XOVER Lipshitz-Vanderkooy

[steph_tsf]

Here is a Flowstone implementation of the Lipshitz-Vanderkooy XOVER.

[tester]

I love, when in the middle of the night, I start a schematic and it hits me and my neighbours with a loud white noise or some scary screaming...

There is another practical aspect of saving schematics as "turned off" by default. From time to time, there are situations when schematics simply crash FS/SM on load (FS is more stable than SM in this area). Turning it off - enables environment to stabilize before you do something.

p.s.: Thanks.

[Magnum Opus]

Does Lipshitz have a little mustachio?
Ah Yes Vanderkooi rings a bell, call me semi ghost.. as in not really anonimous but close.

[steph_tsf]

Hello, here is an update. The scheme looks attractive.

[steph_tsf]

Just updated with the newest FFT-based Audio Analyzer and IIR filter packs. See the three attached .fsm.

[steph_tsf]

Just updated with the 4-channel FFT-based Audio Analyzer. See the attached .fsm

ch1 : reference channel (no need to graph it)
ch2 : lowpass signal
ch3 : highpass signal
ch4 : reconstruction (lowpass + higpass)

reference : ch1
impulse response graph : ch3 (we can also graph ch1 or ch2 impulse response)

For implementing a crossover operating at 2 kHz, we need to configure the Lowpass filter in the LV Xover:
1318 Hz
6th-order
Bessel

We also need to configure the Delay cell in the LV Xover:
14 samples

The resulting graph illustrates the LV Xover behaviour and properties.
Using a 6th-order Bessel lowpass, the Lowpass slope in the transition band looks adequate.
The -6 dB point of the Lowpass is 2 kHz.
The Highpass slope looks adequate.
The -6 dB point of the Highpass is 2 kHz.

When asking the Audio Analyzer to compensate for a delay of 15 samples, we get easy to interpret phase curves.
The Lowpass phase is a typical Bessel one, essentially linear from DC to 2 kHz, appearing horizontal thanks to the 15 samples delay compensation introduced into the 4-channel Audio Analyzer (knob in the left-low corner).
The Highpass phase looks essentially horizontal past 2 kHz.
There is a tiny phase difference between the Lowpass and the Highpass in the transition band, never exceeding plus or minus 5 degree.
Clearly, the speaker drivers operate in-phase inside the transition band. The multiway loudspeaker wont "beam" particular frequencies at particular angles. The radiation pattern stays homogeneous.

Below 2 kHz, there is indeed a phase difference between the Lowpass and the Highpass. The frequency corresponding to a 22.5 degree phase difference (still considered as a benign phase difference) is 900 Hz. At such frequency the Lowpass surpasses the Highpass by 19 dB in magnitude. The contribution of the Highpass can be neglected.

Above 2 kHz, there is indeed a phase difference between the Lowpass and the Highpass. The frequency corresponding to a 22.5 degree phase difference (still considered as a benign phase difference) is 3.2 kHz. At such frequency the Highpass surpasses the Lowpass by 24 dB in magnitude. The contribution of the Lowpass can be neglected.

The Lowpass + Highpass summation delivers a signal exhibiting a flat phase and a flat magnitude, from DC to 22 kHz. Such Xover doesn't distort the phase. A pulse signal will show as a pulse signal, after the crossover. This a decisive advantage compared to the Linkwitz-Riley Xover that we find in most multiway speakers. The Linkwitz-Riley Xover disturbs the phase to the point that a pulse signal will appear as a burst extending over time, after the crossover.

So, what's the catch of the LV Xover?
Is there any drawback? Why are not all digital Xovers proposing the LV Xover modality?
Actually you can't define the Highpass slope per se. Provided the Delay is optimally configured, the Highpass slope is a pseudo 3rd-order Butterworth. See the attached .fsm
If you dare proposing a LV Xover modality in your next digital Xover, you will face customers having selected the "6th-order Bessel LV Xover", pretending that your software is malfunctioning as the Highpass slope is not a 6th-order.

Cheers,
Steph