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

[JeremyFife]

Hi,
I have a question about the impact, in terms of improved power or SPL, resulting from filtering out low frequencies to main speakers. This is for passive speaker systems.

I've been reading about the benefits of integrating subwoofers: applying crossovers and LPF/HPF filters so that the amplification for the main speakers does not have to deal with bass frequencies below the crossover point. The principle is that low frequencies need more power, at a given SPL, than higher frequencies so filtering them out either reduces the power requirement of an amplifier or allows it to drive the remaining frequencies to higher SPL.

Have I made some sort of fundamental misunderstanding?

Is there any (reasonably simple) calculation that shows how much extra effective power an amplifier will gain if it no longer has to amplify bass? I'm thinking specifically about the new WiiM Amp Pro - measured here at 117W into 4Ohms - what would that be if e.g. frequencies below 80Hz were filtered out?

Nicely illustrated by @staticV3 in https://audiosciencereview.com/foru...streaming-amplifier-review.52372/post-2098144

 

[staticV3]

The principle is that low frequencies need more power, at a given SPL, than higher frequencies

Low frequencies do not inherently need more power than high frequencies.

It's just that, in the type of content that us humans tend to enjoy most, low frequencies are present at a higher amplitude than higher ones:


And so they tend to take up most of the power reserves of a given Amp.

so filtering them out either reduces the power requirement of an amplifier or allows it to drive the remaining frequencies to higher SPL.

Have I made some sort of fundamental misunderstanding?

If your Amplifier is clipping, then chances are, by applying a high-pass filter before amplification, you will be able to achieve higher SPL with clean playback.

Similarly, if your speaker is clipping, then chances are, by applying a high-pass filter before or after amplification, you will be able to achieve higher SPL with clean playback.

Is there any (reasonably simple) calculation that shows how much extra effective power an amplifier will gain if it no longer has to amplify bass? I'm thinking specifically about the new WiiM Amp Pro - measured here at 117W into 4Ohms - what would that be if e.g. frequencies below 80Hz were filtered out?

Still 117W, but they can possibly be put to better use by applying a high-pass filter.

 

[JeremyFife]

Thanks, that makes sense - max power is not magically increased.

I get that low frequencies don't in themselves need additional power. My understanding was that they did need more power to achieve a particular SPL, that more power is needed so that we can hear bass at the same level as the higher frequencies. I may not be expressing that correctly.

Perhaps the question becomes: can we quantify the increased SPL available from an amp (into a specified load / speaker sensitivity) if the low frequencies are filtered out?

 

[staticV3]

can we quantify the increased SPL available from an amp (into a specified load / speaker sensitivity) if the low frequencies are filtered out?

Not in isolation.

It all depends on how the Amp handles various frequencies and loads, the type of content that you play and which speakers you play it with.

 

[JeremyFife]

Didn't think it would be easy Thanks for the reply, really appreciated

 

[Old_School_Brad]

You could take that curve and integrate the area below from 30-80 Hz and do the same from 80-20.000 Hz. Assume a 4 ohm resistive load. Assume a sensitivity of 87 dB/W/m from 30-20.000 Hz
I was about to ask chatgpt to do it but it told me its data analysis limit was reached for today.

 

[Blumlein 88]

Well I generated some pink noise, often close to music in general. I implemented a 24 db/octave roll off at 100 hz. Down -3db at 100 hz and rolling off at that rate. The removal of the low end was only close to 3 db in total level.

I did the same for brownian noise. It was a difference of near 8 db.

Did the same for a few bits of music and it was generally 1-2 db. Kind of disappointing.

I need to look at some music and find the loud parts and try the same procedure.

 

[Old_School_Brad]

Well I generated some pink noise, often close to music in general. I implemented a 24 db/octave roll off at 100 hz. Down -3db at 100 hz and rolling off at that rate. The removal of the low end was only close to 3 db in total level.

I did the same for brownian noise. It was a difference of near 8 db.

Did the same for a few bits of music and it was generally 1-2 db. Kind of disappointing.

I need to look at some music and find the loud parts and try the same procedure.

Hmm. I think the difference should be larger.
This is a fresh measurement. The C weighted peak is low frequency and about 20 dB higher than the A weighting. EDM.

 

[Blumlein 88]

Hmm. I think the difference should be larger.
This is a fresh measurement. The C weighted peak is low frequency and about 20 dB higher than the A weighting. EDM.

View attachment 395884

Maybe, but it will always depend upon the music. I could select a solo organ recording and there may be nearly nothing out of the woofer range.

 

[Old_School_Brad]

Maybe, but it will always depend upon the music. I could select a solo organ recording and there may be nearly nothing out of the woofer range.

Of course. It was just to make a different example. It's entirely dependent of the music. And the EQ.

 

[DonH56]

A look at equal-loudness curves shows that deep bass requires 10~30 dB or more to sound as loud as in the midrange. Listening at 80 dB at 1 kHz (very loud), you need roughly 90 dB at 100 Hz, 100 dB at 50 Hz, and 120 dB at 20 Hz to sound as loud as the 1 kHz signal. Every 10 dB increase requires 10 times the power. 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. That is one of the justifications for off-loading deep bass from the main amplifiers and speakers, and use of (often multiple) subwoofers specifically designed for deep bass.

See e.g. https://en.wikipedia.org/wiki/Equal-loudness_contour

HTH - Don

 

[Blumlein 88]

So trying a few other bits of music with drums on them. You seem to get around 3-6 db difference. You can find outliers like the EDM posted above. So I think for general purposes figuring 6 db would be somewhere close to reasonable. 1/4 power. So maybe the upper range speaker would only need 25 watts where it needed 100 watts previously.

 

[Blumlein 88]

A look at equal-loudness curves shows that deep bass requires 10~30 dB or more to sound as loud as in the midrange. Listening at 80 dB at 1 kHz (very loud), you need roughly 90 dB at 100 Hz, 100 dB at 50 Hz, and 120 dB at 20 Hz to sound as loud as the 1 kHz signal. Every 10 dB increase requires 10 times the power. 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. That is one of the justifications for off-loading deep bass from the main amplifiers and speakers, and use of (often multiple) subwoofers specifically designed for deep bass.

See e.g. https://en.wikipedia.org/wiki/Equal-loudness_contour

HTH - Don

I don't think this is the proper way to consider this however. Nothing wrong with the info. The issue is with a given piece of music, if you filter the lows to a sub how much does that free up for the main speaker. Those effects related to how loud something sounds to us at different frequencies is already baked into the mastering done for the music.

 

[voodooless]

This is the average music spectrum of a boatload of music:

So here you see, most of the energy is below 100 Hz. But it also shows still significant energy above 100 Hz and at 200Hz is about 5 to 6dB down so only about 4x less power. A filter at 80 Hz or so would maybe gain you 2dB of headroom. It’s significant, but not massive (about 1.5x less power).

Also note that your sub will need much more power to make up. Usually sub drivers are low efficiency, so need extra power, you can easily need 4x to 10x the power for a sub, especially if you only have one.

 

[DonH56]

I don't think this is the proper way to consider this however. Nothing wrong with the info. The issue is with a given piece of music, if you filter the lows to a sub how much does that free up for the main speaker. Those effects related to how loud something sounds to us at different frequencies is already baked into the mastering done for the music.

I understand, and agree, but most real-time data I have seen shows that recordings tend to include (or boost) the bass to "sound good" to us, which usually means louder, following the curves. In live music in which I have participated, jazz, rock, or orchestral, again the bands (and conductors) tend to work to bring out the bass to serve as a foundation for the rest. A lot of that energy is percussive so steady-state analysis may not reflect the full dynamic range required, of course.

The other issue I have seen over and over is people boosting the bass in their systems, e.g. tweaking levels after running room correction, and often see 10 dB of boost. That's too much for me, but some do even more, preference again.

As for power distribution, I wrote a post about that long ago, but it is a bit hard to follow (I really need to clean it up, just not a high priority for me now).

Bi-amping 101

This article provides a quick look at how bi-amplification works. The conventional scheme is to split the signal into two frequency bands before the power amplifiers that drive the speakers. The bass (LF) amp sees only LF signals and drives only the LF driver (woofer). The treble (HF) amp sees...

www.audiosciencereview.com

 

[pjug]

I would not try to use HP filter to main amplifier as a way to skimp on power. It is hard to know how the peak voltage will change when you do that. Somewhere on here someone showed that a HP filter might increase the peak level in some tracks.

edit: see here https://audiosciencereview.com/foru...r-lower-the-needed-power-for-a-speaker.13779/

I would just choose a power amplifier as if you plan to run it full range.

 

[JeremyFife]

Thanks all, that's really interesting.
Perhaps the biggest benefit of high passing an amp then is to protect it from clipping if you push the volume up.
Also interesting to see @Blumlein 88 's thinking and the possible 6dB boost (if I've read that right)

Love this place

 

[Blumlein 88]

Thanks all, that's really interesting.
Perhaps the biggest benefit of high passing an amp then is to protect it from clipping if you push the volume up.
Also interesting to see @Blumlein 88 's thinking and the possible 6dB boost (if I've read that right)

Love this place

Yes, just a rule of thumb you might get 6 db worth of help.

 

[Old_School_Brad]

So trying a few other bits of music with drums on them. You seem to get around 3-6 db difference. You can find outliers like the EDM posted above. So I think for general purposes figuring 6 db would be somewhere close to reasonable. 1/4 power. So maybe the upper range speaker would only need 25 watts where it needed 100 watts previously.

If we expand our perspective a bit beyond audiophile and classic rock music to include popular genres like hip-hop, EDM, and modern synthesized pop, the differences in bass frequency distribution regarding power requirements are significantly greater. I think a 3-6 dB difference is relatively low when considering contemporary music as a whole.

So, like many aspects of Hi-Fi, the answer is influenced by your personal taste in music.

 

[pjug]

Perhaps the biggest benefit of high passing an amp then is to protect it from clipping if you push the volume up.

But then again the opposite might happen.