R/E/P Community

Please login or register.

Login with username, password and session length
Advanced search  

Pages: 1 2 [All]   Go Down

Author Topic: Slate Virtual Microphone Claims  (Read 9227 times)

klaus

  • Moderator
  • Hero Member
  • *****
  • Offline Offline
  • Posts: 1659
Slate Virtual Microphone Claims
« on: January 20, 2014, 07:50:37 pm »

Received a PM from a client, who wanted my opinion on this:

http://www.youtube.com/watch?v=_rzzoFCm8qk&feature=youtu.be

Towards the end of the presentation, the presenter claims this:
"With the VMS you get the EXACT SOUND of classic and modern microphones and pre-amps".
Logged
Klaus Heyne
German Masterworks®
www.GermanMasterworks.com

aremos

  • Jr. Member
  • **
  • Offline Offline
  • Posts: 59
  • Real Full Name: Ariel Remos
Re: Slate Virtual Microphone Claims
« Reply #1 on: January 20, 2014, 09:39:53 pm »

Didn't Antares (Auto-Tune) put something out like this years ago?

Where you could record with a 57 (or even a Radio Shack mic) and then put the original recording into the plug-in & make it sound like a C800G, C12, U47, etc.?
Logged

Jim Williams

  • Hero Member
  • *****
  • Offline Offline
  • Posts: 575
Re: Slate Virtual Microphone Claims
« Reply #2 on: January 21, 2014, 11:22:46 am »

As I recall they included an Audio Upgrades 'modified' AKG 414 B-ULS in their samples. I always wanted my SM58 to sound like a modified 414.
Logged

radardoug

  • Full Member
  • ***
  • Offline Offline
  • Posts: 100
  • Real Full Name: Doug Jane
Re: Slate Virtual Microphone Claims
« Reply #3 on: January 21, 2014, 02:56:31 pm »

What a load of B**S**!  Quite smart though because you have to buy their hardware which is presumably not cheap.
Logged

klaus

  • Moderator
  • Hero Member
  • *****
  • Offline Offline
  • Posts: 1659
Re: Slate Virtual Microphone Claims
« Reply #4 on: January 21, 2014, 03:30:54 pm »

"Slate's" premise here is to start with a proprietary, "completely frequency-flat, uncolored" microphone", upon which you are then able to superimpose the various mic sounds you wish to emulate through electronic manipulation.

One fundamental flaw in this approach (there may be others) is the mechanical properties of the capsule.
Anyone who has, for example, followed the lengthy discussions about the difficulties to reproduce the mechanical response of a Neumann M7 PVC capsule will understand that, even slightly different diaphragm tension, different substrate material, different sputtering methodologies... will alter the sound of the capsule, for everyone to detect and appreciate (or not).

This is just one of many optional capsules in the mix of hundreds of mics to be emulated.
In simple terms: you cannot electronically emulate or copy the complex mechanical/dynamic behavior of a capsule's movements.

And this is where the whole concept propagated so effectively (three microphone forums currently have threads on this new invention) may fall on its butt.
Logged
Klaus Heyne
German Masterworks®
www.GermanMasterworks.com

Kai

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 314
Re: Slate Virtual Microphone Claims
« Reply #5 on: January 21, 2014, 05:52:58 pm »

... the presenter claims this:
"With the VMS you get the EXACT SOUND of classic and modern microphones and pre-amps".
I have used various ways to model mics:
Antares, convolution and the "learn sound" option of Sequoias FFT Eq.
Modeling my own (say known) mics always yielded useful results.
Was it authentic 1:1 - no, but close if the mics where not too different.
Did antares give an exacting sound of a historic?
I have no idea, I don't know their "original".
If the VMS is decently built, it can be very useful.
I will have an eye on it.
Regards Kai
Logged

soapfoot

  • Full Member
  • ***
  • Offline Offline
  • Posts: 230
  • brad allen williams
Re: Slate Virtual Microphone Claims
« Reply #6 on: January 21, 2014, 06:14:20 pm »

even if it were possible, it sure doesn't sound like much fun.
Logged

David Satz

  • Full Member
  • ***
  • Offline Offline
  • Posts: 122
Re: Slate Virtual Microphone Claims
« Reply #7 on: January 21, 2014, 10:51:27 pm »

Klaus, it isn't necessary for a device or software routine to "emulate or copy the complex mechanical/dynamic behavior of a capsule's movements" since the source material in the type of situation that we're talking about is already the output of a capsule. I think that the main flaw in the idea really lies elsewhere.

It may be possible to emulate very closely the response of a given microphone at one angle of sound incidence. But different capsule designs produce different relationships between their response at any one chosen angle of sound incidence versus their response at other angles.

This is why the manufacturers have to print different polar response graphs for their different microphone models; otherwise they could all use the same set of graphs, no?

So if you only record single-point sound sources in anechoic chambers, then I suppose an emulator could work pretty well for you. But with most real sound sources in most real acoustic spaces, a rather large amount of the sound energy arrives from different angles; I would think that even the best possible emulation would be less successful in that case.

--best regards
Logged

klaus

  • Moderator
  • Hero Member
  • *****
  • Offline Offline
  • Posts: 1659
Re: Slate Virtual Microphone Claims
« Reply #8 on: January 22, 2014, 01:45:53 am »

Klaus, it isn't necessary for a device or software routine to "emulate or copy the complex mechanical/dynamic behavior of a capsule's movements" since the source material in the type of situation that we're talking about is already the output of a capsule.
How so? The source material in this system includes a mechanical-to-electrical transducer (condenser capsule). Let's assume for clarity of my argument, that the capsule they are using in their emulator mic is of sufficiently low enough quality, that its capacity to translate complex acoustical waveforms is severely compromised.

How then, could such capsule deliver to the electronic processor/emulator enough data points, for the processor to read and fill in the missing information to render, let's say, a K47 or MK2 capsule's characteristics? I could understand that this system could somehow work (never mind all the lossy processing from such a complex machine), if the exact dynamic behavior of the capsule was available for rendering. That way, one could possibly fake the sound characteristic of an M49, U47, M147 or M149 with the K47 capsule input available.

But if the timbral impulses of the original capsule-the kind that determine its characteristic personality- are not sufficiently present to trigger a specific pre-recorded mic model algorithm, how can it be faked? Wasn't this the same problem with the Antares mic modeler-where you supposedly could plug in your SM58, to produce the sound of an ELA M 251?

P.S.: if the video isn't an all-out fakery in the first place-no live footage is shown during the A-B tests starting at 2:09- but an actual comparison between a real U47 and the modeler system, I don't think that the result is all that flattering to the product: The sibilance of the modeler, with its thinned-out esses doesn't please my ears half as much as the U47 (with god knows what electronics and capsules inside!)- all this with Sennheiser in-ear phones straight through the Macbook Pro.
Logged
Klaus Heyne
German Masterworks®
www.GermanMasterworks.com

David Satz

  • Full Member
  • ***
  • Offline Offline
  • Posts: 122
Re: Slate Virtual Microphone Claims
« Reply #9 on: January 22, 2014, 08:54:19 am »

Sorry, Klaus, you apparently meant that the complex characteristics of the particular "target" capsule would not be present in the input signal, which is of course true. When you wrote "a capsule's movements" I thought you were saying that capsules generally have special modes of motion that can't be synthesized well enough for this type of tool to succeed at what it claims to do.

But even if you are saying that each particular capsule has such characteristics, that's unfortunately an inherently unprovable statement. There are ways to highlight the limitations of microphone modelers that don't have that problem, and what I wrote is one.

There are other contradictions in the claims made by this vendor. Why are two different, supposedly "completely neutral" microphones being offered--one large and one small? Are they not identical to each other, being both "completely neutral"?

I also wonder at the statement that the plug-in is neither a convolver nor a filter. More specifically I wonder whether whoever wrote the script knows the actual meaning of those terms, because if you're putting in one signal and getting out a non-linearly altered version of that signal, it's a filter. And I'll bet a shiny nickel that it does use convolution based on impulse response, directly or indirectly, whether or not the writer of the script realizes it.

One last point which is petty, I know, but: Isn't the guy singing into the back of the U 47 in the first part of the video? The label can be seen on the side facing away from him.

--best regards
Logged

Kai

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 314
Re: Slate Virtual Microphone Claims
« Reply #10 on: January 22, 2014, 09:42:58 am »

I'll bet a shiny nickel that it does use convolution based on impulse response, directly or indirectly, ...
I hold against this bet  :)

They claim their plugin has zero latency, this is impossible with convolution or any other type of windowing process like FFT.
It's (almost) possible eg. with FIR or IIR filters.
Additionally they seem to use saturation effects.

Of course they use filters of any make, or how else should the frequency response curve be adapted to the mic (and preamp!) they model?

I'm quite interested in their system. As soon as it's available here in Germany I will get one to test.

If you look at it as an investment - that not so good.
As it combines hard and software, you have to rely on, that the company building it will continue to exist and willing to update the software in the future.
Very likely that any time near or far this will no longer happen and render the system (per se) useless.
This will not happen with a real U47 eg.

Regards
Kai
Logged

Jim Williams

  • Hero Member
  • *****
  • Offline Offline
  • Posts: 575
Re: Slate Virtual Microphone Claims
« Reply #11 on: January 22, 2014, 11:21:13 am »

The main flaw is a microphone is an analog device, all these emulations are dependent on a digitized signal.

So far no conversion I've ever heard can emulate the complex, non-band limited details of the analog source, whether a microphone, guitar, voice or orchestra.

A perfect simulator will be dependent on perfect conversion. Neither are available.
Logged

brightmillion

  • Newbie
  • *
  • Offline Offline
  • Posts: 19
  • Real Full Name: dan holman
Re: Slate Virtual Microphone Claims
« Reply #12 on: January 25, 2014, 03:28:38 am »

How would it deal with proximity effect, if their mic is totally flat - how will it know how close I am singing to the mic, since it will in theory record a "flat" representation of my singing.. but when I switch models will it magically know how to adjust the proximity to make it sound like their respective mics "would have" recorded it? (or is this something that theoretically could be EQ modeled in..?)
To me that is the magic of a lot of vocal mics- the way a singer can play with proximity to get the appropriate "vibe" for a song before it even gets recorded.

On a more "hippie" take at the topic- It seems to me that every piece of gear has an (indirect?) quality that comes from the hardware, which in turn affects the final usage and result. Such as a guitar modeler that does strats, les pauls,etc. I find that you inherently play the same musical piece differently on a les paul vs a strat, and that may have some affect on the resulting "sound" we associate of that particular instrument that maybe can't be specified in a model. Or how the layout of the knobs on an EQ might influence how the piece is used (yes, you can always go against the design) but maybe what we all associate with a "sound" of a piece of gear is influenced by how it has most frequently been used?
Maybe its just me, but when I stand in front of a great mic I feel like I owe it to perform better ;) It really does pull out a better performance.. Singing into a 57 just doesn't make most performers reach for that magic. Even if they were to know that we'll "u47 model it in the mix"
Logged

Jim Williams

  • Hero Member
  • *****
  • Offline Offline
  • Posts: 575
Re: Slate Virtual Microphone Claims
« Reply #13 on: January 25, 2014, 12:17:04 pm »

Saw it at NAMM yesterday along with the collection of virtual console screens. It all sounded like NAMM noise to me.
Logged

BradL8068

  • Newbie
  • *
  • Offline Offline
  • Posts: 8
Re: Slate Virtual Microphone Claims
« Reply #14 on: January 26, 2014, 04:27:02 pm »

Sadly I fear this will be "close enough" for new people entering the industry.

I run a three-room large facility in N.Y. Our overhead is incredible. All rooms have monitoring systems that cost about 80 Grand. I have consoles that cost around $500,000 new, and many vintage outboard eq's and compressors, a closet full of vintage microphones. Yet people working in their home studios are "close enough". Heck, I'll bet my expenditures on headphones alone exceeds the budget of many home studios.

MP3's were "close enough" to CDs, which were close enough to vinyl. Digidesign tried with liquid, Antaries tried with their mic modeler. What concerns me with Slate is he is currently respected as a plug in designer. A recent survey of my clients' desires for plug-ins all listed 'Slate'.

The irony to me, as far as his virtual console collection is: if older consoles were that colored they would not have been considered professional.

I wish this entire industry was more focused on compelling music, and capturing that sound, and less on gear and plug- ins.
Logged
Brad Leigh, leighaudio.blogspot.com, leighaudio.com

Piedpiper

  • Jr. Member
  • **
  • Offline Offline
  • Posts: 76
  • Real Full Name: Tim Britton
Re: Slate Virtual Microphone Claims
« Reply #15 on: January 27, 2014, 12:56:56 am »

extremely well said...
Logged
row row row your boat...

Pied Piper Productions

didier.brest

  • Newbie
  • *
  • Offline Offline
  • Posts: 15
Re: Slate Virtual Microphone Claims
« Reply #16 on: February 06, 2014, 04:35:37 pm »

I hold against this bet  :)

They claim their plugin has zero latency, this is impossible with convolution or any other type of windowing process like FFT.
It's (almost) possible eg. with FIR or IIR filters.

Finite Impulse Response (FIR) filtering is convolution.

Quote from: Wikipedia page on FIR
The output y of a linear time invariant system is determined by convolving its input signal x with its impulse response b.

The zero latency claim, which means that there would be no (sensitive) delay between the signal of the actual microphone and the signal of the virtual microphone, is legitimate if the Slate Virtual Microphone is based on minimum phase filtering.

The impulse response may be made dependent on the level of the input signal, which is named dynamic convolution and allows for modelling the non-linear behaviour of the microphone to be emulated.
The Liquid Channel from Focusrite is quite good in doing this for vintage preamps emulation. I agree that emulating a microphone is more difficult than emulating a preamp because of the dependency of the impulse response of the microphone with respect to the direction. But good modelling of the response on the main axis might deliver interesting results at least in some cases (close miking of a voice ?). Who knows ? Not me. I would be interested in blind listening to samples comparing the original and the copy.
Logged
Didier Brest

Kai

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 314
Re: Slate Virtual Microphone Claims
« Reply #17 on: February 07, 2014, 01:07:53 pm »

Finite Impulse Response (FIR) filtering is convolution.
Sorry, I would have needed to be more precise:
I meant convolution based on real world impulse response, which is not as "finite" as one would like it to be for fast processing, not on (maybe simpler) mathematical filter functions.
To my knowledge and experience the computing of such takes too much processing power to be made in quasi realtime, except if you use dedicated DSPs.

In fact a real world impulse response covering the frequency range down to 20 Hz needs to be at least 50 ms long, practically (due to windowing) about double of that.
100 ms (4410 samples @ 44.1 kHz SR) is a lot of data to be convoluted with the input signal.
Now make this dynamic with several IR's blendet -

I would be very interested how they do it, maybe they found a shortcut?

I would be interested in blind listening to samples comparing the original and the copy.
That's exactly what I expect to do.
"All theory is grey".

Regards
Kai
Logged

klaus

  • Moderator
  • Hero Member
  • *****
  • Offline Offline
  • Posts: 1659
Re: Slate Virtual Microphone Claims
« Reply #18 on: February 07, 2014, 03:43:16 pm »

Hello Kai,
Thanks for going into that much detail, but… you lose a lot of readers if these thoughts cannot be put into more simple, easier to understand terms. I will remove this request as soon as you have done so.

Best,
KH
Logged
Klaus Heyne
German Masterworks®
www.GermanMasterworks.com

David Satz

  • Full Member
  • ***
  • Offline Offline
  • Posts: 122
Re: Slate Virtual Microphone Claims
« Reply #19 on: February 07, 2014, 11:14:49 pm »

I think that the presenter has a very convincing manner, but sometimes the ability to be convincing is inversely related to how well one really knows what one is talking about.
Logged

klaus

  • Moderator
  • Hero Member
  • *****
  • Offline Offline
  • Posts: 1659
Re: Slate Virtual Microphone Claims
« Reply #20 on: February 08, 2014, 12:02:49 am »

Is this just a feeling, or can you give an example, where the information given by the presenter is inaccurate or misleading?
Logged
Klaus Heyne
German Masterworks®
www.GermanMasterworks.com

David Satz

  • Full Member
  • ***
  • Offline Offline
  • Posts: 122
Re: Slate Virtual Microphone Claims
« Reply #21 on: February 08, 2014, 08:33:51 am »

?? I and some other people in this thread have been talking all along about ways in which this product can't possibly live up to all the technical claims that have been made for it, since they are logically contradictory. Having already done that, I wanted to make an observation about the persuasive quality of the video.

If postings based on feeling were banned, this forum would save about 80% on bandwidth and disk space, with your own postings being among those most affected. Fundamentally this forum is about esthetics; even the discussions about facts and science are ultimately governed by feeling.
Logged

klaus

  • Moderator
  • Hero Member
  • *****
  • Offline Offline
  • Posts: 1659
Re: Slate Virtual Microphone Claims
« Reply #22 on: February 08, 2014, 12:52:31 pm »

(…) Fundamentally this forum is about esthetics

The discussion of audio esthetics is one of this forum's aspects, equal in weight to other subjects which participants are interested in to have the best possible understanding and usage of recording microphones.

What is sometimes perceived by others, and maybe you, too, as a forum void of, or hostile to, scientific examination of microphones, and audio products in general, is my strong skepticism towards the assumption that, if we cannot measure or quantify an aural phenomenon, it cannot possibly exist, and those who claim its existence are fools.

This attitude had led to decades of delay in the development of better sounding stored music, recording devices and musical instruments based on said storage, and it is still at the heart of the crux in most modern microphone designs: a prevalent attitude that "good" measurements are the key to good sounds and sales.
Logged
Klaus Heyne
German Masterworks®
www.GermanMasterworks.com

Steven Slate

  • Newbie
  • *
  • Offline Offline
  • Posts: 1
  • Real Full Name: Steven Slate
Re: Slate Virtual Microphone Claims
« Reply #23 on: February 11, 2014, 02:31:11 am »

Thanks so much to Klaus for the discussion on the new VMS.

We have worked with some very talented microphone designers to develop two microphones that have an extremely high bandwidth and linear frequency response with a sufficiently high quality capsule.

Then the issue is measuring the reference microphone response.  It took us several months to develop a process that allowed us to capture enough complex measurements that were adequate enough so that we could develop an algorithm that was able to recreate the tone of the specific microphone.

I realize that there are a lot of physical properties involved in the way a mic capsule such as the M7 operates, but ultimately we found that the response was something that could be measured and then recreated algorithmically.

The polar pattern of our microphone shares a similar ratio of off axis frequency deviation to the classic Neumanns, which is aided by the sheer fact that the attenuation of high frequencies in off axis response makes it harder for the human ear to discern minute differences.

This is proven by the fact that when our 47 algorithm was about 80% accurate to the real 47 in its on axis response, it was already indistinguishable off axis.

The goal of this product is to offer another tool for music makers to help create their art, and we're very excited about the progress thus far.  Thanks again for the discussion,

Cheers,
Steven
Logged

bodtbody

  • Newbie
  • *
  • Offline Offline
  • Posts: 22
Re: Slate Virtual Microphone Claims
« Reply #24 on: February 12, 2014, 05:53:08 pm »

You want to sell microphones , why don't you say so.
If this was serious , the software would accept any microphone in the input.
The software could then emulate a " toys are us" microphone to the very fine microphone you say you will sell.

Logged
venlig hilsen JP

Jim Williams

  • Hero Member
  • *****
  • Offline Offline
  • Posts: 575
Re: Slate Virtual Microphone Claims
« Reply #25 on: February 13, 2014, 12:04:50 pm »

Antares has had that product out for years. It didn't seem to slow the sales of expensive mics, though.

For me, there are too many variables to include in a modeling stew, like capsule, internal wire selection, front end mic amp designs, front end capsule input impedance, output amplifier designs- solid-state or transformer- the mic cable itself, the preamp, the pre- amp's input impedance, the pre-amp's input design, s.s. or transformer, the converter, bit depth, sample rate and bandwidth, the converter's dynamic range and THD specs, on and on.

Then you can start the whole thing over again with the playback components. The DAC, sample rate, bit depth, clocking design, analog I/V design, analog filter design, power supply design, so on and so on.

Change just one of those factors and that reference is reset.
Logged
Pages: 1 2 [All]   Go Up