How do you make sure you are hitting Nebula at -18dbfs?

[Sheikyearbouti]

I am even more confused...

I think an audio with -18 rms and -10 peaks will get more benefits through a neb preset than an audio with -10 rms and -18 peak levels

So you support guys are saying we don't have to worry about rms level coming into nebula just be careful about peak levels?

Look, it is just like the hardware world - it depends always on this or that particular hardware/Nebula preset and how it was sampled. There is only one simple rule - do not go over 0dBFS.

The -18dBFS peak is just a guide. You would expect to get more of the subtle Nebula distortion above it. If you want to push it hard - do it - just watch out for the clip indicator.

But also bear in mind that if the preset is from, say, a super-clean, high-headroom modern S*L preamp, you may not hear a change at all at different input levels (as it would be mostly phase distortion that Nebula has captured, not harmonic distortion - the device wouldn't have almost any harmonic distortion at levels from -18 to 0dBFS peak). So it is always preset specific, just like in the hardware world - you must learn every single preset and the best way is - use your ears and don't clip

[Sheikyearbouti]

an audio with -10 rms and -18 peak levels

hmm, this case is not possible...

[kindafishy]

Too much technical, theory supported reasoning, blah blah for me (with the utmost respect for the many of you who love the technical blah blah)... It is interesting, but in the end, it confuses me and doesn't help.

Ultimately, I would like this to be like a microwave to me. I put stuff in, I push a couple buttons, hot stuff comes out. I don't need to know any technical details or theory or science to enjoy what it does. Just in cold, out hot. Happy fishy.

It piqued my interest when giancarlo said that levels should be a -18 peak. Up until this point, ALL of the material that I have read over the years in regards to Nebula has told me to go in at -18 rms.

So, can I assume that everyone was always wrong, and the real way to properly hit Nebula is at -18 peak?

Seriously... final answer... in cold, out hot. What is the correct approach?

[Sheikyearbouti]

Look at the input meters - do you see peak around -18dBFS? OK, this is your starting point.

Do you want to drive it more? Bring the level of the signal entering the plug-in up - do you like what you hear? OK, keep it.

Want even more drive? Keep going up - do you see the clip indiactor on/red? That's not good - trim the input level down.

And all the time - use your ears.

[Sheikyearbouti]

Ultimately, I would like this to be like a microwave to me. I put stuff in, I push a couple buttons, hot stuff comes out. I don't need to know any technical details or theory or science to enjoy what it does. Just in cold, out hot. Happy fishy.

There is a problem here. If you want a plug-in that is like hardware you must learn to use it as such. And even on your microware you can change the power from 150W for defrosting to 700W for cooking - what are you after? And you know that if you try to defrost at 700W you will have a bad result - you will cook parts of what you try to defrost, but others will be unforzen and fine. It is the same thing.

And on top of this, sometimes we use the oven, sometimes we BBQ, sometimes it is a microware, sometimes we boil. A good chef will learn how to use all of these and all of their abilities and limits.

[kindafishy]

Darn it all when someone uses my analogy against me...



Thanks, sheik (still love your forum name). I get your point. I like the approach of starting with the meters peaking at around -18 and then driving (not to the point of clipping of course).

However...

The main problem with this approach is that Nebula is often (on the order of 90% of the time) extremely subtle. On a single track, it is going to be difficult (if not impossible) to ascertain that driving Nebula at -18 peak or -14 peak is making an appreciable difference.

Where the difference will really be heard is when this is done cumulatively on many instances.

The problem is, of course, that Nebula is much too resource intensive to be able to do this in any practical sense.

If I have 30 tracks, each using a program I want to experiment with, that means that I am unfreezing and freezing many tracks between tweaks to get a sense of what I prefer (especially because I have NO tracks with one Nebula instance, there are more like at least 6 per track, plus bus instances). I just can't take an hour each time I want to drive a program a little more, unfreezing, tweaking and refreezing 30 tracks to see if I like the cumulative change (and another hour, and another hour for more experimenting or for going back on an experiment).

It's just not possible. Not until quantum computing (or something similar) arrives in a few years and we don't have to bounce Nebula (although I am sure giancarlo will find a way to burst a quantum computer into a ball of flames as well).

So, this is where I am coming from. I was hoping that there is simply a best practice that I can follow that will give me the best results without having to spend 18 hours getting there.

Up until this point, that place was -18 rms. Now it looks like we are being told that it is -18 peak.

Either one is fine. I just want to know which one to use, and be done with it.


PS - I use cold water to defrost things, as I think a microwave is inappropriate for that purpose .

[giancarlo]

we'll try to make nebula light, even if a bit worse soundwise
At the moment this is the compromise.... you can freeze, you can render offline, you can build a server... and you can just do everything without this guideline. It does not sound bad.
In acqua plugins, ie aquamarine, we are moving this tech to a transparent level. It is possible we'll build a particular nebula configuration for gainstaging controls like aquamarine ones... it is already possible now....

[kindafishy]

we'll try to make nebula light, even if a bit worse soundwise
At the moment this is the compromise.... you can freeze, you can render offline, you can build a server... and you can just do everything without this guideline. It does not sound bad.
In acqua plugins, ie aquamarine, we are moving this tech to a transparent level. It is possible we'll build a particular nebula configuration for gainstaging controls like aquamarine ones... it is already possible now....

Thanks giancarlo. I don't mind the bouncing to a small degree, personally. It is something that I was quite aware of going in, and I think the trade off is worth it, without question. It is just prohibitive in having a fluid creative environment, that's all. Again, that is okay, it just means there has to be a particular way of working (i.e. commit early and live with it - which is actually great!).

Very interested in the gainstaging you just mentioned. Would you please talk about this some more and explain what it is and how it works? If you have already described it elsewhere, a link is most welcome. At your convenience, of course...

Thanks!

[ngarjuna]

What a mess! This is not the simplest subject to begin with but come on Acustica...

First off, dBFS does not automatically imply peak reading. Maybe it did once upon a time (I'm not sure that's even the case) but it has become very common to compare windowed average level readings to Full Scale and end up with calibrations where averaged windows are related versus full scale. But this is all beside the point.

For years and years the suggestion on these forums have been:
-treat Nebula inputs like hardware vis-a-vis metering; in other words VU metering
-various digital VU emulations were endorsed throughout the years for completing this task; who knows how many copies of Klanghelm's VU meter Tony sold because of this forum alone (and it works great!)
-VU meters are not more like peak meters than RMS meters, I'm not sure where that comes from but the opposite is true: the rise and fall times mean by definition it's not an instantaneous amplitude reading, plain and simple. That is the whole point of the VU spec: to use the ballistics of the VU meter to show more of the average signal behavior than instantaneous reading meters. In fact often VU readings are fairly close to RMS readings (although there is divergence depending on different variables in the scale) but it's true that they're not the same thing.

For example:
Here's the response of a VU meter (black line) compared to instantaneous input level (grey area) of a drum beat. Level is in dB and time is in seconds.

And now all of the sudden you want to tell us it's all about peak metering? Either something has drastically changed or...I don't even know what. The good news is: you can continue following the sage advice that's been dispensed around here for years (treat Nebula inputs as you would hardware inputs, as in VU metering) and get very good results. Of course you should never rely on metering to set your level input to Nebula anyway, only your ears can be trusted, but if you're just talking about getting into the ballpark, which seems to be what this discussion has focused on, then VU metering should get you very close to what you'd see on the hardware meters for 99% of these devices.

[giancarlo]

dBFS are referred to full scale so yes, I'm speaking about peak values
Guys please stop this garbage
In most of presets you should refer to PEAK value
I explained it extensively also in our beta forum, maybe I should repeat it again and again
Anyway a little book is on the way

Depending on the envelope follower used in nebula, sometimes you should refer to RMS, anyway PEAK is always safe
Some of them are based on RMS, some of them on peak

We'll publish a table

[giancarlo]

btw, some vu is based on average - not squared, RMS is not automatic!

In general rms is referred to square value, so norm 2 -L2
Average value is L1, so norm 1
Peak is L oo, it is a limit to oo for average

As you can see, peak is the safest one, it is the loudest one

You should match the type described in the proper library manual
Libraries based on gdrive are based on peak

[ngarjuna]

I don't know whether it's the language barrier or what but you're not providing anything useful to your users by speaking this way. People are not asking for a detailed analysis of various metering (of course VU metering isn't the same as RMS metering, VU metering is a loudness scale) they're asking for real world advice on hitting Nebula from the position of having less experience reading meters of different types.

[giancarlo]

I don't know whether it's the language barrier or what but you're not providing anything useful to your users by speaking this way. People are not asking for a detailed analysis of various metering (of course VU metering isn't the same as RMS metering, VU metering is a loudness scale) they're asking for real world advice on hitting Nebula from the position of having less experience reading meters of different types.

no it was not directed to you really!
It is something not new for 3rd party, so if there is not enough info on it I cannot help!
On acustica side, we NEVER applied gain staging - apart acqua plugins, here the reason for our lacking of infos
Each developer applies the gain staging he prefers

[Sheikyearbouti]

I don't know whether it's the language barrier or what but you're not providing anything useful to your users by speaking this way. People are not asking for a detailed analysis of various metering (of course VU metering isn't the same as RMS metering, VU metering is a loudness scale) they're asking for real world advice on hitting Nebula from the position of having less experience reading meters of different types.

The problem is that it is not all so simple - you are working with dynamic devices here.

This is why if you want to make your life simple, hit the Nebula input with -18dBFS PEAK and see what happens. Than you can experiment and go hotter... I most often do this.

What would be the exact level - this is where the engineer becomes painter - it is your call - you have to be the one that choses the brush and the colours! And if you want to use the cooking analogy - you are the one that has to make a choice if the steak will be rare or medium (we currently cannot burn stuff in Nebula).

And you are right that it depends on the nature of the sound - does it have lots of transients or not, is it compressed already or not, etc. This is why I always look both the VU and peak meters.

But as I said before - you have to learn each preset, as all of these are unique regardless of the standard we follow, because they also depend also on the sampled device. Now add to this that some developers will use a different standard. I persoanlly really like how Eric Beam (STN) does his presets offereing preamps at different standards.

At the end there is only one strict rule - don't clip.

[ngarjuna]

no it was not directed to you really!
It is something not new for 3rd party, so if there is not enough info on it I cannot help!
On acustica side, we NEVER applied gain staging - apart acqua plugins, here the reason for our lacking of infos
Each developer applies the gain staging he prefers

I just had a shower, I think I was getting cranky. Apologies for that, no call for it.

I think we're also talking around each other; in reflection of your most recent posts I think now that you are talking only about Nebula's internal metering which, to be totally honest, I never use and was not talking about. That's my bad because it should have been more obvious to me that was being discussed here.

On that basis I believe I gave the misimpression that one should tweak the Nebula internal metering part of the engine; and that's not at all what I'm saying! Don't do that. I only use external meters in the chain prior to hitting my Nebula instances to make these decisions, perhaps because I've been using Nebula since the days of very short scaled metering in Nebula (of course Nebula's metering support has gotten much more elaborate in the last few years).

I do agree with Sheik that both VU/RMS style meters and peak meters can be useful for providing further information than only using one or the other.

But none of what I'm saying has anything to do with tweaking the internal Nebula metering system (which I believe also affects things like the envelope follower? and probably should not be adjusted). On that subject one should definitely follow Acustica's advice.

Sorry for any obfuscation I have added to this already muddled subject.