amplifier does getting to that power level and especially how it runs out of gas beyond that power level. Some designers give huge Voltage headroom above the continuous power, some have huge current reserves, and most amplifiers have rather different overload behavior and sounds, and within those differences we again hear differences in what causes the overload into different speakers. Many amplifiers will sound quite different when pushed too hard into a speaker with a 2 Ohm dip in the midbass, vs a higher impedance load that barely dips below 8 Ohms.

For most amplifiers there are 3-4 dimensions of variations in distortion vs level, and even then the resulting distortion can be quite different.
 

Robh3606

Well-Known Member
Aug 24, 2010
1,471
463
1,155
Destiny
However here’s the biggest flaw I see in speaker efficiency and required power estimates: When is the last time you played a song that is a 1KHz tone? Music is incredibly complex spanning ~20 – ~20KHz simultaneously and often in bursts with dynamic swings. So all the calculations and resulting power estimates above are mostly meaningless because they are predicated on the signal equating to one frequency (and a fairly high frequency at that). Woofers are notorious for drawing significant current and are primarily excluded at 1KHz, so IMO you need much more than 80 watts to achieve 95DB @ 4 meters. Based on impedance and phase graphs for many speakers I’d take a swag and say you need 2x the average power to avoid compression, clipping, etc. For peaks I'm sure significantly more is required. And that, IMO is why power matters.

Hello sbo6

I think are looking at this from the wrong direction and making it much harder than it needs to be. First of all your statement saying all speaker sensitivity is based on a 1 K tone is simply incorrect. Even if a 1K tone was used if the system was full range and +/_ 3db 20-20K of your tone the calculations would essentially be correct.

The easiest way to determine amp power needed is to know what your max SPL requirement is at your listening position and go from there. If you have enough power to meet your Max target SPL you will be fine. All passive speakers are set up typically using the woofer sensitivity as the reference level for sensitivity. You are not likely going to need much more power at 100Hz vs 1K to reach the same SPL.

As far as speakers as loads considering impedance and phase angles again you just pick an amp that can safely drive them.

I do not think running calculations or using available on line calculators is a meaningless exercise. They are actually a handy tool to do a quick snapshot and let you know how much power you need.

Here play with this:

http://myhometheater.homestead.com/splcalculator.html#anchor_13193

Rob :)
 

Alrainbow

Well-Known Member
Dec 11, 2013
3,189
1,387
450
As explained to me , All sensitivity measurements are done at center frequency which is 1K by any measurement i have ever seen not bass .. Also you need more RMS power for bass , more peak power for mid and highs , this peak power is what pushes amps into clipping much more than bass with high crest factor recordings ..

Exception being servo bass which is why Arnie had 3K watts per bass tower in the IRS V
 

Folsom

VIP/Donor
Oct 25, 2015
6,024
1,490
520
Eastern WA
Ultimately it very much depends on speaker efficiency and impedance, room size and your expectations in terms of loudness. It all matters. Also, if as Don says, Peak-to-average in music was measured ages ago at around 17 dB, by my calculations above that equates to ~40x (I know Don has it at PF = 50). So for the scenario above

- If 10W is required for 86DB average power then for a 17 DB average to peak (86DB -> 103DB) ~400 watts is required.
- If 20W is required for 89DB average power then for a 17 DB average to peak (89DB -> 106DB) ~800 watts is required.

That's a lot of power for a fairly typical speaker type and room size.

10w for 86db means the speaker has a sensitivity of about 77db... who uses anything like? And btw it'd be 640w to get to 103db.

A 20w for 89DB is the same as a 10w for 86db.
 

andromedaaudio

VIP/Donor
Jan 23, 2011
8,355
2,731
1,400
Amsterdam holland
Nominal impedance means shit basically .
Every speaker manufacturer should be obliged to deliver an impedance related to freq..graph with the speaker .

Regarding efficiency it should be standarized to draw a medium line through the speakers average response and take the eff . from there .
As 1 khz as a measurimg point is taken and the speaker has a bump there then you dont get an honest nunber .
Some speakers have a + - 3 db response , so go figure
 
Last edited:

Folsom

VIP/Donor
Oct 25, 2015
6,024
1,490
520
Eastern WA
I think this is a technical explanation in support of the subjective view held by me, MikeL, Jeffrey_T, JimFord, PeterA and many others that (if sonic quality remains constant as power rises) seemingly excess or "unnecessary" power or "headroom" is good because it allows for the sound to remain effortless on dynamic peaks.

Once a speaker is sufficiently insensitive that moderate power is required (say, 50 to 100 watts; once the power requirement exceeds SET or PSET output levels) then, I, personally, believe that throwing a lot of power at it (say, 250 watts, 400 watts) is a good thing. (I fully appreciate that current matters as much as watts, depending on the impedance nature of the loudspeaker load.)

V*I = P

Noticed the

=P

at the end.
 

Alrainbow

Well-Known Member
Dec 11, 2013
3,189
1,387
450
Nominal impedance means shit basically .
Every speaker manufacturer should be obliged to deliver an impedance related to freq..graph with the speaker .

Regarding efficiency it should be standarized to draw a medium line through the speakers average response and take the eff . from there .
As 1 khz as a measurimg point is taken and the speaker has a bump there then you dont get an honest nunber .
Some speakers have a + - 3 db response , so go figure

1k is center frequency , you are +/- in relation to 1K. i dont think efficiency is measured there thou only Sensitivity , maybe Duke can chime in about efficiency ..
 

andromedaaudio

VIP/Donor
Jan 23, 2011
8,355
2,731
1,400
Amsterdam holland
+ - is related to the full freq band 20 - 20 khz that off course also includes 1 K.
If you take it only at 1 K it means shit basically..
Because that could be your + 3 db point, while the rest of the band is substantially lower.

See freq responses arent a flat line .

Ps although at andromeda design they are for a very large part lol
 
Last edited:

About us

  • What’s Best Forum is THE forum for high end audio, product reviews, advice and sharing experiences on the best of everything else. This is THE place where audiophiles and audio companies discuss vintage, contemporary and new audio products, music servers, music streamers, computer audio, digital-to-analog converters, turntables, phono stages, cartridges, reel-to-reel tape machines, speakers, headphones and tube and solid-state amplification. Founded in 2010 What’s Best Forum invites intelligent and courteous people of all interests and backgrounds to describe and discuss the best of everything. From beginners to life-long hobbyists to industry professionals, we enjoy learning about new things and meeting new people, and participating in spirited debates.

Quick Navigation

User Menu

Steve Williams
Site Founder | Site Owner | Administrator
Ron Resnick
Site Co-Owner | Administrator
Julian (The Fixer)
Website Build | Marketing Managersing