solderdude
Grand Contributor
It's somewhat similar to room correction (in FR only) up to 1kHz or so (Philips MFB didn't work as high) but on top of that you reduce distortion in the lows (like seen in NC headphones)
The main question is probably audibility. Those large subs have relatively low distortion, but in absolute terms it still not very low. However sub frequency distortion is not very audible. But fact is that MFB gives you much better control of cone movement, so will objectively beat the same system with only EQ. Would be fun to do a comparison!Large subwoofer elements with low distortion and substantial ability to pump air. It combined with powerful Class D amplifiers, which are many subwoofer amplifiers W for little money. Then measure with a microphone and set the EQ. With this, does not MFB then become unnecessary? Or what do you think?
No matter how interesting this MFB technology is, I wonder. Modern subwoofer elements, which in themselves have low distortion. Large subwoofer elements with low distortion and substantial ability to pump air. It combined with powerful Class D amplifiers, which are many subwoofer amplifiers W for little money. Then measure with a microphone and set the EQ. With this, does not MFB then become unnecessary? Or what do you think?
But Velodyne is a well recognized good brand? I suspect the subwoofer designers have their specifications, goals they want to achieve. Then they start from those that are available to achieve the goals. MFB may or may not be part of the solution.I think MFB is provably unnecessary, because many (most) low distortion and high output subwoofers on the market do not have servo mechanisms in them, and yet they still have state of the art performance. A servo mechanism wouldn't prevent me from buying a subwoofer if it had excellent measured performance, and a servo isn't a prerequisite for choosing a sub either. One of my subs is a servo Velodyne, and while it is has good performance for a sealed sub, non-servo sealed subs are available in the same size range that have as good or better measured performance.
Well, that is debatable, specifically for a subwoofer. You can use a significantly cheaper driver if you use MFB. You don’t need a massive motor. Just lots of excursion and a stiff cone. Add a bit of cheap class-D power and you can have a state of art performing subwoofer.Of course also related to the extra cost of adding MFB
Regarding Subwoofer. The laws of physics are not so easy to get around. Good subwoffer = large.I think the idea that it’s only useful for subwoofer is just wrong. There is far more use in smaller drivers with small enclosures.
That sounds wrong. Of course it is if it is too high. Audible distortion that is.One misconception about distortion in the low end is that it should be not audible. But it is, not in the low register but the harmonics clutter the low-mid range. Most of the recording pro's who use the Grimm LS1 in their recording studio traded in the old non mfb subs for the DMF sub. And not only the Pro's....
Has nothing to do with room correction.
But Velodyne is a well recognized good brand? I suspect the subwoofer designers have their specifications, goals they want to achieve. Then they start from those that are available to achieve the goals. MFB may or may not be part of the solution.
Of course also related to the extra cost of adding MFB
No matter if with MFB or not, at some point you're always exceeding excursion limits at signal peaks and hence you need a limiting scheme. This happens already in the driver when BL drops and suspension stiffness increases at some point.
With MFB, you also need some sort of limiting to prevent forcing the driver to do things it simply can't. The best scheme is to limit excursion in the input signal to the power stage to begin with. For this, one needs to transform the signal into the true excursion equivalent, then soft-clip it, convert back to input signal domain. All this only in the relevant LF region, so a split-band approach. Not exactly easy to implement (DSP helps some, though).
By this, only the LF content is occasionally excursion-clipped and the promised driver distortion compensation by MFB is not degraded at lower momentary excursions. Of course the sensor must be linear to start with and have a higher absolute excursion limit than the driver itself.
As simple and elegant as MFB is in principle, in an actual implementation the devil is in the details.
As for using cheapish drivers and fix them with MFB, I'd say this can be rewarding but it's not a given. Cone breakup and suspension woes are your worst enemy for MFB. But you can get away with lesser motor designs (except that you need to put a good sensor somewhere).
Nevertheless MFB, like any other form of feedback, is best used on already quite linear drivers. Let's say a Purifi woofer, with MFB you could tickle the last bit of LF precision out of that already excellent drivers at low and medium levels. At (overly) high levels, it will (and must, as explained) break into lots of distortion quite abruptly once you max out the momentary distortion. Thermal limits need handling here as well, btw...
As long as the amp power fits for the driver used and starts to clip before it fries the driver this is a completely fine solution, I whole-heartedly do agree. A well balanced compromise.A feedback loop will keep the excursion in control until the amp starts clipping. We use the clip output of the Hypex amps to control a limiter in the dsp control loop. Problem solved.
If you play a single note at 50Hz it is very hard to hear a difference between 1 or 10% distortion several papers state. The main problem is that we never listen to a single note. The distortion of a 50Hz note with 10% distortion in real music will mask the content from 100Hz and up.