Prehaps Linear phase EQ in analog way is possible...

[Kohugaly]

I've been recently playing with convolution and EQs ...And got a few ideas I what to share. PLEASE CORRECT ME IF I AM WRONG!

IR of an EQ composes of an short "peak" representing unaffected part of the signal and an "tail" representing the affected frequency range. The tail also causes the change in amplitude and shift in phase for certain frequency.

In linear phase EQ the tail is not only after the peak, but symmetrically also before the peak. This way the first one shift phase backwards and second one forwards, but their effect on amplitude is the same. Because the "pre-tail" would have to start somewhere in the future LP EQs have to cause latency to be practically possible.

That's when i got an idea. If you EQ a sound with nonlinear EQ (Analog or digital - it doesn't matter), you shift the phase in one direction. Then reverse the sound and EQ it using exactly the same settings. It should shift the phases right back. (although bear in mind, that amplitude will be affected twice)

Am I correct? ... I know it probably doesn't even matter, because when someone need and LP EQ he simply use one... It's just an idea, that was bothering me.

 

[RedDog]

Mind = bent.

 

[Kohugaly]

Is there a way to measure it?

 

[etcetera]

That's a really interesting question. I have no idea of the answer, but I'm going to ask around and see if I can find one.

 

[Skinny Viking]

 

[Kohugaly]

YES ...it fckin works.... just tested it on a vectroscope and phase meters... I made a sinesweep, saved it as file "dry" applied EQ saved as "first stage"... than made two copies on one I applied the same EQ, on the second I reversed and apied the same EQ (than reversed back)... I made two stereo files from "dry" on one channel and double EQed sweep on another. The "double EQ" was out of phase, but "EQ-reverse-EQ-Reverse" was perfectly in phase although amplitude was different for both channels. whana hear the files or see the images?

 

[-Noodles-]

Your first post is right in it's thinking, but explanation isn't clear.

It seems that you're reversing the dry signal to imitate the symmetrical properties of a minimum phase filter?
If so, the decay of the signal will be shorter and reduce ringing in the circuit. Could be beneficial in many uses but could really affect some sources (especially low frequency materials). There are a lot of EQs which claim to be minimum phase out there - but I've not tested them methodically. As to the reasons why you'd want a minimum phase 'analogue' EQ I'm not sure

I use my ears and make a judgement on if they're any good!

 

[Firaxis]

I always thought a lot of that "analogue" character was gained from the signals coming out of the gear being slightly mismatched/out of phase. It doesn't really make much sense to me to remove that element of the gear but then again I haven't ever tried it

 

[Kohugaly]

Your first post is right in it's thinking, but explanation isn't clear.

It seems that you're reversing the dry signal to imitate the symmetrical properties of a minimum phase filter?

I apply EQ (which shifts phase forward), then reverse (which also reverses the phase shift) and EQ again (which shifts the phase the same way but actually cancels the first phase shift because signal is now reversed), then reverse back to normal and the result sound is double Amplitude affected, but not phase affected (the phase effects canceled each other).

It is not a minimum phase EQing. This is totally no phase shift happening - just amplitude of EQed frequencies.


I'm not talking it's better or something (actually it is totally unpractical to EQ, reverse, EQ and reverse destructively) It was just an idea, because I've read million times, that linear EQing is not possible in analogue world. It was really just an mathematical/physical note to the world of EQing.

As I've posted already when someone needs LinPhase EQ (in some special tasks "it's not permitted" to f.ck up the phases) in mixing he simply uses one and when he doesn't he simply uses classical EQ because it sounds better in usual cases (there's no need of using LP EQ if it's not needed). The same goes for you Firaxis - you're right - It removes the "analogue" or "coloring" element of the EQ.

By the way... Did you heard of anyone ever used that "my" approach to preserve the phase? Perhaps in the old times when digital Linear phase EQs didn't existed. I'm just curious...