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
Phase Linear EQ with no pre ringing? Equalizer Plugins
Old 1 week ago
  #61
Gear Maniac
 

I did some more testing.

At first I tested the zero phase in pro-Q3 vs linear phase (high). This is quite easy to A/B blind with 100%.

But zero phase in pro-Q3 vs linear phase (low) isn't as obvious. I get only a 40%/60% rate.
Old 1 week ago
  #62
Lives for gear
 

Quote:
Originally Posted by AreYouHuman View Post
So far I never used linear phase on individual tracks as low latency also better fits to my workflow.

But I just tried the new fabfilter pro-q3 for fun. I tried the same setting with zero phase and linear on hihats. So only a mid and high range test! I can't stand pre-ringing on the bass.

The linear phase sounds indeed a bit tighter and with better room. I can a/b it blind without a problem on headphones. But maybe the pro-q3 has just a mediocre zero phase implementation? Who knows.

My setting were bell/12db at 4300hz, 7.6db gain, Q 1.

But I also almost blindly trust FabienTDR opinion when it comes to technical explanations.

Not sure what this means for me now.
you should still try on one band removing some mud, sometimes it can be more transparent with linear phase, you might get pre ringing as low as the noise floor..it might still be earable thought.
Personally i rarely use Lp cause of the latency and cpu, even if sometimes it can sound better .. but to me the most transparent eqs, can only be one that let you adjust the amount of pre ringing to the point it s not a problem for the band used , so using a eq that let you adjust the amount of phase shift between lp and phase.

just do what sound best to you and don t care about technical, the less you know about technical aspect the more you trust your ears and the less you re victim to placebo , evrything your ear is first filtered by your brain and then your judgement come of what you heard, the less you know the less it s filtered i guess.

Last edited by Fred_Abstract; 1 week ago at 11:47 PM..
Old 1 week ago
  #63
Lives for gear
 

Verified Member
in pop music, i cannot see much use for fir filters: sometimes, deliberately introduced phase shifts are getting used to widen the picture and there are way too many things going on (hundrets of tracks, efx etc.- just watch your phase meters while mixing).

for classical music, for mastering and for aligning speaker systems, there is some benefit (besides the typical issues)...
Old 1 week ago
  #64
Lives for gear
 
Greg Reierson's Avatar
 

Verified Member
Linear phase EQ and stereo phase are completely unrelated concepts.
Old 1 week ago
  #65
Lives for gear
 

Verified Member
i'm not talking about phase correlation (which is an issue in at least two of the fields i mentioned), i'm talking about how useful it is to use linear phase filters in different areas/genre...
Old 1 week ago
  #66
Quote:
Originally Posted by deedeeyeah View Post
in pop music, i cannot see much use for fir filters: sometimes, deliberately introduced phase shifts are getting used to widen the picture and there are way too many things going on (hundrets of tracks, efx etc.- just watch your phase meters while mixing).
Allow me a pedantic comment:

FIR is just a filter structure. The most precise filter structure known to humanity.
It can cover everything you want, given sufficient CPU power.

Directly equivalent are so called "real-time FFT" filtering structures. They are equal to FIR, but from a frequency domain perspective (FIR follows the time domain perspective).

The main practical difference is about CPU performance and latency. FIR have lowest latency, but bad calculation complexity. Small IRs are rather fast, but everything beyond ~1000 taps will likely create trouble.

Real time FFT on the other hand imposes a fixed latency and limited precision directly depending of the size of the FFT block (it defines the maximum size of the IR).

Qualitatively, both are equal. Both can theoretically run almost any IR you like. This IR can be linear phase, min phase, or anything between. FIR (and realtime FFT) always have a finite IR length, though.

An IIR on the other hand is a recursive structure (it feeds-back previous values to eternity), so it naturally only supports IRs that are infinite in nature. They accumulate errors, and thus need very precise internal number representation (64bit). Fast like a lightning, but the range of possible IRs is somewhat limited.

Good overview:
Filter Classification

Last edited by FabienTDR; 1 week ago at 01:37 AM..
Old 1 week ago
  #67
Gear Head
Quote:
Originally Posted by Greg Reierson View Post
Linear phase EQ and stereo phase are completely unrelated concepts.
How is it so? If I do minimum phase with mid or side low cuts for example, it does changes to what sounds in left and what sounds in right, the stereo picture is lost, all panned elements are in wrong places. Linear phase does nothing to stereo. Check this on masters with panned left/right elements. Try equailibrium.
Old 1 week ago
  #68
Lives for gear
 

Verified Member
Quote:
Originally Posted by FabienTDR View Post
Allow me a pedantic comment:

FIR is just a filter structure. The most precise filter structure know to humanity.
It can cover everything you want, given sufficient CPU power.

Directly equivalent are so called "real-time FFT" filtering structures. They are equal to FIR, but from a frequency domain perspective (FIR follows the time domain perspective).

The main practical difference is about CPU performance and latency. FIR have lowest latency, but bad calculation complexity. Small IRs are rather fast, but everything beyond ~1000 taps will likely create trouble.

Real time FFT on the other hand imposes a fixed latency and limited precision directly depending of the size of the FFT block (it defines the maximum size of the IR).

Qualitatively, both are equal. Both can theoretically run almost any IR you like. This IR can be linear phase, min phase, or anything between. FIR (and realtime FFT) always have a finite IR length, though.

An IIR on the other hand is a recursive structure (it feeds-back previous values to eternity), so it naturally only supports IRs that are infinite in nature. They accumulate errors, and thus need very precise internal number representation (64bit). Fast like a lightning, but the range of possible IRs is somewhat limited.

Good overview:
Filter Classification
thx fabien, spot on (as always) - i should not have used technical terms (and confusing definitions) but should have put it in much more simple terms (see below). also, i noticed too late that this thread is in the mastering forum... - i should not post minutes before a gig!



[it should have been something like that:
imo it's not worth going the extra mile (and chose specific filters) if things get mixed/'messed up' anyway (in a pleasant way though) - but one could also argue that in exactly this situation one should better use the 'best'/most appropriate filters...]
Old 1 week ago
  #69
Lives for gear
 

Quote:
Originally Posted by Stereo Flux View Post
How is it so? If I do minimum phase with mid or side low cuts for example, it does changes to what sounds in left and what sounds in right, the stereo picture is lost, all panned elements are in wrong places. Linear phase does nothing to stereo. Check this on masters with panned left/right elements. Try equailibrium.
This is bit special case with unlinked M/S EQ.

Choice of min-phase or lin-phase EQ doesn't directly affect stereo picture per se (in the sense of narrowing it besides applied correction itself or steering it towards one side, change position etc.)
Stereo information in M/S system is carried solely by difference (S) channel.
So when you apply some minimum phase filter to S, then of course you'd introduce respective phase shift there, but that shift is then affecting both decoded L and R channels by the same amount.

M/S decoding matrix is like this:
L = M + S * atten. factor
R = M - S * atten. factor

Of course when you combine M/S channels like that in the decoder and one of them is phase shifted, it will affect your outcome and there might be certain perceived difference in sound. Some cancellation or boost will occur according to the shift.. say you might feel loosing overal definition with some filtering, possibly you'd need slightly alter the filter settings to achieve similar apparent results, when comparing lin-phase and min-phase. But the shift won't be different among L and R output channels, so stereo information is not directly affected by that.

Finally, the choice of used filter type (min or lin) might be totally subjective and material dependent.. but it's the same like with other cases (not isolated to unlinked M/S) and IMO it's not possible to say for example, that lin phase is not suited for the task as it affects stereo image.

Michal
Old 1 week ago
  #70
Lives for gear
 

And another comment.. to what deedeeyeah mentioned before.
I don't think, that lin-phase filters would be more useful for say classical than pop or say audio for video post production. From my experience, it's also useful always, when you benefit from preserved phase coherence between tracks or channels regardless of genre.
This might involve treatment of multi miked sources, parallel FX chains.. etc. With regards to latter for example it's very common to use additional filtering to band-limit or shape some parallel chain (with heavy compression, distortion) to be blended with original. Quite often it's really significant difference to use lin-phase EQ there, especially with transient heavy material, where it's rather apparent. Also built-in filters at vast majority of saturators are minimum phase, so it might be worth of shot to bypass it and use another EQ there.

Of course, it's always about balancing of benefits and tradeoffs, so easy comparison is big plus. So I personally really appreciated modern EQ plugins (Pro-Q, Equilibrium, LP10) with easily switchable operating modes with few clicks and lightweight operation, if you don't need long FFTs for increased low frequency resolution. Those have also common preserved magnitude response (that was quite annoying at some older plugins, which has different responses in min or lin mode).
It's a quite significant reason, why I won't be happy to use just dedicated lin phase plugin without chance to switch it to normal mode.

Michal
Topic:
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
Forum Jump
Forum Jump