The standard measurement for a response of a speaker assumes that the speaker will be used all by itself.
For example, if you buy a midrange raw speaker, then you will put this into a system where it is the sole midrange.
And so a measurement of the frequency response at 1 watt at 1 meter seems reasonable.
But what if you are putting 25 of them into a line array. Now you need to see what the frequency response will be at .04 watt at 1 meter. Its well documented that frequency response changes with different amplitudes of gain. But most people who look at it, look at the corruption of the speaker running at 50 watts and the change in the response with all that heat in the voice coil.
I’d like to see what it looks like with really low volume levels. Additionally, if you measure a 2 inch wide range speaker, the coil is likely to get very hot very quickly compared to a 5 incher. It seems that the size of the speaker will give it an advantage right away up front.
I’ve done several hours of search on the Net using many different search phrases, but have turned up almost nothing.
Anyone have any ideas. I don’t want your speculation. I want to see data.
For example, if you buy a midrange raw speaker, then you will put this into a system where it is the sole midrange.
And so a measurement of the frequency response at 1 watt at 1 meter seems reasonable.
But what if you are putting 25 of them into a line array. Now you need to see what the frequency response will be at .04 watt at 1 meter. Its well documented that frequency response changes with different amplitudes of gain. But most people who look at it, look at the corruption of the speaker running at 50 watts and the change in the response with all that heat in the voice coil.
I’d like to see what it looks like with really low volume levels. Additionally, if you measure a 2 inch wide range speaker, the coil is likely to get very hot very quickly compared to a 5 incher. It seems that the size of the speaker will give it an advantage right away up front.
I’ve done several hours of search on the Net using many different search phrases, but have turned up almost nothing.
Anyone have any ideas. I don’t want your speculation. I want to see data.
Last edited:
Can you show some data that displays how much frequency response changes with amplitude? My understanding is that it doesn't change much
At low power levels, drivers are typically in the most linear region for parameters like Cms, so little change should be expected. If you go at such a low power level that stuff like static friction or hysteresis might (I don't know for certain if they do) become significant, then you better be in an anechoic room if you hope to hear anything.
Temperature at 1W and 0.04W will be about the same, close to ambient temperature, so no change .
Who says so?
"Most" people definitely not.
Universally accepted test level is 1W , for practical reasons: repeatable, easy to achieve, and all drivers can handle it.
1W is low enough to basically not affect measurements.
Why do you say so?
Speaker diameter is irrelevant.
What matters is voice coil thermal resistance, which depends on many parameters: VC diameter , winding length , former material , gap width , magnetic system mass and materials , polepiece geometry, VC venting , type and size of dustcap , ambient temperatute inside the cabinet , closed vs. open or vented , .... notice how many relevant parameters I mentioned, yet not cone/basket diameter.
Maybe there´s a reason you found nothing.
Ouch!!! Sorry Master!!!! We´ll try to do it better next time
Measurements of loudspeaker frequency response is usually done at low power, whether they will be used at higher power or not -- the a average power of music peaking at whole watts will likely be in the milliwatts. The 1W/1m reference often doesn't represent the actual measurement level, the graph is just scaled to those terms to show sensitivity. For example, you might measure a speaker at 1V, then scale the result by 20*log(2.83/1).
I fully understand the way things are done, and how engineers(electrical and otherwise) follow the rules. If they didn’t follow the rules then people would get electrocuted and bridges would fall down.
I’m looking for someone who thought outside the box and wanted to know more. Just because it seems that way in a linear system, it has been shown that speakers are not linear no matter how much you want them to be.
Go here:
Speaker frequency response measurements - Speakers - Audio
I quote from that:
Archived from groups: rec.audio.tech (More info?)
"Ethan Winer" <ethanw at ethanwiner dot com> wrote in message news:<3dadnSKtBLKE_E3dRVn-hg@giganews.com>...
> George,
>
> > how the frequency reponse changes with changing SPL <
>
> Theoretically, the frequency response of a loudspeaker will not change at
> all with SPL. Whether it actually does in practice is a different matter,
> but it's hard for me to see why it would change much if at all.
Oh, please. This issue has been EXCEEDINGLY well documented and
and that documentation is readily available to anyone making a
modest effort to find it, Mr. Winer. Consider, for example:
Gander, M., "Dynamic Linearity and Power Compression in
Moving-Coil Loudspeakers," J. Audio Eng. Soc., vol 34,
no 9, 1986 Sept.
You don't have to wade far into this article, because by the third
page, he compares normalized frequency response curves of the same
speaker measured at 1 watt and 100 watt, which show significant
response differences exceeding, at some frequencies, 5 dB.
Consider, as well:
Henricksen, C. "Heat Transfer Mechanisms in Loudspeakers:
Analysis, Measurement and Design,"J. Audio Eng. Soc.,
vol 35, no 10, 1987 Oct.
Also....
https://www.klippel.de/fileadmin/_m...linearities–Causes_Parameters_Symptoms_01.pdf
I’m looking for someone who thought outside the box and wanted to know more. Just because it seems that way in a linear system, it has been shown that speakers are not linear no matter how much you want them to be.
Go here:
Speaker frequency response measurements - Speakers - Audio
I quote from that:
Archived from groups: rec.audio.tech (More info?)
"Ethan Winer" <ethanw at ethanwiner dot com> wrote in message news:<3dadnSKtBLKE_E3dRVn-hg@giganews.com>...
> George,
>
> > how the frequency reponse changes with changing SPL <
>
> Theoretically, the frequency response of a loudspeaker will not change at
> all with SPL. Whether it actually does in practice is a different matter,
> but it's hard for me to see why it would change much if at all.
Oh, please. This issue has been EXCEEDINGLY well documented and
and that documentation is readily available to anyone making a
modest effort to find it, Mr. Winer. Consider, for example:
Gander, M., "Dynamic Linearity and Power Compression in
Moving-Coil Loudspeakers," J. Audio Eng. Soc., vol 34,
no 9, 1986 Sept.
You don't have to wade far into this article, because by the third
page, he compares normalized frequency response curves of the same
speaker measured at 1 watt and 100 watt, which show significant
response differences exceeding, at some frequencies, 5 dB.
Consider, as well:
Henricksen, C. "Heat Transfer Mechanisms in Loudspeakers:
Analysis, Measurement and Design,"J. Audio Eng. Soc.,
vol 35, no 10, 1987 Oct.
Also....
https://www.klippel.de/fileadmin/_m...linearities–Causes_Parameters_Symptoms_01.pdf
When the speaker is tested, the manufacturer does not tell you the test conditions. It may be in an infinite baffle or something different than what you intend (LA). They also omit things like distortion or IR. Most of the smallish drivers used in LAs have 20w or so Pmax and that seems to be sufficient for the LAs I've read about.
There are many threads in this form regarding the performance of LAs. Post your question in one of their threads as they will have experience and hard data for the particular driver they use. More importantly you can compare the single driver performance to the array performance.
Example : The making of: The Two Towers (a 25 driver Full Range line array) or Full Range TC9 Line Array CNC Cabinet
There are many threads in this form regarding the performance of LAs. Post your question in one of their threads as they will have experience and hard data for the particular driver they use. More importantly you can compare the single driver performance to the array performance.
Example : The making of: The Two Towers (a 25 driver Full Range line array) or Full Range TC9 Line Array CNC Cabinet
Right, lets talk about this.
What's happened here is that you've found out that, when speakers are pushed to their limits, the frequency response changes. At high excursions or high power levels, those effects do come in, and can do so in a big way.
However, those only occur in the last 6dB or so of output, give or take a few dB.
Lets take an example of power compression.
For a while, I used some Fane Sov. Pro 12-500 drivers in my PA system. They're pretty good drivers, with a 3" voicecoil and a large ferrite magnet with a vent through the polepiece.
I ran them from 80Hz up to 1.2kHz, and gave them an amplifier that's capable of 90v RMS output. They did well at a lot of gigs. Plenty of output, sounded good, etc etc.
However, the day came where I used them for a rock 'n' roll gig outdoors, which required considerably more power into those drivers, to a point where I touched the limiters a couple of times. Now, those drivers are 8ohm nominal, but drop below 6ohm for large parts of the frequency range I was using them for. 90v input into 6ohm is 1.35KW, which is a lot of power for a driver whose long-term survival rating was 500w. They were hammered like that for a couple of hours.
By the end of the gig, the back of the magnet was pretty hot to the touch, and I found the system actually sounded quite different - there was too much treble and too much bass.
Turned out those Fane units were around 6dB into power compression, and with the large magnet saturated with heat, it'd be a while before that dissipated.
So, you give drivers a hard time for a while, they get quieter.
Next up, excursion issues.
By it's very definition, any driver acting in its linear excursion range will act linearly. The trouble is that the range where the driver is acting truly linearly can be rather small, despite a larger rated Xmax.
If you play a low-frequency tone and increase the level, it'll get to a point where the level increases only give more distortion. There's another frequency response deviation at high power. The difference is that this is instantaneous, while power compression has much larger time constants.
So long as you're not dropping a lot of power into the drivers (ie, enough to cause heating or produce excursion troubles), they'll act linearly.
You seem to be extrapolating. Just because some things happen at high power levels, it doesn't mean that trend continues all the way down to zero power. The things we're looking at are what happens when you push a speaker to the ragged edge, which should never happen in home HiFi.
Chris
This is what I’m trying to avoid. The assumption is that if you don’t drive the speaker to high power levels and force it to do weird things then its “OK”.
My question is one that I haven’t found anyone to try to respond to without extrapolating is what happens to the frequency response when its driven at very very low levels.
Suppose you drive the speaker at .04 watts. That’s 4 hundredths of a watt. You might not even be able to hear it individually. Years ago(like maybe 15 years), I found a reference from someone who did this, and said that at very very low gain levels in the amp to the speaker, all speakers are essentially flat in their frequency response(I wish I had saved it, and now I cannot find it anywhere). And remember, you have to drive the speaker directly from the amp or you will have passive crossover artifacts(so in a system you need electronic crossovers).
No manufacturer is going to do this, especially if it found that at very low levels even a cheap speaker has a flat frequency response. And no one who makes line arrays will do it, because if it were true then they couldn’t charge $8000 for their speakers with very expensive components.
I’m looking for someone who wanted to find this out by actually measuring and then comparing to the circumstances when it was driven at a low level of .04 watt. WESAYSO loves measuring stuff, so I asked him directly. We’ll see what he has to say.
Sounds good to me..... except, good luck finding this kind of data on any raw speaker, even those that cost a fortune. I’ve never seen it.
Fine.... but I’m not talking about measuring single drive performance at the standard 1 watt/1 meter. I’m talking about measuring it at 4 1/100ths of a watt at maybe much closer, or at 1 meter.
I want to know empirically what driver’s do when almost no power is pumped into them. We already know that the distortion drops through the floor---if only by listening to a line array(if you’ve never listened to one, then your thoughts on this don’t count).
Yeah.... someone who did this at 4/100ths of a watt to find out measurements.
Of course manufacturer's don't publish full Klippel results, but unless something's really weird in a driver's design, it'll always be at its most linear when at low power levels.
4/100 of a Watt isn't such a low level, it's only 14dB down from reference efficiency, probably a fair level for a quieter listening session.
4/100 of a Watt isn't such a low level, it's only 14dB down from reference efficiency, probably a fair level for a quieter listening session.
... distortion increases (at least for the pc83-8)
The FR shape seems to stay the same but the distortion increases. The levels were adjusted at the amp and the soundcard, and the mic needed to be placed at 0.1m to get a reasonable signal at -40db from the 1W reference. The increasing distortion is trending at intermediate stages as well.
The FR shape seems to stay the same but the distortion increases. The levels were adjusted at the amp and the soundcard, and the mic needed to be placed at 0.1m to get a reasonable signal at -40db from the 1W reference. The increasing distortion is trending at intermediate stages as well.
Attachments
Last edited:
Since I cannot measure the output of my speakers in watt’s I have no idea the actual listening.
So if .04 is typical of a quiet listening session, then lets break that .04 into 25 pieces and tell me what the speaker is doing at: 0.0016 of a watt.
Many people tell me that their cheap speakers sound great as long as they keep them at a quiet listening level. Its when they turn them up to a normal listening level that an expensive speaker comes into its own, and the cheap speaker is left in the dust.
But I do not have equipment to do measuring, so I have to rely on someone who did.
Show me the measurements to verify your statement. I’m just making a hypothesis.
That makes no sense to me. I don’t have the equipment or I wouldn’t be asking the question? If I could do the measuring why would I even bother to be asking the question????????
You are making my point for me. If it IS the most linear at low power levels, but they only measure at 1 watt/1 meter, then it might be very very linear at .0016 watt at one meter(something that is typical for a line array system with 25 speakers in the sound range(like the mid range for example)).
Go back to composing great piano pieces.
Interestingly, distortions aren't always linear, they can rise as excursion decreases, have a look at some of the distortion measurements Linkwitz took.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.