Was seeing a lot of chatter about this Nebula 3 technology and was wondering what is it? Thinking maybe we can start a discussion about its benefits and what is this change of workflow thing i keep reading about.

 

It's best described as 'a sort of impulse response' as mattmckay said.
But then a lot more .

It captures not only frequency response, but also harmonic distortion added. And it does it dynamically, so it knows to respond differently on a low level signal than on a hot signal.

So, in simple terms, it's a technology that can 'sample' (With a test tone) analog gear and capture the magic that it can bring. It works specially well for preamps, equalizers, reverbs and sampling consoles.

So, are you a person who can hear the difference between a real Neve 1073 mic-pre / EQ and a plugin, than a Nebula 'capture' of the preamp will have the 'thing' you're not hearing.
If you don't hear the difference between the current state of plugins and the real-world counterparts, Nebula is not for you .

The only thing that comes close to the analog feeling in a digital DAW are the (good) UAD-powered plugins, and maybe some Slate Digital plugins.

The way I hear it, it is unmatched in reproducing Reverb's (specially from old reverb tanks, but also vintage studio reverb-units), console and tape 'emulations' and vintage equalizers.

Now, about the workflow: It are 'captures' or 'profiles' from analog gear. Imagine an old Neve 1073 equalizer again. It has a high-pass, low band, mid band and high/air band. So basically, you have one profile of the EQ all set at a certain setting. Then you have another profile with the high raised a bit, then another with the high even more raised, than another profile with the low a bit more and high off, etc...

So you can't really get a plugin with a lots of knobs to tweak. There are some methods for making it easier (they have knobs and sliders in a profile, but still the interface is very bare-bones and will feel nothing like a good plugin) so you have to use multiple instances to get the same effect.

Another thing is that with great audio quality comes great resource consumption :P. If you have a nice modern fancy Core i7 processor you might think that you would never hit the CPU limit... until you start using nebula.
If I on a track start with preamp emu, console in emu, tape emu, console insert-emu, eq emu I'm pretty much on the limit on a single core of my Core i7 @ 3.5ghz. And the profiles made of an EQ which sampled all the bands and knobs and sampled it at proper 96khz/24bit (for example) can quickly use up to 500mb of ram or more .

So for workflow: The interface is bare bones and doesn't really 'invite' to experimentation (my opinion) and you'll quickly hit CPU limits. So this requires bouncing / rendering tracks more. See it as old-skool P) working with tape: You have to commit more during mixing.
Me for example, I clearly have a 'tracking' face where I do all the recording takes and put the song together. After that I make a copy of the project including all recordings. I apply nebula to _every track_ to mimic an old console (like I said, preamp, console, tape, console, eq) and bounce the tracks. And _then_ I start mixing with the bounced tracks with the 'nebula effect' already printed on them.

They are working on stuff (sometimes they release a product as a standalone VST plugin, in cooperation with popular profilers / samplers). Those plugins often have more optimisation and an interface with more knobs to be like a normal VST plugin, and they're working on a skinning interface who can link multiple instances together (so you get a 1176 skin that can control which capture of a real 1176 to use depending on how you put the knobs ).

But that brings me to the company behind it: The technology is great (if you're into analog stuff), but please remember that it's still a simple company with one or two developers and maybe on moderator / support guy on the forums. They're from Italy and the communication and their english is not always good, even so far as sometimes they answer questions in such a technical and fail-english way that it confuses more than that it answers.
Also, they can say that they're working on certain stuff, but they make no promises. They didn't get accepted by Avid for instance, so they have no Protools SDK so no RTAS or AAX. Also, they have said for almost a year know that they're working on a Mac AudioUnit (AU) x64 version, but as of this moment still nothing. A lot of Mac users on their forum are angry about it.
So, if you're still interested, let it be for what they have right now, not what they say will come later. Who knows, you might be waiting years for it. The VST plugins (even for mac) seem to work fine, and for me as a Windows user everything is fine.

So, I'm a great Nebula fan and user, but also the first to warn that it is not for everybody. You must love it for the sound to sit through all the negatives, that's basically it. Also, if you're happy with your current crop of VST plugins and are not whining that they're nowhere close to the real thing, than leave Nebula behind .

 

hey, dipje, hey, mattmckay,

very thorough descriptions... *yes*

all four of the thumbs of this here me are up, up, up... *yes*

deluge of gratitude particles flowing from the thank you system...

 

You have to receive the license from Avid to make your plugin work in Protools. If you don't get accepted, you're not allowed to make an AAX version. What the criteria are, I don't know.

Onqel - from TSE audio http://www.theserinaexperiment.net/ - had made the well known free X50 and X30 guitar ampsims. He's about to release the big version 2 that makes it commercial (and into a proper complete guitar suite instead of one loose amp sim), and he IS accepted by Avid and thus can make an AAX version. Why Nebula can't while they are a company with paying customers since .. 2008 or something? (I see news articles about Nebula in google from 2007 / 2008). Maybe a language thing, maybe a Mac thing, who knows. One the reasons I despise anything ProTools / Avid related

 

Oh really? Acoustica is from Italy? God...
Is a German Company baby... You messed up the things, because third party libraries for Nebula from developers like AlexB, now that's italian stuff, because Alessando Boschi make that libraries.

I'm a advanced Nebula user, i use it everywhere in my mixes/masterings, and you and the first guy who answered, you guys said only the half of the truth about Nebula.
You can't have problems with Nebula if you use PRESONUS STUDIO ONE as your main daw, because it's freezing option called "track transform" is like a editable freezing, like you would bounce in place the track after you loaded your Nebula stuff on it and you can go back to realtime state when you want to tweak something on Nebula's interface. I even use "timed kernel" mode with up to 30-40ms which improves more the accuracy of Nebula's processing.
The first guy saying that Volterra Kernel Technology is a Impulse Response thing, he just underrate it. You described it a little better, but Volterra Tech it's not only capturing the freq' response like a IR thing would do. It captures all the nuances of the unit, the harmonic distortion, including even the dynamic behaviour.
Nebula's tech is like stealing/capturing the hardware's soul/behaviour, or cloning it, in a accurate way capturing the way that piece of hardware behave.
I'm just too lazy to write more details here, because i actually work on some mastering tasks, but please, don't scare the topic owner.

About the skin, there's plenty of skins on the internet, you even get them for free when you purchase the third party libraries.
As long as he combine Nebula with third party libraries and a daw like Studio One, you mix with nebula like you would mix with the algo plugins. I also combine the nebula instances with algo instances, getting the best of both worlds.

 

Acustica (spelled exactly like that, because it's in italian word), is indeed an italian company, run by a guy named Giancarlo del Sordo. There is a german company named Analog in the Box (which seems to be out of bussiness now) that's related to Nebula, but they were only library developers, just like Alessandro Boschi (aka AlexB) is.

So, you can get up from the floor now

 

Why? Because is more fancy and shiny to have the real deal in in the same room with you, so that way you can impress the clients, you really didn't predicted this?
About "mathematics" on Volterra Kernel Tech, i know it's all about 1 and 0, but tell me, when you mix a vocal through a real hardware compressor, the signal getting back through your converters then into your daw, that channel or the final output of your project (as wav/mp3/whatever) it's not mathematically created as well? *yes* We're not living in the vinyl era listening or recording to tape, so even the real hardware ends up as a "virtual format".
We all release digital music, so the Volterra thing captures the real behaviour of a hardware with that test tone that returns back in the converters so a good Nebula 3rd party program reacts exactly as the real hardware would react, because mathematically as you say it "records" all the behaving and reactions of a real hardware when it's hit by a test tone. Well, that test tone is sent to the hardware clean, but it turns back different, now that details that makes the difference are exactly the things that get's recorded by Nebula, and it's not only Freq Response, but also the other noances and dynamic behaving.
Check for example the VM-Comp program (sampled Vari Mu compressor) and try to match up some levels comparing it with a real hardware and see how big is the difference between the real Vari Mu and the sampled Vari Mu Even the AlexB 4KD (which is the 4000E Gcomp) sound's great when you tweak the program correctly to your master bus in the "timed kernel" mode. If you didn't got into 3rd party libs that much, that doesn't mean that Nebula can't sample a compressor behaviour and flavour.
And about company's nationality, i thought is german because i knew one of their german developers, called Mirco Reimer.

 

The analog signal which the real hardware unit process is only before hitting the converters, just before turning back to your daw. Look, i.e. a vocal track in your project is routed outside of your daw right to a analog compressor, the signal reaches the compressor and all that processing inside the compressor is analogical, but when it turns back to your daw it becomes mathematically again because there's no other way a computer can read a analog signal than converting it to 1 and 0 inside the ad/da converters.(That's why all the 3rd party Nebula programs developers don't forget to specify the converters used for the sampling process of that unit).

Nebula to capture what hardware units does to the sound, it sends a test tone to that particular unit (the way you would route a vocal outside your daw through that unit) and imagine that test tone as a "clean empty bottle".
When the test tone signal "empty bottle" is routed to that hardware unit, it gets processed, then gets out from it back to the converters, Nebula analize that signal "empty bottle", and now, imagine that the "empty bottle" returns full of liquor. Now Nebula can read the "recipe" of that liquor (which is made by 1 and 0 too as long as it reached the converters/computer). Now think about Nebula like a "Drinks Machine" that steals the "recipe" made of 1 and 0 from a hardware unit's processed signal that reaches the daw after did the processing of that test tone, and prepare that "recipe" for other digital "empty bottles" in your daw, using 1 and 0 too as ingredients.
The only analog advantage you get these days from a hardware unit is only it's processing, the smooth way that it scoulpts the sound without making it harsh, but after the proccess of "scoulpting" is done, the analog signal has to go back to your daw so you can hear what that unit did to the sound but the only way your computer can read that analog signal is to convert it through your converters into 1 and 0 digital signal.

Nebula is not that used by majority due to the reasons like is unintuitive, CPU Hug in certain daws (excluding Studio One), and the fact that the parameters of the sampled unit is splited on multiple programs which requires multiple instances in some cases but not all the time.(which some would consider killing their workflow).
I don't suggest to be used only Nebula in the mixing/mastering process if you really want to use it, but you can use algo plugs as usually and loading some Nebula instances only on the main/highlighted tracks of your project(i.e. vocals), when you'd like to add some flavour that usually a real unit would add. For example, if you have a pretty harsh softsynth melody, you can add a Studer 15IPS program (by CDSoundMaster, in their Reel 2 Reel product) and the harshness is smoothly gone but also the sound gets improved.
Even the Moog Minimoog filter got sampled by AlexB and it works great. Also sampled eqs hardware units like Siemens 295, you can try to boost some highs with a nebula program of that unit and just pay attention on how smooth the highs can sound compared to a algo plug eq, and btw, there's no algo eq plug to match the quality of Nebula's program eq which are based on the real deal. Though i won't bet for all the Nebula's sampled programs, but on the EQ and Reverbs side it just kills the competition with the algo plugs.

Totally agree!

 

Thanks to all who posted. You guys really gave thorough enough responses, complex as it may seem i get the idea. I think ill just wait until the technology catches up with itself and my ears are able to distinguish between a real neve1073 and a plug in version. As for the 1 and 0 digital argument it is a valid one, machines can come close to analoge but due to the fact of its mechanical nature of 1 and 0 it can never duplicate a smooth analog frequency curve. Given most ears can't even distinguish between the difference i think many algorithmic plug ins have came close enough while others as experts have said dont even come close. I will keep my eyes on this nebula stuff though as it develops.

 

I recommend you to just start testing the 3rd party products like tapes (Reel2Reel and TapeBooster+ from CDSoundMaster which works together, Reel2Reel in the Pre and TapeBooster+ in the post.) Eqs like AITB MammothEQ (Manley) which beats the crap out of UAD emulations or other algos, AITB DocFear (Pultec), Siemens W295b EQ (CDSoundMaster), Black Master EQ (AlexB) which i use every time for low-end boost in the mixbus and is the only tool that makes the low end thick, round and more feelable which i couldn't achieve the same feeling with a algo. By the way, for bass tracks and not for busses, you can also try the Vintage MPeQ (AlexB) which can make any bass sounds full and awesome, even over a clean sub bass, it makes it sizzle in a orgasmic way due to harmonic distortion of the eq. Also the W492 Neumann EQ has sweet and great mids for vocals, etc.
For reverbs check the EMT 140 Reverb, AKG BX15 - BX20, all by VNXT, also 6k Halls and 480 by STN.
For compressors you can check the Vari Mu comp Fenix EQ (by AlexB) or VM-Comp(by STN) which works great on vocals for that focusing effect.

There's a plenty of quality programs libs out there and its worth trying.

 

I'm asking this to be sure, as I have barely tested Nebula:

Whatever "profile" you use is just that. Right? Like, let's say, I have a sampled 3-band equalizer (that means the hypothetical hardware unit has, for example, three bands to boost or cut at 1k, 5k and 10k). Then, if I'm a perfectionist I would have to sample the response at a flat state, then boost the 10k by a percentage or let's say, 0.1 db and resample that as many times as the movements/adjustments, then the 1k or whatever etc. I can make and take "snapshots" (much like samples), so that the software can analyze what the hardware does at every state.

Then, it will be mirrored but just in that state. I mean, the software doesn't create a "generic"/"estimated" profile based on those snapshots to try and re-create the behavior of the unit as a whole. So, if I have sampled the +0.1db that doesn't mean that it can "guess" the +0.2 based on the flat or +0.1 or both, or the materials of the unit. And I mean that, as in "by design". It's not a physical modelling plug-in. Right?

I think the same stands for WAVES Q-Clone, no?

And, anyway, I also wanted to draw a potential drawback from this. And, that is, how flexible it can be. The hardware may have a specific sound but that also comes from not only being hardware (which can mal-function, get "burned-in", get old, get built from different materials [like a software can have different approach or code] etc. and thus change some characteristics even slightly) but also that it has "that sound" because it works in a very specific way and in a very specific area.

My EQ may have a 1k boost or cut which I may extremely like for whatever reason but what about if I want to add or cut at 2k? Then, you can feel how limited it can sound/be. So, I'll have to go to a different EQ or leave it as is. So, that's not very handy... I'd most likely need to use an emulated (and not as accurate?) software equivalent or just try another software EQ. Plus, the software EQ would probably have way more filters, curves, Qs, frequency bands, etc. while not having the "character" of the hardware. But, it would get the job done even if not "ideally".

So, it's not a simple matter of being "better" or at least, preferable (and that is, based on subjective preferences).

Same goes with Reverbs where you can have different algos, frequency EQ'ing, room modelling, different approaches, way more settings (like pre-delay, reversing etc.) That doesn't necessarily mean that one is better than the other, rather than both can fulfill different needs and different tasks.

That is why during my testing I used a console emulation as a "tone shaper" but I couldn't get myself to using the EQs or Reverbs because they felt very limited. Ok, that reverb sounded full and awesome but what if that isn't what I'm after?

That is the very reason I'm mostly leaning towards psysical modelling for instruments or a combination of sampling and modelling because you can simply make it your own. Yes, it may not be very good or accurate but it opens possibilities. You could have a perfectly sampled piano but it could also be way different from what you're after. So you' still load it with effects or go to the core and create your "own piano". Honky, wildly detuned, totally unnatural but that doesn't mean it would be "worse" in every sense. It's preferences, context, creativity, whatever... It can be an enhancer, a shaper...

Same goes for "Which DAW is the best". (You can have things that can be measured and listed [different features] but it's not the only thing that matters)

I think you get my point. Nothing I'd leave out but also nothing I'd deem "godlike". And hopefully, I'm not missing something major.

 

Even a old "burned-in" piece of gear affects the sound you send through it in its own way, Nebula perceive what exactly that unit did to the sound (and yes, usually these developers use their own hardware units which they use everyday in their studio, so they doesn't buy a new one especially for sampling it with Nebula, and AlexB even improves his hardware units before sampling it, adding custom better tubes and etc.) All the nuances that the hardware unit adds to that test tone in the sampling process is captured by Nebula, but in a dynamic and organic way.
Let's not go again in theoretical details, just why don't you test further a 3rd party Nebula program like the Mammoth EQ, set the program on "timed kernels" with a value of 40ms and boost some highs on that particular raw track you'd like to use for test. Bounce that track and name it "Nebula Processed", then route that raw track to a real unit of Manley Massive Passive using the same boost amount you used in Nebula but also make sure to match the volume levels of both files. Tell me what you hear . You can do the same for the bass or mids too, is up to your choices. So the real world tests are more important than the theoretical details, that way you can see for yourself. Also make sure you have some good monitors to hear good enough.

Nebula is more advanced than Q-clone, because Q-clone only preserves phase/filter response, while Nebula is far more advanced using clean kernels and a different further sampling approach, the kernels making Nebula process the sound in a more organic/natural and dynamic way.
Saying it shortly: Q-Clone reacts statically while Nebula reacts dynamically. *yes*

 

To change to timed mode, first you have to hack your nebula.xml file, on the LTIMED change the value to "100000" so you remove the limit.(but you won't need to go over 50ms in the "kernel page" on the interface of nebula anyway). Then in your Nebula's interface after you open the plugin, go to "Kernel" menu and there should be between "Timed" and "FREQD" some small arrows which represents the "Clean", "Even" and "Odd" parameters. Click that arrows and they should go to the "Timed" side, then change the ms over 10 (whatever number you like, 20, 30, 40, but the more you rise it, the more cpu will be eaten). These parameters represents the "flavours" of that hardware unit, even the imperfections details.

Take in consideration the fact that some people can't hear the difference on crappy monitors, but it do make a little difference in quality/spatiality. I usually use timed kernels due to the fact that my Studio One daw allows me to go hard with the cpu with no problems doe to the "Track Transform" feature.
Nebula can't be used for tracking, but if you want to lower the CPU on the reverbs, go to "Master" menu in the interface and change the "dspbuffer" to a higher value. I rised it from the default 512 to 2048 (make sure you change the value on Nebula Reverb plugin, not on the normal Nebula) and the cpu utilisation got improved significantly, using the cpu like a normal algo reverb would do.

 

I'm not trying to make an argument against Nebula. As I said, everything has its uses. I just strive to have a balanced view. Anyway, here are some good resources:

Nebula F.A.Q.: http://www.acustica-audio.com/phpBB3/viewtopic.php?f=4&t=396 (From that it seems that what I said pretty much stands based on the question about pre-delay)

Tweaking MAST page #1: http://www.acustica-audio.com/phpBB3/viewtopic.php?f=11&t=1847

Tweaking MAST page #2: http://www.acustica-audio.com/phpBB3/viewtopic.php?f=25&t=1858

List with free Nebula libraries: http://www.acustica-audio.com/phpBB3/viewtopic.php?f=10&t=370

Nebula VST Plug-in Tips – Switching To TIMED Kernels: http://www.learndigitalaudio.com/blog/nebula-vst-plug-in-tips-switching-to-timed-kernels

Review: http://blog.subvertallmedia.com/2011/04/11/nebula-setup-and-basic-use/

Review #2: http://blog.subvertallmedia.com/2012/11/30/nebula-19-months-later/

It just seems that what he figured after a year, I though after two projects. He makes some of my points but from a different point of view. Whatever... I find it is good reference material. (I had that "unregistering" issue happen to me btw)

 

wow thanks man That was my next thing was if i want to try it out where do i start. imma try it out imma imma try it out!!!

Do you need a base program to run these plugs in other words do u need audio acustica program to run a nebula 3 plug or is this self contained within each plugin?

found a good video about gain staging

 

Nebula 3 Pro is just one plugin, the sampled units i mentioned above are actually sorted as "programs" or "libraries" that you fill the Nebla 3 Pro with it and use them inside the plug. After you bought a library, you get all the installing instructions in the same package with it, and it's not that complicated though.

 

Thank you for the GREAT explanation with Nebula. I never use this technology but i will try the demo products. Can we use the Free version of Nebula and buy the sample products?

 

More infos :
http://www.learndigitalaudio.com/blog/nebula-vst-plug-in-tips-switching-to-timed-kernels

Cheers

 

No, i'm sorry. Usually most of the 3rd party developers specify the fact that you must use the commercial version of Nebula to be able to use their Nebula programs.

 

I recently changed the computer and ran into problems with Nebula. No matter what I do with the xml, I get crackles and artifacts both in a 64bit Studio One and a 32bit Live. I use the non-reverb plugin for mojoing and sparse eqing. No reverbs and very rarely comps. I guess I have a setting problem, since the machine (I7 3770k proc) should handle easily all this.
As many of us probably use the same software version I was wandering if some of you guys who got it working could post here the xml settings you use and maybe any specific parameter tweak on the most used libraries.

Anyway, any tips as to how to set asio buffers or anything else really for Nebula to run smoothly is much appreciated. Thanks in advance!

P.S. I post here cause I don't want to bloat the forum with too many Nebula threads, but this seems a tricky and much important issue although some are getting pissed with all this recent hype on the topic...