Calculating power for two channel setups seems to be fairly straightforward. But, what about multi-channel systems? If watching a movie with expected peaks of 100dB, should all channels be expected to reach 100dB? Or, will the center channel and rear surrounds peak at lower levels?

I've seen articles on how sitting distance from the speakers affects decibel levels, but I haven't seen anything on how the room size affects levels. With all things being equal (i.e., same equipment and speaker distance from listeners), will the decibel level change if the room is larger versus a smaller room? If so, what formulas are used to account for room size?

cpplain, i can tell you from my own experience, room size is a HUGE factor.. the current house i am in resulted in my upgrading from a single 5 channel amp. to 3 separate amps.. My current Great room is 2,400cu/ft.. But, since it's a Great room it opens to several other rooms within the house...


I went from 135W per channel, to 250W per channel... When i listen to music, the average sound level is about 95/96DB.. Movies are about the same volume. So, pretty close to your stated 100DB. For a given amount of power the room size will play a role in the over all DB level within the room.


Others will chime in as well. hope this gets the conversation rolling.

Chris, no, all channels don't peak simultaneously. This is essentially a non-event in actual movie and music program material, although of course it can take place during laboratory testing with test signals applied to all channels equally. This is why an "all channels driven" power spec is unrealistic for real world home listening and the FTC amplifier power regulations require a two-channels driven basis, explicitly rejecting a proposal to require the all channels standard.

Both listening distance and room size are factors in establishing power requirements, but in typically sized home rooms distance is by far the more significant factor. Studies by Dr. Toole(as reported in his book "Sound Reproduction")and others show that the combined effect of distance and room size lead to a reduction of sound level of about 3dB per doubling of distance, taking into account the contribution of room reflections added to the direct sound. This is sufficiently accurate for estimated power at various sound levels, and I use this for my calculations. For example, a speaker with a typical sensitivity of 89dB at 1 meter with 1 watt would have a comfortably loud 85dB level at a 9-10' listening distance(86dB at 2 meters and 83dB at 4 meters)with that 1 watt, and would use 100 watts at a brief 105dB peak.

John,

I think I might not have been clear enough in my first question. I'm not necessarily asking if all channels will peak simultaneously, but rather, should it be expected that all channels can peak, at some point, to the same level. For example: if reference level is 105dB max, then should it be assumed that the center and rears could reach that level? Or, would it only be the front L/R to reach 105dB?

Alan's article on speaker power requirements (http://www.axiomaudio.com/power.html) lists 6dB loss per distance doubled. Is that an outdated or too conservative number?

When you're estimating power, do you use anechoic or in room speaker sensitivity ratings?

I have a rather large room of approximately 5500 cu/ft when taking into consideration the cathedral ceilings. How do you think the size will affect the power calculations if at all?

Sure, any channel can at a given instant be the loudest one. That's why amplifiers have equal power capabilities in each of their channels.

A 6dB loss per doubling of distance is correct for an omni-directional sound source in anechoic conditions. As mentioned above, the result in actual home listening rooms with reflections added to the direct sound is about 3dB per doubling.

Anechoic sensitivity ratings are used, since the 3dB figure takes into account the room contribution, and it shouldn't be counted twice.

Rooms aren't "filled" with sound at an equal level, despite the term being often used. The cathedral ceiling has a small effect on the level of room contribution from reflections. There's no easy way to determine how small the effect is, other than actually measuring with an SPL meter in your room at various distances to determine how close the loss in level is to the nominal 3dB per doubling figure.

I have similar size listening room in terms of volume; my listening position is about 12 feet from the main speakers. I regularly listen at levels between 95 - 100 dB.
With M80 as main speakers, and an amp putting up 300W (all channels driven), I have a lot of power to spare. My preamp volume goes about half-way to reach that sound level.

I would say 200W on all channels should do, unless you have very, very inefficient speakers or you sit 20 feet away from your speakers.

Depending on the amount of loss per distance doubled, I've calculated I would need 100W-200W to hit reference level. It looks like the best thing for me to do is to take some measurements to determine exactly how much my room affects decibel levels.

Would an iPhone SPL app be accurate enough for these types of calculations?

right, i think that you should use the iPhone SPL;
it may not give exact actual reading, but in this case, it's the relative level you're looking after, and it should do an adequate job.

I used two iPhone SPL apps. There were fluctuations in readings from both apps, but they were consistent enough to determine I've got about an 8dB loss from 1m to 4m (4dB loss per distance doubled).

Back to the gist of my first question. If I have a 5.1 system and each speaker requires 100W to reach reference level (105dB peaks), do I need an amp/receiver cable of 100W per channel, all channels driven?

No, Chris; as mentioned before, "all channels driven" isn't a realistic standard outside a test lab in the real world of home listening. When 1 or 2 channels happen to simultaneously hit a 105dB peak(which for most users may never happen, since reference is considered too loud for home use and a 10dB lower setting is common)the other channels aren't at that level.

I would suggest you look at one that is capable of that power, regardless of any of your findings, the worst thing you can do for your speaker is under drive it and to achieve reference volumes 100W/ch is at least the minimum I would look at and as others have previously stated, the chances of your system ever requiring all channels driven to the same volume level is extremely rare, if ever, and as you noted you will more than likely see each speaker at some point in a movie could require 100W to achieve refeernce levels, so having an amp that is capable of driving 100W/channle is a very good idea, too much power is a far better thing than too little.

My Denon 3808 is capable of 100W/ch. as shown here

John, just playing the devils advocate...is it not possible that there could be a movie that would produce a brief peak of 105dB across all channels? In that case: if the receiver was not able to produce enough power and went into clipping, would that damage the speakers?

Just a quick question, you do realize how loud 105 db actually is and that most HT system are never played back at those levels. The system is calibrated so that you can achieve reference levels when the volume is set to '0'. I listen at -30 to -10 depending on the film and my room is ~4000cf, vaulted ceiling etc.

And to answer your question to John, IMO, it is always possible, anything is possible, probable is another story and yes clipping is what damages speakers.

Chris, the devil doesn't have a clue; it's unknown for program material to simultaneously require max power in all channels; and incidentally, clipping isn't necessarily damaging whenever it occurs.

I was looking for a lab test on the 818 to show you, but apparently there isn't one published yet. The closest I found was this test of the 1009, which is a 9-channel unit with essentially the same amplifier as the 818, with the same 135 watt rating.

Like you, I wasn't able to find lab tests for the 818. I did see the 1009 test you referenced but wasn't sure how close the 818 would be.

At what point does clipping cause damage? Is it a matter of duration?

Yeah, duration and the sturdiness of the voice coils of the speakers involved.

master and commander chapter 4.

Good material to show what quality amplification can do, and if you have "enough." Tests all channels vigoruously, and I can't think of another track off hand that really peaks all channels almost simultaneously. Sure there are many others.

Duration as well as degree both play into a failure caused by operating a speaker in this manner....


I was unwilling to take this risk. Which, is why i bought more power as soon as i noticed my pervious amp being driven into the clipping region.

I know this is an exception, but I have seen it with a friend of mine: he absolutely loves music on "7 channel stereo" (which I don't like personally); he does have the same speaker all around. According to his receiver specs, this means: center channel= a sum of main L+R, and the rest is split equally between all L and R speakers.
So it all comes down to what your intentions are, and you budget.

AHH, but even in stereo the demands on the left and right channels are virtually never the same, so not technically all channels driven, very close but not exactly the same. If it was in 7 channel mono then you have the same demands on all channels driven.

True; the "all channel driven" spec makes no difference. I was getting confused with something else.

Jason and Adrian, it can also be pointed out that selecting an "all channel" mode(even all channel mono)doesn't in itself create additional power requirements. When more than one sound source is playing simultaneously, the sound power level is increased by 3dB for each doubling of the number of sources(1 to 2 speakers, 2 to 4, etc.). This relationship is illustrated here .

The bottom line is that to keep the same overall sound level with, for example, 4 speakers rather than 2, each speaker would require only half, i.e., 3dB less, power, so no more total power is used with 4 rather than 2 speakers playing simultaneously. If the total sound level was to be increased, of course more power would be used, but this would also be the result if only one speaker was used to create the higher sound level.