Subjective Blind ABX Test of EnABLed FF85WK - Round 6

[Bigun]

...the doppler effect where the cone is producing say 100 Hz and 3000 Hz simultaneously ...

If the cone is pistonic then both the 100Hz and 3000Hz are movements referenced to the magnet, a fixed assembly. The 3000Hz can not be said to be riding on top of the 100Hz and there is no Doppler effect. The question is whether the cone is pistonic at 3000Hz or if it's in break-up mode where indeed, it is riding on the cone.

 

[freddi]

speaker with 9 "full range" drivers and active EQ

if one takes a little FR outdoors and measures it with two mixed sine waves, will there be sidebands? and if so, are they FMD?

 

[keantoken]

If the cone is pistonic then both the 100Hz and 3000Hz are movements referenced to the magnet, a fixed assembly. The 3000Hz can not be said to be riding on top of the 100Hz and there is no Doppler effect. The question is whether the cone is pistonic at 3000Hz or if it's in break-up mode where indeed, it is riding on the cone.

If the cone is moving pistonically at 100Hz, it doesn't matter whether the 3KHz is pistonic or not (this only determines where on the cone it radiates from and how much is radiated), it still radiates from the cone at it's position as determined by it's place on the 100Hz cycle. It does not radiate from a fixed position as that would require the cone to be still. It would have to be radiating from behind the cone or in front. Even in 3KHz breakup mode, the cone position at 100Hz is the same, otherwise it is by definition not pistonic.

 

[Bigun]

I don't understand what you are saying.

I'll say in a different way what I was saying.

If the 3kHz signal applied to the voice-coil causes the cone to move in a well controlled manner at 3kHz relative to the fixed magnet, the 3kHz signal is not 'riding on top of' the 100Hz tone (it is 'riding on' the fixed magnet). There will be no Doppler.

If the 3kHz signal applied to the voice-coil excites a 'parasitic' 3kHz oscillation of the cone surface, relative to the cone surface (not the fixed magnet) then the parasitic oscillation is 'riding on top of' the 100Hz tone. There will be Doppler.

 

[wesayso]

How do you apply two tones at once to a single driver without having the higher note riding on top of the lower one.

To produce 100 Hz, the voice coil has to move, combine that 3 KHz tone and it will oscillate faster while its still moving steadily to produce the 100 Hz tone.
Just look at the wave shape, that will show how the voice coil will be asked to move.
That will mean the 3 KHz tone can't "only" move relative to the fixed magnet. As it's already asked to move to produce 100 Hz.

(Sine) Wave shape of a 100 Hz signal with a simultaneous 3 KHz Sine Wave:

Analysis of the above wave shows the 100 Hz and 3 KHz peaks:

See the similarity to freddi's graph? Only that graph shows the 'by products' when the cone is asked to do both of these tones simultaneously. The harmonic distortion.

If, to produce the 3000 Hz tone, the voice coil would only move relative to the magnet, what movement would create that simultaneous sound it is asked to do at 100 Hz? They can't take turns...

Just a quick test of 2 tones to show it's wave shape. Please let me know why the 3000 Hz tone shouldn't be considered as "riding on top" of that 100 Hz tone.

 

[keantoken]

You are using terminology in a confusing way and I think you are suffering for it.

The source of motion in a loudspeaker is the force of the voicecoil against the magnet. No matter what frequency, the motion starts at one point, that is the voicecoil. Whatever position the voicecoil is at, the sound originates at that position, relative to the magnet. So when the voicecoil is 1/2mm forward, 3KHz sound will be coming from that same position. When it moves back 1/2mm, so will the 3KHz source. There is no other point where the force can come from.

 

[Bigun]

Yeah, it's still in need of better words, I'm not expressing myself at all properly for a native speaker am I :-(

Let me put it another way then - I see a difference between these two cases:

a) A woofer cone moves at 100Hz relative to the woofer magnet. The listener is stationary wrt the woofer magnet so she hears the 100Hz as a clean 100Hz tone. If the listenerr were moving wrt the woofer magnet she would hear a distorted tone. Now glue a tweeter to the woofer cone so that it's dragged around at 100Hz and ask it to produce a 3kHz tone. The tweeter dome moves at 3kHz wrt the tweeter magnet. The tweeter magnet is moving at 100Hz with respect to the listener so she hears a Doppler distortion of the 3kHz literally riding on the woofers 100kHz motion.


b) Alternatively, consider a perfect single full range driver that produces both tones. The cone follows the waveform that wesayso posted above, relative to the magnet. The listener and the magnet are not moving wrt to each other, so the listener hears the 100Hz and 3kHz tones undistorted, just like the FFT wesayso shows.

 

[wesayso]

There is strong evidence that distortion, unless it is gross, is simply not audible. There are many reasons for this. Masking is big one. See attached figure (from Toole's "Sound Reproduction"). Masking works upwards, so that a bass frequency can mask higher frequency sounds.

But let's put that aside for a while and assume that there is gross intermodulation distortion going on in our full range driver. So much that it is clearly audible. IMD can be caused by the motor, spider, surround and other elements of the driver, but the dots surely cannot affect anything other than the cone. And if we are talking about the cone, then we are talking about IMD introduced by the doppler effect where the cone is producing say 100 Hz and 3000 Hz simultaneously (quite likely in a full range driver, less of a problem in multi-way systems).

In such a situation, the 3000 Hz tone will be modulated by the 100 Hz tone. The modulation happens because the 100 Hz tone is making the cone move at a certain velocity and the 3000 Hz is riding on top of the 100 Hz tone. How then can the dots affect this situation? The distortion arises from the low tone and the high tone and their relative speed at which the cone is moving. I just don't see how the dots can affect this situation.

There is other mysterious story that is quoted by the DDR proponents about low-level sounds being reproduced better in the presence of high-level sounds. Again, masking plays a significant role here. A -40db 3 kHz signal will be completely masked by the 0 db 2 kHz signal. Doesn't matter if the loudspeaker faithfully reproduces it or not. You won't hear it. And again, it is difficult to see how the dots can have any effect on the outcome.

But putting all the theory aside, one can measure the IMD in drivers. It should be pretty easy to do. My bet is that it will not show any difference. And the ultimate test is what X was trying to do here. We can only hope that he can get his hands on two identical drivers, one with and one without the dots.

I'd dare to say in some full range drivers, distortion can attribute to the sound of that driver.
Normally it won't if distortion levels are down far enough, but just check some drivers out there (mainly metal cones) that exhibit a peak in 3rd order distortion. The shape of that 3rd order distortion can be found on the top end FR trace of that driver, creating a sense of "detail" that isn't present in the original piece of music played trough it.

An example of such a full range driver would be an Aurasound NS3:

Look at the shape of the 3rd order peak at 2.2 KHz (purple trace, partly hidden behind the red THD trace). Now look at the shape of the FR curve starting at ~6.5 KHz... try and compare the 2 shapes closely. So each time this driver is asked to play something at 2.2 KHz, it will alter (enhance) the tones it is asked to play around 6 KHz. This may "seem" to bring out extra detail, while actually it is a by-product of its level of 3rd order distortion.

Now if adding dots can bring that peak down, that would/could provide hearing more actual detail.

This is a case where for me, this driver would be considered to have too much distortion. And by far this isn't the only one showing behaviour like this.
This plot is borrowed from: http://medleysmusings.com/aura-sound-ns3-193-8a1/

 

[norman bates]

yup, Frequency modulation distortion (FMD), bass making voices sound garbled, the enemy of many of us here.

And harmonic distortion, Even a single frequency has harmonics and sub-harmonics (I think) from the cone compared to what is leaving the amplifier.

More drivers = less cone motion = less harmonic distortion.
I think I read in a jbl pdf that when you double drivers, excursion halves, and distortion drops to a quarter.

I know 9 x 4" is way more detailed than a 12" with a whizzer, but the F3 of the 12" will be usually way lower than the 4". And 9 x 4" really lights up the room with sound (can be a good thing).

But yes, modulation distortion, even I hear it past normal listening volumes (a bit past tv levels) on my 12" run wide open with 4db boost under 200hz. I can cross it over at the receiver and it will clean up the mids, but then the phase wrap sounds odd (especially on bass guitar and drums) to my ears.

 

[keantoken]

a) A woofer cone moves at 100Hz relative to the woofer magnet. The listener is stationary wrt the woofer magnet so she hears the 100Hz as a clean 100Hz tone. If the listenerr were moving wrt the woofer magnet she would hear a distorted tone. Now glue a tweeter to the woofer cone so that it's dragged around at 100Hz and ask it to produce a 3kHz tone. The tweeter dome moves at 3kHz wrt the tweeter magnet. The tweeter magnet is moving at 100Hz with respect to the listener so she hears a Doppler distortion of the 3kHz literally riding on the woofers 100kHz motion.

Where the magnet is doesn't really matter much because it's just what the voicecoil pushes against. The same way the road is what a car pushes against. It's the cone, or the car that is moving and producing the signal. A car produces doppler distortion because it is both making noise and moving at the same time relative to the listener. It is no different with airplanes even though they are not pushing against any road.

Separate woofers and tweeters will not cause doppler distortion, but we don't need to know anything about the position of the magnets to see this. We only need to know that the sound sources are each stationary.

b) Alternatively, consider a perfect single full range driver that produces both tones. The cone follows the waveform that wesayso posted above, relative to the magnet. The listener and the magnet are not moving wrt to each other, so the listener hears the 100Hz and 3kHz tones undistorted, just like the FFT wesayso shows.

The magnet is not the sound source, it's job is pretty much the opposite, to stay in place hard enough to push the cone around. Wherever the cone is, that is where the sound source is, and if it is 1/2mm forward at the time, cone vibrations at that moment will radiate sound from that position. And when it swings to the -1/2mm position of the 100Hz sine wave, any 3KHz vibrations of the cone at that moment will come from a -1/2mm position.

The forces applied by the voicecoil are not delayed if it gets further from the magnet. If input was delayed by the distance from the magnet at the speed of sound, then your argument would be correct. However the force on the voicecoil is instantaneous, there is no delay between current applied to the coil and force applied by the magnetic field.

 

[norman bates]

oops.
post #109 of mine.

Inter-mod distortion, that's where bass (say a 100hz sine wave) makes the voice on a full range driver sound garbled.

 

[freddi]

fwiw I ran the old weak motor and long discontinued Beta15cx in a reflex the size of the original Karlson K15's rear chamber and system tuning (~48Hz) vs the same speaker in K15 with 32Hz + 160Hz 1:1. The Karlson's sidebands were at least 10dB lower than that of the reflex. Although K15's fb is around 48Hz, there's a near cone null at ~ 37Hz. Cone motion of a M151 in K15 at 37Hz, 20vrms was less than half that of the reflex. I don't think that shows in X's Akabak sims (?) I don't believe in speaker magic but ...
:

 

[rjbond3rd]

Wherever the cone is, that is where the sound source is, and if it is 1/2mm forward at the time, cone vibrations at that moment will radiate sound from that position. And when it swings to the -1/2mm position of the 100Hz sine wave, any 3KHz vibrations of the cone at that moment will come from a -1/2mm position.

As the cone is in motion from the 100Hz tone, isn't the 3kHz tone modulated by the swinging? Or are you saying that 1mm is too small to be noticed?

 

[keantoken]

1/2mm is just a number I made up for the sake of explanation. The point is, we have a sound source that is moving from point A to point B. How can it do that without causing doppler distortion?

 

[rjbond3rd]

Everyone is in agreement on that.

 

[keantoken]

Check post 107.

 

[technophile50]

There is also the effect of changes in the strength of the magnetic field that the voice coil is in with displacement - usually decreases either side of the neutral position - which will amplitude modulate a high frequency signal being reproduced while the cone is undergoing high excursion low frequency signals - e.g. http://www.alpine-usa.com/assets/images/products/SWR-T10_12ThinFactorForce.jpg

The slope of this speakers BI through Xo will cause modulation (sidebands) at +- f; the concavity will cause sidebands at +-2f.