Sensitivity, Efficiency, and Room Gain

[Colin]

all true, which is why I got two different responses with my Khorn measurements. Don't forget that the tube amp is capable of only 6 watts maximum output, while the solid-state Pioneer puts out ten times more - 60 watts. Just like loudspeakers, the RMS rating of amplifiers at their low THD points is also NOT the true story either. Even with the crude approximation of the pre-amplifier dial, the tube Paramours are pumping out 1.5 watts to attain these SPL levels, while the Pioneer M-22 is putting out about 15 watts! Strange, isnt it?

[DrWho]

on top of all of the concepts discussed so far, you've got the "frequency response of the ear"...i don't have any charts available on my pc at the moment (i can get some for those interested), but i do know that our ears are most sensitive at the midrange frequencies (1k-3k is the peak) and then a measured 30db louder at like 40Hz and 16kHz sounds just as "loud" (im going off of memory here in case im way off on the numbers).

all that to say (and i know it could be more clear) is that it is for this reason that the same measured value could be the same, and why one song (or speaker) could sound louder than the other. (a speaker that is louder at 3k would sound louder than a speaker that is quiet at 3k, but the overall output could be measured the same)

another phenomenom is how compression can make things sound louder too. i'll try to paint an example to express this (but i have no idea how right i am either). sit in a room and listen to some totally uncrompressed music and increase the volume...eventually what happens is the room sounds full and we enjoy this volume a lot. crank it some more and it gets too full...this is an acoustically created example of what compression sounds like...as we overload the room, everything as far as dynamics just gets squished until you either go deaf, or you can barely tell it's still music playing. this "sound" that we hear is an extreme example of something we might call "loud"...because we know what a room that's "overloaded" sounds like, music that is compressed will appear loud too; even though the measured dB output is the same!

all that to say, as far as loudness goes, you're left in the hands of the recording engineer who has EQ'd for the frequency response of his ears, and the mastering engineers who are applying all sorts of enhancements and compression to give it that crystalike sound...i guess what im really trying to say is loudness has more to do with it than just the measured dB. there are more factors than just the two i've mentioned...for example, by putting compressed music into a room, you've got the acoustical compression of the room to contend with again

(btw, does anyone know how well what i call "acoustical compression" relates to a normal compression unit? to me, both things sound almost exactly the same...just that the room one has to be a lot louder to experience.)

another thing to throw in there is the concept behind line arrays...the closer you are to a direct radiator, the soundwaves move out more in a plane than in a sphere...if i remember right, this means that you lose 3db every meter with the plane, versus 6db with the sphere (im too lazy to do the math to check if im correct, don't got my chart with me, and it doesn't matter cuz the concept is the same). in a room (especially a corner), i have no idea what the "true" drop per meter is...perhaps this is what the room gain compensation is referring to? either way, the compensation is not a linear value, because the rate at which the dB is dropping is changing based on how planar or spherical the wave pattern is...how important is this? we prob couldn't hear the difference, but we're going for perfection right?

well im sorry my post was more conceptual than getting into the actual math and equations and stuff...i was hoping some1 else here might be more knowledgable and they would probably do a better job throwing their numbers around for me

i would say though that performing these measurements with an RTA and SPL meter at the same time would be a better approach because it will show the frequency response with the overall output measured...but them toys can get expensive

[DrWho]

i almost forgot to add...you need to mount your metering equipment on a stand to obtain accurate results. i've had instances where moving an inch causes a 12db change in measured output (oh man, gotta love all the phase issues rooms create)...i think we should all listen to music outside, or create some kind of wall that is acoustically transparent, but everything else opaque? something like that

[sfogg]

" Also, you need to either use a reference plot of the speaker impedance, or you must measure impedance at the frequency you test. Clearly "2.83V" is only "1W" when the speaker impedance is at 8 ohms - which is RARELY!"

Actually, a lot of tests use 2.83v no matter what the impedance is. So the test isn't really testing the sensitivity to 1w, it is testing the sensitivity to 2.83v to keep the voltage input consistent between tests.

Shawn

[sfogg]

"because the rate at which the dB is dropping is changing based on how planar or spherical the wave pattern is...how important is this? we prob couldn't hear the difference, but we're going for perfection right?"

Usually you can hear the difference here if you try matching something like a ribbon tweeter (which is a line source over some of its bandwidth) with a conventional woofer. The SPL drop off by distance isn't the same between the two driver units. So at some distance the twos level will be nicely balanced. But further from that distance the speaker will sound brighter as the high end isn't dropping off as quickly as the low end.

Shawn

[Deang]

Now Colin, shame on you, for that sounds like some really SET fuzzy math to me.

Your post really exemplifies the whole problem, and in spite of the fact you indicated you understood the import of what Shawn was saying -- you missed it completely!

This is where I hoped we'd end up: The difference between sensitivity and efficiency. They are NOT the same, and a speaker may be "sensitive" but not very "efficient". In fact, you could have a sensitive speaker that actually sucks MORE amplifier power at a given frequency than a less sensitive one! So, to me -- the sensitivity rating seems like a smoke and mirrors spec.

What we want to know is how much acoustic power do our speakers generate from 20 to 20K with our amps.

[sfogg]

"Even with the crude approximation of the pre-amplifier dial, the tube Paramours are pumping out 1.5 watts to attain these SPL levels, while the Pioneer M-22 is putting out about 15 watts! Strange, isn't it?"

If you are 'figuring' power output simply by looking at the dial on your pre-amp you aren't even *remotely* close to estimating the power your amps are putting out. That is why your results are strange.

Shawn

[Colin]

I dont think I missed the point completely. I think I made it. That is why I included the different amplifiers used with each test, because solid-state and tubes are different; I was getting different results. I have a lot more tests on many conventional loudspeakers, taken with the same solid-state Pioneer M-22, but those are on a PC that is NOT on my network right now. My point was that I got widely varying results. A quick check of the big ole Klipsch corner horns just now, with the sweet and delicate sounding flea-powered Paramours instead, shows the same 1KHz test tone (Stereophile Test CD #2) with a 79dB SPL at my listening seat (half the volume). But yes, there is a subjective and objective difference between sensitivity and efficiency. I dont know how to objectivity measure the acoustic power our loudspeakers put out through our amplifiers in our rooms. I do know that we can all share the same crude RS SPL meter and a 1kHz test tone.

[sfogg]

Colin,

I assume above you mean the volume control on your pre-amp is of a linear taper, not the more traditional log (sometimes called audio) taper.

The reason the position of your pre-amps dial is meaningless for power output is as follows.

Consider two amplifiers. Doesn't matter if they are tube or SS.

Both have a max output of 10 watts into the same load.

We hook the first one up to your pre-amp and to hit 90dB at your listening position you need to turn your pre-amp up 1/2 way.

With the second amp to hit 90dB at your listening position you need to crank your pre-amp up as far as it goes.

Does this mean the second amp is putting out twice as much power?

No.

The very fact that the speaker is putting out the same SPL shows that if anything the actual power output from the amps are the same... and if you measured it with a volt meter you would find that to be the case.

So why did the second amp need its volume cranked further then the first amp?

Simply because the input sensitivity of the second amp was different then the first.

Amp 1 may need 1v input to reach its full output of 10 watts.

Amp 2 may need 2v input to reach its full output of 10 watts.

So you would see that for the second amp to output the same wattage as the first amp (no matter what the wattage out from the amp is) your pre-amp needs to output twice as much voltage. IOW your volume control would have to be much higher up even though the end power output of the two amps are the same.

You can even have the case where your pre-amp doesn't have enough voltage ability to drive an amp to full power. This can happen in the case of passive pre-amps fairly easily with the wrong amps. The plus side is you *can't* clip the amp in this sort of setup.... the downside of course is you can't use all of your amps available power if/when needed.

The same thing happens even with your sources too. Find a 1kHz test tone on a CD recorded at @-20dB and compare it against the same test tone recorded at -30dB. To reach the same SPL level on the -30dB CD you will have to turn the volume up higher.

Heck, if you used two different CD players and the *same* CD, pre-amp, amp and speakers you could easily find that to reach the same SPL you need two different volume settings on your pre-amp for the two different CD players. That is because the CD players output voltage may differ as well.

All of the above is why the position of your pre-amps volume control is totally irrelevant to the amount of power your amp is delivering.

" I do know that we can all share the same crude RS SPL meter and a 1kHz test tone."

That is not enough, you also need a volt meter to measure the amps output into the speaker. Otherwise you end up with usless completely irrelevant data.

Shawn

[hurdy_gurdyman]

I'm continously amazed at how many people seem to think the settings on the volume control have any meaning at all connected with amp output. All the volume control does is to cut back the signal coming in, thus contoling the amout of signal going to the amp. That's all it does. Different signals will need different settings on the volume control to get a given wattage output. And as mentioned in previous post, different amps require different amounts of voltage to reach the same given output wattage. Different speakers react differently to different amps. All this means that the setting of the volume control is irrelevant to testing speaker sensitivity.

Dave

[Colin]

I was merely trying to keep as many things as equal as possible with the few simple but crude tools that I have. I know the volume control knob is NOT a true approximation of the actual power output of either the line or power amplifier. My vintage PAS 3 does use a A-taper audio potentiometer. Rated output is 5 Vrms, 10 Hz 80 kHz, Maximum is 12 Vrms, 20 20 kHz.

[cwm]

This is all interesting, here is what I did once with my LaScalas. In the process of relocating some equipment I had an opportunity to put the LaScalas at the opening to a garage doorway i.e. firing out into the open outdoors.

I used my ANSI type 2 SLM (just calibrated) and a pink noise sound track on a test CD. For measuring I used the power scale on a McIntosh MPI 4 scope and set the level slowly up to the 1 watt level (I know this is dependent on the scale accuracy but it is fairly accurate). I then measured at 6ft and got a reading of 99DBc which should be 98dbc (6db down from 104dbc) I think the descrepancy would be from ground reflection, so the rating on the horn is fairly on the mark. I reasoned that the pink noise with equal level per octave would allow for slight response anamolies in the speaker and the scope/SLM themselves. In the listening room?----who knows?!! But the bottom line is those horns will "rock with a watt"!!!

[cwm]

I'll correct myself slightly---make that equal energy per octave with the pink noise source.

[maxg]

Seems my original summation was right - the figures mean almost nothing in the real world.

Reading this thread has helped explain a lot to me, but, not why the RF3's are so loud relative to the Heresys / Sansuis and my own speakers.

That is, unless you are all saying that the impedance of the RF3 is so radically different from the other speakers as to make up the difference...

[djk]

"Errors in this multiply fast. The difference between 8 ohms and 6 ohms is already 2.5dB."

Times 10 when comparing power.

1.25dB is the difference between 2.83V across 8 or 6 ohms.