Hi everyone, as I wrap my head around the differences between linear phase eq and minimum phase eq I'm looking for online audio examples of the differences between these two tools.

Can anyone point me to some audio examples of pre-ringing when using a linear eq? I've yet to come across any in my searches.

Thank you in advance.

Here you go. One file uses a minimum phase EQ, the other a linear phase (see file names). I made these very obvious, to create a worst case example, using a strong transient snare sound with fairly low frequency cut at narrow Q.

Hi Robin, that's fantastic, thank you very much. I hear it very clearly.

Do you personally find this to be an issue when working? Do you have a general preference for either type of EQ?

The problem of a downbeat creating a "swoosh" leading up to it it is a reasonably rare occurence, but it does happen and is something to watch out for whenever using linear phase EQ. Usually linear phase pre-ringing rather shows itself as a perceived difference in tone, since the pre-ringing is much shorter with higher frequency and broader EQ changes. To me, that different tone, along with a slight perceived softening of transients, is enough to put me off linear phase most of the time.

Like 24-96 said, you usually don't get a noticeable RINGING so much as the perception of softened transients. Sometimes a linear phase EQ is just the ticket for a certain problem, particularly in the high end. Most of the time, I stick to minimum phase.

For the exact same reasons above, LP in the higher end of the spectrum can be great for bringing out 'inner details' and harmonic textures etc without pushing the main transients in that area so much.

good point!
you can also benefit in low end, when LP 'distortions' actually can sound good

The temporal distortion of the transient is more significant the lower the frequency. It has a relationship to the Secant function of the unit circle, where time is delayed almost exponentially as a function of frequency. It is best to use as wide Q as possible, and combine with minimum phase EQ for sharper moves.

The LP10 EQ by DDMF has the function to adjust the resonance from minimum to linear to maximum phase, with maximum being equivalent to a TIIR or IIR in reverse. It can set an inbetween, where resonance is biased to either side of the impulse. While I am not fond of FFT processes in general (except analysis), YMMV.

I do agree that it really brings out the details in a very transparent and silky smooth way. In some situations this is exactly what is called for.

Just writing so it is in my history.

Sent from my HTC Desire

Which makes sharp linear-phase HPF's the worst possible place to apply it.


DC

For mastering, yes. There is no benefit to LP lowcutting unless the material is already loose. Using a sharp LP lowcut on the master is just bad practice.

In sound design, there are not such simple rules. As an in between process with particular treatment afterwards, LP lowcut can make very interesting sounds.

The amount of ringing really depends on the signal: whether the signal contains any significant energy at the cutoff frequency of the filter. If the transition band of the filter does not overlap any significant energy in a signal, then a linear-phase HPF is the best choice because it keeps the signal above the cutoff intact. For example, if your signal lies all above 70 Hz and there's a hum at 50 Hz, I would use a linear-phase filter to perfectly separate them, without any phase distortion to the useful signal.

At any rate, when I have audibly compared these filter families minimum-hase always sounds better to me. At every frequency. It actually seems to do less damage.

Although I haven't tried your system that allows for an in-between phase response to be phase response, which does seems interesting.

DC

Reviving this thread: Is there an easy way to visualize and quantify FIR pre-ringing in VST plugins?

I wouldn't say easy but if you understand some of the principals you can test plug ins very effectively with this - https://ddmf.eu/plugindoctor/

It's a little buggy and if you don't know a little about how these things work you will get confusing results.

What are you using within Plugindoctor to do that? The oscilloscope?

Yes. Time behavior of equalizers is well described in impulse response snapshots. Use something like RXs reconstructed waveform view to get an idea of what your DA converter does to represent those sample points.

A perfect impulse is zeroes only with a single sample at max level. Leave enough time around the lone peak to capture the event you’re looking for. Run this through any process and the output will be the IR for that system.

Normal EQs will only show things happening after the excitation impulse. Linear phase will have equal amounts of activity before and after the impulse. Mixed phase will (usually) have more happening after the impulse, and some things happening before the pulse.

The quanta you’re looking for is the RMS value for the before/after parts. Linear phase have the same area under curve before/after the impulse. Normal equalizers / minimum phase will have infinite difference as there is nothing happening before the impulse arrives. Mixed phase systems will typically give some lesser RMS numbers for the before parts and larger RMS figures for the after part. The difference between the before/after can then again be used as a ratio for dB figure if you so pleases.

Awesome, thanks for this

Well, I tried to test Waves LinEQ with a Dirac pulse but for some reason the pulse was perfectly preserved with no ringing -- is Waves LinEQ smarter than me?

That's what I was expecting, thanks for confirming it.

You could try using a “pulse” of DC. (That’s what I do fwiw)
(ie...step response)
This guy shows example in his blog > https://cravedsp.com/blog/linear-phase-eq-explained

Thanks. For those that need one, I'm attaching a half-square wave impulse.

There are some excellent examples in the RME ADI2 pro FS manual when discussing their different AD/DA filters

https://www.rme-audio.de/download/adi2profs_e.pdf

Page 77

Short Delay = Minimum Phase
Slow = long roll off
Sharp = sharp roll of

They use some confusing terms but take a look at that and will explain how different filters in LP and MF effect the transient response as well making cuts within the pass band.