The impact of high-pass filter on the step response

[ernestcarl]

So when I look at the phase traces in the 4 measurements in your set, I'm bewildered about their phase traces. They aren't anything I know how to work with.

I created a synthetic minimum phase version IR from the sub's target curve (orange trace) to serve as a very simplified control comparison.

In reality, I do not use the "flat" sub equalization version (blue trace) shown below since I equalize for multiple seat positions in the room at large and not just a single MLP spot. It's just there to illustrate what the equalized in-room magnitude response might look like if the very LF adhered to the idealized target more rather than less.

"Align IR start" function was used to auto-adjust the time offsets between the three separate IRs so as to appear closer with each other.

Turns out, things aren't nearly as bad as it seems if you try to see through past the room-induced kinks in the response:



1/6 smoothing default and Log (dB SPL) was used in the wavelet spectrogram.

 

[ppataki]

Finally I had time to perform some more measurements!
I am attaching them for reference

Here is what I did:
- measured the system at the MLP with low shelf applied; no high-pass
- then applied high-pass filter at 23Hz with 24dB/octave slope using CraveEQ, Acon Equalize and Fabfilter Pro Q3
- each EQ was applied first in minimum phase mode then in linear phase mode (with the exception of Acon where an additional mixed phase mode is there)

In short: the results are all pretty much the same regardless of which EQ I used
MP HP filter causes a lot of post-ringing and also shifts the energy curve to the right whereas LP HP filter barely impacts the spectrogram (very light addition of pre-ringing, not audible)

Feel free to inspect the measurements, I hope I did everything right
Any further comments or suggestions are welcome

https://drive.google.com/file/d/1UB1VKM6YBb0PKZfDk0sXH4KdewGFmznq/view?usp=sharing

 

[ernestcarl]

In short: the results are all pretty much the same regardless of which EQ I used
MP HP filter causes a lot of post-ringing and also shifts the energy curve to the right whereas LP HP filter barely impacts the spectrogram (very light addition of pre-ringing, not audible)

"A lot of post-ringing", BTW, is really only a thing in the zoomed-in graphs. A more balanced macro look at things says that the effect of your (not very steep) filter(s) is not all that significant -- to me.

 

[Head_Unit]

It is the hp filter causing ringing in the time domain...is it audible? Yes, indeed. A very low hp at 10hz is audible, on a full-range system, with the right content. Bass will sound sort of slower, sluggish, and there seems to be more bass. But it is not audible, if the content does not have energy in the affected frequency range, and any negative effects of the filter is insignificant compared to distortion caused by over-excursion.

- This expresses what I've thought about ported boxes, and maybe why I prefer them under-tuned
- Ah...similar would apply to sealed boxes? Low low resonance better? Got other stuff on my mind today, no room for that higher math...
- Hmmm...you have ringing from the speaker, then ringing from the room which *might* overshadow the speaker if the speaker is tight?

 

[Kvalsvoll]

- This expresses what I've thought about ported boxes, and maybe why I prefer them under-tuned
- Ah...similar would apply to sealed boxes? Low low resonance better? Got other stuff on my mind today, no room for that higher math...
- Hmmm...you have ringing from the speaker, then ringing from the room which *might* overshadow the speaker if the speaker is tight?

Yes, typical resonances from the room makes any ringing from the speaker itself insignificant. You may still hear a difference between a low-q roll-off and a steep high-q, but compared to removing a removing a room resonance using eq, this difference has no practical influence on the sound overall.

The speaker roll-off can also be fixed using eq, if the room is very dead in the low region, it may make some difference.

The problem with closed boxes, especially smaller ones, driven by drivers with high moving mass and weak motors, is the non-linear behavior that causes the response you see measured as a small-signal system to change dramatically when output is increased. This has a far more significant effect on transient response than a slightly larger ringing from a portet design, because the acoustic loading from the port increases capacity and reduces cone travel.

 

[ppataki]

Finally I have managed to actually be able to hear the difference that I visualized in my first post.
Here is how to do it:
- Download this file
- Play it and listen to how it sounds without any HPF
- Now apply a non-linear-phase traditional HPF let's say at 20Hz with 48dB/oct
- You shall hear that the peak will sound like it is 'doubled'; there will be a delayed ringing audible; instead of 'tam-tam-tam', you will hear like 'tadam-tadam-tadam'
- Now change that same HPF to linear phase - the sound will go back to 'normal'

So that post-ringing introduced by the non-linear-phase HPF is indeed audible under certain circumstances. I am not saying that I can hear it when listening to normal music but this test shows that there could be an easily audible difference....
On the other hand the pre-ringing that shall be present with the linear-phase filter is not audible - or at least I cannot hear it
I have tried this with multiple solutions: Jriver's own PEQ, CraveEQ, Fabfilter Q3, etc. and the results were always the same

Let me know what you guys hear; I would be very much interested in your experience too

 

[dasdoing]

just because you have something in the negative on the wavelet, doesn't mean it's pre-ringing.
here is how a dirac impulse looks like

the peak energy time curve is what matters

 

[ernestcarl]

Finally I have managed to actually be able to hear the difference that I visualized in my first post.
Here is how to do it:
- Download this file
- Play it and listen to how it sounds without any HPF
- Now apply a non-linear-phase traditional HPF let's say at 20Hz with 48dB/oct
- You shall hear that the peak will sound like it is 'doubled'; there will be a delayed ringing audible; instead of 'tam-tam-tam', you will hear like 'tadam-tadam-tadam'
- Now change that same HPF to linear phase - the sound will go back to 'normal'

So that post-ringing introduced by the non-linear-phase HPF is indeed audible under certain circumstances. I am not saying that I can hear it when listening to normal music but this test shows that there could be an easily audible difference....
On the other hand the pre-ringing that shall be present with the linear-phase filter is not audible - or at least I cannot hear it
I have tried this with multiple solutions: Jriver's own PEQ, CraveEQ, Fabfilter Q3, etc. and the results were always the same

Let me know what you guys hear; I would be very much interested in your experience too

I’ve done listening tests with different HP filters with my sub before. 48dB/oct is definitely audible to me and degrades the sound a little bit. Hmmn… I do not really remember hearing this double beat you speak about… will test your track later when using my PC and headphones. Nowadays, if I can get away with it, I’d much rather use either 12 or 18dB/oct. However one can also compensate with an FIR — even if only partially.