Watts per channel needed??

JerryLove

JerryLove

Audioholic Samurai
I would disagree slightly with hd, just slightly though. 2.83V/m is technically better/more electrically correct, but the W/m is better in one way. That is, if the manufacturer specified its speaker's sensitivity 88 dB/w/m, then you know it is 1W for 88 dB output at one meter regardless of the impedance.
The problem is, though, that it's not.

Assuming that the measurement itself is accurate (88db @2.83v) then the actual wattage depends on frequency.

Let's assume that they measured 1w at nominal impedence of 8ohm on that speaker.
If at 12kHz it's 16ohm, then a sinewave at that frequency will only draw 0.5w to get 88db.
Conversely, if there's a 4ohm dip at 60Hz, that will pull 2w for 88db.

An ideal amp would scale the same way; so an amp wattage would be completely valid.
In the real world, few amps double power each time the resistance halves; though basically all increase somewhat.

There is no single w=spl relationship because no speaker has a flat impedance curve.
 
C

CajunLB

Audioholic
I'd put that in the same sense that the guy driving a bus is an engineer....
When I mowed lawns for money I’d tell the ladies at the bar I was a land maintenance engineer. It broke the ice by making them laugh when they would ask what it meant.‍♀
 
Verdinut

Verdinut

Audioholic Samurai
When I mowed lawns for money I’d tell the ladies at the bar I was a land maintenance engineer. It broke the ice by making them laugh when they would ask what it meant.‍♀
Did that sometimes lead to a pussy clipping? :D
 
lovinthehd

lovinthehd

Audioholic Slumlord
I would disagree slightly with hd, just slightly though. 2.83V/m is technically better/more electrically correct, but the W/m is better in one way. That is, if the manufacturer specified its speaker's sensitivity 88 dB/w/m, then you know it is 1W for 88 dB output at one meter regardless of the impedance.

That popular calculator we linked everyday twice on Sunday is based on W/m, so yes the user would have to make adjustments for speakers that have nominal impedance other than 8 ohms. I think I posted a conversion table recently for that purpose.
I understand your point, but most of what I read originally referred to 2.83V as the standard. It does cause some confusion as you mention if the speakers being compared are different impedaces. Then again sometimes sensitivity isn't in an anechoic space or measured over a wide enough range of frequencies. Would be nice for the industry to have a rock solid standard everyone stuck to.
 
lovinthehd

lovinthehd

Audioholic Slumlord
I don't know about the US, in Canada one can be sued for that if not actually a licensed engineer lol..
From what I've read the US is a bit unique in that....believe many countries have strict/proper definition of engineer.
 
slipperybidness

slipperybidness

Audioholic Spartan
From what I've read the US is a bit unique in that....believe many countries have strict/proper definition of engineer.
Dunno about the rest of the USA, but Texas indeed has strict definition of "engineer", and when that title can/can't be used.

I am not an engineer, so could miss a detail.

To use that title in Tx, it is something along the lines of passing a test to get an engineer license.
 
P

PENG

Audioholic Overlord
The problem is, though, that it's not.

Assuming that the measurement itself is accurate (88db @2.83v) then the actual wattage depends on frequency.

Let's assume that they measured 1w at nominal impedence of 8ohm on that speaker.
If at 12kHz it's 16ohm, then a sinewave at that frequency will only draw 0.5w to get 88db.
Conversely, if there's a 4ohm dip at 60Hz, that will pull 2w for 88db.

An ideal amp would scale the same way; so an amp wattage would be completely valid.
In the real world, few amps double power each time the resistance halves; though basically all increase somewhat.

There is no single w=spl relationship because no speaker has a flat impedance curve.
My point has nothing to do with impedance curve or power output etc.. It was just about the use of the linked Homestead calculator.

That calculator requires the following inputs:

- Speaker Sensitivity in dB/1W/1M (M is meter)
- Amplifier Power in Watts
- Distance in feet
- No. of speakers
- Speaker Placement (relative to walls or corner)

So when the user enters the input "Speaker Sensitivity", he/she has to be careful because if the speaker's specified sensitivity is in dB/W/M, then there is no need to be concerned about the "nominal impedance" of the speaker. If the specified sensitivity is in dB/2.83V/M, then the user has to know the "nominal impedance" of the speaker in order to do the necessary adjustment, simply because that calculator expects the sensitivity to be in dB/1W/1M.

As @Beave mentioned before: "Amps are designed to be constant voltage sources, not constant power sources." So like @lovinthehd , I also prefer sensitivity to be specified in the dB/2.83V/1M standard, but reality is that many manufacturers are still using the alternate dB/1W/1M standard. My additional point to HD's is simply that in a very narrow sense, dB/1W/1M is better if the regularly linked Homestead calculator is used, because then the user would not have to make any adjustment to the calculated result, or in entering the input field "Speaker Sensitivity. That's all..
 
P

PENG

Audioholic Overlord
I understand your point, but most of what I read originally referred to 2.83V as the standard. It does cause some confusion as you mention if the speakers being compared are different impedaces. Then again sometimes sensitivity isn't in an anechoic space or measured over a wide enough range of frequencies. Would be nice for the industry to have a rock solid standard everyone stuck to.
Agreed, I actually developed my own calculator using Excel that allow inputs for sensitivity in dB/2.83 V/1M, or dB/1W/1M and for the dB/2.83V/1M format/standard, you can enter the specified "nominal" impedance of any value.

I did it because as you said, would be nice for the industry to............., but reality is many, including some big names such as Monitor Audio, DefTech still use the dB/1W/1M, while some, such as Polk Audio and others don't even tell you what it is, i.e. just XX dB.

Example of the one that don't tell:

The OP's Polk Audio S4 specs stated:

Impedance: Compatible with 8 ohms outputs
Efficiency: 91 dB

Really strange impedance spec, and the "Efficiency" must have been incorrectly used and likely meant to be "Sensitivity", but then it never provide the conditions, i.e. 91 dB under what condition, 91 dB/2.83V/1M, 91 dB/1W/1M, or whatever... Can't really win on this one.:( But then it is Polk so I am not sure what to expect..
 
I

IansDad88(Don)

Junior Audioholic
how skewed are the sensitivity ratings from the companies in home audio?
I know in the car audio industry they just throw the numbers around and when you calculate it a woofer boasting a 91db rating realistically have 87db. obviously these numbers mean pretty much nothing to us as we tend to over power our woofers with 1.5-3x the rated power to combat impedance rise but were just trying to get loud lol.
That's what I've been saying..! Haha
 
lovinthehd

lovinthehd

Audioholic Slumlord
That's what I've been saying..! Haha
IIRC many don't measure sensitivity below 400hz, so not much use for the woofers generally....tho you probably won't see a sensitivity spec for the majority of active subs for home audio either, usually that's more a diy thing.
 
JerryLove

JerryLove

Audioholic Samurai
My point has nothing to do with impedance curve or power output etc.. It was just about the use of the linked Homestead calculator.

That calculator requires the following inputs:

- Speaker Sensitivity in dB/1W/1M (M is meter)
- Amplifier Power in Watts
- Distance in feet
- No. of speakers
- Speaker Placement (relative to walls or corner)

So when the user enters the input "Speaker Sensitivity", he/she has to be careful because if the speaker's specified sensitivity is in dB/W/M, then there is no need to be concerned about the "nominal impedance" of the speaker. If the specified sensitivity is in dB/2.83V/M, then the user has to know the "nominal impedance" of the speaker in order to do the necessary adjustment, simply because that calculator expects the sensitivity to be in dB/1W/1M.

As @Beave mentioned before: "Amps are designed to be constant voltage sources, not constant power sources." So like @lovinthehd , I also prefer sensitivity to be specified in the dB/2.83V/1M standard, but reality is that many manufacturers are still using the alternate dB/1W/1M standard. My additional point to HD's is simply that in a very narrow sense, dB/1W/1M is better if the regularly linked Homestead calculator is used, because then the user would not have to make any adjustment to the calculated result, or in entering the input field "Speaker Sensitivity. That's all..
That is why the calculator will always be an approximation of need and cannot be an accurate measure of need... sensitivity cannot be measured (accurately) in watts... at least not without factoring in ohms.
 
lovinthehd

lovinthehd

Audioholic Slumlord
Yes but easily enough to manipulate input data to get different pictures of results...
That is why the calculator will always be an approximation of need and cannot be an accurate measure of need... sensitivity cannot be measured (accurately) in watts... at least not without factoring in ohms.
 
P

PENG

Audioholic Overlord
That is why the calculator will always be an approximation of need and cannot be an accurate measure of need... sensitivity cannot be measured (accurately) in watts... at least not without factoring in ohms.
Keep in mind when a speaker manufacturer claims 90 dB/W, they likely did measure the spl output and the applied voltage, not watt, then they would calculate/convert whatever dB at 2.83V to 90 dB/W as though the speaker was an 8 ohm resistive load but that it actually wasn't. That is, sort of creating a mathematical equivalent only. As you alluded to, you really can't measure spl per watt unless the load is a resister, but a resister won't produce much spl just by buzzing..

In reality, looking at the whole "power/watt" thing for speakers is technically fraud as they are not resisters so they don't really consume a lot of watts like highly resistive loads do It would be more meaningful to compare amps based on their rated current output at their rated voltage output, or by rated VA but specify either the rated/maximum voltage or current. If we know any two of the 3 variables, V, I, VA, we can calculate the 3rd easily. Ideally, amp output specs should also include the max phase angle it can handle at max VA output without being overheated.

You see, that's why I only wanted to address the question of "how much power do I need...." and the use of online calculators that are for approx. calculations only because the whole thing is based on the false premise of "power" instead of voltage and current to begin with.

Note (I know you know, this is for those who don't know) that power in watt does not equal volt x ampere that is VA, not watt. For watt, you need to know the phase angle. As an example, we can apply 100 V to a big heavy coil/inductor and measured current flow of 100 A, the power in VA would be 10 kVA but power in watt would be near 0 watt because the phase angle would be near 90 degrees.

This is just one example to show why it isn't quite right to look at how much power in watts do we need, when the load is a moving coil loudspeaker.

By the way, 2.83 V was probably chosen because when applied to an 8 ohm resister, the power dissipated would be 1 watt, making it easy to for conversion back and forth with the dB/1W/1M format as many speakers are specified 8 ohms nomimal. I mean, otherwise a lower or higher voltage could have been used for measuring speaker sensitivity.

Again, I only mention this in case some may not know of this fact.
 
Last edited:

newsletter
  • RBHsound.com
  • BlueJeansCable.com
  • SVS Sound Subwoofers
  • Experience the Martin Logan Montis
Top