Does low frequency content lower your headroom? Why?

  • Thread starter Thread starter cordura21
  • Start date Start date
cordura21

cordura21

New member
If you multi-compress is it better to compress harder on the low part, or the reverse?

Cheers, Andrés
 
Only compress as needed (no rules to follow really). So, if you've got some wild low end stuff going on, then compress it more on the low end, if you've got some sibilance or crazy high end, then compress more in the high range.

Do what sounds good.

And to answer the question in your subject...Yes, low frequency content will suck up headroom if you you've got low end stuff that isn't useful. A kick drum will produce frequencies down as low as probably 20hz, but do you want those frequencies present? If not, don't compress, filter it out with a high pass filter at the frequency where you want to start cutting things off.
 
thanks for the reply.

I asked cause I remember, but can't a find, a thread of Sonusman -I guess- where he said about separating into his and lows and compressing one or the other. Anyways, I read that in his big thread and in others that too much low end will eat the headroom more quickly than other frequencies, and tried to understand it better.

Andrés
 
Recording meters don't really give a hoot WHAT frequency is producing energy, they just show how much they are producing.

Bass frequencies take many times more power to appear to sound as loud as midrange frequencies, thus, when you are trying to balance your sonic landscape in a mix, about 80% of the overall energy is dedicated to frequencies 200Hz and below. Inversely, stuff 4KHz and above only make up about 2% (50% on modern recordings it would seem! :)) of the overall energy.

Relative loudness. A great study for any aspiring engineer.

Creepy
 
thanks, I think it has to be in part cause of the Fletcher Munson effect, right?

You say meters don't tell you much, how about spectrograms? There are some that even offer a "power" view.

Sorry, I know my reasoning is flawed cause I never got really right what "power" is, and have vague ideas about it.
 
cordura21 said:
thanks, I think it has to be in part cause of the Fletcher Munson effect, right?
Click to expand...
No, not primarily, though this is a factor.

The main reason is because our perception of loudness approximately scales with the amount of energy contained in the sound wave, not just the amplitude (height of the wave) . The energy is a function of both the amplitude and the frequency. Hi frequencies have more energy for a given amplitude than low frequencies. So for the loudness to seem even across the frequency spectrum, the low frequencies must have larger amplitudes than the hi frequencies.

barefoot
 
Yes, the Relative Loudness Curves have EVERYTHING to do with this.

barefoot, what you just described IS what the relative loudness curve is all about. Our ears are most sensitive in the 2-7KHz region. If you played a 2Khz test tone at 40dB, a 200Hz test tone would have to be like 55dB to be percieved as loud as 2KHz is at 40dB to our ears (This is not an exact comparison from the relative loudness curves. I don't know them by heart!)

In recording, we have a ceiling on how loud you can record something. Obviously, if you have a lot of bass content in the mix, you will eat up a lot of "headroom". Compared to a mix with less low's, this mix will sound much quieter because everything has to be turned down on the bass heavy mix to keep the input of the next device from clipping.

Creepy
 
Actually, I think you guys are both kinda right, but not all in the correct places. The F-M Curves refer to the way that the ear perceives the relative loudness of different frequencies AT DIFFERENT OVERALL LOUDNESS LEVELS of the program material.
So while this does manifest itself more in the lower frequencies more than anywhere else (trying turning a bass-light mix really loud and see for yourself what our friends Fletcher and Munson were talking about), it doesn't really have an affect on the original question to this thread.

I think what Looney Tunez said about filtering out the low lows that we won't hear anyway is right on. F-M Curves enter into this question only as an interesting side story. Meters don't care about either one of them.

Ken Rutkowski
www.OuterLimitRecordingStudio.com
 
Creepy said:
barefoot, what you just described IS what the relative loudness curve is all about.
Click to expand...
No, the first order principle is that loudness perception is essentially an intensity (energy/time/area) perception rather than just an amplitude perception. This directly addresses the question of why bass frequencies take up more headroom than high frequencies. This is also why pink noise sounds much more spectrally balanced than white noise. Pink noise has a flat intensity spectrum.

But, pink noise doesn’t sound perfectly spectrally balanced, because our ears are not simple linear intensity meters. So, equal loudness curves like Fletcher Munson help refine the first order principle. Of course, Fletcher Munson contains the basic "loudness relates to energy" principle, but this is not what it’s really trying to convey. It’s trying to convey the deviation from the first order loudness principle.

Even Fletcher Munson needs refinement though, because it doesn’t account for the temporal dynamic aspect of loudness. For example, a short duration pulse will sound quieter than a long pulse of equal intensity and spectral balance.

Anyhow, this is bit of a semantic argument. Look at it either way you feel more comfortable.

barefoot
 
great answers, thanks everybody.

So from a practical point of view, what's a better approach when doing multiband compression?

Here's my guess: it's better to compress bass so the amplitudes are not so highs, yet the rms or average signal is less dynamic and more constant. In this way, the bass is there yet it doesn't it all your headroom.

Please correct me if my reasoning is wrong, I am learning a lot from this thread.

Cheers, Andrés
 
since we have a gen-u-wine speaker designer and "real" engineer here, let me ask barefoot about a related topic...

if recorded material has a high level of sub-sonic content, in the struggle to reproduce those tones, will the typical speaker driver end up distorting frequencies that are higher and more easily audible? In other words, if we are hearing apparent distortion at, say 160 hz, could it be in fact caused by what is going on at 40 hz ? (And would we tend to hear the effects at a harmonic "multiple" of the fundamental?)
 
cordura21 said:
.....it's better to compress bass so the amplitudes are not so highs, yet the rms or average signal is less dynamic and more constant. In this way, the bass is there yet it doesn't it all your headroom....
Click to expand...
Yes and no. Bass frequencies are always going to eat up most of your dynamic range if you want your music to sound like…well.. music. It’s just a fact of the principles of loudness we discussed. On the other hand the largest short duration peaks tend to come from kick drums for which the bulk of the signal lies in the low frequencies. So compressing the lows will tend to smooth out the large peaks more than compression in other bands. Of course you can use a single band compressor or a limiter to smooth out large peaks as well. Sorry, there are really no hard fast rules. A lot of it comes down to what sound and feel you’re aiming for.
littledog said:
...if recorded material has a high level of sub-sonic content, in the struggle to reproduce those tones, will the typical speaker driver end up distorting frequencies that are higher and more easily audible?....
Click to expand...
Absolutely. Speakers have really horrible levels of distortion compared to everything else – especially in the low frequencies, and especially for ported or passive radiator (like Mackies) speakers working below their tuning frequencies. If you’re trying to make a 6” or 8” driver output any significant amount of energy below, say 60 Hz, I guaranty that you are creating truck loads of distortion – upwards of 10% or even 20%.

Most of the distortion is harmonic and can interfere with other higher frequency content, beyond just misrepresenting the fundamental. Higher order (more complex) distortions like intermodulation distortion (f1-f2, f1+f2, f1-2*f2,…) are less intense, but are particularly offensive to the ears.

barefoot
 
barefoot said:
Speakers have really horrible levels of distortion compared to everything else – especially in the low frequencies, and especially for ported or passive radiator (like Mackies) speakers working below their tuning frequencies. If you’re trying to make a 6” or 8” driver output any significant amount of energy below, say 60 Hz, I guaranty that you are creating truck loads of distortion – upwards of 10% or even 20%.
Click to expand...
i've played 33hz through 8's.. they weren't distorting at all..
even while mixing..i don't have any problems with distortion through 6”. i don't have 1 song that doesn't have a lot of low bass.. i have tracks that go down to 18hz...
no problems with distortion at all..
my only problem is ppl that have crap like rockford, JL and kicker subwoofers.. they blow easy when i play my songs
 
c9-2001 said:
i've played 33hz through 8's.. they weren't distorting at all..
even while mixing..i don't have any problems with distortion through 6”. i don't have 1 song that doesn't have a lot of low bass.. i have tracks that go down to 18hz...
no problems with distortion at all..
Click to expand...
Well, then the volume was either very low, or you don’t know what a pure 33Hz tone sounds like. And I’m not being condescending. Most people don’t have any experience at all with unadulterated deep bass, unless perhaps they’ve heard it through good headphones. I have little doubt that at normal volume you were listening to at least 10% distortion.

I’ve had success getting less than 1% distortion at reasonable listening levels in medium sized rooms with pairs very sophisticated, very expensive 18" drivers in "quasi-passive" configurations. You can get away with smaller drivers using active feedback techniques like those used in Velodyne subs.


barefoot
 
Thanks barefoot - it's a pleasure to have you around as a resource. :cool:
 
barefoot said:
Well, then the volume was either very low, or you don’t know what a pure 33Hz tone sounds like. And I’m not being condescending. Most people don’t have any experience at all with unadulterated deep bass, unless perhaps they’ve heard it through good headphones. I have little doubt that at normal volume you were listening to at least 10% distortion.

I’ve had success getting less than 1% distortion at reasonable listening levels in medium sized rooms with pairs very sophisticated, very expensive 18" drivers in "quasi-passive" configurations. You can get away with smaller drivers using active feedback techniques like those used in Velodyne subs.


barefoot
Click to expand...

now my volume RIGHT NOW is quite high...
yes i know what 33hz sounds like... i build custom stereo system with the best quality products, that usually pick up down to 25hz...
i can send you a quick sample of a song that goes down to 18hz.. the 6"s just just about reach peak xmax at high level, they can't pick up 18hz.....
 
check out these 3d freq.analysis i did on a few songs....
look at
"IF YOU DARE"
"BIG TRUCKS ON 22's remix"

then look at 3 6 mafia's late night tip...
if you ask anyone, late night tip has one of the lowest bassline ever.. and can just about blow any speaker made... and that song can't touch any of mine...

with the big trucks on 22's remix.. i've shattered front windshields(which is damn hard to do) broke rear view mirrors, broke odometers... you name it... not to mention blew apart a few subwoofers and subwoofer boxes... ;)
 
c9-2001 said:
yes i know what 33hz sounds like... i build custom stereo system with the best quality products, that usually pick up down to 25hz...
i can send you a quick sample of a song that goes down to 18hz.. the 6"s just just about reach peak xmax at high level, they can't pick up 18hz.....
Click to expand...
Recording the signals isn't the problem - reproducing them accurately is.

Which drivers are these? Did you actually analyze the signal coming out of the speaker? What test methods and equipment did you use? If what you say is true - no distortion at 18Hz from a 6" driver - then you should apply for a patent immediately, because you've obviously made quite an astonishing breakthrough. (Btw, xmax is far from the whole story with respect to distortion)

I'm not interested in continuing a pointless debate. But, I just can't keep quite when misinformation is being spread. Believe what you want to about what you're hearing C9, but the numbers don't lie.

barefoot
 
should i cut everything below 40hz?

i'm into making hip hop music, most which requires a lot of low end bass. the type that's loose not often tight, sort like listening to bass on 12 inch woofers compared to 6inch.

if the lows below 35hz create a lot of distortion, should i cut it?
 
You must log in or register to reply here.
Back
Top