Hi all,
I'm a loudspeaker designer, among other things, and I have a couple of simple suggestions for how to improve the facilities on the DSP. The Linkwitz transform is great to have but its usefulness could be extended quite a bit - and quite easily.
As it stands one can exchange one second order high pass filter with another at a, presumably, lower frequency. You can also set the Q of your new high pass. But what you can't do is tailor a fourth order high pass. Or, at least not a particularly accurate one. The additional high pass section that you can add doesn't specify what Q it has. Is it a 2nd order Butterworth, with a Q of 0.707. In some respects this may be OK by itself, but it doesn't allow one to construct any of the standard 4th Order filter alignments. For instance, if one wanted to construct a 4th Order Butterworth high pass one would need a Q of 0.54 on one section and a Q of 1.31 on the other. With Bessels and other alignments the picture becomes more complicated, with the frequencies of the two sections being different.
The only technically correct filters than one can currently do at the low frequency end are third Orders; this being done by setting your output Q from the Linkwitz transform and adding a first order high pass. (Incidentally the labeling of the fields in the Linkwitz transform could usefully be improved to show explicitly which are the input and output parameters.)
The simple addition of a Q parameter to the 2nd Order high pass would mean that we could have any low frequency transfer function we wanted. This can be very important in terms of cone excursion, especially in combination with the transform. It also has implications maintaining resolution when using the Linkwitz transform since one would then roll off the amount of additional gain required below a certain frequency. (On which subject may we please have some more clarity on the defeat function, and what is the internal volume control. This would also have a bearing on where one sets the LF cutoff.)
I'm not sure what the high pass function, as it presently exists, is generally used for. It may be that it is mostly used a very low frequencies. I can imagine that if one is playing something that originally came on vinyl this could be very useful for taking out warp frequencies and reducing cone flap. In that case it is probably perfectly adequate, though from the perspective of technical correctness it should ideally have a Q parameter anyway. If it did, and was kept as a standalone filter with another high pass being added, then one could have sixth order high pass alignments at low frequencies, which could be interesting.
So, in summary, my suggestion is (along with more thorough documentation and better labeling) that you at least add a Q parameter to the existing second order high pass or, better still, provide an additional high pass function which also has a Q parameter when used as a second order. The second suggestion here would have an additional benefit of allowing one of the filters to continue being used at VLF.
