Hello Lbstyling and friends,
Guillaume of LUPISOFT kindly responded to me informing;
"EKIO uses IIR filters. The processing is done using a cascade of second order transposed direct form II biquad sections. Every calculation is done using 64 bit floating point numbers."
Since I am not an expert in mathematical and theoretical design of audio crossover, the theoretical comparison between EKIO and other XO solutions would be far beyond my knowledge and capabilities.
I would like to share, therefore, only these two articles;
http://www.differencebetween.net/science/difference-between-iir-and-fir-filters/
and
https://www.advsolned.com/difference-between-iir-and-fir-filters-a-practical-design-guide/
I like the "Advantages, Disadvantages" descriptions in this article, especially "Analog equivalent advantage for IIR";
in this article, this portion would well fit for what Guillaume of LUPISFT informed me;
View attachment 65496
I would really welcome some of you may join this thread to describe about the theoretical, mathematical and physical basics of EKIO in comparison with other software XO solutions, even though the actual algorithm of EKIO's internal processing is not open for public.
In any way, I am always much impressed by EKIO in terms of functionality, simplicity, GUI design, unlimited I/O channel numbers, flexible I/O in full ASIO routing, fast and light CPU processing, very small physical memory consumption, and most importantly its wonderful sound quality including phase features.
IIR induces post ringing (less audable as the sound masks the ringing) FIR minimum causes preringing (ringing sound comes before the actual sound event- more audable), FIR linear phase (perfect). As far as i'm aware, that's basically it. Given the choice, FIR linear is the one you want, but it is only possible with reasonable amounts of computing power. Some MiniDSP units can do a limited amount of it. Most PCs can do all you wish.
As to how audible this is and what the threshold is, I personally have tried all 3 with the same crossover and EQ settings and to me, I am comfortable it is clearly audible, but evidence for this conclusion (double blind tests) are not conclusive as far as I am aware.
To use it, I would recommend JRiver and the 'convolution' filters exported from Pos's RePhase software. (he is a member here). If you use JRiver's EQ filters, they are FIR minimum phase, so I get a 'softening' of the sound, and some strange effects when too much EQ is applied. Subjectively, the effect is much worse with narrow Q filters particularly when I go over 7.
Last edited: