High/Low Pass Filtering... Errr.. What?

P

PizzImperfect

New member
Hi All,

If you're aware of a guide somewhere which explains this alittle, please just link it up and don't worry about explaining it all.
I record pretty basic tracks, usually just my vocal and my guitar.. sometimes a drum machine.
I've just been reading about 'fine-tuning' a vocal sound and it VERY vaguely touched on using high/low pass filtering to get rid of frequencies that are too high and obviously low...

Just wondering, how is it generally done? And.. if the frequencies are too high/low to hear.. why cut them out ?

Apologies if its a n00b/repeated question!
 
There's the rub -- If there is nothing to speak of (or hear) above/below those frequencies, (A) you're sort of wasting your time and (B) taking enough of those frequencies away in those tiny but "naturally occurring quantities" is what can easily lead a perfectly decent mix to "smallness" (for lack of a better term).

If you're cutting "just to cut" I'd start very conservatively...
 
PizzImperfect said:
Sensible advice there.
So is this essentially just.. EQ'ing out the unwanted frequencies?

Thanks
Click to expand...

yeah, pretty much.
Low Pass allows the lows to pass which is similar to a High Cut.
and vice versa...

An example of where I'll use Low Cuts (similar to High Pass) would be my distorted rhythm guitar. There's not much below 100 hz that I want to keep. To my ears, it just adds a "rumble" and messes with my bass guitar. So I'll set my "Q" at around 100 hz (maybe up to 150) and cut out some of the bottom end of it.
I've found it to add clarity to the mix.
One thing is to make sure you're not tryin to make the track sound good solo'd. Ya don't care how it sounds solo'd. What counts is that the track plays well with the others.

When in doubt as to where to cut (which btw...I rarely boost, mostly cut) you can take the track that's flippin ya shit (solo'd this time) and start by sweeping a BOOSTED notch of EQ across the spectrum. As you sweep this boost across from your lower EQ to the upper end, you'll hear a section that REALLY sounds like ass. That's where you want to cut.

Trial n error but you'll get it. And this is just how I do it. There's a lot of other dudes on here that are much better at this than I am.

Hope somethin helps. :drunk:
 
Quick definition:

High-pass:
Allows the HIGHer frequencies to PASS. In other words, cuts out all the frequencies lower than the one you chose. So for example, if you apply a High-pass filter at 100hz, all the frequencies HIGHer than 100hz will PASS, and all the frequencies lower than 100hz will be cut out.

Low-Pass:
Allows the LOWer frequencies to PASS. In other words, cuts out all the frequencies higher than the one you chose. So for example, if you apply a Low-pass filter at 5k, all the frequencies LOWer than 5k will PASS, and all the frequencies higher than 5k will be cut out.
 
PizzImperfect said:
Sensible advice there.
So is this essentially just.. EQ'ing out the unwanted frequencies?

Thanks
Click to expand...

Correct. But I think the important thing is KNOWING which frequencies are unwanted. You should have a reason for every cut/boost you make. And, "I read it on a web site/book some where" doesn't count as a reason.
 
Sweep like dog breath said and cut the unwanted freqs and boost the few that you want to stand out. A lot of it is artistic so you'll have to experiment and see what you like and works for your music. I takes some time but after a few tries you'll get something that sounds at least decent.

Just for reference when we say sweep, set a narrow Q (range if freqs that will be cut or boosted) pull it up 10 db and sweep across to hear what sounds good and what doesn't then cut the bad.
 
TripleM said:
Correct. But I think the important thing is KNOWING which frequencies are unwanted. You should have a reason for every cut/boost you make. And, "I read it on a web site/book some where" doesn't count as a reason.
Click to expand...
Amen. You can use a high-pass to cut rumble at 40Hz, or you can use it to shape a vocal track all the way up to 800Hz (WELL beyond the fundamental and even the first harmonic) in some cases.

Know what it does, know how to hear it - and it'll tell you what to do.
 
As stated several places above, a low pass filter essentially cuts out high frequencies and allows low frequencies to pass while a high pass filter cuts out lows and allow the high frequencies to pass. Seems kind of confusing but it's not. I often cut low frequencies below 40 htz with a high pass filter. Even though you may not hear them, they restrict your dynamic range when you use a mastering compressor. You can get a much louder mix for example with out going into distortion, by getting rid of these frequencies. Even though you don't hear them, your mastering software figures them into the equation in gauging overall loudness. Hope that makes sense. I have also had the experience of having low frequencies that I couldn't hear on my near field studio monitors come back to haunt me as ugly growling sounds on other speaker systems. I have found it good practice to always use a high pass filter set to eliminate stuff below 40 htz for those reasons.

Dave
 
Q. Why does using a high-pass filter make things seem louder?


If you are going to apply a LPF to an entire track , then use the best eq you've got !!! ( maybe a good linear phase ; just bounce in place or freeze it ) I would'nt go using overly steep slopes either ; a little goes a long way , there is no reason to just abruptly make a steep chop at one end of the spectrum.

The spectral editor in wavelab shows frequency and intensity on a time line ( 3 axis) which is a real good way to see how some tracks end up with resonances or rumble.... but usually it's just in a few places... ( which is why treating the whole track is a little radical sometimes ....) And it happens on stuff where you really wouldn't expect it ( like a cymbal with spikes happening below 500hz or crazy stuff like that )
 
flatfinger said:
Q. Why does using a high-pass filter make things seem louder?


I would'nt go using overly steep slopes either ; a little goes a long way , there is no reason to just abruptly make a steep chop at one end of the spectrum.
Click to expand...
Yaeh but isn't that what a low or high pass filter does? It cuts everything off abruptly, no? If not, then it's not a high or low pass filter, it's just a high or low shelf. Correct me if I'm wrong, but that's how I always understood it.
 
RAMI said:
Yaeh but isn't that what a low or high pass filter does? It cuts everything off abruptly, no? If not, then it's not a high or low pass filter, it's just a high or low shelf. Correct me if I'm wrong, but that's how I always understood it.
Click to expand...



a 12db slope is gentle and slow ..... 48db slope is not real appropriate because it,s
much harder to execute without much more of a phase smearing effect...... every filter design is different ...like the old saying " the devil is in the details" ..........
 
fil3.gif


As you can see by the frequency response graph of the HPF filter, all frequencies below the cutoff frequency are not totally rejected. Instead the filter attenuates the signal more and more as the frequency drops lower and lower. How fast attenuation is as the frequency lowers is called the filter's slope. The slope specifies how many additional attenuation (in dB) that there will be as the frequency shifts beyond the cutoff frequency.
Click to expand...



filter3butterworth8n100.png
 
So,now i'm getting confiused again.Ok...

low pass=high cut...allows the lows to pass cutting out the highs at such and such frequencies

high pass=low cut...allows the highs to pass and cuts the lows at such and such desired frequencies

so then what is a low shelf,peak,and high shelf

I'm using sonar LE 6 and my parametric eq has 4 bands.It has the low shelf,peak,and high shelf settings.I just realized that i could set each band with it's own shelf/filter.I used to think that the filter was for the whole eq.Unfortunately my help button won't work for my plug ins.It don't work with my Vista version of Windows.Maybe it's because i'm using the LE.

Also does each band represent it's own band of frequenecies? I thought each band was independent and you could sweep across the entire bandwidth cutting out several honkers.It seems that when i sweep the number 4 band,the highest,it's noticeably different then sweeping the number 1 band.
 
mcmetal said:
so then what is a low shelf,peak,and high shelf

I'm using sonar LE 6 and my parametric eq has 4 bands.It has the low shelf,peak,and high shelf settings.I just realized that i could set each band with it's own shelf/filter.I used to think that the filter was for the whole eq.Unfortunately my help button won't work for my plug ins.It don't work with my Vista version of Windows.Maybe it's because i'm using the LE.

Also does each band represent it's own band of frequenecies? I thought each band was independent and you could sweep across the entire bandwidth cutting out several honkers.It seems that when i sweep the number 4 band,the highest,it's noticeably different then sweeping the number 1 band.
Click to expand...

I don't know about Sonar, but here's one eq that I use. It's a 4-band parametric with a simple display. In the picture I put a low shelf cut with a 24dB/octave rolloff, a 1 octave wide low-mid boost, a wider high-mid cut and a high shelf boost with a 6dB rolloff. The two inner filters are peak/notch.

eq.jpg

On this eq the filters are all independent. You can move any one to any frequency. But on some eqs the filters only cover limited ranges, overlapping enough to get it done most of the time.
 
PizzImperfect said:
And.. if the frequencies are too high/low to hear.. why cut them out ?
Click to expand...

I'll offer my $.02.

When I reach for a HP or LP filter, it's usually to help multiple tracks sit together. I tend to need it more when multiple tracks were used on the same source (drums, live band, etc.). For example, I might have a bass cabinet mic'd and DI'd. I might HP one of the signals so the lows do not get muddy. I also might use it in a situation where I have a main stereo pair plus spot mic's (like on a live jazz group or something). I just recorded a big band and got a great drum sound in the main stereo pair, though just a little too far back. I used a low-pass on a single overhead drum mic to take the cymbals out and just add a little beef to the drums in the main pair. Worked great.

I tend to avoid it on entire mixes if at all possible. I think Massive's comment about making a mix smaller is articulated perfectly. I believe you do hear/sense those frequencies more than you might think. I think good spectral balance is critical to getting a good mix, and I would certainly shape the high or low end if need be, but not with a HP or LP filter unless something is really FUBAR.
 
bouldersoundguy said:
I don't know about Sonar, but here's one eq that I use. It's a 4-band parametric with a simple display. In the picture I put a low shelf cut with a 24dB/octave rolloff, a 1 octave wide low-mid boost, a wider high-mid cut and a high shelf boost with a 6dB rolloff. The two inner filters are peak/notch.

View attachment 65565

On this eq the filters are all independent. You can move any one to any frequency. But on some eqs the filters only cover limited ranges, overlapping enough to get it done most of the time.
Click to expand...

That a nice simple layout that shows all four band's filters.My sonar doesn't show them all together.I read some older posts and got a little better understanding of a shelf compared to a cut/pass filter.

Seems a shelf is just like having a knobs on a stereo.It'll roll off a frequency to a certain degree and shape the sound.A low cut/high pass filter will drop off at your desired frequency and not let anything below it through.A shelf will only go down as far as you set it and then level off.You can pull it down with a big cut in gain and use it as a low cut filter if your looking to reduce rumble.

I'm not sure how well it would work for the higher frequencies though.You could set a high shelf to level off at a certain degree but probably couldn't completely eliminate certain frequencies from passing through.

A peak filter just gives a bump up or down in the desired frequency then? There would be no sharp slope up or down just a reduction or a boost in gain?

I'm getting it for the most part so i'll just need to work with it some more now that i got a better understanding of what i'm working with.
 
Using filters can help to get more headroom in your mix (it helps to get better separation and you'll be able to push your mixe to an higher output levels)
 
You must log in or register to reply here.
Back
Top