Quantcast
When is MIC MODELLING really uptospeed ? - Gearspace.com
The No.1 Website for Pro Audio
When is MIC MODELLING really uptospeed ?
Old 24th January 2009
  #1
Gear Addict
 
🎧 15 years
When is MIC MODELLING really uptospeed ?

I only know of Antares Mic Modeller, and reviews were pretty bad.

Since there are now lots of Guitar Amps & Speaker sims, guitar pedal sims, and all kinds of EQ/Comp plugs... when is it time for real sounding Mics ?

Say, record something with a clean mic and preamp and then make it sound like a U87, a M149 etc. etc.

Is it already existing with high quality sound results ?
Old 24th January 2009
  #2
Lives for gear
 
Volodia's Avatar
 
🎧 15 years
Quote:
Originally Posted by androne ➑️
I only know of Antares Mic Modeller, and reviews were pretty bad.

Since there are now lots of Guitar Amps & Speaker sims, guitar pedal sims, and all kinds of EQ/Comp plugs... when is it time for real sounding Mics ?

Say, record something with a clean mic and preamp and then make it sound like a U87, a M149 etc. etc.

Is it already existing with high quality sound results ?
The recording is only as good as what's already there , you cannot recreate dynamic response , details,transient response when it's not there to start with .

Mic modelling has nothing to do with ampsim .
Old 24th January 2009
  #3
Lives for gear
 
RCM - Ronan's Avatar
 
🎧 15 years
Quote:
Originally Posted by androne ➑️
I only know of Antares Mic Modeller, and reviews were pretty bad.

Since there are now lots of Guitar Amps & Speaker sims, guitar pedal sims, and all kinds of EQ/Comp plugs... when is it time for real sounding Mics ?
Still waiting on real sounding amp sims.

My take is that the whole concept of mic modeling is flawed. how can one emulate the proximity effect and off axis characteristics of a mic after the fact. How can you record with a mic that does not really capture elements of a sound and then make it sound like a mic that accentuates those same elements.

Just use the best mic you have and then use a good quality EQ to change tonal balances to what you are trying to hear.
Old 24th January 2009 | Show parent
  #4
Gear Head
 
🎧 10 years
Quote:
Originally Posted by rcm ➑️

My take is that the whole concept of mic modeling is flawed. how can one emulate the proximity effect and off axis characteristics of a mic after the fact. How can you record with a mic that does not really capture elements of a sound and then make it sound like a mic that accentuates those same elements.

Just use the best mic you have and then use a good quality EQ to change tonal balances to what you are trying to hear.
I completely agree with this. This is also probably why Antares Mic Modeler was discontinued entirely.
Old 25th January 2009 | Show parent
  #5
Gear Guru
 
🎧 15 years
if you took a fuzzy photo of an office building, you would not be able to use a "Lens Modeler" to zoom all the way in on a single window and see who is there.

Yes it happens every week on CSI, but only because the writers don't care.

You can go the other way. You can use modeling to discard information, but not to add information that was never captured in the first place.
Old 25th January 2009 | Show parent
  #6
Gear Guru
 
Ethan Winer's Avatar
 
🎧 15 years
Lightbulb

Quote:
Originally Posted by rcm ➑️
My take is that the whole concept of mic modeling is flawed. how can one emulate the proximity effect and off axis characteristics of a mic after the fact.
Exactly. In theory one could account for those, but the software has no way to know if you were 2 inches from the microphone or 12 inches. All it knows is that you're trying to EQ one type of mic to match another, which is kinda brain dead anyway. It makes more sense to just EQ what you have to sound like what you want, independent of what brand and model mic was used. Or wish was used. heh

--Ethan
Old 25th January 2009 | Show parent
  #7
Lives for gear
 
🎧 10 years
While I agree with everybody here, I found nebula's sampled mic.s useful sometimes (used it on vocals and guitars). Basicly you're adding an eq preset but it can be practical while trying to find the sound you want (and it works way better than the antares plug-in).
Old 25th January 2009 | Show parent
  #8
Gear Head
 
bump's Avatar
 
🎧 15 years
Hi there,
Useless for what they were (under) designed to do, however, I use them to run any old track thru just to give a contrasting or new texture to someting that might be lost in the mix. (Or just cuz its fun.............)
Old 17th November 2009 | Show parent
  #9
Lives for gear
 
imaginaryday's Avatar
 
2 Reviews written
🎧 10 years
i found a good use for this plugin:

take a steel string guitar that was recorded direct (DI), run antares as follows: source in bypass and choose a ribbon mic model or a tube mic of your choice.

it seems to warm-up the signal and rids it of that cold/shrill metallic sound. give it a try and you'll hear what i am talking about.

cheers.
Old 17th November 2009 | Show parent
  #10
Lives for gear
 
duckoff's Avatar
 
🎧 10 years
Haven't used it for years, but I used to like it as a kind of random Eq, just to flick through 'models' & come up with a bit of tonal variation.
I guess I could do the same with some random eq presets......
Old 22nd November 2009 | Show parent
  #11
Lives for gear
 
Melgueil's Avatar
 
🎧 10 years
Thumbs up

Quote:
Originally Posted by rcm ➑️

My take is that the whole concept of mic modeling is flawed.
Right there with you on that.
cdlt
Old 22nd November 2009 | Show parent
  #12
Lives for gear
 
sleeper1400's Avatar
 
🎧 10 years
mic modeler is dumb...lol.
Old 22nd November 2009 | Show parent
  #13
Gear Guru
 
theblue1's Avatar
 
🎧 15 years
Quote:
Originally Posted by androne ➑️
I only know of Antares Mic Modeller, and reviews were pretty bad.

Since there are now lots of Guitar Amps & Speaker sims, guitar pedal sims, and all kinds of EQ/Comp plugs... when is it time for real sounding Mics ?

Say, record something with a clean mic and preamp and then make it sound like a U87, a M149 etc. etc.

Is it already existing with high quality sound results ?
There is far more to a specific mic's particular characteristics than a simple frequency response curve or even nonlinear dynamic responses (and nonlinear dynamic response is something extremely difficult to program into something convincingly, at that). You might get some modifications to your sound that prove useful -- but I think it's highly unlikely that over a range of recording experiences, that mic sim software would produce satisfyingly similar results to the real thing.


Quote:
Originally Posted by Volodia ➑️
The recording is only as good as what's already there , you cannot recreate dynamic response , details,transient response when it's not there to start with .

Mic modelling has nothing to do with ampsim .
I wouldn't say nothing to do with amp simulation. But they definitely require different approaches, since you are attempting to simulate very different devices.

And, while there are those satisfied with amp simulation -- I would suggest that I suspect that most of those folks are people who use additional guitar FX in their chain, real or virtual, rather than clean tone guys. I've played through a fair number of sims and even have a POD XT that a friend parked over here and -- speaking strictly for myself -- while there are things that are useful and amp sims are a good thing to have a couple of in your toolbox -- for me, amp sims are a thin and altogether unsatisfying stop gap.

Quote:
Originally Posted by imaginaryday ➑️
i found a good use for this plugin:

take a steel string guitar that was recorded direct (DI), run antares as follows: source in bypass and choose a ribbon mic model or a tube mic of your choice.

it seems to warm-up the signal and rids it of that cold/shrill metallic sound. give it a try and you'll hear what i am talking about.

cheers.
And that's the kind of usefulness that I'm thinking of. Left-field, unintended uses seem like perhaps one of the more satisfying avenues for further exploration.
Old 17th December 2009 | Show parent
  #14
Lives for gear
 
MonkeyAdam's Avatar
 
🎧 10 years
I would be willing to bet that "mic modeling" is a little more complex than EQing, while much less complex than physical modeling. To go along with rcm, I don't believe that mic modeling purports to give you the pickup pattern of those mics. I suppose in terms of advanced physical modeling beyond what is possible now this would require room modeling, mic and performer placement modeling as well. Not a great idea really, but I bet it might happen in the next hundred years.

Back to the present topic, I would imagine it to be prerequisite that you start off with a clean neutral recording. Surely it is a prerequisite that the recording be in a dead room, so the pickup pattern doesn't really matter. If I had made this plugin, this is what it would be: impulse responses of each mic in an anechoic chamber. Then different impulses for different distances, low-cut filter settings, etc.

Ideally, the impulse response contains data regarding how the mic handles all frequencies over time. Now you can use FFT convolution apply the impulse response to your own signal. So it is a complex multiply (FFT), compounded for the duration of the impulse. This means it deals with spectrum over time. Please don't get lost here... if you were thinking about saying, "But frequency is time and vice versa.", then don't. You're wrong. Other things happen over time other than frequency. Sorry, some people on GS don't quite get that.

Convolution reverb works in the same way. Not to take away from joeq's CSI analogy, which I like, I'm sure that he wouldn't tell you that reverb only subtracts from the signal and doesn't add anything. So no, it won't give you extra detail, but that's not to say it won't add character.
Old 17th December 2009 | Show parent
  #15
Lives for gear
 
3rdeyeKnight's Avatar
 
🎧 15 years
The microphone just might be the only piece of studio front-end that can not be emulated successfully.

In the day of Melodyne DNA and robotic prosthetic limbs being controlled by thought, I dont doubt that somebody, somewhere, is gonna create some serious uproar within the GearSlutz community sooner or later.


Quote:
Originally Posted by MonkeyAdam ➑️

Back to the present topic, I would imagine it to be prerequisite that you start off with a clean neutral recording. Surely it is a prerequisite that the recording be in a dead room, so the pickup pattern doesn't really matter. If I had made this plugin, this is what it would be: impulse responses of each mic in an anechoic chamber. Then different impulses for different distances, low-cut filter settings, etc.

Ideally, the impulse response contains data regarding how the mic handles all frequencies over time. Now you can use FFT convolution apply the impulse response to your own signal. So it is a complex multiply (FFT), compounded for the duration of the impulse. This means it deals with spectrum over time. Please don't get lost here... if you were thinking about saying, "But frequency is time and vice versa.", then don't. You're wrong. Other things happen over time other than frequency. Sorry, some people on GS don't quite get that.

Convolution reverb works in the same way. Not to take away from joeq's CSI analogy, which I like, I'm sure that he wouldn't tell you that reverb only subtracts from the signal and doesn't add anything. So no, it won't give you extra detail, but that's not to say it won't add character.
Sounds like you're on to something there.
Old 17th December 2009 | Show parent
  #16
Gear Nut
 
🎧 10 years
Quote:
Originally Posted by Ethan Winer ➑️
Exactly. In theory one could account for those, but the software has no way to know if you were 2 inches from the microphone or 12 inches. All it knows is that you're trying to EQ one type of mic to match another, which is kinda brain dead anyway. It makes more sense to just EQ what you have to sound like what you want, independent of what brand and model mic was used. Or wish was used. heh

--Ethan
You could concievable use amplitude to drive a proximity effect. E.g. louder = closer -> boost bass. Sort of an envelope follower that modulates a tone control.

That could probably simulate an omni mic response reasonably. Start with an extremely flat electret or something, it might not be too bad, between EQ-ing to sim a particular mic, and emulate the proximity effect.

However, once you throw directional into the picture, can't think of a good way to drive that change from just a single mic.
Old 17th December 2009 | Show parent
  #17
Gear Guru
 
Ethan Winer's Avatar
 
🎧 15 years
Lightbulb

Quote:
Originally Posted by flmason ➑️
You could concievable use amplitude to drive a proximity effect. E.g. louder = closer -> boost bass. Sort of an envelope follower that modulates a tone control.
But without having a way to calibrate a given volume against a given amount of bass boost, it's still not accurate. After all, one can sing louder and softer while remaining at the same distance from the microphone. So how would the modeler know which is which?

--Ethan
Old 17th December 2009 | Show parent
  #18
Lives for gear
 
3 Reviews written
🎧 10 years
Cool

Quote:
Originally Posted by duckoff ➑️
Haven't used it for years, but I used to like it as a kind of random Eq, just to flick through 'models' & come up with a bit of tonal variation.
I guess I could do the same with some random eq presets......
yea, it was actually a really good product for two reasons imho:

gave you instant real-world-understandable eq simulations that really did approximate the tonal characteristics of the mic in question, at least enough that you could kill sibilance or add warmth or a little bit of pseudo tube harmonic distortion in really musical ways. I actually used it for a while and enjoyed it's musicality, if not it's lack of accuracy. I don't recall why I stopped using it....

and the other reason it was good was to teach people without a lot of experience with different mics and micing positions just what one could expect to achieve with proximity effect, choosing different mics and different designs of mics. it was really great as a tool for inexperienced wannabe or newbie recording musicians.

So I actually think it's a good tool. but as Ethan and others wisely pointed out, it can't model a mic when the original mic's positioning and the room aren't taken into account. so it was only really a vague approximation, but sometimes it really did surprise me with how close it sounded to me using the real mic in question.

so I think it's better than maybe some others here think. however, as noted, I don't recall why I stopped using it. maybe I just found that I really didn't need it since I already could dial in some of the desired characteristics in my actual recording/mixing chain.
Old 10th January 2010 | Show parent
  #19
Here for the gear
 
🎧 10 years
microphone modeling

[Ideally, the impulse response contains data regarding how the mic handles all frequencies over time. Now you can use FFT convolution apply the impulse response to your own signal. So it is a complex multiply (FFT), compounded for the duration of the impulse. This means it deals with spectrum over time. Please don't get lost here... if you were thinking about saying, "But frequency is time and vice versa.", then don't. You're wrong. Other things happen over time other than frequency. Sorry, some people on GS don't quite get that.]

It's likely Antare did what you describe. Hard to imagine their engineers didn't think of a time-shifting FFT recalibrated for different distances and angles. But this approach would only work for a linear system. You're trying to model the new mic as the sum of weighted frequencies (thus a linear system). But nearly all real systems have nonlinearities; with a mic there would likely be interactions between frequencies of 2nd and higher order. You need to use nonlinear methods to do the transformation right.

Impulse-response functions can be used to model a system only when the system is fully described by its first-order Volterra term, i.e. when it's perfectly linear. Some people don't quite get this.
Old 10th January 2010
  #20
Lives for gear
 
🎧 10 years
Quote:
Originally Posted by androne ➑️
I only know of Antares Mic Modeller, and reviews were pretty bad.

Since there are now lots of Guitar Amps & Speaker sims, guitar pedal sims, and all kinds of EQ/Comp plugs... when is it time for real sounding Mics ?

Say, record something with a clean mic and preamp and then make it sound like a U87, a M149 etc. etc.

Is it already existing with high quality sound results ?
Don't hold your breath. It's simply impossible to create anything like it. Ever.

Since at the output of the mic capsule the information, which angle of incidence a particular pressure change had, is lost.
Old 10th January 2010 | Show parent
  #21
Lives for gear
 
MonkeyAdam's Avatar
 
🎧 10 years
Quote:
Originally Posted by thestrangeblue ➑️
You're trying to model the new mic as the sum of weighted frequencies (thus a linear system). But nearly all real systems have nonlinearities; with a mic there would likely be interactions between frequencies of 2nd and higher order. You need to use nonlinear methods to do the transformation right.
Great input! Perhaps an element of physical modeling would be well suited to cover those higher-order characteristics. At some point, it starts to sound like a lot of trouble for something that is ultimately a silly little "effect". It seems the general sentiment is that it's a neat idea but a flimsy one at best, while there are many other ways for engineers to get whatever sound they require or desire. The weakness of the idea being that the effect is layered over the end user's mic (I know this is not new input, I'm just summarizing).

Quote:
It's likely Antare did what you describe.
I think that's what I was trying to say (I was just riffing really), BUT now that I think about... It would be much less expensive (CPU expensive) to have performed some analyses and fashioned some EQs around the test data. I'm kinda leaning towards thinking this is what they did.
Old 10th January 2010 | Show parent
  #22
Lives for gear
 
Timur Born's Avatar
 
🎧 10 years
I don't think you need to have that "neutral" input mic with the "dead" room. But the software needs to have a model of both the mic it emulates and the input mic and you have to tell it which mic you are using.

That way the software can "translate" between the two. Obviously it cannot make up for anything missing from your input mic, but it could try to come up with some educated guesses (by knowing the weaknesses of the input mic).

It wont replace a real mic, but it could be quite useful and convinient.
Old 24th April 2010 | Show parent
  #23
Gear Maniac
 
🎧 10 years
I've played around with it and I think it is very dependent on the source mic. It works well with a CAD e100. Now I don't know just how accurate the models are but, it gives you a really nice set of choices. It isn't magic but, it does give you a lot more sound options.
Old 24th April 2010 | Show parent
  #24
Lives for gear
 
toneguru's Avatar
 
1 Review written
🎧 15 years
Quote:
Originally Posted by rcm ➑️
Still waiting on real sounding amp sims.

My take is that the whole concept of mic modeling is flawed. how can one emulate the proximity effect and off axis characteristics of a mic after the fact. How can you record with a mic that does not really capture elements of a sound and then make it sound like a mic that accentuates those same elements.

Just use the best mic you have and then use a good quality EQ to change tonal balances to what you are trying to hear.
As usual RCM hits the nail on the head.

Try and get yourself the real thing (58 Fender Bassman 4/10, 58 Fender Deluxe, Ampeg Fliptop, or Bock 507, 47fet, KM54, 44BX) you won't regret it.

Cheers
Old 25th April 2010 | Show parent
  #25
Gear Addict
 
🎧 10 years
This can change with the use of fractal math. This is used in graphics to enlarge a picture and fill in with information without losing quality. It makes sense that this kind of tech could be used along with algorithms in order to create mic flavors, pre-amp flavors etc by sampling what is there and then filling in what should be there. At that point though we should even be able to accurately model other people's voices. it all comes down to math anyway.

Has anyone seen the new engine in photoshop that fills in data after you select and remove an object? It can fill in the background based on surrounding background.

We are only limited by our beliefs about whether or not something is possible.
Old 25th April 2010 | Show parent
  #26
Moderator
 
psycho_monkey's Avatar
 
🎧 15 years
Quote:
Originally Posted by husky band ➑️
I've played around with it and I think it is very dependent on the source mic. It works well with a CAD e100. Now I don't know just how accurate the models are but, it gives you a really nice set of choices. It isn't magic but, it does give you a lot more sound options.
If you don't know the original models, you can't really comment on how successful something is, or how well it works!

If the resultant changes are useful, great - but just don't call it a mic modeller, call it an eq or something.
Old 25th April 2010 | Show parent
  #27
Gear Guru
 
🎧 15 years
Quote:
Originally Posted by Dove ➑️
We are only limited by our beliefs about whether or not something is possible.
this is not true, if you stand on a train track, no amount of "belief" that the train is not coming will prevent it from squashing you like a bug when it comes roaring through. There are basic laws of physics, math, information theory that are violated with the premise of mic modeling because a microphone is not a processor, it is a capture device.

Quote:
Has anyone seen the new engine in photoshop that fills in data after you select and remove an object? It can fill in the background based on surrounding background.
Nobody wants or needs a mic modeler that can remove a voice and add hiss and room tone. Downhill is easy, it is the 'uphill' claims of mic modeling that are false.

You can not add accurate detail that was not captured in the first place. The best you can do is add some fractal garbage that from a distance resembles the missing detail. It is not recreating the background where these programs fall down, it is recreating the foreground.

No software is going to model your iPhone camera image into the shot taken by a Nikon D90. There isn't enough detail there for it to work with.
Old 25th April 2010 | Show parent
  #28
Lives for gear
 
🎧 15 years
Yep, there are fundamental reasons why this cant work using existing mics. It's not a question of computer processing power or coding. It's about the inital capture of the sound that will then be altered to create the model. Now I suppose that what you could do is develop an anechoic environment with some sort of perfectly even sound capture within the space (laser?). Algorithms could then simulate mic, position and room sound. The technology for that sort of capture is a long way off and just creating the environment would cost a lot more than just going out and buying a nice collection high end microphones.

It's something interesting for the future but I dont think a periodic check of Antares' website is going to be the answer to your prayers on this one.

Jack
Old 25th April 2010 | Show parent
  #29
Lives for gear
 
Dysanfel's Avatar
 
4 Reviews written
🎧 15 years
A mic modeler is as stupid as a plugin that would try to make your $99 squire sound like a Les Paul Standard, or a plugin to make your voice sound like Etta James.

However, you can get interesting effects out of these products.
Old 23rd September 2011
  #30
Gear Head
 
🎧 10 years
Wish they would have one that magically made the PDI-09 do other mic sounds. Once you know what the impulses and speaker sims do, you can almost win a better sound with an EQ...probably the same for the mics. Gigapulse is most likely based on the test tone neutral going through a speaker/mic, then of impulse differential is made from the "better" mic. Impulses aren't bad, but you need to know exactly what you are trying to accomplish to get results. Random playing around with plugins and impulses will consume all the time you might ever spend on getting anything lengthy recorded.


πŸ“ Reply

Similar Threads

Thread / Thread Starter Replies / Views Last Post
replies: 87 views: 44660
Avatar for ray_subsonic
ray_subsonic 24th November 2011
replies: 60 views: 19395
Avatar for dibravibra
dibravibra 19th October 2020
replies: 143 views: 32971
Avatar for chrisdee
chrisdee 30th September 2020
replies: 6381 views: 809581
Avatar for Wilks832
Wilks832 1 week ago
Post Reply

Welcome to the Gearspace Pro Audio Community!

Registration benefits include:
  • The ability to reply to and create new discussions
  • Access to members-only giveaways & competitions
  • Interact with VIP industry experts in our guest Q&As
  • Access to members-only sub forum discussions
  • Access to members-only Chat Room
  • Get INSTANT ACCESS to the world's best private pro audio Classifieds for only USD $20/year
  • Promote your eBay auctions and Reverb.com listings for free
  • Remove this message!
You need an account to post a reply. Create a username and password below and an account will be created and your post entered.


 
 
Slide to join now Processing…

Forum Jump
Forum Jump