Skip to content

So you want to use your amp sims better..

or

Audio reference level best practices and how they relate to guitar analog emulation in the digital domain.

Author: Ted Onyszczak

Do you feel like you’re on a never-ending quest for perfect gain-staging and understanding levels in your digital signal chain? Then this in-depth article that provides both basic understanding and how to apply that understanding to your digital signal chain could be for you.

The article is written by Ted Onyszczak who have 25+ years experience of working with audio. It is reproduced on the site with his kind consent. The original content has been edited for clarity and presentation. For more info see the tab “Credits & license”.

1. Introduction

Image from Audient

So just mess around with it till it sounds good.

Aaaaaaaaaaand … done.

Haha just kidding.

If you’re reading this, I presume you’re the kind of person who hates that answer on public forums and Youtube videos. Yes, you can get a good tone by just using your ears. But you can take it to the next level in sound, playability, response and usefulness if you expand your knowledge of how the world of digital audio and analog emulation work. It is possible to be faithful to an analog chain of gear, but it’s not easy or as intuitive.

Not everyone has to be, or needs to be, this pedantic about their digital signal chain. I’m a firm believer in doing whatever works for you. Quite frankly, for most people ‘keep the peaks at -6’ will do it. But there are people who seek the reason behind things.

This is for those people.

For those who spent a long time learning about their pickups, pedals and amps, and want that experience in the computer it can be infinitely frustrating. Many people complain that digital simulations don’t respond like a real amp does, or the tone is too harsh. Level practices play a fairly significant role in overcoming those issues and improving the sound of digital emulations. With some help you can wrestle the wealth of available guitar software options into a more flexible, usable, system with an amazing tone comparable to the real thing. It takes a bit of learning and doing, but once you truly digest this knowledge it will step up your sound to the next level.

I’m not answering all questions for everyone. And nothing I’m saying here is genre or tone specific. And in the first version of this I’m not going to cite too many examples. This is for bringing awareness to a range of technical issues relating to getting the best basic practises for recording and using digital guitar products in general.

I’m assuming a basic knowledge of audio, guitars, and recording. While I go into some depth on the technical terms that this is directly about, I’m not going to break down every basic concept and explain them all. Working knowledge of your DAW of choice is a must. I use Pro Tools but I’m familiar with most. I’ve sunk probably $16-20k into software over the years at least. Most direct DAW references I will make are for Pro Tools, but I try to keep it pretty generic. I will have to assume you can infer and extrapolate for your DAW of choice. Also I’m an old fuddy duddy. I know that people want examples, walk throughs and ideally a bunch of youtube videos. But I worked in audio post for 25 years and I’m not prepared to make videos to the quality I would demand of myself. Plus time, and desire. If there is a demand I may be coerced in the future.

I’ve been wanting to write this for years. I started a bunch of times but kept getting pulled away. But the main hurdle I had was coalescing my thoughts, experience, and knowledge into a coherent, concise form. Talking about audio can lead to many tangents. You have to find a way to make a non linear story linear. Like any author I now suspect. I endeavoured to do that. Results may vary.

I would suggest going chapter by chapter and only moving on when the basic premise of the current chapter has sunk in. Paragraph by paragraph even. If this creates more questions than answers I find that to be a success. That is essential how I learned what I know reading stuff way back when I started. I would pour over something, make ‘a ha?’ moments and go back and pour over it again getting farther each time.

If you absorb what I have organized here, that is only the start of the journey. A foundation. Where it goes from there is your choice but you will have more tools to steer that journey. And it’s a trip that never ends. I worked in audio for 25+ years because every day I found something new to learn. But I also recorded a lot of great projects and made sure the rubber met the road and stuff got done. I try not to let technical desire interfere with just getting stuff done. But if you have it as a foundation, the fun stuff happens quicker and better too.

Enjoy.

Buckle up.

And good luck.

2. RMS vs. Peak

The history of digital audio as a consumer and a professional domain, has been fundamentally flawed since it’s widespread inception in the 1980’s. Not in the way you think I’m going to say. Done right digital audio is a transparent medium better than analog in many regards. And once developers grasped it’s inherent weakness those weaknesses have been largely overcome. Digital sounds good is what I’m saying. Anyone who has a problem with it’s sound quality is having a problem with how it interfaces with them. And that is valid, you do better work if you like the medium and interface. That’s why we all have favourite guitars. So you work with what you like best and you will get your best results. But digital is a fact of life and so ubiquitous as to be essential to any working musician and recording professionals.

No, it’s digital metering and level practices have been flawed and greatly mis-represented throughout that time. This was due to a combination of cost, and ease of use. Manufacturers wanted to sell as much product as cheaply as possible, while simultaneously making it more accessible to people who didn’t have as much training and knowledge as previous generations of professional audio engineers might have had. As a result for many years, and even up to the present, peak reading meters were the main means of presenting recording digital levels. They can make for safer higher quality digital recordings in the short term for someone with only a basic understanding of audio. In the long term they are greatly misleading. That is because they do not adequately represent how the human ear hears sound. It’s ironic that something discovered and researched in the 1920’s, at the very inception of recorded audio, practically got lost by the 1990’s.

The human ear is not as sensitive to peak audio as it is to average volume level. What we perceive as louder or quieter is actually an average of the last few hundred milliseconds. Luckily for the bulk of early audio recording, direct to disc and then to electrical analog audio tape, this coincided with the limitations of the medium. Analog recording media has a tendency to round off peaks and not reproduce them accurately at high volumes. So the fact that we’re not as sensitive to this distortion made those mediums workable for most people for a long time. In the 1920’s at Bell Labs they determined that average, or RMS volume, Root Mean Square for math geeks, and yes I’m aware of the minor differences, is how we perceive volume. So they devised a means of visually representing this so that audio engineers could better utilize their recording equipment and mediums. The humble VU meter was born.

The VU meter is an electromechanical means of representing something pretty close to average volume. It’s still a little peak sensitive. And most modern versions will also contain a handy peak LED indicator for those situations that call for it. It also happened to be a great way of maintaining a desired analog voltage level. This is key as analog equipment, especially in the pre WWII days, was not as high fidelity as it can be today.  Proper balance between noise aka hiss, at quiet volumes, and distortion at loud volumes, was needed. So the VU meter was calibrated to help engineers maintain that optimal level. More on that later…

In modern times there are many ways to try to more accurately represent average or RMS levels. As we saw, in the early days of digital almost all outboard digital hardware use peak meters. These had the advantage of being cheaper to implement, and served one very useful function. Unlike analog, Digital mediums distort in ways that are not usually pretty if you go over a certain maximum volume. This is called clipping. Analog mediums tend to ease into distortion over a range of volume, giving you some cushion before you’ve gone too far. Also, analog distortion tends towards being harmonically related to the source material, meaning the distortion can be subtler, and even useful at times. But digital systems and converters have a hard ceiling. Over that ceiling the volume cannot go up at all. The signals waveform gets clipped off. This can be very harsh sounding and if recorded that way, undoable the original signal that got clipped is lost completely like the library of Alexandria. Though in modern times the use of hard clipping has even come into vogue. So peak meters do one thing: tell you if you’re clipping. Very handy, but a bit of a one trick pony. As well, I think there was a presumption that if you were recording into digital gear in the early expensive days, you had Vu and RMS meters on your console outside of the recorder so the critical thing was peak for reporting overs.

Before digital some analog equipment and consoles used peak meters. Engineers recording drums in particular found them more useful than VU meters as they could make sure their analog tape wasn’t shaving off the transients of drum hits, Something that reduces their impact In the first 10-20 years of digital, finding RMS meters on digital gear was a rarity, and I struggle to come up with an example off the top of my head.

Flash forward to Digital workstations and computers. Peak meters continue to be the norm. For much the same reason as hardware. They use less CPU power, something at a premium initially in the 90’s. And they keep the dreaded peak clipping away. Pre floating point math this was a critical component of digital bus summing as well. Floating point math, and what it does for summing in digital systems being for another day’s myth busting. As computers became more ubiquitous and more powerful they started replacing most other forms of recording, and emulations of the vintage ones they replaced. As they did so some DAW manufacturers started adding more sophisticated forms of metering. All while leaving the lowly peak meter as the default usually.

Now many DAW’s offer Peak, VU ballistic, RMS and LUFS forms of metering. All but the peak meters are ways of more accurately representing the average volume. None of them are perfect and each one serves slightly different purposes. But learning to use them over peak meters is key to moving forward with gain staging. In the order I mentioned them, they are of increasingly longer average times, With LUFS being up to infinite time. I’m not going to go into each one in detail, but suffice that a VU or RMS meter is the best practice for displaying the actual level of what you’re recording and hearing.

Many of the standards adopted for them come from the audio post world. Post audio is the generic term for all audio for picture. In audio post, consistent levels from one program/show to the next are key. And setting levels that allow for whisper to an explosion are also key. So audio post have had more accurate metering and reference levels for a long time. In fact in the US and Europe it was set in law in order to combat commercials and program material from jumping up in volume. Post has its own loudness war that necessitated those laws, they’re no better than music people and don’t let them say otherwise.

You will notice when switching to VU/RMS meters that they seem to behave slower than peak meters. That is to be expected due to the nature of how they’re representing sound. But you learn to watch them and get a sense for how they are reacting and what the ‘average of the average’ is. In this manner you learn to tell that they are ‘averaging’ a certain level even if they still fluctuate quite a lot. Something similar can be done with peak meters but it’s very difficult and less accurate as the meters move too quickly to often get any sense of an ‘average’. They are solely concerned with the loudest ‘bit’ at any given instant. But on some systems this is still the only option without using third party plugins. Also to note, most VU/RMS meters have a way of indicating a peak of over 0dBFS, eliminating any need for peak meters. Something I hounded into my students when I was a teacher, to the point where if I run into one now and ask ‘what are peak meters good for?’ they will instinctively respond ‘Not a ****ing thing.’ And while not strictly true, %99 is close enough.

I felt I had to start this primer here, as this is a fundamental of audio most self taught home recorders never truly learn on their own. I didn’t learn in depth until years into my career when I made the leap from home digital studio to post audio professional. I settled on the Pro Tools, my DAW of choice, Venue RMS meters. They change colours brightly to yellow at the reference level of -20 RMS, the North American Post Standard, this will come up later. They also display red well in advance of clipping. Those give me the most clear readout or me when working with large amounts of tracks at one glance. I also use the Waves VU meter plugin and ones from similar companies, especially as they can be recalibrated. The drawback is inserting them on tracks and having enough screen space to show more than a couple channels. The takeaway from this chapter is that using some form of averaging meter is critical and essential. You settle on the one that works best for you and your situation. But you cannot move forward to understanding decibel scales without first understanding that they are in reference to analog RMS levels, not peak. As with each chapter, further research, experimenting and practice is on you, I’m just the tour guide.

3. dB scales

Wow is this a can of worms….How do you mathematically measure and represent volume. And then how do you make it accessible to the ‘layperson’?….Well they figured that out a long time ago too, the measuring and representing part anyway. Scienticians figured out quickly that volume is what’s called a relative logarithmic scale. They pop up a lot in music, frequency being another example. This made the math a little tricky. As each doubling of perceived volume was an order of magnitude larger in voltage  or pressure values. It wasn’t a question of 2 being twice as loud as 1, but a much larger number. So a unit that represented the quiet sounds would get unwieldy at large volumes, and vice versa. Much like it’s futile to use light years to measure a coffee table or kilometers to measure interstellar distances. 4 dimension space-time general relativity aside. So units had to be created that could adequately represent all potential volumes. Enter the Decibel. I’m not going to explain all the technical details. I want to try and make the unit more accessible and usable.

Initial recording systems used actual air and sound pressure to move a needle and scratch out a recording. But with the advent of electricity in recording and transmission of sound, volts became the unit of choice. And the range of voltages used is fairly large. As well, early electrical recording and transmission equipment had very tight tolerances of voltages they could operate optimally in. Go too quiet and the inherent noise in the system would overpower the wanted signal. Go too loud and the signal would distort, eventually to the point of incompressibility. You could throw more power at a system to get higher voltages and therefore more maximum signal, but this quickly gets cost prohibitive, dangerous, and not optimal for many applications. So initial experiments at Bell labs, because much of this was researched for telephone transmission, and other places led to the creation of a standard. A way of measuring voltages and representing them as something relating to level. And a way of maintaining optimal voltage levels in audio equipment. Enter the Decibel. The equation that makes the decibel involves logarhythms, which is what converts an unwieldy doubling of powers into numbers more usable. It scales the values up and down to make them easier to digest. In decibels every 6dB is a doubling of perceived volume. That’s an important takeaway many self taught audio people don’t initially grasp.

The actual math is not relevant here but worth investigation once the practicality of what they represent has sunk in. You have to also eventually get into the nitty gritty of the fact that it’s a relative scale. And that it’s technically ‘one tenth of a bell’ but the short version is that a decibel means nothing without some set standard reference point. Each type of decibel scale has its own reference point. And they can all be interchanged to a certain degree in various ways as we’ll see.

First up is dBSPL. This measures actual air pressure. This is the classic scale people use to talk about jet engines damaging your hearing, over 120 dBSPL BTW. And average conversation, about 60. etc. At a certain point for our purposes SPL aren’t a huge component of what we do. But Post audio has a reference point that equates these physical volumes to levels inside audio systems. They chose a level in voltages, which we’ll get to shortly, that equates to the best average level SPL in a theatre to understand spoken dialog. While still leaving space for quieter whispers above a noise floor, and explosions before distortion or pain and hearing damage occurs. That level happens to be 85dBSPL. The reference point for SPL is the quietest sound you can hear. That would be 0 dBSPL. This is what all the other levels are measured against. I won’t get more technical, but dB as a number is meaningless without a reference point. So each scale has one and is implied when you state a number using that scale.

The most critical dB that is used in recording and audio equipment is the dBu. It was determined that the best practice was to set a reference level of 0dBu equating to .7746 volts average. There’s a lot more to it involving power and wattage, but we can speed ahead without issue. Then they determined that for professional audio gear, +4dBu which averages 1.228 volts was actually the best compromise between all the issues that plague analog audio. Again note this is average, not peak, the peak voltage is higher depending on the type of signal. They determined that this reference level gave the best middle ground between noise and distortion. You can go louder than this, provided the power supply on your gear allowed it, more on that later. This is an average point. And optimal level. Not absolute quiet, as in dBSPL, or the maximum volume, as in dBFS, which we’ll get to.  So VU meters on consoles and recording equipment were calibrated so that they read an average of 0 at 1.228v RMS. And provided the gear had enough headroom, typically about 12dB, so +16dBu for those keeping score at home, for tape and 20-24dB, +24-+28dBu, for professional recording equipment, you got a pretty accurate noise free representation of the sound as voltage.

This leads to a discussion on headroom. Headroom is the amount of signal available over your reference level before distortion. Well, before major destructive distortion occurs. Analaog systems all have some minor distortion, it’s what most people now love about them. For a +4dBu system with 20 decibels of volume available over that point, you would say the headroom is 20db. More headroom requires the system to provide more voltage to accurately represent the signal. More power means bigger power supplies. So a compromise for cost is always needed. So all equipment has varying degrees of headroom. At the other end of the signal way down quiet is the noise floor. This is the quietest signal that can be represented before it is buried in noise from which you can’t discern or use it. For most tape and vinyl systems the noise floor at the other was around -50dB. This gives most recording systems a range of about 70-80 usable decibels with decent headroom provided. The other gear in a studio has typically a much lower noise floor so under ideal circumstances about 100-120 decibels of usable range can be available.

Enter the second analog voltage dB scale. In the 1950’s home hi-fi stereos and the very early stages of semi professional recording equipment came into being. These systems didn’t need the higher voltages that professional gear used. The cost of maintaining the signal at those levels would have made home recording and stereos too expensive. So a different scale was created. dBV. This scale had a slightly different reference point. 1 volt equaled 0dB. They also do some jiggery pokery to get around wattage and impedance, but again, not critical here. But 1V was too much so they chose a reference point of -10dBV as the optimal level for pro-sumer gear, as it came to be called. This voltage is only .316 volts. So allowing for a decent amount of headroom, you only had to have a power supply that provided a few volts rather than dozens. Much much cheaper to produce. And the tolerances on the equipment got better, reducing the noise floor so that the lower voltages used didn’t equate to too much more noise. But home systems such as cassettes have much less usable dB range and headroom compared to pro gear even still.

Now both of them being logarithmic scales referencing a different 0dB, the difference in levels between -10dbV and +4dBu is not 14dB. It’s actually 11.78 dB. Cause one reference .775 and the other 1V. This becomes important when interfacing gear in a modern studio that may contain both types of reference levels. In the 90’s it led to running levels too hot or too quiet between pieces of equipment. In the 2020’s it has to do with your audio interface and how it gets those voltages into the computer, but more on that in a minute, we need one more scale to continue…

So you have all this range to play with? How do you hook it up to an analog system optimally? That is the million dollar question. You cannot make +4dbu equal to 0dBFS. You would have literally no headroom whatsoever. So you have to pick some arbitrary point on the dBFS scale to be your analog equivalent reference point. There are a few standards that were set over the years, particularly in audio post production for movies and TV. But there is no consensus. Post audio used -20 dBFS in North American and -18dBFS in Europe for instance. These are the two most common. And this segues nicely into our next chapter….

4. Audio interface conversion levels

So you go with the North American audio post standard and have a reference level of -20dBFS. RMS to +4dBu. This gives you 20 dB headroom. The same as most professional consoles and gear. More than enough for most purposes. So keep your recording levels in your DAW averaging -20 and you’re golden. … Not so fast.

Not all recording equipment is built the same. 20 db of headroom in an +4dBu system requires more than 22V of power to accurately represent. That is a powerful and expensive power supply. The same voltage would be required by your audio interface to receive it. If you have a decked out Pro Tools HD system or similar, fine. The converters are powerful, accurate, and calibratable. So you can set them to -20, or -18 in Europe, and you’re good to go. Provided you have a boatload of money.

Home USB/Thunderbolt/Firewire interfaces face many limitations. Particularly if the USB connection is providing the sole power. USB provides about 5V of power. You can play games with hardware to double that, but then you’re getting into amps and suffice to say, nothing comes for free, steal more voltage, you have less amps and those are important for reasons outside the scope of this paper. Also, that voltage is DC, direct current or one way, and analog audio is AC, alternating current or both directions. That actually equates to about 2.5V AC as you can easily split it in half, a form of biasing btw, to create AC. Plug that into dBU equations and that comes to about +10dBu, give or take because we’re dealing with a peak limit but close enough. Only 4-6 dB of headroom, or less. Less than half that of professional analog tape. And less than a third of other gear. So how about using -10dbV? Well, that gives us almost 18dB of headroom depending on factors so we’re kinda getting somewhere.

So what does my interface do it’s conversion at? That’s a good bloody question. Most manufacturers don’t give out what reference level they calibrate to. I’ve never been sure as to why, but I suspect ease of use and not wanting to confuse entry level customers. If you know what to look for you can infer from the specifications what your interface is capable of. The manufactures will usually list a maximum input and output level in either dBu or dBV. This will tell you the maximum your system can take or put out. It can also give a clue as to where in dBFS you’ll land.

…For this we must digress and cover a little history relating what we covered in the first chapter. Peak meters were good at telling us when we hit the dreaded clipping point. So wisdom for early digital recording was to never hit the peak and you’re fine. But…for every 6db below peak you are functionally losing 1bit of recording resolution. So if you peaked at -18 your 16 bit recorder was now a 13 bit recorder. 16 bits being the standard for the first 10+ years or so due to many technical issues. This made people panic. So they wanted to record as hot as possible to get the maximum number of bits. So as close to 0 without going over and you were golden. But look in the manual of the first ADAT’s, the recorder of choice in pro-sumer studios in the 90’s and it would tell you that -18dbFS RMS, average not peak again hint hint, equated to +4dBu. Giving you 18db of headroom. Close to a pro console and more than analog tape. With most signals if you averaged -18 your peaks would be around -6 peak give or take. But if you pushed it louder, to ‘preserve bits’ you were going to go over the average of -18/+4dBu. You would be pushing your analog gear out of its sweet spot and adding distortion potentially. As well not all power supplies are created equal and they become non-linear at the farthest extents of their range. Adding more ugly distortion. So if you pushed your digital recording levels you were pushing your analog gear, and the power supply in the ADAT to ugly points and added pretty ugly distortion. I suspect this gave digital it’s early reputation for being cold and digital. Not any actual flaw in the medium. People who used best practises back then got amazing sounding recordings. Cause 13 bits below the sweet spot is still 78dB of usable signal, significantly more than analog tape and vinyl ever could produce. So it was a needless worry.

But this recording method carried over into interfaces for digital audio. So people would tend to push as close to 0dBFS in a DAW regardless of whether this was best in the DAW or the interface converters themselves. Hint: it was best for neither. So what level should you record at? Well, we still haven’t gotten close to an answer, it’s still that complicated. I’ll use a hypothetical simple USB interface powered by the USB bus as an example. If it’s max input is +10dBu and we take the output of a professional mic preamp and run it into that at an average of +4dBu, we will quickly and constantly be overloading the input because of the lack of headroom. So we turn down the preamp output? Well now it’s not working optimally. There is no simple best solution to this. Some form of padded input on a line level input reducing the signal by about 10-12 dB is usually the start of a solution, being the difference roughly between dBu and dBV reference points. But you can still lead to level imbalances. It’s best if you’re dealing with professional audio to get an interface capable of professional levels so that you don’t have to add any negative gain stages.

But let’s simplify even further for those who are not going to go the moderately expensive rout and take the nice preamp out of the equation and use the preamp available on most budget USB interfaces. It might not have the sexy tube or class A warmth, but by and large they are all pretty excellent these days. Especially if you’re old enough to remember how bad it used to be in home studios. Now you just have to set a volume on the preamp that corresponds to your chosen reference level inside the computer. -18 or -20 dBFS in most cases. This will give you 18-20 dB of headroom, again more than tape, and comparable to a pro studio. Your peaks on average signals will be around -6 or so, but not usually near the dreaded 0 except on spikey percussion and dynamic instruments. So reading the RMS meters in your computer you get the signal to hover around the reference level you want. Turning the preamp up and down until you reach that optimal level. For microphone and line level instrument recording, this is the safest practice. Not technically completely optimal, but close enough for rock ‘n’ roll. If you start to incorporate outboard gear like mic preamps, a higher quality interface that can handle proper +4dBu signals with headroom quickly becomes a must. And while also improving signal quality, they can make life easier for interfacing equipment in complex setups.


So you get your recording done and as long as you don’t hit the red and average your optimal dBFS level you’re golden… unless you’re a guitarist….

5. Input impedance in relation to guitar tone and levels

Electric guitars are quiet creatures. It takes a preamp and amp to make them the loud snarling beasts they become. How you get to that point is a real trick of electronics. The average single coil pickup puts out about 100mV, in dBu this is about -18, so about 22db Below optimal level. Humbucker and active pickups are proportionally louder and put out more voltage but it’s still on the order of hundreds of mV.

But wait there’s more. Guitar pickups are a magnetic system containing coils of wire. This produced a fair bit of what’s called resistance. Or Impedance in AC systems. Functionally what this means is that the input impedance of anything you plug your guitar into should be much higher than the guitar output impedance. If not, it will change the colour of the guitar output. Mic Preamps have very low input impedances as microphones are low impedance outputs. If you plug a guitar straight into a mic preamp your tone will get dull. So most interfaces or DI boxes have what’s called a hi-Z input. Typical on the order of about 1Mohms, same as a classic Fender amp under ideal circumstances. This is high enough that the output of the guitars pickup will not be affected….


But guitarists are fidgety beasts. All those pedals and toys they play with aren’t made equal. Some of them have low impedances cause they’re built cheaply, some intentionally. A fuzz face for instance has a very low impedance. This is intentional, dulling the sound before distorting it so that the high frequencies don’t make the sound brittle. Wah wah pedals are similarly low because of the parts used to make the signature sound. But a typical buffered boss pedal has an input impedance of about 1MOhm and that became a defacto standard of sorts. Don’t get me started on buffered and non buffered pedals…jesus are you mad!?!? Some amps, like Marshal, offer inputs with a lower impedance of typically 500k ohms to intentionally dull bright input signals in order to give you tonal options..

So if you have a nice plugin emulating a low impedance pedal like a fuzz face, your 1Mohm input might not be the best idea. UAD, Avid and IK Multimedia all made boxes that do this on input. They have means for varying the input impedance of the recording box to deliberately change your guitar tone to better emulate those pieces of gear. Be warned they record them that way permanently and it cannot be undone, but that’s yet another can of worms about workflow. Some DI’s by Radial, and the Z-Box from MOTU have similar functions. It’s also a solution for wireless guitar transmission systems changing the tone as many with a good ear will note if they’ve ever put a wireless system before their fuzz box.

Can you simulate it? Sort of. There’s no plugin or eq equivalent out there that I know of right now. And it’s a complex interaction that produces resonant frequencies. But if you play around with a resonant low pass filter from 1-10khz you can sort of get there if you know what you’re shooting for. I spent a fair amount of time for instance recording through my eleven at different impedances and then trying to emulate it with fabfilter EQ. With varying degrees of success, but in all cases it made a real difference at least. Some combination of roll off with a slight bump of 3-6db at the roll of point gets pretty close, and you move the frequency down to emulate lower input impedances.

Suffice to say if you’re chasing the prefect amp sim tone, this is a critical step to consider whether you need it or not.

So what level to record guitar at finally? Yeah you’re getting tired of ‘it’s complicated’, but it is. You can record to one of the standard RMS reference levels we talked about previously, -20/-18. And that’s a pretty safe play and your peaks should be well clear of the dreaded 0. That too depends on the material you perform. But it all depends on what plugin chain you plan to use. You may choose to record quieter or use some form of clip gain/plugin to modify your signal due to those plugins. The Avid Eleven interface for example in direct mode, will average below -30 on a system calibrated to -20dbFS/dBu. I have done tests on mine, as they are particularly tight on specs. I find it is calibrated so that -20dBFS is +4dBu, being a large interface with wall power it has a good power supply. So when you track your guitar through it’s DI input and record the raw signal, you are performing no volume change to the signal whatsoever. So the quiet signal you record is the exact raw signal. I find that a hand starting point. For many amp simulators that is the preferred level, below -30 we’re still talking about. With the resolution of modern systems, this is a perfectly adequate level even if you subsequently have to turn it up later in the computer to -20/-18. So this is how I typically record. But sometimes you just have to get stuff done and are in an unfamiliar studio. I will often just record to a safe level of around -20 knowing I can modify it down later without problems. Being aware is the critical part. And being aware of what your plugins of choice require is key. Which brings us to our next chapter.

6. Manufacturer plugin reference levels, in particular analog emulation plugins

As we have seen, in the analog days of recording, there were standards. Reference levels that by and large all professional gear adhered to. +4dBu being the biggest one. With pro studio gear you knew you could interconnect devices and if you adhered to that average level, you’d be safe. Some gear had less headroom than others, tape recording being the big one. But engineers would learn the tolerances of their gear and work within those. The digital revolution in the late 80’s and 90’s democratized recording to a certain degree. It lowered the cost of getting in for one thing. Even a $3000 ADAT was a bargain compared to a 24 track analog reel to reel that would set you back hundreds of thousands. But digital also allowed for a certain amount of leeway in recording practices. By and large if you didn’t go into the red, you were safe. You could get too quiet, but the noise floor on digital gear and recorders was a lot lower than analog so you had more forgiveness there. Sure, as I explained earlier there were still best practices you could follow to improve your sound. But it was seemingly harder to mess it up.

Then along comes computers. As they got more powerful they took over recording. You now had orders of magnitude more power available to you in a 2010 laptop than most 1980’s professional studios had at all with their millions of dollars in equipment. As computers became more powerful a world of high quality plugins opened up. Initially they were fairly generic digital utility plugins in the late 90’s. But more power and desire brought about plugins that emulate the analog gear computers were replacing. You could now have a seemingly infinite number of instances of your favourite vintage piece of gear at a fraction of the cost. For the next 20 years from 2000 on the quality of those emulations got better. Hampered initially by lack of power, once computers had the juice, a world bloomed and research into emulation only improved them over time.

Initial efforts at guitar amp simulation in the digital realm were pretty iffy. I had a Boss SE-50 with some algorithms claiming to be amp sims. For the type of hard industrial music I made, they actually worked. But not because of their accuracy but because of their harshness. The warmth and subtlety of analog distortion is difficult to emulate in a computer. There are some programming issues to be worked around. Oversampling and faster computers got around the biggest hurdles, aliasing and latency, but it’s been a 20 year process of refinement. Guitarists can be a fickle bunch and prone to scrutinizing their tone. Many eschew digital recreations as being harsh, or having responses not like the original gear being emulated. I suspect, like ADAT recording in the 90’s, there are some practices that are causing some of the discomfort with digital amp simulations. And like recording to ADAT’s some best practices with levels can go a long way to improving the tone and recording quality.

Any emulation of an analog device in plugin form has to have some reference level to an analog voltage. As we saw, with digital dB scale, 0 is the maximum level you can achieve before distorting, where 0 in analog is the best average level. But while digital recording democratized it for more people, some standards were left hanging in the wind. There were no set universal level standards. Post audio had two very common ones, -20 dBFS and in Europe -18dbFS. But in music recording, there was no standard. Each converter in any piece of gear was free to choose whatever reference level they wanted. Recording into a DAW with pro-sumer interfaces and plugins only worsened the dilemma.

Nobody is sure who’s working to what level.

With analog emulation plugins this can be potentially disastrous. They emulate both the distortion and in many cases the noise floor of vintage gear. So you can run into the same problems as engineers did using those original units. Go too loud, you distort. Too quiet you lose the signal to noise. Plus you have the added caveat that at a certain level many emulations will break down completely at too loud a signal and distort in ways not from the original gear, clipping and so forth. Most designers of analog emulation plugins settle on either -20 such as UAD, or -18 such as Waves, and antelope I just found out. But there are a whole range of options. Some UAD tape emulations use -12 for instance.

With guitar amp emulations this problem is even worse. You are dealing with signals upwards of 20 db quieter than your +4 levels. Do you emulate your amp to take that relative signal, or make it take some form of more common -20/-18 level? The short answer is, it varies from plugin to plugin. And most makers of guitar amp emulation software don’t release that information I suspect with some smaller ‘bedroom’ operations, they don’t actually know and are kind of winging it. With painstaking research I’ve whittled out some answers and information but the results make using plugins a mess. A difficult mess. But this is critical because even tiny differences in input volume and impedance can have a significant effect on the tone. My experience has shown me that even a couple dB before the emulation could change just the harshness of an overdriven amp even if the overall distortion was not noticeably more or less.

For instance, I own an Avid Eleven rack. They lovingly took the guesswork out of this by making all their modules react to the same level that was calibrated to the input of the device. Kemper and Helix do likewise for the most part too. I can plug a guitar into my unit, and set up a chain of effects and know that it will respond much like the real devices. Unless I deliberately go out of my way, I won’t get any unwanted level surprises from module to module. They even handily emulate the input impedance of the first device making it that much more subtle and realistic. And at it’s max sample rate of 96k, the aliasing of the distortion is pretty minimal. All well and good, until they discontinued the hardware. UAD did much the same with their Unison technology by matching the impedance to the first plugin in the signal chain. Motu’s Z-box does a simple analog passive impedance lowering. And some DI’s and IK Multimedia’s AXE I/O allow for varying the input impedance as well..

Left in the lurch after it was discontinued, I was back to using the plugins. Handily they included all of them for Pro Tools users. But my ease of use was gone. I bought the Eleven Rack because at the time my favorite amp simulation plugin was eleven free. I’m an advocate of keeping it simple and knowing your gear inside and out. I own one amp and one pedal board of analog gear and I know them like the back of my hand. I do the same digital. I was a fan of the Boss SE-50, then the early Steinberg Hughes and Kettner Warp VST plugin. I eventually settled on Eleven free, even though I forked out for the full version, because it was in every Pro Tools system I used, and quite frankly the amps did everything I needed. So getting an Eleven rack only furthered this deep dive. But I discovered something upon having to partially abandon my Eleven rack. I had been using the plugins wrong. The eleven effects plugins have a switch in their interface that converts them from -20 to around -30. I did some phase null testing and the number was between 10-11db lower. And the -30 reference level corresponded to my eleven rack versions. I tested that as well.  But the eleven amps did not have this switch. After testing I realize they were set to the -30 level. I’d been using them at -20 previously as that was my general reference level for everything else. My waves plugins are -18 but I just hit them 2dB quieter, not worth the fuss for 2dB. This was like permanently having a 10dB clean boost pedal on. Not the end of the world, but it changes the response of the plugin.

Once you find the correct reference level, the plugin behaves differently. Gain knobs have smoother travel and wider response. The high end of the distortion can be less harsh, a common complaint. So I went looking into other plugins. But finding the answers proved hard. Not owning a UAD box I found the manuals for their guitar system enlightening. They use a lower level, around -30 as well. I discovered the Plugin Alliance releases of the same emulations do too through experimentation. I initially demoed some of them and didn’t like the tone and play. I figured a Dumble or Diezel, amps I hadn’t played analog, should be more responsive and subtle. Once I tried -30 the amps opened up. I suspect the code was a straight port of the UAD version. But Plugin alliance wouldn’t answer my direct questions. At all. It was actually a little disheartening.

So the end result of my research was sadly inconclusive. There are many small amp simulation companies doing amazing work and the field has exploded in the last couple years. But try as I might I cannot get straight answers out of many of them. Some have vaguely admitted to a -20 level, such as some Nembrini products, however he’s not as certain as I’d like. Some, like Neural DSP, never answered me at all. Antelope was unable to get that information from the developers of their amp sims I just got informed. However being aware of this reference level discrepancy can be enough to overcome many issues. You now know that it can be a problem and if the plugin you’re trying isn’t quite responding the way you’d expect or like, this may be part of the issue.

7. How to play with your plugins like you do real pedals and amps

So how do you set up a chain of plugins to behave like your pedal board and real amp? Or like you Avid Eleven or Line 6 Helix? The answer is…good luck. I’m really helpful aren’t’ I?

If you are aware of the level discrepancies mentioned previously, you can create a chain of plugins that will behave more like a real analog chain. Most amp and pedal emulations have input and output gain knobs. These are usually independent of the emulation inside the plugin. They can be used to adjust the input and output levels to the correct reference point. Either up or down. This will play havoc with presets, much like threshold does with compressor presets. But with some judicious awareness and time you can make your own that will be more useful. Some plugins allow those knobs to be set independent of preset, but this is nowhere near a universal feature, just a little bonus for some plugins.

But let’s start at the start. What level do you record at. As far as I’ve been able to test, my beloved Eleven rack is calibrated so that -20dBFs is +4dBu. So when I turn all the gains and modules off or go direct, my guitar signal comes in an average of below -30. Quite quiet. No gain or modifying of the amplitude of my guitar signal is being done at all. I like that. Makes me feel safe. With 24 bit recording this is still safe as far as noise floor and bit depth. In fact, if I were to turn it up, I’m just turning up the noise anyway with it. Can’t escape that. I believe the Helix does much the same thing. One thing I do find disconcerting is many Eleven presets output near or above -20, a significant boost to the internal level that could be a sign of mild improper gain staging.

So for many guitar plugins, if I run that signal straight in, I’m going in potentially 10-14db too quiet. Gain in those plugins won’t respond the way you’d expect. So I do one of three things. If it’s after the recording, I can turn the clip gain of the file up so it plays back 10+dB louder. Or, I often make a gain utility plugin my first plugin and turn it up there. This has the advantage of working real time while I’m tracking. Lastly I can change the input gain of the plugin. If I’m using a chain from different manufacturers I will often use this method to save on plugins instantiated. The drawback being switching presets, so I avoid this if I’m preset hunting after the fact most of the time. Though you can easily make chain presets and presets in utility gain plugins to quickly combat these level matching issues outside of the plugin.

I can end up with a few utility gain plugins between emulations to do the same things. But this can get unwieldy. What it leads to is my predilection for KISS, Keep It Simple Stupid, and knowing my simple chain well. I stick to a few plugins and know them very well. If I can, from the same company to insure compatibility, hopefully. No guarantee there. But I’m a guitarist. So I also like to play with toys. And I’m finicky and particular about my tone. So yes, I own like 10 Tube screamer plugins. At least, I’m actually just guessing. If you’ve ever compared comparable plugins like Tube Screamers you’ve undoubtedly come across the differences, especially in output volume. They are not all created equal.

So I have a set level I record at for all of my guitars. The same gain regardless of instrument. In the case of my Eleven rack, no gain, nice and safe. If I didn’t have the Eleven I would use a DI into a preamp and record at some set reference level that I would then gain up or down in the computer after the fact, usually just below -30 if I’m in a pro studio. I would endeavour to find a 0 gain neutral input stage like the Eleven if easily possible with the equipment at hand. Using a DI and line level input can often achieve this. But if I’m going into a mic pre, I look for what combination of gain and pad give me the closest to zero gain. If it has an output level knob I can roll that down to achieve zero gain too if need be.

If I switch guitars I don’t change this setting. My level should be well clear of digital peak. As well, if I change guitars to humbucker pickups or active pickups, the level will go up proportionally but still be well clear of clipping or issues and still well below even a -20 level. Just like it would into a real analog pedal and amp chain. Perhaps for some presets or settings this might be too hot or at least too different, but if the intent of the designer was realism, the emulation should handle level changes the same as a real amp would. This is a critical concept. When people modify the gain in amp sims, or in their recording before the sim, ‘so that you’re in the yellow’ as several popular developers say, your single coils and humbuckers would be calibrated to output the same level. This is not how they behave in your analog pedal and amp chain. The humbuckers should be a little hotter, active pickups hotter still. Once you establish a level you keep the settings the same regardless of guitar. It has to do with your input device. This is one of the biggest takeaways you should get from this. You cannot set all your guitars to be the same level, they aren’t.

One last note, this setup allows me to record through my analog pedals and get the most benefit from them. If I use an analog boost pedal with ‘colour’, my recorded signal will be hotter, but still behave exactly as the boost would into a real amp if the sim is programmed properly. Also, using a combination of buffers and DI’s I could potentially record the dry signal and the output of my pedal chain and have the option for both in the computer. But to be honest I usually just commit to a sound and re record it if I change because I would play the part subtly differently anyway with a different sound. The purpose here is to enhance the nuance of my playing, not neutralize it. So if the differences change my performance that’s a good thing. The intent.

If manufacturers provided calibrated input reference level information, those who want to set up their system this way would not have any issues. We could find the level we want and set it to that. But sadly most do not. I have been trying to get this information for years from plugin developers. Or through experimentation. Most have not provided it. Until they do I experiment till I find what I think is the safe input reference level, and then for that plugin or developer I do not change my gain settings once I have established it. Then my guitars can go through a chain of plugins safely like the real thing. And when I use the gain inside a plugin, not the input/output gain, I’m getting the tone shaping the developer and original emulated hardware intended.

8. Conclusion

So in the end if you make these practices part of your routine and ingrain them in your workflow, and then you can just mess around with it till it sounds good.

And done.

Not kidding this time.

Hopefully this made some sort of sense. And while it may have opened up a world of questions and things to investigate, I hope it answered many basic questions you may have been struggling with. I am a self taught professional audio engineer. I always sought the root knowledge and reasons for why and how audio works. I was never content with just ‘getting by’. I feel that once I absorbed some new piece of knowledge and fully incorporated it to the point where it was subconscious second nature, my recordings got better and I got faster at getting the results I sought. I hope this does the same for you. The intent was to expose some core concepts many people are maybe only passingly familiar with, not so much to provide examples as just information for a starting point.

Credit & License

Creative Commons-licens

The article “So you want to use your amp sims better..” is written by Ted Onyszczak and is licensed by CC BY-ND 4.0.

The original text in pdf version is available for download on Google Drive

Author: Ted Onyszczak
Version: 0.99
Date: April 03, 2021

Tags: