Thanks a lot Dave !
I will read carefully his site hoping it will be not too scientific.
Kind regards,
gino
Which of the 'best' speakers produce 10-15% distortion and a what level and frequency?
Just asking because two loudspeaker manufacturers which do give THD figures (Neumann and ME Geithain) show less than 0.5% above 100Hz at 96dBspl (1m).
Even my ancient Tannoys state less than 0.5% at 90dB from 50Hz to 20kHz.
That is not what i said. I said nothing specific about distortion but that on the whole speakers have only reached that far in their potential.
dave
Sorry, I misunderstood and as per usual jumped straight in both feet forward. ;-)
That said: What do you think is or might be releasing the rest of the potential?
I mean may be current technology is already realizing 95% of the potential moving coil drivers are capable of and all that is left in that area are small variations in the set of compromises we choose.
That said: What do you think is or might be releasing the rest of the potential?
I mean may be current technology is already realizing 95% of the potential moving coil drivers are capable of and all that is left in that area are small variations in the set of compromises we choose.
Last edited:
Even in a perfect loudspeaker you will get IMD peaks when playing tones of 50Hz and 300 Hz. As soon as you ask any single diaphragm to produce two tones of differing frequency, IMD distortion will occur.
The great Paul Klipsch did very readable articles on this, and they are available for free at richie00boy's site. I learned a lot from them.
Read Research - Articles
I think we can still go quite a ways by improving moving coil tech.Speaker driver manufacturers are still pretty conservative.
But new tech is probably required.
Most importantly a much better understanding of (and new) measure techand a better understanding of how the ear/brain works and how the 2 mesh.
Also to appreciate much better speaker tech we will need better recording tech.
dave
Definitely. We're still waving bits of paper around with magnets.
I'm still amazed we get anything resembling music at all...
Charles, there's probably still some mileage in moving coil drivers. New materials will keep cropping up, allowing us to shave off a few more grams and stand a few more watts. Mark's been making some impressively thin cones for a while now, for example.
Hi yes of course there is nothing like zero distortion.
But the values are different.
But someone has already explained to me that distortion figures are of doubtful use .. with speakers with higher distortion resulting in a better sound than speakers with a lower distortion ... so i calmed down on the all issue.
Thanks a lot for the valuable suggestion.
I have understood the complexity of the topic.
Of course it would be nice to have a set of measurements usuable for establish the sound quality of a speaker.
I read an article suggesting the drivers selection by ear.
So this is the only way in the end.
Thanks and regards,
gino
It depends on the distortion imo. I remember Geddes saying something along the lines that if a driver measures below a certain threshold then he considers it defective. Now he didn't put it quite like that but the meaning was such that he'd probably consider a lot of the poorer performing hi-fi drivers defective.
As has been pointed out, Geddes designs are inherently very low distortion to begin with. Most pro audio drivers used in a domestic environment will exhibit low distortion simply because they are never going to be used near their operating limits.
If Geddes were to be constrained and forced into designing a loudspeaker using the run of the mill 'hi-fi' drivers I wonder how he'd pick which ones to use. Some drivers exhibit significantly lower distortion vs some others and usually it's the better ones that perform better when pushed closer to their limits. This is very important if you are going to be using your drivers up to their limits.
I mean it's one thing to say that hunting for the lowest distortion driver is meaningless if all the drivers you'd design with are low distortion to begin with. It's another if there are lots of products on the market that are of questionable performance to start with. If Geddes would consider these defective by default (and therefore never use them) then this doesn't dismiss the validity of running distortion sweeps if the goal is to weed out the defective drivers. I too wont routinely buy or design loudspeakers with mid, mid/bass drivers that don't contain shorting rings, or otherwise motor linearising mechanisms, because to do so is a waste of time when those with advanced motors can be had for relatively little.
As has been pointed out, Geddes designs are inherently very low distortion to begin with. Most pro audio drivers used in a domestic environment will exhibit low distortion simply because they are never going to be used near their operating limits.
If Geddes were to be constrained and forced into designing a loudspeaker using the run of the mill 'hi-fi' drivers I wonder how he'd pick which ones to use. Some drivers exhibit significantly lower distortion vs some others and usually it's the better ones that perform better when pushed closer to their limits. This is very important if you are going to be using your drivers up to their limits.
I mean it's one thing to say that hunting for the lowest distortion driver is meaningless if all the drivers you'd design with are low distortion to begin with. It's another if there are lots of products on the market that are of questionable performance to start with. If Geddes would consider these defective by default (and therefore never use them) then this doesn't dismiss the validity of running distortion sweeps if the goal is to weed out the defective drivers. I too wont routinely buy or design loudspeakers with mid, mid/bass drivers that don't contain shorting rings, or otherwise motor linearising mechanisms, because to do so is a waste of time when those with advanced motors can be had for relatively little.
"Distortion" is a very general term. There's several types in both the linear and non-linear categories. Many speaker drivers have a lot of distortion compared to specs on amp circuits, so listing actual distortion figures might reduce sales of drivers that are actually relatively competitive. Standardization of test methods would be difficult as well.
Hello and thanks for your very valuable reply
Only to say that this is exactly my feeling ... of course money is not a problem for a rich man.
The drivers he selects are surely extremely low in distortion in normal listening condition. This is a sane approach i think.
So distortion is not a problem because there is no distortion to speak of.
I am now wondering if the right way is to go with HE drivers, even of the like of those used for PA purposes, that i imagine have very low distortion figures.
The main critic is that the sound is unrefined ... what this means i do not know.
I feel that at least some PA models can sound very good indeed once that their crossover for instance is modified using better quality parts.
One case i have in mind.
An audiophile was mentioning how good the bass of a little Kef speaker was.
Then i found in the italian magazine Audio Review a measurement that shows how this speaker has really no bass at all !!! (... once the maximum accepted IMD is set a 5%).
I think that that audiophile does not know what real bass is.
Thank you very much indeed and kind regards,
gino
An externally hosted image should be here but it was not working when we last tested it.
Last edited:
Yes and most measurements are done in an anechoic chamber already, at 1 meter with different drive levels (1 watt vs 2.83v RMS). I mean they already have the baffle, the room and the mic distance, all they'd need is to settle on a reference level. SEAS used to do all their measurements at 96dB, seems like a good enough figure for me. Then if the drivers were supposed to be capable of ridiculous SPL levels, like the pro audio stuff, they could choose to add in a 110dB measurement too.
It appears that the frequency response on the KEF LS 50 was done at an extremely high SPL, 110db. I don't think this speaker was really designed with that kind of level in mind. Wonder what the curve would look like at, say, 96?
You can't just look at one graph and say, "it has no...", you have to look at the whole thing.
You can't just look at one graph and say, "it has no...", you have to look at the whole thing.
Absolutely and 96dB would do me too.
I find it quite frustrating that the few companies that publish distortion curves do so at wildly different levels.
For example Beyma consistently uses 1W/2.83V while BMS measures somewhere close to Xmax or max power, this could be 25W for compression drivers or 1000W for some woofers.
Mind you anything is better than nothing but I don't know what conclusions I could draw from the curve at near full power as to how the driver performs at more sane, domestic levels.
SoundStageNetwork.com | SoundStage.com | NRC Measurements: KEF LS50 Loudspeakers
According to this test the LS50 reaches or exceeds 10% THD+Noise at 90dBspl for frequencies below 100Hz.
Indeed. And the curve is quite different from the one done at 110 db. I certainly wouldn't expect any speaker with parts as small as an LS 50 to give large amounts of fundamental bass below 100Hz. That is why I use subs under my 110mm two-ways...
I've not heard any current KEF systems, but I suspect that a listening test would show reasonably clean bass above 100Hz, but not deep bass or powerful bass. Not what it was designed for.
I agree that many audiophiles appear to mistake upper bass for "real" bass. I remember demoing early Magnepans, and having customers remark on how good the bass was... Not deep at all.
I've not heard any current KEF systems, but I suspect that a listening test would show reasonably clean bass above 100Hz, but not deep bass or powerful bass. Not what it was designed for.
I agree that many audiophiles appear to mistake upper bass for "real" bass. I remember demoing early Magnepans, and having customers remark on how good the bass was... Not deep at all.
Hi and let me explain the graph. It is called MOL (Maximum Output Level).
The curve represents the max SPL at 1 meter that the speaker is able to give with max 5% of IMD.
We can discuss if 5% is too low ... if we increase at 10% the curve should be a little higher, i do not know how much higher.
I have checked some of these curves and it seems that more or less (i am talking of the MOL below 200 Hz mainly) woofers with same size have similar curve, more or less. 5.25 " is the woofer size of the Kef ls50.
I would pay to see graph like this in the drivers datasheets. and i am not giving away money easily ...
For me they are very telling ... they fix the max SPL i could get.
I love this test.
Thanks and regards,
gino
Only to add the woofer size is important for a great bass ... high SPL with low IMD at low frequencies ... 😉
At least a 12" ... with 15" better 🙄
Have a nice evening
gino 😀
At least a 12" ... with 15" better 🙄
Have a nice evening
gino 😀
The difference in the bass level output at high vs low SPLs could quite easily be attributed to port linearity. As the air speed increases the output from the port also proportionally decreases re output level of the driver, such that you have less bass at higher output levels.
110dB is pretty loud for a tiny driver too. That looks like an octave-band, pink noise stimulus, room averaged type measurement. 110dB with such a stimulus is quite demanding and especially so for a small driver. It is reasonable to assume that they might have bass limited the pink noise a little to prevent driver damage. Then again that might have defeated the point of the test if it has been done at a specific level of distortion. Do you know what frequencies they ran the IMD tests at?
110dB is pretty loud for a tiny driver too. That looks like an octave-band, pink noise stimulus, room averaged type measurement. 110dB with such a stimulus is quite demanding and especially so for a small driver. It is reasonable to assume that they might have bass limited the pink noise a little to prevent driver damage. Then again that might have defeated the point of the test if it has been done at a specific level of distortion. Do you know what frequencies they ran the IMD tests at?
- Status
- Not open for further replies.