Can active DSP speakers play louder when they are high-passed?

[kemmler3D]

I would expect a DSP speaker to play louder if you HP the input in almost any case. Even if you only limit at the input, your overall signal level is going to be lower after the highpass, meaning you can increase the gain and make the rest of the signal louder.

I am actually having trouble thinking of a scenario where you wouldn't get a net gain in the mids/highs this way... possibly if the bass is already heavily compressed / limited / filtered so a reasonable HP doesn't get you any headroom. Possibly if there is an overall SPL limiter that is (for some reason) already so low that bass output is already not a limiting factor. (would be very odd.) Basically, highpass should always give you a little extra gain as long as the speaker is well implemented. IMO...

 

[KSTR]

Protection can be done in many different ways and the idea in studio monitors is to keep them alive during all kinds of typical mishaps, things that will simply happen in a studio, like microphone feedback, a synth playing 3kHz square wave at full power by accident, measurement sweep at full blast, any sort of setup error (it's easy to get an electronic feedback loop in a large studio console with the wrong routing), etc.
Usually the protection via limiting will not necessarily catch extreme and long-term abuse but still there should be basic protection, muting the speaker or switching it off before it gets cooked too much.
The thermal modeling I mentioned often is simply looking at the amp input or output voltage, squaring it roughly (as power is V^2/R, impedance variation often can be neglected), integrating it and use that as the threshold for a hard-knee compressor. For simplicity, often a global VCA before the XO is used.
(Mid-)Woofers often can run without any special limiting protection when the amp max power is choosen wisely. In an extreme emergency it will simple shut down before the voice coil is in danger, and modern woofers are mechanically protected and cannot be destroyed by overexcursion.
That scheme can be expanded to more complex analysis of the state a driver/amp is currently in but the net benefit isn't that much compared to the increased cost and design effort (plus, the more auxiliary electronics you have, the more it is prone to failure). With DSP you have a lot more options without increasing cost too much.

 

[Jukka]

From my experience with my DIY dsp-active speakers: all woofers need EQ to reach target bass extensions, so if you can high-pass the signal within the speakers own amp, it can play louder with less power. Depending on your device and it's capabilities you may also attain higher max SPL since less or no EQ is eating the signal sensitivity away.

I have no idea how active speakers are generally made, which is the SPL limiting factor, but my speaker elements are rated for higher power than the amp modules can give, so my limit is the voltage gain of the amps. So if you high-pass signal going into the speaker, it may consume less power, but can still hit the limit of voltage generation ability and clip.

-EDIT- PS. Regarding limiting of different frequencies, my speakers has a very high sensitivity high frequency compression driver with a horn that has sensitivity of 110 dB/W, so while the mids and woofers may have better power handling, the HF will always be able to play louder.

 

[Frgirard]

Normally I'd be the person answering this question, but I'm stumped.

Traditionally if you take your passive speakers and filter out the bass, you can then let them play much louder than they would otherwise. You can accomplish this filtering using any number of methods from DSP to a sock, but they all work.

However, for contemporary monitors from say Genelec or Kali, will sending them a high pass signal enable them to play any louder? My understanding is that these speakers have limiters implemented in the digital processing, so sending any loud signal will trigger the limiter.

Am I mistaken? If I use a high passed monitor will it actually be able to achieve a high spl? I ask because in theory some of these speakers should be able to provide significant output, but I don't know if the limiters prevent this or if they are smart enough to know that no bass is present.

it's depends
The KH 420 has differents protection.
One is a subsonic filter at 9Hz.
Other is a thermal protection.
so if you play loud at 30 Hz what happens ?

 

[gnarly]

Regarding the bandwidth of the limiter, it seems logical that it would just look at the total level being sent to each driver after the crossover.

Yes, each driver/amp section typically has its own voltage limiter(s), preferably with both rms voltage (thermal) and peak voltage (excursion) limiters.

In other words the dsp splits the signal in 2 or 3, then monitors total output on these channels after all filters are taken into account. That would explain why the entire tweeter comes down when it limits.

But it would be rare indeed to limit all the sections, based on one section going into limit. Each section is limited for it's individual protection only, as who knows what the spectral content of the music is (or dropped mic !)

It's interesting that the woofer is often the driver holding back the speaker back. We really do ask a lot of these little high excursion midwoofers.

Amen !
Imo, one of the least understood and most important fundamental facts about speakers is the simple physics about driver displacement requirements vs frequency, for equal SPL.
The need for 4X the displacement for equal SPL, for each octave decrease, needs to be burned into our mellons!
It does 90% of explaining why speakers sound strained when the volume is turned up....the spectral balance simply has to shift up, when the low end can't keep up.

* 4X per octave decrease for normal piston-driven direct radiators. Horn loading can of course increase efficiency, as a substitute for displacement.

 

[gnarly]

The thermal modeling I mentioned often is simply looking at the amp input or output voltage, squaring it roughly (as power is V^2/R, impedance variation often can be neglected), integrating it and use that as the threshold for a hard-knee compressor. For simplicity, often a global VCA before the XO is used.
(Mid-)Woofers often can run without any special limiting protection when the amp max power is choosen wisely. In an extreme emergency it will simple shut down before the voice coil is in danger, and modern woofers are mechanically protected and cannot be destroyed by overexcursion.
That scheme can be expanded to more complex analysis of the state a driver/amp is currently in but the net benefit isn't that much compared to the increased cost and design effort (plus, the more auxiliary electronics you have, the more it is prone to failure). With DSP you have a lot more options without increasing cost too much.

Good clear explanations.

Limiters have been one of my fascinations, working with DIY "hi-fi" PA muti-ways. I've put built-in amp limiters on the scope in roll-mode, and studied predictive modeling of driver sizes / ratings, to try to gain confidence in making my own settings in DSP.
Current processing platform (q-sys) allows three voltage limiters. RMS and peak both use traditional voltage level, time above level to enact, and hold time. Instantaneous uses 1 sample limiting. There also the ability to do any manner of side-chain limiting (as below for example)

Here's probably my best unusual limiter design. I built a vented sub using a Faital 18FH500 18" driver, designed for f-3 at 35Hz.
I knew it would need a high-pass filter being vented, but i was surprised to find over excursion could also occur higher up centered around 60Hz. (when really crankin ! )

So, I put together a peak limiter, that is frequency dependent, and with selective frequency limiting.
Worked like this.

What I think is cool about it, is there is essentially no high-pass in effect at moderate listening levels.
And then limits according to potential excursion. (The usual RMS thermal limiter is also in play)

I went on to verify the predictive line level limiting from the DSP was working properly out of the amp.
Here's the amp's output at various signal levels. Relative drive levels are in upper left.

And the measured peak voltage amp output, per frequency on the right.

I think it get's all there is to get from this sub, cleanly and safely. Hope this was of some interest .......

 

[617]

Would it be reasonable to assume that an inexpensive DSP monitor has the processing ability to do some kind of predictive limiting on each amplifier channel separately?

@KSTR 's comments on the reasons for limiting on monitors is interesting, he points out that monitors are subject to bizarre and unintentional signals (feedback, poorly configured synthesizers, XLR cables plugged into a wall socket) and the limiters are as much for that as they are for people just playing them too loud.

@gnarly thanks for summarizing how limiters work in the digital domain. My question is this - by what mechanism does the speaker know to limit LF voltage spikes more than HF? Also what kind of maniac runs a 18" pro sub to excursion limits at home at 60hz??

 

[gnarly]

Would it be reasonable to assume that an inexpensive DSP monitor has the processing ability to do some kind of predictive limiting on each amplifier channel separately?


@gnarly thanks for summarizing how limiters work in the digital domain. My question is this - by what mechanism does the speaker know to limit LF voltage spikes more than HF? Also what kind of maniac runs a 18" pro sub to excursion limits at home at 60hz??

The mechanism for actives, is that there is a separate limiter for each the LF and HF. LF limiter(s) catches LF thermal & spikes. The HF does the same for HF.
A limiter is typically the last component in line level signal flow, before the signal goes to the amp channel/driver.
So for a multi-way active, the limiter works after the xovers and EQs and whatnots in the DSP, and only effects one driver/amp channel.


Now for a passive, a predictive line level limiter presents a dilemma like you ask about.
Since the line level signal feeds the entire passive speaker, a choice has to be made what to protect.
A common setup is to set the limiters for the sub (you pretty much have to, to get the sub to be able to play up to its capability).
And find some way to protect HF drivers after the amplifier output.
Iow, some type of limiting device between the amp and the HF driver. Like light bulbs in the old days.
And gizmos like these nowadays...https://4donline.ihs.com/images/VipMasterIC/IC/LFSI/LFSI-S-A0011119789/LFSI-S-A0011119904-1.pdf?hkey=6D3A4C79FDBF58556ACFDE234799DDF0


maniac? i resemble that remark

 

[617]

Here is additional information from Kali:

“It’s triggered by how much the input voltage is being amplified. There is thermal protection if the amp runs too hot but that’s a completely different issue and will just shut the entire unit down until the temperature drops. The limiter only cares about the voltage.“

So it seems that explains how the limiter is implemented more or less.