Effects of EQ below 20hz on headphones and amps ?

[odyo]

Hello there. I'm curious about this matter. I hope more knowledgeable people can shed light on this. Since we generally use low shelf filters we boost inaudible below 20hz as well. This frequency range also consumes lots of power. Generally audiophile headphone specs states the capability of 5-10hz.

What happens when we boost this range ? Power consumption should be dramatically increased. I imagine distortion and overall driver performance affected negatively from this as well.

 

[Doodski]

Boosting the low frequencies does boost the power output. Myself I have the 32Hz low frequencies toned down to the point I can just hear the bass lessening and then I dial it back a bit the other way. If I can't hear or feel it then there's little to no reason I can see for even amplifying that. I use a motherboard for headphones and need all the power it can deliver for audible tones.

 

[odyo]

Boosting the low frequencies does boost the power output. Myself I have the 32Hz low frequencies toned down to the point I can just hear the bass lessening and then I dial it back a bit the other way. If I can't hear or feel it then there's little to no reason I can see for even amplifying that. I use a motherboard for headphones and need all the power it can deliver for audible tones.

The thing is we can only hear 20hz but the drivers tries to produce below 20hz. So im curious if we are lessening the driver's performance for nothing.

 

[Doodski]

The thing is we can only hear 20hz but the drivers tries to produce below 20hz. So im curious if we are lessening the driver's performance for nothing.

The drivers mechanically roll off. The lower the roll off the more power req'd and better construction too. Of course attempting to produce low freq while going quasi full range is a bad thing.

 

[solderdude]

What happens when we boost this range ? Power consumption should be dramatically increased. I imagine distortion and overall driver performance affected negatively from this as well.

Not recommended when playing vinyl as the excursions get too big. Because of this, when using dynamic headphones, they get out of linear range and the music riding along on inaudible rumble will get distorted.
With well recorded digital recordings it is not really a problem because there isn't much signal there anyway. There just seems to be when looking at FFT's of music.
With some recordings there is a benefit of EQ'ing in (when there is a roll-off) some subbass.
Examples are movies, some piano recordings etc., organ music.
One would think that for piano this isn't needed and in most cases it isn't. Yet I found that it adds some realism when subbass is there. My suspicion is that you may get to hear the mechanical 'sounds' (very low thuds) when a pedal is pushed

I wouldn't worry about power consumption, you probably don't use as much power as you think.
When you need to EQ more than 10dB I wouldn't bother. It means the driver really won't like it. Safe to do with planars. With something like an LCD-2 or similar you don't need more boosting than +5dB or so. Besides planars usually can handle that even at high SPL.

For any SS amplifier it really does not matter if sublows are boosted as long as you don't clip the amp (you'll hear that). For tube amps it may be another matter.

 

[RayDunzl]

Does the power of a sine wave of the same voltage differ with frequency?

I don't see a frequency component in the calculations.

 

[RayDunzl]

Since we generally use low shelf filters we boost inaudible below 20hz as well. This frequency range also consumes lots of power.

Only if there is a signal in that range, which I think is rare, for the preponderance of content.

 

[wrigglycheese]

I'm VERY new to all this but I dramatically roll off below 20hz using Equalizer APO and Peace GUI thinking it would help put less strain on the headphones etc for the sake of what I can't even hear.

Has it turned out i am on the right track here ? No pun...

I mean if one has super HQ super expensive headphones maybe don't worry about it but if like me I just have the cheapo DT990 atm so it may clean up the sound by basically ditching having them trying to reproduce <20-23hz - this is my theory and crazy to find a thread about just this!

I dramatically roll off 20k+ also, tbh.... Letting my headphones only need to reproduce sound I can HEAR.

I'm here to learn so, tell me I'm an idiot if that's the case.

 

[Zensō]

High pass filters are commonly used in music production. It’s not uncommon to filter out most of the content below 25-30 hz during mixing and mastering, so it’s unlikely your music has much energy below that level.

 

[odyo]

Not recommended when playing vinyl as the excursions get too big. Because of this, when using dynamic headphones, they get out of linear range and the music riding along on inaudible rumble will get distorted.
With well recorded digital recordings it is not really a problem because there isn't much signal there anyway. There just seems to be when looking at FFT's of music.
With some recordings there is a benefit of EQ'ing in (when there is a roll-off) some subbass.
Examples are movies, some piano recordings etc., organ music.
One would think that for piano this isn't needed and in most cases it isn't. Yet I found that it adds some realism when subbass is there. My suspicion is that you may get to hear the mechanical 'sounds' (very low thuds) when a pedal is pushed

I wouldn't worry about power consumption, you probably don't use as much power as you think.
When you need to EQ more than 10dB I wouldn't bother. It means the driver really won't like it. Safe to do with planars. With something like an LCD-2 or similar you don't need more boosting than +5dB or so. Besides planars usually can handle that even at high SPL.

For any SS amplifier it really does not matter if sublows are boosted as long as you don't clip the amp (you'll hear that). For tube amps it may be another matter.

Thanks for the reply. How about extreme eq ? Are there any differences between 1db or 30db eq theoretically ? Lets say i have 2 headphones. Headphone a lacks 1db in 500hz, headphone b lacks 30db in 500hz. So i corrected them with eq. What happens to music in this case ? Is dynamic range harmed ?

 

[TomB19]

You'll feel the output below 20Hz, with a good set of headphones. The headphones will wiggle a tiny bit. It's a bit disorienting, actually. I like it.

I've never tried boosting the low end but I've tried cutting it. My old Sonys used to need a high pass, set to about 40Hz. Low frequency content ruined the entire experience.

I have Sennheiser 58X and had a pair of Hifiman HE4XX for a while. Neither seem to require a high pass. I just run them flat. Headphones designed today are amazing.

BTW, I sold the Hifimans to a buddy who uses them every day. He loves them. I heard them again a couple of months ago and they sounded fantastic (definitely much better than when I had them). The Sennheisers also sound far better than when I first got them. The point being, I believe in break-in so you will want to dismiss my opinion, at the very least, or mount a personal attack. lol!

 

[solderdude]

Has it turned out i am on the right track here

There is no need to sharply filter LF unless you are listening to vinyl, The same for >20kHz. There is no need to filter that.
Most music is CD quality so sharply rolled off anyway. For Hires there is very little 'energy' there so no need to roll that of as well.
Those using filterless NOS DACs run on 44 or 48kHz should use such a filter.
Besides, sharp filters means phase shifts and perhaps 'ringing' in that frequency band.

Lets say i have 2 headphones. Headphone a lacks 1db in 500hz, headphone b lacks 30db in 500hz.

Let's take the extreme one, Lifting 30dB won't help when it is a dip in the FR.
Usually a sharp dip in an FR curve does not mean there is no output there.
When you would be boosting 30dB your pre-amp should be -26dB at least. This means you loose 26dB of output signal and thus your amp would have to have a very high gain if you want to play loud.
When you have a DAC with 100dB S/N ratio it would become 74dB S/N ratio and might become audible in silent passages when playing loud.

Sharp dips should not be compensated, when you need to don't go over 6 - 10dB. Instead buy another headphone that doesn't need such compensations.

 

[waynel]

Does the power of a sine wave of the same voltage differ with frequency?

I don't see a frequency component in the calculations.

No but the amplitude has to be fairly high to hear at 20HZ

 

[odyo]

There is no need to sharply filter LF unless you are listening to vinyl, The same for >20kHz. There is no need to filter that.
Most music is CD quality so sharply rolled off anyway. For Hires there is very little 'energy' there so no need to roll that of as well.
Those using filterless NOS DACs run on 44 or 48kHz should use such a filter.



Let's take the extreme one, Lifting 30dB won't help when it is a dip in the FR.
Usually a sharp dip in an FR curve does not mean there is no output there.
When you would be boosting 30dB your pre-amp should be -26dB at least. This means you loose 26dB of output signal and thus your amp would have to have a very high gain if you want to play loud.
When you have a DAC with 100dB S/N ratio it would become 74dB S/N ratio and might become audible in silent passages when playing loud.

Sharp dips should not be compensated, when you need to don't go over 6 - 10dB. Instead buy another headphone that doesn't need such compensations.

It was theoretical question. Not that i wanna fix sharp dips. I'm curious what are the downsides other than power. What happens to music source since eq does change the amplitude of frequencies. If i understand correctly you are saying preamp reduces the sinad right ? So if i have -40db preamp my 120 sinad setup will be 80 sinad. So we lose power and sinad but do we harm the fidelity of music source ? How precise are these eq softwares ?

 

[solderdude]

The answer was also theoretical of what the effects could be. My other point was that such extremes are not possible IRL.

So if i have -40db preamp my 120 sinad setup will be 80 sinad

It is better to speak of Noise floor, that is part of the SINAD. The noise floor comes closer to the max. signal strength. Still 80dB dynamic range is more than enough for music reproduction so in practice SQ will remain the same but when the amplifier is turned up noise may become audible.
Distortion products SINAD are already so low that they 'drown' in the noise floor or are very close to it.

How precise are these eq softwares ?

That will depend on who wrote the software. I don't use software EQ and have no idea who does it 'best' because it is not in my interest zone.

 

[wrigglycheese]

That will depend on who wrote the software. I don't use software EQ and have no idea who does it 'best' because it is not in my interest zone.

Sir, sorry to hijack. How much audio fidelity is lost do you speculate using something like Equalizer APO - as unless I EQ out some frequency peaks at the top end I can handle very few headphones at anything bar VERY modest volume. I am not one to blast things out, but I am sensitive to high frequencies and only through the discovery of EQ APO am I able to finally enjoy my Beyer DT990 and even the HD560s need some top end peaks sorting out.... And I won;t get started on the T50rp.....

Would love your thoughts.

 

[solderdude]

I think the benefits of EQ (digital and analog) generally outweigh the theoretical loss in fidelity.
So any small artifacts that may arise during the EQ process will be much smaller, if audible at all, than the increase in SQ because of improved linearity of the frequency response.

On the HD560S I only EQ down the 4kHz region about -5dB with a BW of 3kHz on either side. Sounds great to me then. No need to address the treble.
The T50RP (depends on the model and when lucky to have matching drivers) needs quite complex EQ.

 

[RayDunzl]

How precise are these eq softwares ?

Very precise...

I would posit that the EQ functions are (can be) as precise as the Measurement functions, as I am assuming they use the same mathematical models (maybe in different directions) to achieve their results. Modification and Measurement being two sides of the same mathematical functions.

I can throw in a few measurements if you want some "proof".

I've not seen anything unexpected with what I have to work with.

Well, I have, but we aren't talking about Operator Error, are we?

disclaimer: I get to be wrong, so, somebody have at me...

 

[wrigglycheese]

I think the benefits of EQ (digital and analog) generally outweigh the theoretical loss in fidelity.
So any small artifacts that may arise during the EQ process will be much smaller, if audible at all, than the increase in SQ because of improved linearity of the frequency response.

On the HD560S I only EQ down the 4kHz region about -5dB with a BW of 3kHz on either side. Sounds great to me then. No need to address the treble.
The T50RP (depends on the model and when lucky to have matching drivers) needs quite complex EQ.

Thanks for a great reply appreciate it.

 

[RayDunzl]

How precise are these eq softwares ?

I can throw in a few measurements if you want some "proof".

Inconvenient to do a clean measure right now, since the little woman is watching The Demise of American Democracy right now.

The original post was asking about sub-20Hz eq, and I can do that right now.

Here is some low level (barely audible - about like as if the sink faucet were running in spray mode in the kitchen) white noise overlaid on the TV audio (to provide a baseline), with the 18Hz area boosted by 16db with a Q of 20 digitally at a miniDSP OpenDRC-DI, and recorded via an analog output (RCA) on the preamp to a channel on the Focusrtie 4Pre USB:

So, there's "Software EQ" generating the change, with the change being passed into the analog chain after the DAC, and the analog signal being converted back to digital and measured digitally, with good visual correlation of what was requested with what was created, at a frequency less than 20hz.

Signal chain:
REW white noise generator -> USB -> Focusrite Ch 1 (digital) -> Focusrite S/PDIF out (also carrying the TV audio) -> miniDSP OpenDRC-DI with 18Hz digital boost -> DEQ2496 (no EQ active) -> Benchmark DAC2 -> XLR -> preamp -> RCA out -> Focusrite Ch5 input -> USB -> REW measurement