I don't see the words "sensitivity" or "efficiency" mentioned in that statement. If anything they are suggesting that at one time efficiency was used as one measure of a loudspeaker's capability, and now sensitivity is used instead.
It never was that way. Sensitivity and efficiency are two difference things. Stating that at one time they were the same is like stating that at one time the terms volume and area meant the same thing, but now they're different. Because a manufacturer incorrectly uses one term instead of another does not make it the same.
Here's another thing to think about - the quoted efficiency of a speaker is basically a "ratch" - because it varies with impedance and the speaker's actually impedance typically varies in its passband, and this impedance also varies with frequency. The sensitivity rating however does not vary with impedance. The term "nominal impedance" was actually dreamed up to assist with being able to state one figure for efficiency, rather having to do so for every frequency the speaker would see, it being the "effective impedance over that part of the working frequency range where the average power is likely to be greatest". I think this definition dates back to the 1970's if not earlier (Benson makes mention of it in his "Theory and Design of Loudspeaker Enclosures".
The terms are often mixed up unfortunately. Even in the LDC, which is like the "Bible" for loudspeaker design.
And that's exactly where I differ from you, because I believe it's possible to take a reasonable guess as to how much difference they would make, if one is familiar with the impact of such things.
Let me give you an analogy. Take a look at my vehicle. It's top speed is about 100MPH. With added turbo and a few other tweaks to the engine, I can probably get it to about 125MPH. But there's no way I can get it to 200MPH with those simple tweaks. Something more drastic will need to be changed. So, if someone comes along and tells me that they managed to get vehicle that's the model as mine to go to 200MPH with just adding a turbo, I would view that with a great deal of skepticism
.
Well, that's a bit disappointing. What if the "stub" at the end is significantly longer than pictured? What if there are more internal panels in the horn than those that are pictured?
If the picture of the horn's internals are not accurate, then any sim based on it will also not be accurate.
You don't see the word "sensitivity" in there because that word is over to the left. This is their definition of the word "sensitivity".
An externally hosted image should be here but it was not working when we last tested it.
Listen I don't care what the formal definition is. JBL and Harman both say I'm right (or at least I was right until the industry changed the definition). Maybe JBL, Harman and I are all wrong. I couldn't care less. As I pointed out, even in your own link, the guy said sensitivity changes with impedance so impedance must be specified. Remember I pointed that out?
That's from your own link. So apparently Harman, JBL, and this guy from your own link are all wrong. But whatever. This isn't a big deal. I'll continue to reference all my graphs to impedance and call it sensitivity.
My point is that it's unintuitive, mostly useless and downright dumb to measure sensitivity at 2.83 V regardless of impedance. IMO of course. This doesn't help anyone. A 1w/1m rating is incredibly helpful because it evens the field and you can compare like to like instead of comparing vastly different things. The only thing a 2.83 V rating is useful for is for people that can't figure out how to turn the volume knob and need to know what the spl is at 2.83 V. IMO of course.
Last edited:
I'm not sure I follow why you think the speaker's efficiency rating is more useful.
1. If you take two speakers with the same efficiency, but with one having half the impedance of the other, and hook them up to an amp, the one with the lower impedance will be 3dB louder than the other.
2. If you take two speakers with the same sensitivity and different impedances and hook them up to an amp, they will have equivalent SPL output. That's because the sensitivity rating is not dependent on impedance.
Seems to me that the sensitivity rating is a much more useful one to use.
Yup, again, let's throw all the data we have out the window because it might not be 100 percent accurate. I'm being honest and trying to help TB46 here.
The truth is that when there's 50 pics or more on the internet showing the internals and they all agree to a very high percentage of accuracy it's very likely that there's at least some truth there.
Your approach on the other hand seems to be that the data isn't important at all and we should just start drawing things into the plans that clearly are not there in an attempt to make an wildly inaccurate sim match a measurement that also may or may not be accurate.
We know very well that Danley uses low compression ratios and fairly traditional flare shapes. There isn't likely to be a bunch more panels than are shown in the pictures I posted.
I dunno, it seems that someone who appears to be most interested in "accuracy in simulation" wants to get one done on an internal diagram that may or may not be accurate.
What's that saying you used again? Oh, I remember: "garbage in - garbage out"
.So, you're still choosing to ignore then the fact that it produces an Fb that's just one 1Hz off of Danley's TH115 that supposedly shares the same layout?
Spec sheets are a marketing tool aimed at consumers. Consumers speak in watts, not volts. Consumers know how many watts their amps put out and how many watts their speaker can take (at least based on the specs the manufacturers give, which are not always realistic). Consumers want to know how loud the speaker is with 1 watt input, not 2.83 V input. Consumers have no idea what their amp voltage rails are or what the voltage capacity of their speakers is. For the most part consumers don't know how voltage applies to speakers at all. They see a higher sensitivity rating and they think that speaker is better because that's what they have been told is better. It's really that simple. You disagree and that's fine.
I showed half a dozen links that say the same thing - sensitivity used to be based on a 1w/1m measurement, regardless of whether this was the true definition of sensitivity or not - these links all say that used to be the definition before the industry changed it and it was DEFINITELY how they presented the sensitivity spec regardless of the actual definition of sensitivity. I think the industry changed it to put higher numbers on spec sheets and confuse the consumers. Regardless of definitions, they used to use one standard and now they use another, this new standard conveniently allows them to use much higher numbers. The old way was better even if it doesn't fit your definition.
Last edited:
I dunno, it seems that someone who appears to be most interested in "accuracy in simulation" wants to get one done on an internal diagram that may or may not be accurate.
What's that saying you used again? Oh, I remember: "garbage in - garbage out"
What do you propose then? Buy a cab and cut it up to see what's in there and get perfect measurements? I agree that would be best but realistically why not use the data we actually have instead of making things up? TB46 showed that the two drawings I posted are within a high degree of agreement, doesn't that suggest to you that there might be something there that could be of use? Why not throw ALL the data out and just redraw it by hand and make guesses? Add V plates and bifurcated paths going back and forth through the throat and just assume that's more accurate than the drawings on the internet that all almost perfectly overlay? Seriously, what do you propose? We KNOW for a fact that Danley likes low compression ratios and fairly traditional flare shapes but you insist on making up stuff that isn't in the plans that would violate these imperatives to account for a difference between a measurement and a sim that isn't even close to accurate.
I don't ignore it at all, in fact it should be pretty similar. The 115 doesn't have nearly as much mass loading as the 118, there's no stub, the second mounting baffle is completely absent leading to less path length through the throat area and less mass loading in the last segment, the driver is a lot smaller, and it's entirely possible that the 115 was measured without the grill and the 118 was measured with the grill.
There are some details that can't be known but that doesn't mean you can't take the data at hand and properly simulate the cab based on this data. Maybe it still won't match the measurement and there are several valid reasons why that may be the case but it will surely be more accurate than your sim that has several glaring errors and deficiencies. I thought you were interested in an accurate evaluation of the cab and you absolutely cannot get there by guessing.
Last edited:
Hmm, going back at least 60 yrs [and a lot more from memory, but don't have the really old docs anymore], speaker efficiency has 'always' been defined as X dB @ Y [m]watts @ Z distance, while sensitivity was an amp rating: X V RMS @ Y W output, so sensitivity has 'always' been independent of impedance in ratings AFAIK.
I guess Just Big & Loud got successful enough to make a change and the industry followed.
GM
Last time I checked, there are little or no amps that have their volume control levels specified in either watts or volts.
...and they don't need to. With the efficiency rating, they have no idea how loud one speaker would be compared to another when the volume control is at the same setting because the efficiency rating is also impedance dependent. If however they use the sensitivity rating, a speaker that's 3dB more sensitive than the other will measure 3dB more sensitive at the same volume control setting, assuming that the amp is not exceeding its max power output, of course.
Finally, as speakers get more and more non-linear as power input to them increases, there really isn't no easy way to determine how loud a speaker would get based on its efficiency and the amp's power output levels. If we want to talk about misleading things, I'd say right up there would be that a one speaker that's more efficient than another would achieve higher peak output levels, because we know that there's a lot more involved than just the speaker's rated efficiency.
Interesting. In a previous post you suggested that since I did not appear to use the same flare rate as in the picture, my sim would be off. Which one is it?
I referred to the likely impact of these things previously. They will have some impact, but having a 20% impact on a layout that you now agree should be "pretty similar" seems to me to be a bit of a stretch. Of course I base my position on the supposed accuracy of the pictures wrt possible driver position and stub size, which it now turns may not be accurate at all.
C'mon, that's grasping at straws now!
If there are some details of the build that can't be known, then how can the cab be properly simulated? It seems to me here that you're being a bit choosy about what can and what cannot affect the accuracy of the sim.
Like I said, I'm not arguing formal definitions.
But I think this is what most people are used to seeing on a spec sheet regardless of formal definition. This is from PE, one of the biggest US distributors, this is what people are used to seeing, and this rating is more useful than a 2.83 V rating specifically because this is what people are used to seeing and they understand it.
An externally hosted image should be here but it was not working when we last tested it.
Sometimes it's called sensitivity, sometimes it's called spl at 1m, sometimes it's called other things but it's almost always 1w/1m.
Anyway, with so much disagreement on the topic (I showed 6 links that said it is or was based on 1w/1m, one link said there's many ways to determine sensitivity and Brian's own link said sensitivity changes with impedance) it's best to just give the customer something they can understand and not need to do mental math to compare 1w/1m for different impedance speakers.
Your sim is off. It's not accurate. "Pretty similar" doesn't mean accurate, it means pretty similar.
Your "pretty similar" sim is for the 115 which has shorter path length, and way less mass loading in several areas in the cab. I don't have a notarized document from Danley saying the drawings I posted are 100 percent accurate, but realistically do you think they are that far off? They agree so closely with one another and other drawings on the internet, how far off can they be? They are a lot more accurate than you wildly inaccurate 118 sim that much is certain.
I'm not suggesting grill on / grill off is a factor here, I'm suggesting it COULD BE. These are details that can't be known. Were you there when the cab was measured? If not don't you think it's prudent to consider all the options?
You are trying really hard here to twist my words and intentions when I've made it really clear that the only thing I care about is accurate simulations. And using the best data instead of wild guesses is appropriate IMO. No I can't guarantee the drawings I posted are 100 percent accurate but I will go out on a limb here and state that they can't be very far off. A LOT closer than your wildly inaccurate 118 sims that has way too little path length and several areas of mass loading not included.
This is really getting ridiculous. We don't agree. You aren't even remotely interested in trying to accurately simulate the cab you are trying so desperately to figure out. You prefer to invent things that are not in the plans, things that go against Danley's style completely to make a wildly inaccurate sim match a measurement that may not even be accurate. I don't even get it.
you can't get that from a picture that's "pretty similar" to the actual layout. You need the actual layout. To me, it sounds like you're being pretty selective by what you mean by "accurate".
Just an idea: the driver in the TH118 may be mounted a bit further back in the horn than the one in the TH115. This would shift S2 closer to S1 and move the "pinch" further up the horn because it's now no longer at S2. That's going to make it impossible to sim in HornResp I think, as HornResp does not provide a means to insert a pinch at that point.
I'm off to a club tonight (party time, though I'm likely still going to "talk shop" with the owner about upgrading his subwoofers, LOL). Tomorrow I'll have a closer look at this change to see if a reasonable approximation can be made for this by starting the horn at the pinch and treat everything afterwards as a "throat chamber"....
By "accurate" I mean using the available data to be as accurate as you can possibly be, not making things up and simulating path length as chambers and disregarding features of the cab like mass loading in several areas. This should be exceedingly clear at this point. Your sim has several problems and it can't possibly be accurate. I already asked what you propose as an alternative. Buying a cab and cutting it up and measuring? No. Clearly you want to go the other way completely, and disregard the dimensioned drawing that are readily available and instead make up your own flare, use chambers instead of properly simulating the flare, and completely ignoring the fact that there a large amount of mass loading in several areas of the cab.
I'm going to need to go to bed, but maybe tomorrow we can make another 100 posts that all say exactly the same thing again. Same time tomorrow? You can twist my words some more and continue to pretend the dimensioned drawing is less accurate than your wild speculations on what might be in the throat that the drawings don't show. It'll be fun for the whole family.
I'm going to need to go to bed, but maybe tomorrow we can make another 100 posts that all say exactly the same thing again. Same time tomorrow? You can twist my words some more and continue to pretend the dimensioned drawing is less accurate than your wild speculations on what might be in the throat that the drawings don't show. It'll be fun for the whole family.
Lol, this cab is impossible to sim with Hornresp regardless of where S2 is. You absolutely cannot make a reasonable approximation by fudging in chambers. Where's the reasonable approximation for the several areas of mass loading? This needs to be done in Akabak and no amount of tweaks to your sim is going to be remotely accurate.
Hi just a guy,
I just wanted to illustrate as clearly as I could, how Hornresp takes the throat chamber length into account in offset driver simulations, and how the chamber can contribute to the effective overall horn path length. I guess I was a little concerned that the comment in Post #83 ("in an Od sim the throat chamber is in parallel with the horn path") could perhaps confuse some people. The throat chamber is certainly not treated that way in the Hornresp simulation model, and it is probably better to think of it as entering the horn at a 'T junction', as shown in my previous attachment 4.
Kind regards,
David
Ok, T junction is a fine way to put it. By parallel I just meant it wasn't "in series" or part of the path itself. My point was they both stubs can't add length.
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.