Warm Audio WA-19 (AKG D-19 clone), mechanical HPF, phase issue?

[Paul W]

Warm Audio has a new WA-19 dynamic mic ($200), which they make some interesting claims about. It's supposed to be mostly a recreation of the AKG D-19 (but made better, naturally).

They say it has an adjustable acoustic high-pass filter which avoids electronic phase shift from high-passing in your preamp. (I'm guessing that's basically controlling rear venting.)

It's been my impression that such mechanical filtering properties that determine frequency response of a capsule have equivalents in simple LRC circuits, and imply similar phase shifts, so that there's no free lunch here. You might avoid an electronic phase shift, but you'd still have an acoustic phase shift for analogous reasons.

Is that correct?

They also make some claims about reduced proximity effect and in one of their videos make it sound like that's because it's a dynamic rather than a condenser. Do dynamic mics have less proximity effect, or do normal mics of either type have similar proximity effect for a given polar pattern?

The D-19 and WA-19 have a row of little slots going down the body (handle) of the mic, and I'm wondering if they do something similar to Electro-Voice's "Variable D" trick to reduce proximity effect, and if not, what those slots are for.

Edited to add: the Amazon page for the WA-19 says "Additionally, ventilated slots on the side of the WA-19 body reduce proximity effect, eliminating boomy bass associated with placing the mic closer to louder sources," so I guess that it is something like Variable D.

Edit 2: Sound on Sound article about the D-19 and D-24, confirming that: https://www.soundonsound.com/reviews/golden-gear-akg-d19

 

[kingkorg]

It's been my impression that such mechanical filtering properties that determine frequency response of a capsule have equivalents in simple LRC circuits, and imply similar phase shifts, so that there's no free lunch here. You might avoid an electronic phase shift, but you'd still have an acoustic phase shift for analogous reasons.

Is that correct?

Exactly. Regarding the rest of the claims, I don't believe a thing they are saying. Just marketing...

This alone is hilarious. Almost as if Chat-GPT generated it, only worse. Absolute nonsense, every part of it.

"The WA-19 has a dynamic humbucking coil that helps the mic deliver elite off-axis rejection, especially in the bass/lower midrange with minimal noise, making it easy to grab tight snare bottom tones without aggressive kick drum bleed."

 

[garp]

I don't think Warm ever designed capsules by themselves. They made a great job at capturing the look of the D19, really not sure about the capsule design employed here, no doubt it is made somewhere far east, no doubt that the margin they make on it is huge, no doubt that lot of geeksters will love it.

 

[Voyager10]

It's been my impression that such mechanical filtering properties that determine frequency response of a capsule have equivalents in simple LRC circuits,

Yes and no. Many acoustic/mechanical filters are resonant systems (https://en.wikipedia.org/wiki/Resonance) and as such have RLC equivalents.

However, it's entirely possible to build an acoustic system, e.g. based around a time delay, which doesn't have an RLC equivalent. All sane filters come under the general category of "linear time-invariant systems" which mean there are mathematical laws which relate frequency response and phase response. You can't choose one without putting limits on the other; you can't try to escape by making it mechanical rather than electronic.

(By "sane" I mean: a Leslie speaker is not an LTI system, I'm presuming you don't want this on your microphone?)

 

[Paul W]

Yes and no. Many acoustic/mechanical filters are resonant systems (https://en.wikipedia.org/wiki/Resonance) and as such have RLC equivalents.

However, it's entirely possible to build an acoustic system, e.g. based around a time delay, which doesn't have an RLC equivalent. All sane filters come under the general category of "linear time-invariant systems" which mean there are mathematical laws which relate frequency response and phase response. You can't choose one without putting limits on the other; you can't try to escape by making it mechanical rather than electronic.

(By "sane" I mean: a Leslie speaker is not an LTI system, I'm presuming you don't want this on your microphone?)

I thought that LTI systems only had to be well-characterized by an impulse response, which would include FIR things like delay lines (with an impulse response that's just a spike at the delay time) or I guess multi-tap recirculating reverbs, which must have IIRs.

Did you mean to say "minimum phase"? (Not that I have a good handle on exactly what that is, just that it's related to LRC circuits.)

My impression (from comments around here by Ricardo) is that normally microphone capsules act very much like "minimum phase" systems, at least for sound not coming from the back.

I don't really understand that, though, given the (short) acoustic delay networks in capsules, and I don't know if the interference tube on a "variable D"-type mic would count.

 

[abbey road d enfer]

They say it has an adjustable acoustic high-pass filter which avoids electronic phase shift from high-passing in your preamp.

As is often the case, phase is the magic token that explains everything. Actually shows that the person who wrote it has no clue.
Phase-shift, so what?

They also make some claims about reduced proximity effect and in one of their videos make it sound like that's because it's a dynamic rather than a condenser.

Pure nonsense. Just like "dynamic mics are less prone to feedback than condensers".
Both directivity and proximity effect are defined by acoustic construction, not the transducing principle.

Do dynamic mics have less proximity effect, or do normal mics of either type have similar proximity effect for a given polar pattern?

The latter is true.

The D-19 and WA-19 have a row of little slots going down the body (handle) of the mic, and I'm wondering if they do something similar to Electro-Voice's "Variable D" trick to reduce proximity effect, and if not, what those slots are for.

Of course.

 

[abbey road d enfer]

I thought that LTI systems only had to be well-characterized by an impulse response, which would include FIR things like delay lines (with an impulse response that's just a spike at the delay time) or I guess multi-tap recirculating reverbs, which must have IIRs.

Mathematically, although a delay line provides a nice impulse response, it doesn't count as Minimum Phase because it's not invertible.
A 1ms delay results in 360° phase-shift ar 1kHz, 3600° at 10kHz, not zero.

Did you mean to say "minimum phase"?

I guess that's what Voyager 10 meant. An LTI is not necessarily Minimum Phase.

My impression (from comments around here by Ricardo) is that normally microphone capsules act very much like "minimum phase" systems, at least for sound not coming from the back.

It's commonly admitted that the response of any system can be analyzed as a MP system combined with a non-MP one.
It's also commonly admitted that for audio, a system can be considered MP within a certain BW.
Ricardo's experiments led him to conclude that he'd never seen a mic that wasn't MP. I guess this was true, within the limits of his test equipment.

I don't really understand that, though, given the (short) acoustic delay networks in capsules, and I don't know if the interference tube on a "variable D"-type mic would count.

Clearly, there is a limit where the length of the delay line exceeds the wavelength of signal, where MP cannot be assumed. In most cases, the non-MP behaviour happens well over the higher limit of audio frequencies.

 

[Paul W]

Mathematically, although a delay line provides a nice impulse response, it doesn't count as Minimum Phase because it's not invertible.

I don't know what "invertible" actually means in this context. A simple delay seems trivially invertible... just run the signal backward through the same delay to push the signal back "forward."

NVM, I guess when they say "causal and invertible" causal means that both the forward function and the inverse must go forward through the signals presumably with some bounded constant delay to reconstruct the original.

Even so, I don't know what that ends up really meaning about what kinds of functions have such inverses.

 

[Voyager10]

So yes, all LTI systems can be defined by their impulse response.

In a minimum phase LTI system, frequency (amplitude ) response and phase response are locked together - if you fix the frequency response you have no choice about how the phase comes out.

All other (non-minimum phase) LTI systems can add additional phase shifts (i.e. delays) compared to an MP one with the same amplitude response, but they can't remove them.

TL;DR all filtering, whether mechanical or electronic, results in phase shifts. And WA's marketeers need to learn what a humbucking coil does.

 

[garp]

I have a bunch of EV664 that i really love. The internal acoustic chambers and the overall design of the capsule are mindblowing, you can clearly see all the maths and science involved in this microphone, and the result is really stunning. Today with far more computing sources and better machining we are still making clones of past designs made and calculated by hand, so what happenned to the R&D of the great brands ? Are they just there to ensure accurate copies of past designs with modern (cheaper?) materials ? When do we stop to develop really inovative designs and slip to a lazy era where tweaking around known designs is considered as "research" ?
I see more research here on group DIY than in the products available at the NAMM each year, and that is a F....g paradox IMHO.

 

[abbey road d enfer]

Even so, I don't know what that ends up really meaning about what kinds of functions have such inverses.

Invertible means that the transformation applied to the signal can be cancelled by applying another reciprocical transformation.
Typically that would exclude transfer functions that have zeros of pure phase-shift (i.e. without the usual correlated frequency response changes).
A bell EQ is invertible. A shelf EQ is invertible if you exclude 0Hz and infinite Hz.
An APF (All Pass Filter) is not invertible. A delay is not invertible (the inverse of a delay is an anticipator - one that invents it is bound to make a fortune).

 

[RuudNL]

"The WA-19 has a dynamic humbucking coil that helps the mic deliver elite off-axis rejection".

 

[garp]

I was always tempted to open a thread about the evidence of drug abuse in the advertising business.

 

[Wordsushi]

Warm Audio is a great case study of the effect good marketing has on the market for a prosumer product posing as professional in an environment where urban legends and fairy tales are the coin of the realm.

Tell me you wouldn't be shocked if you found out Bryce from Warm starred as Prof. Henry Hill in his high school production of "The Music Man".

 

[abbey road d enfer]

"The WA-19 has a dynamic humbucking coil that helps the mic deliver elite off-axis rejection".

Yes, but the capsule is made out of METAL, even the magnet.
Ain't it extraordinary?

 

[garp]

You mean real metal like the soda cans ? That's real luxury !

 

[Voyager10]

Invertible means that the transformation applied to the signal can be cancelled by applying another reciprocical transformation.
Typically that would exclude transfer functions that have zeros of pure phase-shift (i.e. without the usual correlated frequency response changes).

Right, so to give some more examples:

• A comb filter (e.g. signal + 10ms delayed signal) has zero output for certain input frequencies. It is not invertible; you simply can't tell from the output whether these frequencies are present in the input. It's not MP.
• A simple delay is invertible, but the inverse transform involves predicting the future so it's not causal. So it's not minimum phase (both normal and inverse transforms must be causal)
• An RC filter is invertible, but its phase shift is a consequence of it integrating the input signal, not delaying it. The inverse operation is differentiation, which can be done causally. So it is MP.

I'll shut up now. Let's talk about microphones instead.

 

[Paul W]

Right, so to give some more examples:

• A comb filter (e.g. signal + 10ms delayed signal) has zero output for certain input frequencies. It is not invertible; you simply can't tell from the output whether these frequencies are present in the input. It's not MP.
• A simple delay is invertible, but the inverse transform involves predicting the future so it's not causal. So it's not minimum phase (both normal and inverse transforms must be causal)
• An RC filter is invertible, but its phase shift is a consequence of it integrating the input signal, not delaying it. The inverse operation is differentiation, which can be done causally. So it is MP.

I'll shut up now. Let's talk about microphones instead.

For me this is very much about microphones because I'm trying to figure out what Ricardo meant and what the implications actually are for being able to "correct" the frequency response of one mic to match another, and have the phase response automatically fall into place as well. (BTW I do realize that won't fix the off-axis sound if the two mics' off axis sound doesn't bear the same relationship to their on-axis sound.)

It seems intuitive that simply rolling off the highs or lows to match an overall FR curve could work that way, with a capacitor or inductor having similar effects on both frequency response and phase to physical things that limit frequency response, like the venting of a capsule or the mass of a diaphragm.

But suppose you have a delay path in the mic that creates a weak comb filter, which puts a few notches in the frequency response, but they don't go to zero at any frequency.

Would that be invertible, and if so, would mindlessly using a parametric EQ to boost at each of the notch frequencies do the right thing WRT phase? Or do you need some other way to find a different inverse function, which WOULD correct phase?

 

[Timjag]

Can we just take a minute to study the fact that an audio company has brought back the D19…. THE D19???!!!!
I’ve got several some work some don’t, they’re very old “obviously” I kinda like them, sometimes, but I like a 57 too.

Of all the mics!!?? The fiddly old d19 that falls apart easily and has funky acoustic chamber that changes the mic from sounding like a normal dynamic to sounding like you put a 57 the wrong way around!

Not the D12? Or the D20? The D12 a mic everyone is obsessed with a mic that clearly people want judging by the insane prices they go for and poor old panman has to spend his life repairing….

I’m blaming Keyhoe and his Beatles book for this! Who would have cared about a D19 otherwise, who’d have known apart from old staff, that the abbey road engineers used them for everything and only because they were cheap and easily replaceable when using them on that godawful rock n roll business which was clearly a fad.

All I’m wondering is if the case is the same size so I can fix a few of my broken ones

 

[abbey road d enfer]

I tend to agree, but what Beatles book by Keyhoe?
The problem with the D12 is that nobody really knows what the AKG engineers were pursuing when they designed it.
I believe Warm are not concerned with tis aspect of things, as long as the cosmetics are identical.
And of course, nobody knows how a brand new D12 is supposed to sound.