The No.1 Website for Pro Audio
 Search This Thread  Search This Forum  Search Reviews  Search Gear Database  Search Gear for sale  Search Gearslutz Go Advanced
EQ Phase-shift
Old 23rd April 2007
  #1
Here for the gear
 

EQ Phase-shift

I'm tying to understand exactly how analogue audio filters work but i can't find anywhere that will tell me.

I'm not sure how off the mark I am, but is it just that the filter puts a different delay at different frequencies to change the amount of phase cancellation when added to the original input signal? So for no gain reduction, there's no delay but for infinite reduction, the filtered signal is 180º out of phase to the original? Everything else just slots in between.

Any help will be great thanks!
Old 23rd April 2007
  #2
Gear Guru
 
u b k's Avatar
 

Quote:
Originally Posted by Graeme86 View Post
So for no gain reduction, there's no delay but for infinite reduction, the filtered signal is 180º out of phase to the original? Everything else just slots in between.

it seems like you might have the gist of it, but your wording is a little off.

if you bandpass a section of the spectrum and add that back in with the original, you've got a boost of the bandpassed freqs. if you flip the polarity of that bandpassed section and add it back in, you've got a cut. the greater the gain of the bandpassed signal, the greater the cut or boost.

phase shift tends to come into the picture with analog eq because of the nature of the process. you take a dry signal, split it in two, run one side thru filters and blend it back with the other, that filtered side is a little delayed. some circuits have a more pronounced smear around the filtered freqs than others do, and often this is what we hear as the 'sound' of the eq.

i'll go further and say that this is why cuts sound better when accumulated in a mix: the affected freqs are lower in level than the rest of the spectrum (they've been cut), so what you tend to hear in the mix is mostly the unsmeared stuff. with 24 channels of boosted eq, you're hearing phase artifacts all over the place. ime, a desk where most of the knobs are turned backwards tends to spit out a more focused, musical sound.

as always, exceptions apply.



gregoire
del
ubk
.
Old 23rd April 2007
  #3
Here for the gear
 

Thanks u b k,

I understand what you've said, and its really helpful. I think what I'm really after is how a simple RLC circuit can actually filter at different frequencies. For example, how does it "bandpass a section of the spectrum"?

Maybe I'm reading too much into it or perhaps the fact that it just does should be enough. Perhaps this is more at home in the Geekslutz forum.
Old 24th April 2007
  #4
Gear Guru
 
Ethan Winer's Avatar
 

Graeme,

Quote:
what I'm really after is how a simple RLC circuit can actually filter at different frequencies. For example, how does it "bandpass a section of the spectrum"?
These articles might help:

Audio Filters -- Theory and Practice
Mike Pads and Other Small Gadgets
Equalizers and Phase Shift

The basic idea is to combine a direct signal with a copy that's been phase shifted. With the RLC you mentioned the L and C have opposite shifts, so depending on how they're arranged you either filter out the center frequency or filter out everything except the center frequency.

--Ethan
Old 2nd May 2007
  #5
Lives for gear
 
luctellier's Avatar
Quote:
Originally Posted by Ethan Winer View Post
Graeme,



These articles might help:

Audio Filters -- Theory and Practice
Mike Pads and Other Small Gadgets
Equalizers and Phase Shift

The basic idea is to combine a direct signal with a copy that's been phase shifted. With the RLC you mentioned the L and C have opposite shifts, so depending on how they're arranged you either filter out the center frequency or filter out everything except the center frequency.

--Ethan
Really nice article! Thank you! thumbsup
Old 3rd May 2007
  #6
Registered User
 

I have been wondering this and asked at other forums but didnt' really get a concrete answer.

People say that cutting sounds more natural than boosting, what is the reasoning for the difference in sound?

My best guess is: The phase shifted part of the signal is louder in volume on a boost. As in, you look at pink noise in a RTA at -20db. You eq the pink noise +3 and the phase shifted part of the signal is a -17cb at it's peak. You cut and the phase shifted part of the signal is -23db at it's peak. It's phase distortion is lower in volume, thus less noticeable to the ear.

Am I understanding this correctly? There is not a different mechanism going on to achieve a cut that makes it more natural, it's just the volume of the phase distortion?
Old 3rd May 2007
  #7
Gear Guru
 
Ethan Winer's Avatar
 

Quote:
Originally Posted by briefcasemanx View Post
People say that cutting sounds more natural than boosting, what is the reasoning for the difference in sound?
The answer has nothing to do with phase shift. heh

There are a few common problems that people use EQ to solve. One is overall lack of brightness, and in that case EQ shelving boost is the most direct solution. Another is overall low frequency muddiness, and for that an EQ LF shelving cut or high-pass is appropriate.

Just as common is what people sometimes call "surgical" use of EQ - to reduce offensive resonances. Often when something sounds not bright enough the real culprit is a nasal or boxy sounding resonance. When you find and reduce that resonance, all of the sudden the sound becomes brighter and cleaner and often fuller by comparison.

So it's not that cutting is generally preferred to boosting, or vice versa, but rather simply identifying the real problem. I'd say that identifying and squashing offensive resonances is one of the more important skills for a mixing or mastering engineer to develop.

I'll also add that cutting resonances is best done with narrow (high Q) settings, but for boosting a shelving or broad shape is better. Otherwise you're adding resonance which is often obnoxious. Unless of course you want a resonance. The other day I recorded myself playing a Djembe. I'm not much of a drummer, and my touch is not very good. So I didn't get enough of that cool ringing tone Djembes are known for. So I swept an EQ and found the fundamental pitch, then boosted that a few dB with a very high Q. Voila, then I had a nice round tone.

--Ethan
Old 22nd September 2011
  #8
Here for the gear
 

Quote:
Originally Posted by Graeme86
So for no gain reduction, there's no delay but for infinite reduction, the filtered signal is 180º out of phase to the original? Everything else just slots in between.
There can be more than 180° of phase-shift on any signal. For infinite reduction of any particular frequency, it is dependant on the transfer function - the steepness of the curve / -3dB cutoff point / all the other things I don't understand. The term 'Time-delay' is only used when all frequencies are delayed by the same amount. The term 'Phase-shift' is used when the delay amount is frequency dependant, as it is in filter circuits.

Filter components:
Capacitor (Series) = High-pass (Blocks DC / allows higher frequencies to pass.)
Capacitor (Parallel) = Low-pass (Takes higher frequencies down to ground, only allowing the lower frequencies to pass.)

Inductor (Series) - Low-pass (Higher frequencies find it harder to pass.)
Inductor (Parallel) = High-pass (DC and 'lower' frequencies taken down to ground and high frequencies allowed to pass. Rarely used in audio circuitry as you would need an extremely high value inductor to prevent audio frequencies from being shorted to ground. Mostly used in RF circuitry.)

Resistors are usually used to slow the current flow into the capacitor to alter its 'time constant' (how long it takes to charge/discharge it.) This alters the frequencies that are passed and this technique is used to 'fine tune' circuits to operate at specific frequencies.

Because a capacitor's voltage takes time to change, any input signal will be time-delayed a certain amount on the other side of the capacitor. This time-delay is different at different frequencies, therefore it is measured in 'phase-shift'.

An inductor resists changes in current flow, which also causes a different time-delay at different frequencies and is also measured in phase.

The phase-shifting for both of these components is usually an approximate inverse of the way it filters. i.e. For a high-pass filter, the lower frequencies will be phase-shifted more. For a low-pass filter, the higher frequencies will be phase-shifted more than the lower frequencies.

For more complex filters such as band-pass and band-reject, the phase response is usually a little more complicated. Most boost/cut filters pass a full bandwidth signal and 'mix in' a band-passed or band-rejected signal to create a boost or cut at the chosen frequency.

In terms of adding/cutting EQ, there shouldn't be much of a difference in terms of phase-shift. Some people believe that cutting doesn't add any phase-shift, but it does. Cutting might sound better for some things when you want to eliminate 'problem frequencies', but boosting can be more useful when you want to enhance certain characters of instruments. The only things I'm very careful with are very high order low-Pass and high-pass filters. High order high-pass filters can make the bass transients on kick drums sound boomy, or high order low-pass can make high frequency transients on vocals sound like a sine-wave sweep zipping upwards.

If in doubt, use a digital linear-phase EQ

[Apologies if there are any inaccuracies in the information provided. I'm still learning!]
Old 22nd September 2011
  #9
Gear Guru
 
Ethan Winer's Avatar
 

^^^ Nice post Jim, looks accurate to me.
Post Reply

Welcome to the Gearslutz Pro Audio Community!

Registration benefits include:
  • The ability to reply to and create new discussions
  • Access to members-only giveaways & competitions
  • Interact with VIP industry experts in our guest Q&As
  • Access to members-only sub forum discussions
  • Access to members-only Chat Room
  • Get INSTANT ACCESS to the world's best private pro audio Classifieds for only USD $20/year
  • Promote your eBay auctions and Reverb.com listings for free
  • Remove this message!
You need an account to post a reply. Create a username and password below and an account will be created and your post entered.


 
 
Slide to join now Processing…
Thread Tools
Search this Thread
Search this Thread:

Advanced Search
Similar Threads
Thread
Thread Starter / Forum
Replies
loglinear / Mastering Forum
28
loglinear / Music Computers
2
Shaman / So Much Gear, So Little Time
4
jbuntz / Geekslutz Forum
3
Jules / So Much Gear, So Little Time
33

Forum Jump
Forum Jump