Stop the High Pass Filters Madness

[Ottostrom]

I do not completely trust this guys advice so I was wondering what you guys think.
Why would i need to include the information below 50hz on a piano track, or guitar or whatever?
He keeps referring to frequencies like 200hz but isn't the normal tip to cut everything from something like 80hz or under, where you cant even hear it?
And doesn't the phase shifting only occur with some eqs and not for example with the ssl channel eq?

[selig]

IMO, you only need to remove what's causing problems. I would imagine many piano samples already have the sub-sonics filtered out, so no need to do it again. Same for tracks that were recorded in a studio where the filtering was done when recording. This is another reason why full-range monitoring is essential - how can know what's causing issues if you can't hear it? Common practice when removing low 'trash' is to increase the filter frequency until you hear it do something you don't like, the lower it a bit. This is done to make sure you're not removing something important (unless you want to!).

THEN, there's the more creative uses of removing some low energy, which covers just about any other time you cut low frequencies. Most of the time cutting lows and low/mids is done to prevent a buildup of too much energy in that range. I call this a creative decision, mainly because most EQ is done at the discretion of the engineer. There are no rules, except that EQ'ing a frequency that doesn't exist is generally a waste of time!


Sent from my iPad using Tapatalk

[esselfortium]

For me, it varies depending on what the sound is and what sort of song I'm working on. I use a high pass on most things that aren't specifically intended to be bassy sounds nowadays, especially in busy mixes that I really want to have a sharp, well-defined bass section in, and just crank up the highpass frequencies however high I can get away with without noticeably harming the sound.

[TritoneAddiction]

I do not completely trust this guys advice so I was wondering what you guys think.
Why would i need to include the information below 50hz on a piano track, or guitar or whatever?
He keeps referring to frequencies like 200hz but isn't the normal tip to cut everything from something like 80hz or under, where you cant even hear it?
And doesn't the phase shifting only occur with some eqs and not for example with the ssl channel eq?

I've seen the video before. I use high pass filter on pretty much everything, many times on bass and bassdrums too (only set way lower, 20-40Hz). Why keep a bunch of low mess on tracks that doesn't need it.
If I remember correctly he's mostly referring to recorded music. It's not obvious from the video though. Virtual synths is a completely different thing. I think he said that in the comment section somewhere.

[Ottostrom]

IMO, you only need to remove what's causing problems. I would imagine many piano samples already have the sub-sonics filtered out, so no need to do it again. Same for tracks that were recorded in a studio where the filtering was done when recording. This is another reason why full-range monitoring is essential - how can know what's causing issues if you can't hear it? Common practice when removing low 'trash' is to increase the filter frequency until you hear it do something you don't like, the lower it a bit. This is done to make sure you're not removing something important (unless you want to!).

THEN, there's the more creative uses of removing some low energy, which covers just about any other time you cut low frequencies. Most of the time cutting lows and low/mids is done to prevent a buildup of too much energy in that range. I call this a creative decision, mainly because most EQ is done at the discretion of the engineer. There are no rules, except that EQ'ing a frequency that doesn't exist is generally a waste of time!

Sent from my iPad using Tapatalk

But lets say I would always remove everything under 20hz from every channel i create. That would not have any negative effects would it? Even if it would not be needed on most of the tracks.
I feel like the videos is saying that even this would cause problems.

[Benedict]

As Selig says, some sounds have activity under 50hz, some even under 20hz. Sometimes that activity is a useful, sometimes not. The decision has to be made by the engineer, in the mix. Use your ears and know your monitoring and everything should work out ok.

[selig]

There's not often a need to HP everything that's not bass, and that's my takeaway from the video. I'm a "less is more" guy, so if there's no reason to do something, I don't do it "automatically".

I've also come across more than a few folks who HP everything and have what I consider to be overly "neutered" sounding mixes. Plus, there's other more subtle ways to reduce excessive low end besides filtering it. Low shelf EQ is excellent for this purpose and is more likely to sound more subtle in my experience. As a 'shelf', you're simply lowering a range of frequencies rather than filtering them out more radically. You have control over exactly how many dB you reduce the frequency, and over exactly what range you're affecting. Sometimes a parametric does the job even better.

But I tend to reserve filters for fixing problems rather than shaping tone, and there's certainly no rules here!

[dana]

Putting a high-pass at around 30hz is something i will do on all sounds if i don't want to use an eq across the stereo bus for mastering (and a low-pass around 17khz)

I find the mix tends to sound better, i use that as a starting point. Sometimes i will boost the low shelf then put a high-pass on it.. The same with the high shelf, there is no rules. Just listen to the mix, notch out resonances, bring presence to the right elements, reduce dynamics where its needed.

There usually is conflicting information, harmonics from synths usually.

[Ottostrom]

Interesting to hear different views on the subject. Thanks for your input guys!

[Marco Raaphorst]

I keep an eye on everything below 100 hz just because most laptops are not able to reproduce these.

[Creativemind]

As Selig says, some sounds have activity under 50hz, some even under 20hz. Sometimes that activity is a useful, sometimes not. The decision has to be made by the engineer, in the mix. Use your ears and know your monitoring and everything should work out ok.

Well explained there Benedict I think.

deadmau5 did say in a recent video that you should cut everything below 20hz unless it's the bass or the kick drum. This was for electronic music though if it makes any difference. You only really hear anything though on decent studio monitors though don't you.

When would the activity under 50 or 20hz be useful to leave in?

[Marco Raaphorst]

high pass can cause a peak at resonance point. so instead of cutting it, it will simply increase the bass.

[selig]

As Selig says, some sounds have activity under 50hz, some even under 20hz. Sometimes that activity is a useful, sometimes not. The decision has to be made by the engineer, in the mix. Use your ears and know your monitoring and everything should work out ok.

Well explained there Benedict I think.

deadmau5 did say in a recent video that you should cut everything below 20hz unless it's the bass or the kick drum. This was for electronic music though if it makes any difference. You only really hear anything though on decent studio monitors though don't you.

When would the activity under 50 or 20hz be useful to leave in?

When there's nothing there in the first place! In many cases there's nothing there, so the idea of cutting everything doesn't make sense to me.


Sent from my iPad using Tapatalk

[selig]

high pass can cause a peak at resonance point. so instead of cutting it, it will simply increase the bass.

Not sure you mean that every HP filter has resonance - if so, using a 6 dB/Oct filter will prevent this. But we may be talking about different things…


Sent from my iPad using Tapatalk

[Ottostrom]

When would the activity under 50 or 20hz be useful to leave in?

When there's nothing there in the first place! In many cases there's nothing there, so the idea of cutting everything doesn't make sense to me.

But that would not be "useful" to leave in, it would just not make a difference. I know it would save time not to highpass everything (even though it only takes a few seconds) but I think the question here is: when would you want the activity that is there WHEN there is something? What benefit would it bring to the mix?

[selig]

When would the activity under 50 or 20hz be useful to leave in?

When there's nothing there in the first place! In many cases there's nothing there, so the idea of cutting everything doesn't make sense to me.

But that would not be "useful" to leave in, it would just not make a difference. I know it would save time not to highpass everything (even though it only takes a few seconds) but I think the question here is: when would you want the activity that is there WHEN there is something? What benefit would it bring to the mix?

I often find myself asking the question the other way around - what detriment is it to the mix? If there's no detriment, and there often isn't, why take it out? There are many factors that would determine this - the level of the LF energy, the frequency, the correlation (is it a part of the sound or background noise?), etc. For some sounds it can add extra 'heft', especially in a more open and acoustic mix.

My reasoning is that filters cost DSP, and there is no free lunch - filters aren't magical and DO have side-effects. Again, less is more: when in doubt, leave it out. Not trying to change anyone's mind, knock yourselves out with HP filters if it works for you. Just sharing what works for me!



Sent from my iPad using Tapatalk

[Benedict]

Again I am with Selig. Use your ears and judgment in a case-by-case manner. After all that is what separates you as a craftsman from the next guy. This is why a Martin Hannett mix of a Joy Division record is special. If Bob Clearmountain had mixed the same record it would sound different. Apply the filter and use your judgment as to what works better with a straight A/B and even some back later (on other speakers) to see what speaks best.

Chasing the "magic formula" is missing the point (and boat) as great art is about individuality, and all the uniqueness this brings. Something is "wong" only if it reduces the ability to convey the message. Consider something like Meatloaf/Jim Steinman "Bat out of Hell" which breaks just about every rule of having a hit record - too long, too detailed, too show-tuney and a mix that is far from sparkling top or bottom - but is still an extremely popular record (after being one of the best sellers of the 70's). If you want people to want to hear your song again this is the best approach by far (IMO of course for those offended by anyone with a passionate opinion).

[Creativemind]

When would the activity under 50 or 20hz be useful to leave in?

When there's nothing there in the first place! In many cases there's nothing there, so the idea of cutting everything doesn't make sense to me.

But that would not be "useful" to leave in, it would just not make a difference. I know it would save time not to highpass everything (even though it only takes a few seconds) but I think the question here is: when would you want the activity that is there WHEN there is something? What benefit would it bring to the mix?

It was what I was thinking but then just realised it was everything below 20hz I saw. You don't hear that, you only feel it in some instances.

[Marco Raaphorst]

high pass can cause a peak at resonance point. so instead of cutting it, it will simply increase the bass.

Not sure you mean that every HP filter has resonance - if so, using a 6 dB/Oct filter will prevent this. But we may be talking about different things…


Sent from my iPad using Tapatalk

Yes I mean the resonance bump at the cutoff frequency. Probably caused by steep curves, right?

[selig]

high pass can cause a peak at resonance point. so instead of cutting it, it will simply increase the bass.

Not sure you mean that every HP filter has resonance - if so, using a 6 dB/Oct filter will prevent this. But we may be talking about different things…


Sent from my iPad using Tapatalk

Yes I mean the resonance bump at the cutoff frequency. Probably caused by steep curves, right?

Just to be clear, when you talk about filters like in the SSL EQ section, there is no resonance bump - or are you talking about something different? Even those filters that CAN add a resonance bump don't necessarily HAVE to do so. It would be up to the user to add this intentionally, right? Sorry if I'm totally misunderstanding your point…


Sent from my iPad using Tapatalk

[Marco Raaphorst]

high pass can cause a peak at resonance point. so instead of cutting it, it will simply increase the bass.

Not sure you mean that every HP filter has resonance - if so, using a 6 dB/Oct filter will prevent this. But we may be talking about different things…


Sent from my iPad using Tapatalk

Yes I mean the resonance bump at the cutoff frequency. Probably caused by steep curves, right?

Just to be clear, when you talk about filters like in the SSL EQ section, there is no resonance bump - or are you talking about something different? Even those filters that CAN add a resonance bump don't necessarily HAVE to do so. It would be up to the user to add this intentionally, right? Sorry if I'm totally misunderstanding your point…


Sent from my iPad using Tapatalk

Yes you are right, without resonance this is no issue. The SSL EQ won't do that indeed.

[Ottostrom]

Just to be clear, when you talk about filters like in the SSL EQ section, there is no resonance bump - or are you talking about something different? Even those filters that CAN add a resonance bump don't necessarily HAVE to do so. It would be up to the user to add this intentionally, right? Sorry if I'm totally misunderstanding your point…

But the filters on a ssl channel does have a resonance bump according to this video.

[selig]

All I can say is a picture is worth a thousand words: (SSL Filters at default settings, 187 Hz and 3.54 kHz)

[Marco Raaphorst]

Just to be clear, when you talk about filters like in the SSL EQ section, there is no resonance bump - or are you talking about something different? Even those filters that CAN add a resonance bump don't necessarily HAVE to do so. It would be up to the user to add this intentionally, right? Sorry if I'm totally misunderstanding your point…

But the filters on a ssl channel does have a resonance bump according to this video.

Video is confusing. SSL doesn't peak but the HP of the Kong does peak.

[selig]

Kong doesn't HAVE to peak - here's the default setting with Resonance lowered to zero: