Consider this blog post part-review, part explaining where I'm going with my studio.
A Brief History Lesson
Just as every genre of music has its standard bearers and "standard works," so too does professional audio have its "standard gear." Perhaps more than any other technical profession, music production is unusually beholden to antique tools.
For example, the Neumann/Telefunken U47 microphone (first introduced in 1947, hence the name), is the de facto gold standard for condenser (if not all types of) microphones in recording studios. Why? Probably because it was the first microphone of its type that worked really well. All it took was its appearance on a few great recordings to cement its place in the pantheon of recording studio legends, which has perpetuated itself to this day.
Certain other microphones, also engineered in the 40s-60s, are commonly used in modern recording studios as well. They're so well-known that one can simply bandy about certain two- or three-digit numbers with any professional audio engineer - numbers such as 67, 251, 414, 12, 87, 49, 84 - and they'll know exactly which microphone you're talking about. (Coincidentally - or perhaps not coincidentally - all of the aforementioned microphones were made in Germany or Austria.)
However, there are some significant problems with this nostalgia-driven phenomenon.
First and most obviously: these are antique electronics, up to 70 years old in some cases. They have a tendency to be unreliable.
Second, because of their recording-pantheon status and their limited supply (they are, for all intents and purposes, discontinued - there are reissues, but "they don't make 'em like they used to"), they are ridiculously expensive to acquire and to maintain. A U47 in excellent condition can easily sell for more than 20K, and quickly.
Thus, only the big studios and stupidly rich people (or regular people who don't want any other nice things in their life) can afford them. I do not fit into any of those criteria.
Technology To The Rescue
Hang with me for a paragraph or two:
Digital audio is the unquestionable winner of the analog vs. digital debate - AT LEAST in terms of reproduction and fidelity. However, a significant number people prefer the sound of recordings that were captured and mixed using vintage analog equipment. AMAZINGLY - digital audio can reproduce all the nuances of these analog recordings after the master is converted to digital form. So it stands to reason that, given enough processing power and engineering expertise, digital signal processing can achieve effectively identical results when employed during the recording and mixing stages. (Objections to this assertion are plentiful, of course, but those who hold them always speak in highly subjective, ethereal terms to describe the differences they think they hear between the two.)
DSP has been widely deployed for over two decades to model the pleasant non-linearities of analog signal processors like compressors and equalizers. And with the advances in computer processing power as well as DSP coding techniques, it has become so good that many engineers end up using the digital versions of these analog processors, even though they own the real thing in some cases! The convenience of digital audio software, in practice, outweighs whatever minute differences there may be between a DSP plugin and its hardware counterpart.
With this power and expertise out there, some enterprising companies have applied their smarts to modeling microphones - most especially, the aforementioned standards.
However, this is kind of a "meta" problem, because microphones and microphone preamplifiers are, and always will be - by necessity - analog devices. (FYI - there actually is no such thing as an all-digital audio signal chain - digital hardware is only utilized for encoding, processing, and decoding - in the end, it's always analog in, analog out, no matter what.) So how can an analog device model other analog devices?
It's actually not terribly difficult to wrap one's head around how this ends up working in practice. Take a microphone with as flat a frequency response as possible, run it into a preamp with a similarly flat frequency response, measure the total frequency response of that signal chain - and with that knowledge, compare the differences between the flat microphone and the microphone to be modeled, measure the difference, then apply processing to the output signal of the microphone - first - before doing any other processing.
Now, in spite my faith in the abilities of digital signal processing, I have still been cautiously skeptical of jumping onto the microphone modeling bandwagon. My recording philosophy leans hard on microphone technique - and thus quality microphones and preamps - above all else. Then minimal digital processing where absolutely needed. Like my audio engineering hero, Al Schmitt, I like to consider myself a "microphone freak." (Except he has way more toys.)
However, I had heard many demos of the real studio "standard" microphones compared with their models, and I thought I could hear the difference between the real thing and the modeled version of it - in a way that didn't necessarily favor the modeled microphone.
But I also recognized that it's hard to be discriminating about recordings that were done with musicians, studio spaces, and signal paths that I am unfamiliar with. Suffice it to say, my curiosity has long been piqued. So I took the path of least resistance, and bought a couple of relatively inexpensive small-diaphragm condenser microphones from Slate Digital to record with in my studio.
The microphones themselves actually sound quite good on their own. They are un-hyped, as one would expect a "flat" microphone to be. But to make a very long story short, once I applied the mic modeling to the signal, I was astonished at what I heard. My skepticism should have biased me against the results. I half-expected there to be an audible "DSP haze" draped over the original unaffected signal; but instead, it just sounded "right."
For example, though I have never owned a pair of Neumann U67s (another über-expensive vintage microphone), I have recorded with them, prefer them for acoustic guitar, and know their sound very well from many great recordings that have been done with them.
Yet with these two inexpensive little modeling microphones in front of my guitar, I was getting the "right sound" right off the bat, that I would expect from U67s. The recording had all the richness and pleasant color that I have never been able to get with my "other" microphones. And this from DSP!
In five minutes, it was clear to me that I no longer had use for my existing microphone collection. It was that obvious. I couldn't believe that I was even thinking it, because one of my cardinal rules about gear has always been "NEVER sell a microphone." They're small, they don't take up a lot of space, and provide a great way to diversify one's sonic palette on the analog end of things.
And yet, I have put my entire mic locker up for sale, and have replaced every single last microphone with ML-1 and ML-2 microphones from Slate Digital. If I wanted to recommend a product more highly, I not sure I would know how.
Why the Slate VMS and not others?
Those who are in the know are probably aware that there are two other manufacturers in the microphone modeling game, Antelope Audio and Townsend Labs.
Antelope makes four different microphones that can model instrument microphones, cardioid (unidirectional) condenser microphones, multi-pattern condenser microphones, and stereo multi-pattern condenser microphones. (The latter two mics require two and four (!) inputs, respectively.)
The Townsend Sphere microphone models multi-pattern condenser microphones, and subsequently, also requires two input channels to work. This, to me, doesn't provide a practical solution for my needs, given my current setup. However, I do see a couple of Townsend Sphere mics in my future.
(Also, Antelope Audio has burned a bridge with one of the biggest pro audio retailers in the world, Sweetwater Sound [my former employer], due to reliability issues with some recent products - and for the foreseeable future, Sweetwater will not carry any Antelope Audio gear. Knowing what I know about how Sweetwater works, that's a pretty damning indictment of Antelope.)
Much has been said about the Slate VMS microphone's inability to model the off-axis frequency response of the mics that Slate Digital has chosen to model. However, as I am not working in a large, live-sounding room, and don't often work with whole ensembles recorded live - this is not an issue for me in practice. And, truth be told, I find nothing objectionable about the native off-axis response of the microphones themselves. Recording drums with a half-dozen ML-2s presents no issues with bleed between the spot mics and the overheads, for example. Because the microphones themselves are subsequently simpler, they cost less. And from what I can tell from listening tests, the difference between the Slate products and their competition is negligible to the point that it comes down to aesthetic preference.
Lower cost + less inputs required + similarly great results = no-brainer. Slate it is.
The Bottom Line
There are a bunch of microphones modeled by the Slate VMS that I have never recorded with. So I have no basis for comparison in terms of determining authenticity. (Though there are several meticulous "shoot-outs" that have been performed to compare modeled mics to their parent microphones. Sweetwater has the most thorough shoot-out, found here. Again, the results speak for themselves.) But to obsess over minute differences between the real thing and the not-real thing is to completely miss the point.
What is the point? No other microphone that I know of provides as much value and versatility as a microphone used with modeling software.
The value isn't merely found in that I could effectively record a drumset entirely with U67s - a proposition that would cost 60K to 100K+ in the real world (a feat that is probably only capable in a number of studios that I could count on one hand - and would not be technically feasible since the U67 can only handle a maximum SPL of 116 dB).
It's that I can change the "microphone" with a couple of clicks, rather than taking 3-4 minutes to physically change it out. It's that I have a nearly-complete palette of highly desirable microphone sounds available to me that I could never afford without selling my soul - regardless of whether or not they are spot-on accurate in their representation of the original microphones. And I don’t have to have microphones which only serve one purpose. I only need one mic for every preamp input, and I’m pretty much good. That saves money and storage space. Most of all, they just sound good, and let me get to the business of making music much more fluidly - and quickly.
Look out for some demo/finalized recordings with this new setup. I am going to be busy!
Earlier today I was on the website of one of my favorite DAW plugin authors, iZotope. Specifically, their page for the latest iteration of Nectar 3, which is a complete vocal processing suite in one plugin. There are lots of cool things in this software, most notably a frequency unmasking feature which works with other plugins in their software suite by communicating across channels and letting the user know which instruments are clashing at similar frequency regions. They have been leading the charge with this particular type of mixing aid for a couple of years now, and I used to be pretty excited about it when it first debuted. But now I'm like:
The problem behind the problem:
This is a really simple one. Frequency overlap between tracks in a mix is due to one problem and one problem only:
For example, let's use the extremely simple confessional-boy/girl-at-the-piano genre: are you going to play notes/chords that are in the same octave that you're singing in? NO. (Or, at least, you shouldn't.) You will leave musical space around the vocal so it fits in.
What's more, this idea of how to make a great mix has been understood by composers and musicians since the 18th century, if not longer. You want to hear the bassoon in a symphony orchestra? Then make most of the orchestra shut up, bring the volume level of the supporting instruments way down, and make sure they stay out of the range of whatever the bassoon is playing - and boom, there's the bassoon, no DSP/AI magic required. And it will always, Always, ALWAYS sound better.
A great example of good arranging:
Click play to listen*. Now, while you're listening and reading, take note that this album was recorded with minimalist microphone techniques (by Bob Katz). I don't know exactly how he did it, but I'm guessing he used 3-4 microphones at the most. So there is very little wiggle room to compensate for poor arranging by this group. With drums that are capable of deafening SPLs in the same space as a delicate classical guitar, it's a recipe for a musical disaster. Yet they pulled it off. How? Listen and find out!
And now, here is an example of arranging that could be... better:
And yes, I'm fully aware that I just made an example out of a song from an album that won the Grammy for Best Engineered Album (Non-Classical) this year, by one of the most respected recording artists of our time. But don't be distracted by any of that (I'm also leaving the artistry of the music itself alone - maybe for another time). Another reason I reference this song is for anybody who is running Logic Pro X v10.4 and later, who may or may not know that they have access to the complete session file for this song in their template chooser. If you are one of the blessed, you can see the 130+ tracks that make it up, and how many of them, at any given time, are effectively inaudible unless you solo them in the software. (So why bother?)
And even at that, the four or five elements that are going at any given time are all competing in the same frequency space (the speech intelligibility range - 1k-5k - no less); and while I can't speak for anybody else, this mix is like fingernails on a chalkboard to me. NO AMOUNT OF MIXING, no matter how expert, can fix it. Except maybe some creative use of mute buttons.
(In partial defense of the Grammy's, the other final nominees for the BEA-NC category were actually very well done and deserving of that honor. Check them out.)
IT'S THE ARRANGEMENT, SMART GUY.
* Please sign up for an Apple Music membership (if you haven't already) if you love to listen to music. Or feel free to search for these songs on whatever platform you prefer to use. But you would still do well to consider using Apple Music, as it is currently, to my knowledge, the most equitable service in the streaming market, in terms of compensating artists fairly for their streams. No, the streaming game is not perfect, mind you, but A.M. is the best thing out there given the current (sad) financial reality that faces most musicians. Please also note that I am not a brand ambassador for Apple. This is just free consumer advice.