The recent article on sensitivity - Misconceptions/Contradictions, or what?

[PENG]

..all speakers are designed to be driven from voltage source amplifiers and have a flat frequency response when so driven.

That's hard to believe. Let alone "all", where can I buy one like that, flat freq response ?

The reason for quoting sensitivity rather than efficiency is that the true efficiency of a speaker varies wildly with frequency because of its impedance variation. The efficiency quote would be fairly meaningless, so loudspeaker engineers measure and refer to “sensitivity” instead.

The first part does not seem totally correct if we restrict the frequency range to the same one he (presumable Andrew Jones but I could be wrong) used to derive the "nominal impedance" that he referred to "Whenever sensitivity is quoted, the nominal impedance must also be stated. ".

The second part seemed to imply he might be referring to the dB SPL/W/M as "efficiency" which is true to some extent because by definition one can convert that to efficiency that is defined as the sound power output divided by the electrical power input, but the dB SPL/W/M is also used to indicate by some reputable manufacturers (e.g. Wilson Audio, ATC, PSB (shows both), Monitor Audio, Linn,Whafedale, Martin Logan, Canton) as sensitivity. I see no issue with that as it does almost the same as the dBSPL/W/M if the "nominal impedance" is defined and included. We know P=VI*PF so P,V,I,Z,Phase angle, are all mathematically related.

Because a speaker is a voltage-driven device, we would be much better off to move away from amplifier power to amplifier voltage, specified in dB relative to 2.83 volts. At the same time, the minimum impedance of the load that the amplifier can maintain at this maximum voltage should be quoted. Using this, we could directly calculate how loud the speaker could go. An example will illustrate this: suppose we take a conventional 100W amp into an 8 ohm load. 100W is 28.3Vrms, which is +20dB relative to 2.83V. If the speaker is 90dB sensitivity, then max SPL is 90+20= 110dB. So, we quote the amp as a +20dB amp. That’s it...we now know how loud it will play with any sensitivityspeaker. Easy!

I find this puzzling/confusing, let's take a look here, an 100W amp and 8ohm load combination is apparently used as an example to show why it is better off to use the 2.83V scheme?

Base on what he emphasized earlier, the 8 ohm would have to be"nominal", that means whether you use 2.83V or 1W, the sensitivity isstill going to be 90 dB either way, so what the heck is the difference as longas the nominal impedance is also given as he insisted that it must be given?

Question from Worthless Marketing Slug: Now, impedance varies all the heck over the map. What--if any--is the frequency spectrum or single frequency point where sensitivity is measured? If, say, it's 'industry standard' (is there such a thing for sensitivity measurements?) to measure at 1000 Hz, then shouldn't the speaker under tests be measured for impedance at 1000 Hz, then the input voltage adjusted so it comes out to be 1 watt? So ifit's 6 ohms, then the input voltage is 2.44V, if it's 8 ohms, it's 2.83, if it's 5 ohms, it's 2.23, etc. Shouldn't we hold the input wattage constant sowe're comparing apples to apples all the time?

· Answer from Andrew Jones: It’s the varying impedance issue which is responsible for moving away from “efficiency” towards“sensitivity.” 1000Hz has no special significance as a frequency for making thecalculation, so we instead ignore this, use sensitivity, and average this over a range of frequencies.

This is funny, not that I would agree with the worthless marketing slug, but the pro's answer sounded like a circular argument to me. The fact is, if you average things (such asthe nominal impedance I guess..) over a range of frequencies to come up withthe dB SPL/2.83V, since we know by now he would always quote the nominal impedance, then anyone could convert it back to dB SPL/W, or you can take the same approach, keep the input power at 1W over the same frequency range that is used to derive the nominal impedance, say 300 to 3000 Hz or whatever his choice is, measure the SPL, average it (now I know this will be controversial); and then convert it to the 2.83V equivalent mathematically. The two are so interrelated such that you can convert them back andforth, as long as you define and derive the nominal impedance.

· Signal Strength: (2.83V, 1 watt, etc) - 2.83V is preferred

I am happy to see this bullet included in the conclusion of the article. IMHO, it does boil down to just that, "preference". One way is not absolutely better or worse than the other. In fact, not long ago, I posted, ADTG (may rememberthat as I PM'ed him on that topic) about my preference of the 2.83V approach,citing the notion that as long as we know the impedance characteristic. I thought it was a better scheme for much of the same reasons cited in that article. Well, I changed my mind recently, to the pointI now I slightly prefer the 1W scheme. I have had a lot fun reading thatArticle and found it extremely well written and informative. I was just a little surprised by the abundance of seemingly contradictory statements/claims made in some of those Editorialnotes, when the two engineers (assuming they both are) tried to explainwhy they preferred the 2.83V scheme. I wonder what other's opinions/preferences are?

 

[JerryLove]

I don't think we should give nominal nor average anything. If I am interested in sensitivity: I am interested in knowing the *worst* sensitivity because that's what limits my speaker/amp.

Though *worst* can vary depending on how your amp deals with lower and higher resistances.

 

[PENG]

I don't think we should give nominal nor average anything. If I am interested in sensitivity: I am interested in knowing the *worst* sensitivity because that's what limits my speaker/amp.

Though *worst* can vary depending on how your amp deals with lower and higher resistances.

I agree, if they say 90dB/W/M 8 ohm, I would look at the impedance curve, pick my own and do my own conversion and size my amp accordingly, just for argument sake. Same idea if I am given dB/2.83V/M

 

[gene]

That's hard to believe. Let alone "all", where can I buy one like that, flat freq response ?



The first part does not seem totally correct if we restrict the frequency range to the same one he (presumable Andrew Jones but I could be wrong) used to derive the "nominal impedance" that he referred to "Whenever sensitivity is quoted, the nominal impedance must also be stated. ".

The second part seemed to imply he might be referring to the dB SPL/W/M as "efficiency" which is true to some extent because by definition one can convert that to efficiency that is defined as the sound power output divided by the electrical power input, but the dB SPL/W/M is also used to indicate by some reputable manufacturers (e.g. Wilson Audio, ATC, PSB (shows both), Monitor Audio, Linn,Whafedale, Martin Logan, Canton) as sensitivity. I see no issue with that as it does almost the same as the dBSPL/W/M if the "nominal impedance" is defined and included. We know P=VI*PF so P,V,I,Z,Phase angle, are all mathematically related.



I find this puzzling/confusing, let's take a look here, an 100W amp and 8ohm load combination is apparently used as an example to show why it is better off to use the 2.83V scheme?

Base on what he emphasized earlier, the 8 ohm would have to be"nominal", that means whether you use 2.83V or 1W, the sensitivity isstill going to be 90 dB either way, so what the heck is the difference as longas the nominal impedance is also given as he insisted that it must be given?



This is funny, not that I would agree with the worthless marketing slug, but the pro's answer sounded like a circular argument to me. The fact is, if you average things (such asthe nominal impedance I guess..) over a range of frequencies to come up withthe dB SPL/2.83V, since we know by now he would always quote the nominal impedance, then anyone could convert it back to dB SPL/W, or you can take the same approach, keep the input power at 1W over the same frequency range that is used to derive the nominal impedance, say 300 to 3000 Hz or whatever his choice is, measure the SPL, average it (now I know this will be controversial); and then convert it to the 2.83V equivalent mathematically. The two are so interrelated such that you can convert them back andforth, as long as you define and derive the nominal impedance.



I am happy to see this bullet included in the conclusion of the article. IMHO, it does boil down to just that, "preference". One way is not absolutely better or worse than the other. In fact, not long ago, I posted, ADTG (may rememberthat as I PM'ed him on that topic) about my preference of the 2.83V approach,citing the notion that as long as we know the impedance characteristic. I thought it was a better scheme for much of the same reasons cited in that article. Well, I changed my mind recently, to the pointI now I slightly prefer the 1W scheme. I have had a lot fun reading thatArticle and found it extremely well written and informative. I was just a little surprised by the abundance of seemingly contradictory statements/claims made in some of those Editorialnotes, when the two engineers (assuming they both are) tried to explainwhy they preferred the 2.83V scheme. I wonder what other's opinions/preferences are?

Not sure why you felt the need to start a new thread for this article rather than post on the existing one?

You need to go back and re-read the article b/c you misunderstood several points. The Ideal Voltage Source/Flat Frequency response was talking about the amplifier, NOT the speaker.

1 watt driven only applies to 8 ohm (2.83V), keeping the voltage constant rather than varying it based on impedance keeps things consistent. In the case of 1 watt for 4 ohm that would be 2V instead of a 2.83V drive which with a 4 ohm load would be 2 watts.

I used to be in the camp of varying the voltage based on amplifier impedance by taking an average impedance and setting the voltage to get 1 watt. 2.83V drive (regardless of speaker impedance) is a better option and it's a standard practice most manufacturers adhere to now, especially since its per IEC. The way manufacturers still cheat would be when they test fullbandwidth. I found this to be the case with Klipsch speakers. They are so bass and treble intense that unless you bandwidth limit to 300Hz to 3kHz, you can see 4-6dB higher SPL than a speaker that tends to measure flatter. I personally use b-wt pink noise when level matching speakers for comparative tests.

 

[PENG]

Not sure why you felt the need to start a new thread for this article rather than post on the existing one?

You need to go back and re-read the article b/c you misunderstood several points. The Ideal Voltage Source/Flat Frequency response was talking about the amplifier, NOT the speaker.

1 watt driven only applies to 8 ohm (2.83V), keeping the voltage constant rather than varying it based on impedance keeps things consistent. In the case of 1 watt for 4 ohm that would be 2V instead of a 2.83V drive which with a 4 ohm load would be 2 watts.

I used to be in the camp of varying the voltage based on amplifier impedance by taking an average impedance and setting the voltage to get 1 watt. 2.83V drive (regardless of speaker impedance) is a better option and it's a standard practice most manufacturers adhere to now, especially since its per IEC. The way manufacturers still cheat would be when they test fullbandwidth. I found this to be the case with Klipsch speakers. They are so bass and treble intense that unless you bandwidth limit to 300Hz to 3kHz, you can see 4-6dB higher SPL than a speaker that tends to measure flatter. I personally use b-wt pink noise when level matching speakers for comparative tests.

On the other hand, if you re read my post, no I did not misundestand the points you cited. Of course I understand the 1W 8 ohm 2.83V, and 2W 4 ohm 2.83V, I know P=VI*cosine (phase angle) and its other various forms. In what I quoted, he did refer to the speaker, not the amplifier. How could I be wrong when I quoted it directly?

Anyway, let me quote the whole paragraph, from that article:

Plus, all speakers are designed to be driven from voltage source amplifiers and have a flat frequency response when so driven. Therefore, sensitivity does give a direct indication of how much louder one speaker will sound on a direct comparison (disregarding the capability of the amplifier to drive a speaker that cheats on impedance).

Note, he said driven, not drive so he referred to the speaker, not amp.

I started the thread because I wanted to quote some of the editorial notes in that article but I guess yes I could have stick with your thread, sorry I was thinking. Can you move it and close this one?

 

[gene]

I think he meant the speaker would produce a flatter response driven by a constant voltage source rather than varying it based on impedance to try to always keep it at 1watt. I will ask him to clarify the next time I speak with him.

 

[JerryLove]

I think he meant the speaker would produce a flatter response driven by a constant voltage source rather than varying it based on impedance to try to always keep it at 1watt. I will ask him to clarify the next time I speak with him.

That would seem to be an accurate claim.

Voltage determines volume. Voltage adjusted for impedance determines wattage. A constant volume on a non-constant impedance would be a constant voltage but varying wattage.