Paul's Posts — 21 February 2013



Reviewer Keith Howard sent me an interesting note a few days ago that really was an eye opener for me when it comes to understanding how loud music actually gets.  I am going to reprint it here.

I would like to preface this note by pointing out that a lot depends on the type of music you listen to.  For example, classical music is generally very dynamic with average levels quite low but filled with big dynamic peaks.  Rock music, on the other hand, is generally the opposite: lots of loud average levels with very few peaks and valleys in the sound.

This observation is important when it comes to figuring out what kind of amplifier you need and we’ll start to cover some of these needs in the coming days.  Meanwhile, here’s Keith’s post.

“This is a subject close to my heart and has been for some years – so I eagerly look forward to reading what you have to say about it.

You may be aware that I wrote an article about one aspect of this for Stereophile some years ago (, following a three-part feature on various aspects of the subject for Hi-Fi News. In the course of my experiments, listening at only a little higher than I would consider ‘realistic’, I clipped Musical Fidelity’s most powerful amplifier of the time, rated I recall at 700W into 8 ohms continuous, when driving a pair of B&W 805 stand-mounts in a room of about 64m^3.

I also find it ‘amusing’ that if you record a tambourine with a calibrated microphone, you find that it produces short-term peaks of around 113dB SPL at 1m. The amusing part comes when you calculate the amplifier output required to achieve this with a tweeter sensitivity of, say, 90dB.

Not many of us listen to solo tambourine, I know, but it seems not unreasonable to me to expect a hi-fi system to be able to reproduce single instruments without clipping, let alone larger musical forces. The argument that hard clipping of such short-term peaks is inaudible is hogwash – I can hear it when the tambourine recording is processed to clip at 3dB below peak level, and anyone would easily hear it at -6dB clipping. And this without any amplifier latch-up or other misbehaviour when recovering from clipping.

This was done using a piece of freeware I’ve made available on my web site (, scroll down to Clipping Simulation and But you know what? I suspect that hardly anyone has tried the experiment for themselves and then come to terms with what it means.

A 5W SET does not cut the mustard even with 95dB sensitivity loudspeakers!”   Thanks Keith.

In yesterday’s post I mentioned that to reach 100dB with a 90dB efficient loudspeaker, you would need 500 watts.  What I left out is some valuable information: distance from the speaker.  To get that 500 watts I moved the listener away from the 1 meter mark without telling you.  So let me straighten this out.

Here’s a handy list you can use to help you figure out needed watts for playing music, not sine waves.  These figures are all based around a 90dB efficient speaker, meaning 1 watt of sine wave power gives you 90dB at a distance of 1 meter as our reference.

  • 50 watts are needed to get 90dB of music.  90dB average, 107dB peaks
  • 200 watts are needed to get 96dB of music.  96dB average, 113dB peaks (our tambourine now works).
  • 500 watts are needed to get 100dB of music.  100dB average, 117dB peaks.
But, most of us don’t sit 1 meter away from our loudspeakers, now do we?  I don’t, I am about 3 meter away from my loudspeakers.
For every doubling of the distance we need 6dB more power to reach the same level – so we need 10 times more power for a 3 meter distance.  So now let’s look at out chart:
The next set of figures are all based around the same 90dB efficient speaker playing music, but this time measured at a 3 meter distance, meaning 10 watt of sine wave power is needed to give you 90dB as a reference.
  • 500 watts are needed to get 90dB of music.  90dB average, 107dB peaks
  • 2000 watts are needed to get 96dB of music.  96dB average, 113dB peaks (our tambourine now works).
  • 5000 watts are needed to get 100dB of music.  100dB average, 117dB peaks.

If your speakers are 87dB efficient, double the above numbers.  If your speakers are 93dB efficient, halve the above numbers, 96dB halve them again and so on.

How many tweeters can handle all that power?  Not many, still the numbers don’t lie.

A ridiculous amount of power by anyone’s standard – except that is unless we were famed audio designer Dick Burwen whose audio system runs on 20,000 watts of power.

But then you knew this series of posts was going into la la land.

We’ll return to something reasonable tomorrow.

Forward to a friend and help us engage more readers Forward to a friend and help us engage more readers

Get new and fresh stories like this each morning by joining the folks reading Paul's Posts. Click here

Related Articles


About Author

Paul McGowan is the CEO and co-founder of PS Audio Inc. a Boulder Colorado design and manufacturing company of high-end audio products and services. McGowan has been designing and building high-end products for nearly 40 years. Hobbies include skiing, music, hiking, artisan bread baking, kick boxing and cooking. He lives in Boulder Colorado with his wife Terri and his 4 sons.

(18) Readers Comments

  1. Paul, if the Nudell designed speakers which you are lusting after are line sources then they would need a lot less watts to reach the described sound levels wouldn’t they?

    • I don’t remember what their efficiency is – I think it may be around 88dB, perhaps as high as 90dB – but otherwise they will be the same as any speaker with the same efficiency.

  2. Excellent series. The SET side of the discussion is unsurprisingly getting some attention elsewhere. I posted a link to Paul’s series over on the Decware forum [I own several pieces of Decware gear; great stuff], and it has sparked some very engaging discussion.

  3. I have been enjoying this series of posts, Paul. I have been using high-powered amps since the 1970s. To me, they just sound better . . . less stressed, more open, more like real music. I listen to string-quartet music, which is surprisingly dynamic. I have heard some expensive systems that, whether because of insufficient power or because of terrible room acoustics, couldn’t make a string quartet sound remotely natural.

  4. I wonder if the extreme transient dynamic range of the recorded tambourine imposes issues in audio even further up the chain. For instance, what is the ability of even the finest microphone to capture the true dynamic energy of the leading transient spikes you talk about here. Is a great deal of this energy lost right at the start? Curious to me. I would imagine a high-hat in a drum set being pounded would introduce similar challenges? Thoughts?

  5. Cookie’s post, referring to one of the best SNL skits ever, left a bad mental image of Will Ferrell at his not-so-sveltness gyrating around, putting his whole being into pounding that cowbell.

    A good example of how much dynamic range there is in just everyday orchestral music, I submit for your approval a wonderful recording of Ravel’s “Bolero” by Valery Gregiev and the London Symphony Orchestra (LSO0693).

    Okay, now for many of you the image of Will Ferrell is replaced by that of Bo Derek (ya’ll can thank me later). Try to push that image out so you can hear this piece. I know it is difficult as Bo had such delightful… ah,,, corn rows. The challenge is to set your volume so that you can actually hear the very soft snare drum at the beginning and not have blood coming out of your ears at the end. This is an amazing 16 minute continuous crescendo and even the average level at the end is very loud. As a side bar, what amazes me about this recording is how the single snare drummer (he is joined by a second later as the volume picks up) can keep a perfect sequence repeated so many times at all volumes and not lose concentration or have a muscle lock up.

  6. Correct me if I am wrong, but I think The Inverse Square Law – ie. The intensity of sound in a free field is inversely proportional to the square of the distance from the source – does not apply linearly in a closed room.
    Whenever I measure SPL levels in a room, there is very little difference in the measured level between the 1 meter reading and the 3 meter reading due to the fact that most of the energy is reflected and stays in the room.
    I do agree that large amounts of power are needed for accurate reproduction of dynamic peaks, but I think your calculation that 5000 watts are needed for 117 dB peaks would only apply if you were listening outside, away from any reflecting surfaces.

  7. I have to admit I was thinking the same thing but I assume that Paul could not make so crass a mistake; everyone knows you need massively more power to make loud music in the open air.

    Please confirm Paul that you were giving the figures for the power required to make a certain volume music in a room.

    • I am not sure what a “crass mistake” is but it doesn’t sound good. :)

      Indeed, the inverse square law is taken from measurements in an anechoic chamber or outside if you will. I’ll explain in tomorrow’s post that while the info I gave out is correct, there are some dependencies about how we get to the numbers: this being one of them.

  8. Paul,
    I’m really enjoying these recent posts but I must say that I’m bewildered by some of the content especially the electrical and math computations regarding wattage, spls and listening distance. What confuses me is the fact that I have Paradigm Studio 100 speakers (not the most efficient) and a Marsh 200 wpc power amp. I listen to mostly symphonic music at what I would call concert levels (whatever that means) from a seated position of about 3 meters in a 20 x 20 room. I still have enormous amounts of volume left that I could apply but it would become audibly uncomfortable. Based on these recent articles, I shouldn’t have enough power to do the job let alone have a reserve. Am I missing something you might previously have written on this topic or will something you haven’t written yet answer this puzzle?


  9. Thanks Paul
    I have also reall enjoyed the latest series it confirms what I had always believed true you can never have to much power no matter what speakers rated input recommendation is .
    I have also have noted that you can double the output of an almost any AB amplifier and it will pale in comparison to a pure class A design at half or even a quarter of the same wattage AB unit .
    Can’t also thank you enough for that outstanding link
    Wow talk about taking music / sound to a new level ! What a true labor of love.
    How many audiophiles can say they built their house around there sound
    system? Starting in the sixties !!

    • Did you like how he’s using a Steinway grand for a diffusor in one of the horns?

      • Really Paul I thought that was a typo or something ??
        One would think you could find a better diffuser then a Steinway Grand but who am I to question such a genius .
        It would be like asking Stradivarius hey that violin it would make a great diffuser for this speaker I am designing ??
        Your term “Mad Genius” would be most apropos here !
        Yourself and Keith deserve kudos , for this one post alone should get you 3,000 + followers :-)

  10. I have owned my Maggie MG-IIIa’s for 20 years and listened to all sorts of music at sometimes excessively loud volumes (often, alcohol is involved in the excessively loud volumes:) ), and I have yet to blow a ribbon tweeter fuse in my Maggie’s, even though my old Adcom GFA-535 amp is rated at a mere ~100w/ch. into 4 ohms. If Paul is correct and 500 or more watts is appropriate (assuming the old audio axiom that clipping fries speakers, not raw power) then why aren’t my ribbon fuses always blowing?

    • A lot depends on the room, the program material and your individual tastes. I have had to replace the fuses in mine multiple times, actually moving to a larger fuse at one point – but perhaps I have a much bigger room with more damping requiring more power.

Leave a Reply