What is the impact on amplifier power / SPL from filtering out low frequencies - crossover, subwoofer, LPF

[pjug]

I've come across this video which is shocking in terms of the demands of low frequencies at same SPL as high frequencies.

It's the only analysis, if you can call it that, that I've seen. It has now become the de facto bible of power needs of all audiophiles and yet only has 19,000 views. Does anyone know of others?

Based on this, a crossover of 100hz with a sub may reduce power usage from 600 watts to 40 or less.

I am not sure how you are calculating. All that 40-60Hz energy in the track means that you would relieve the main amplifier of a lot of work by high passing like you describe. But when you apply a 100Hz high pass filter to this file, the peaks are only reduced about 1 dB. So short term power requirements don't change much and you still need an amp that can meet that.

edit: Actually this is another example of the HP filter causing an increase in peak levels, not reducing peakks. I was checking peaks in Audacity with the amplify tool and I missed the negative sign indicating clipping in the filtered track.

 

[xanalog]

That means that if you are using 1 W at 1 kHz, then to sound equally as loud you need 10 W at 100 Hz, 100 W at 100 Hz, and 10,000 W (!) at 20 Hz if the speaker was equally sensitive at all those frequencies.

Hi, @DonH56,
Am I having a double-vision problem or is there a typo (or two) in the above statement?

Appears to be a sound justification for bi-amping, as explained in the webpage for Topping MA4.

 

[voodooless]

Hi, @DonH56,
Am I having a double-vision problem or is there a typo (or two) in the above statement?
View attachment 461432
Appears to be a sound justification for bi-amping, as explained in the webpage for Topping MA4.

What justification? It’s an active crossover system, therefore needs an amp per crossover channel. The rest is mostly marketing fluff. The tweet has a less powerful amp because it’s more efficient. This has absolutely nothing to do with equal loudness.

 

[Sokel]

You can easily find the power needs of each way using line-level signal or the recording itself with REW's RTA.
A nice,clean loopback, voltage calibration and go from there.

You just aggressively (or something resembling your own crossover's settings) HP/LP the section you want and let REW show you the power needed at line level.
Most think that it would be very low. With my music the combined power of 100Hz-300Hz is always far higher than 20Hz-80Hz.

 

[Sokel]

Just a quick and dirty one with custom Pink noise (although 20Hz-300Hz normally looks close to White) :



20Hz-80Hz


100Hz-300Hz

That's why is always better to look at the combined power.

 

[DonH56]

Hi, @DonH56,
Am I having a double-vision problem or is there a typo (or two) in the above statement?
View attachment 461432
Appears to be a sound justification for bi-amping, as explained in the webpage for Topping MA4.

Typo; looking at the original, should be 100 W at 50 Hz, if I read it rightly. Not at my computer right now but the rule for power is a 10 dB increase requires 10x the power, 20 dB 100x, 30 dB 1000x (goes as 10^(power in dB/10) ).

 

[techsamurai]

I am not sure how you are calculating. All that 40-60Hz energy in the track means that you would relieve the main amplifier of a lot of work by high passing like you describe. But when you apply a 100Hz high pass filter to this file, the peaks are only reduced about 1 dB. So short term power requirements don't change much and you still need an amp that can meet that.

edit: Actually this is another example of the HP filter causing an increase in peak levels, not reducing peakks. I was checking peaks in Audacity with the amplify tool and I missed the negative sign indicating clipping in the filtered track.

Doesn't a HP filter mean that the frequencies will not be amplified? Won't that essentially turn electronica into a violin quartet from the amp's perspective? It'd be quite something if my Class A/B amp is spending 600 watts for frequencies that I'm sending to my sub so now my system is using 1,200 watts of power for 600 watts of bass.

 

[pjug]

Doesn't a HP filter mean that the frequencies will not be amplified? Won't that essentially turn electronica into a violin quartet from the amp's perspective? It'd be quite something if my Class A/B amp is spending 600 watts for frequencies that I'm sending to my sub so now my system is using 1,200 watts of power for 600 watts of bass.

Here is the 5-Pan_Sonic_-_Laptevinmeri_Laptev_Sea track in the video. I reduced amplitude by 6dB since filtering causes clipping. The top is without filtering and the bottom is after a 100 Hz 24dB high pass filter was applied. The track has a lot of energy below 100 Hz so you can see how much is removed. But still, some peaks are not reduced and in fact the peaks are higher after filtering. Peak signal corresponds to peak power so you haven't reduced that. You have reduced the average power by a lot so you aren't taxing the amplifier nearly as much, but you still need to be able to produce the peaks at about the same level as before filtering.

 

[DVDdoug]

Doesn't a HP filter mean that the frequencies will not be amplified?

It's exactly what it says. A high-pass filter is a low-cut filter.

Won't that essentially turn electronica into a violin quartet from the amp's perspective?

No, but it depends on the filter frequency and the frequency content of the music. A violin only goes down to around 200Hz. It's also limited on the high end, and it's very dynamic. Electronic music usually covers most of the audible range and it's usually very intense (undynamic) so it tends to require more average power. Your amplifier will likely get warmer with electronic or rock music, even with the bass filtered-out. Classical music may actually require more power on the dynamic peaks because otherwise it might be too quiet.

If you have a full-range speaker, fully capable of delivering bass, adding a subwoofer & crossover won't have any effect on the sound (theoretically) but the power load will be shared.

Actually this is another example of the HP filter causing an increase in peak levels, not reducing peaks.

That's probably the result of phase shifts. Filters usually introduce phase shifts, with different frequencies shifted differently. Since there are normally many simultaneous frequencies, they are combined differently, making some peaks higher and some lower. I believe this "problem" is worse when the audio is "artificially" compressed and limited (and most music is compressed & limited to some extent). Without any processing, the different frequencies should combine "randomly" (uncorrelated) so you might get different peaks after phase shifting but everything should average-out about the same.

Another possibility is that some filters create bumps or ripples in the passband. But I don't think that's common with digital filters. (Digital filters aren't perfect but you can usually make them better simply with more processing power so there can be fewer compromises.)

 

[techsamurai]

Here is the 5-Pan_Sonic_-_Laptevinmeri_Laptev_Sea track in the video. I reduced amplitude by 6dB since filtering causes clipping. The top is without filtering and the bottom is after a 100 Hz 24dB high pass filter was applied. The track has a lot of energy below 100 Hz so you can see how much is removed. But still, some peaks are not reduced and in fact the peaks are higher after filtering. Peak signal corresponds to peak power so you haven't reduced that. You have reduced the average power by a lot so you aren't taxing the amplifier nearly as much, but you still need to be able to produce the peaks at about the same level as before filtering.

View attachment 463455

Is there a watt meter I can put on my outlet and monitor wattage. I'm actually curious to see the use because all my gear is on a 15A outlet and I might be maxing it at times.

 

[pjug]

Is there a watt meter I can put on my outlet and monitor wattage. I'm actually curious to see the use because all my gear is on a 15A outlet and I might be maxing it at times.

What makes you think that?

 

[techsamurai]

What makes you think that?

Audioholics found that the Denon 8500h lost 10% of its power when connected to a 15A outlet vs 20A and the review also pointed out to the fact that it might have affected the 8012 review in the same manner (which is practically same as my toroidal)

I have the following gear:

TV (55 inch) max of 150 watts - that could go up if I switch to OLED and increase size but I prefer focal quality vs peripheral.
Console - PS5 and Series X - max of 250 watts (no PC thank God but that could hit 800 watts)
Class A/B amp (Marantz SR8002) - 5 speakers but technically 7 as I'm bi-amping LR so very heavy LLCRR configuration with very low crossover - I'm guessing I might be hitting 800-1,000 watts briefly.
Class A/B pre/pro (Denon 4800h) - I've no idea what it uses in pre-pro mode. Let's go with 50 watts.
Subwoofer (SVS Micro 3000) - 800 watts RMS (claimed 2500 watts peak)

The outlet is 15Amps (1,800 watts).

I'm connecting a 800 watt (claimed 2,500) Sub and a console and a TV to a 15A outlet that can probably easily be maxed out by my Marantz SR8002.

I should talk to SVS and Marantz about this to hear what they have to say.

 

[pjug]

Audioholics found that the Denon 8500h lost 10% of its power when connected to a 15A outlet vs 20A and the review also pointed out to the fact that it might have affected the 8012 review in the same manner (which is practically same as my toroidal)

I have the following gear:

TV (55 inch) max of 150 watts - that could go up if I switch to OLED and increase size but I prefer focal quality vs peripheral.
Console - PS5 and Series X - max of 250 watts (no PC thank God but that could hit 800 watts)
Class A/B amp (Marantz SR8002) - 5 speakers but technically 7 as I'm bi-amping LR so very heavy LLCRR configuration with very low crossover - I'm guessing I might be hitting 800-1,000 watts briefly.
Class A/B pre/pro (Denon 4800h) - I've no idea what it uses in pre-pro mode. Let's go with 50 watts.
Subwoofer (SVS Micro 3000) - 800 watts RMS (claimed 2500 watts peak)

The outlet is 15Amps (1,800 watts).

I'm connecting a 800 watt (claimed 2,500) Sub and a console and a TV to a 15A outlet that can probably easily be maxed out by my Marantz SR8002.

I should talk to SVS and Marantz about this to hear what they have to say.

I am no expert on circuit breakers but I would think if you haven't tripped the breaker then you don't have a problem. And that the 15A breaker will allow more than that for brief periods.

 

[Oddball]

Circuits seem to be more sensitive upon turn on or turn off of the gear. Otherwise once circuit should be able to handle what you have unless driving it beyond reference levels. If you have an issue you can install 16A circuit that would be more resistant but still good enough to prevent malfunctioning of the wiring in the walls that is a major pain.

Let me guess the response from Marantz and SVS - install separate power line.

 

[DVDdoug]

Audioholics found that the Denon 8500h lost 10% of its power when connected to a 15A outlet vs 20A

A voltage drop would from the resistance in the wiring. Resistance is a combination of wire gauge and wire length and I assume wiring standards require heavier gauge for 20A circuits but we know nothing about the length or the actual wiring. If the wiring to the breaker box is short, it will make almost no difference. If the wire gauge in the wall happens to be the same there will be no difference. (And you can't get a power drop or current drop unless you have a voltage drop.)

MOST audio electronics have regulated power that which maintain constant DC voltage when the AC it's connected to varies by 10-15%. And some point, it can drop too low and the power supply will drop out of regulation.

I'm guessing I might be hitting 800-1,000 watts briefly.

Of course, you can't continuously get more power that comes out of the wall. But power supplies have capacitors that briefly store power that can be used on the peaks. And even if you "pull" a little more than the breaker's rating for a few milliseconds, breakers don't blow instantly.

...It does surprise me that some pro amplifiers are rated for 3000 Watts and they only have "normal" power connections. Maybe you have to look at the labels on the back where the power comes-in....

 

[Sokel]

High amperage amps use single or double IEC C20.

Really high amperage ones use professional 32A ones with the equivalent wiring obviously:

It's one way to tell what's going on under the hood.

 

[techsamurai]

I got a watt meter and here's my testing.

Devices
TV - 55 inch Sony 100 watts
Streamer - not that much wattage
Denon 4800h (Pre/Pro) - 60 watts idle
Marantz SR8002 (amp) - 50 watts idle bi-amping LCR for LLCRR.
SVS 3000 Micro - 15-25 watts idle (I saw different values depending on prior usage or just turning it on)

Speaker Crossovers
Fronts Full Range
Center 60hz
Surrounds 80hz


TESTING

Movies
I did hit 600 watts briefly watching Disney+ Rise of Skywalker in the beginning scene with Palpatine and Kylo Ren. My sub was at -20db - I removed it by lowering volume to -60db (no bass). And it dropped to under 300 watts. So the AVR is idling (more on that later). I didn't measure levels but sound travels a lot through my house and everyone was upset at me - it was 80db to 90+db with highs.

So most of the wattage usage came from the sub.

The TV, the AVR, and AMP idle consume 230 watts so ~70 watts for DD+ processing and sound with bi-amped LR Full Range from the AVR and AMP. That surprised me a bit and I don't feel I'm getting the full monte in movies as I used to with my LR. I have room extension to nearly 20hz so those things should sound like subs or at least play 40hz well.

Music
I also played stereo music (82db for 2 meters away) with a lot of bass (Das Boot DJ Mellow Mix 5th minute) and it was using 250 watts with the sub on. I'll do more testing and maybe a more demanding tune. The family will have to be away.

Gaming
PS5 Astrobot main planet - 480-500 watts at decent levels.


Conclusions
I need to test music more and figure out why there is such a drop without the sub which is expected but my system is using 4 channels full range and should be drawing more power. I would have expected the AVR to be using a lot more power in full range mode.

For most people who do not bi-amp with crossovers at 80hz or 100hz, the demand on the AVR is almost minimal unless you play very loud. I'd worry about the 85 inch TV using more power than your AVR.

As for overloading the 15A outlet. Apparently, the 15A outlet (assuming it's 15A as my electrical panel has lots of 20A, 30A, and even 50A outlets but I'll figure that out too) is thermally limited (not power limited to 1,800 watts), meaning it can exceed 1,800 watts briefly until it overheats and the circuit trips. Since it has never tripped in 20 years, that means I've never exceeded its limits, at least not long enough to do that. So not much point going to 20A since I don't exceed the limits but I still have lots of equipment (console, TV, streamer, pre/pro, AVR, sub). I'll also have to check if the circuit handles the entire living room which would include a lot more devices and light fixtures.

 

[techsamurai]

I did more testing with music - it was uncomfortable even at 82-85 db from 3 meters in a large room especially since I have to be close to the speakers to see the wattage which adds another 6-9 db.

Turn Down For What
The intro to the song hit 400 watts (Pre/Pro - AMP - Sub).

Bad Guy
The intro also hit 400 watts without sub at very uncomfortable volume - I had to lower it 6-8 db. I think the volume increased the demand there.

Idle power is 133 watts.

I'm bi-amping so I'm using double the wattage and full range which is the worst-case scenario.

The AMP is Class A/B which are 50-70% efficient so actual watts for sound are much lower than power drawn. A Class D would draw much less power.

The conclusion is that bass and higher volume definitely need more power.

In movies which can go 20db over normal listening volumes, I can imagine it can hit my LR pretty hard. The AMP has been tested at ~180 watts for 2 channels and 120 watts for 5. If I use 4 channels for LR, that's ~500 watts which in power terms can be 700-1000 watts on a class A/B since they need more power to generate it.

I haven't yet tested a crossover to check the difference of full range vs 100hz and 80hz crossovers but I may do that next.