Why Are Neural DSP Gaslighting Us About Plugins and Input Gain?
ฝัง
- เผยแพร่เมื่อ 5 ก.ค. 2024
- For my Presets visit: johnnathancordy.gumroad.com/
For ArtistWorks courses: artistwork.prf.hn/click/camre...
For TrueFire courses and All Access Memberships: prf.hn/click/camref:1100lqzr3
Use the code JNC40 for a 40% discount off all courses, or JNC100 for $100 off all access membership!
So Neural DSP changed their advice on this - which is good.
00:00 Intro jam
1:05 Overview
1:38 What Neural DSP USED to say
2:08 How did we get here?
2:40 Plugins have more GAIN than expected?
2:55 Ed S has been trying to get sensible responses for MONTHS
3:24 Single Coil players who READ the manual ADD GAIN
4:00 Folks were telling us traditional Gain Staging was the right approach
4:30 Neural DSP changed their online documentation since the videos
4:50 Isn't this "OBVIOUS"
5:10 Here's why People Would Turn Up the Interface Gain
5:38 BEST PRACTICE
7:20 It's not just me and Beebs....
7:49 BEST PRACTICE HAS CHANGED
9:18 @RhettShull helped get the word out too
9:45 This is actually only Half of the problem anyway
• How to set the optimal... check out this video from Ed S!
/ johnnathancordy get my lesson tabs and backing tracks here
www.buymeacoffee.com/johnnath... - if you like what I do and
only if you wanted, you could buy me a coffee!
Get my PADs bundle here: www.paypal.com/cgi-bin/webscr... (this folder will grow, and I'd encourage you to ask for different keys/songs and stuff that might suit this type of preset?)
I've decided to make it possible to grab both my Helix/HX Stomp bundles (the expression bundle with freeze presets has always been separate) together - www.paypal.com/cgi-bin/webscr... - I will then email you a link to both bundles!
Try my general patches for Helix or HX Stomp in this bundle using this link - I will then send out the patches! www.paypal.com/cgi-bin/webscr...
You can get my EXPRESSION patches in this bundle using this link - I will then send out the patches! www.paypal.com/cgi-bin/webscr...
Try my patches for Pod GO using this link - I will then send out the patches! www.paypal.com/cgi-bin/webscr...
/ johnnathancordy get my backing tracks here
www.buymeacoffee.com/johnnath... - if you like what I do and
only if you wanted, you could buy me a coffee! - เพลง
Shout out to Rabea for shedding some more light on this - if you want to get to the source and figure out how to Match YOUR interface with the right Calibrated Input levels for Plugins - you should check out @eds4754 Ed S - as he is the guy getting real answers from the plugin companies about this stuff
th-cam.com/users/shorts4rn4pXpNzg4
Bea didn't shine more light on it. He just said everyone was dumb for touching their interface. I was not impressed.
I appreciate the effort you and others have put into sorting this out.
Ed S has changed how I use plugins so much, I already loved his shootout videos before he started this.
That bea vid was bollocks..
THIS video shows that the whole industry and community was not doing/been told what was „apparently“ sooo „obviously correct“….
wtf.
I was lucky enough to find this out by accident about a year ago when my toddler was playing with my interface knobs and turned the input all the way down. Way to go bud!
Yet another advantage to having kids. As well as all of the other moments of joy they bring, they can help you dial in great tone.
Damn I gotta have some kids, thanks guys@@DrProgNerd
I watched the Bea vid yesterday and was a little annoyed by the gaslighting subtext in the video. His attitutude was like, "I don't know why someone would increase the Input gain on their interface."
I did it....because that was what they told me to do....and subsequently every other plugin and mixing video did too.
I've almost always hated the Artist Presets on Neural plugins. I couldn't understand how my favorite artists would create such 'raggedy presets'. Now I know why. Time to go back and play them ....the right way.
Felt the same even he was my guitar influence, for me it was arrogant of him to say that.
I felt like IK was shaming me for
Suggesting this is how to set up your interface (the clipping method)
I was going to comment, but you said everything I was going to say. Exactly. Why? Because we were all told to do that....
Ah, the artist presets still suck when using line level tbh
I'm not convinced every artist preset was made with that in mind
Thanks man. Just to clarify, I wasn’t pointing any fingers, I’d just seen a bunch of videos recently and I was confused! Nothing more.
Just to say, either when using single coils or humbuckers, I don’t change the input signal. I just don’t add gain. Seems it’s been a series of bits of info over the years when using VST’s.
Anyway, wasn’t my intention to suggest people were being wrong or making anything up!
Also yeah, different interfaces are all different but most offer a line/inst level input.
I just didn’t understand why turning up the DI input signal was an option because it would affect the end result. I’d always suggest using input on the VST itself.
Ed commented on my video, and yeah he was very in depth!
Anyway, I hope that makes sense. No offence intended incase anyone felt that way!
Hey Rabea!
Hopefully this cleared it up a bit, since until we made the videos this was explicitly "the Best Practice" that Neural DSP themselves have told plugin users -
taken from "Tips For Using your Plugin"
"To Set up the input level, you need to adjust the gain knob on your audio interface. The best practice is to set the input level as high as possible without causing any distortion or clipping"
Since they've only recently changed this after we made the videos in January, it looks like best practice has changed, and they've consequently updated the site to reflect this.
Hopefully this might mean that bois using single coils like me, Beebs, Tom Quayle and Rhett will get better results from the Neural stuff which was kind of the point of the discussion in the first place and why we were kinda like "why have we been doing this all along".
It isn’t obvious advice because almost all companies tell you NOT to do it. Which we now know was errant information.
I find your comment here some what disingenuous, the video is how you're confused why anyone would do it any other way. If it was that obvious some of the best guitar TH-camrs wouldn't have been confused just like the rest of us. Just my two cents no disrespect intended.
@@akaerik1thanks man. I don’t really comment unless I’m being transparent. There is no gain in being disingenuous.
I was confused why this is something that is coming up now.
In my entire experience using VST’s I’ve never gained up a DI. Nor have I read any ‘how to’ or watched any ‘how to’ videos. I just go with it and use it. But it never occurred to me to gain up a DI. I just leave it as is.
I see other users do, and tbh I guess when it became a problem. That’s when I made a video, just to clarify things from my perspective. I don’t see what’s disingenuous about that.
And no worries I don’t feel any disrespect man! Everyone is entitled to their own opinion. But I can assure you, I’m a firm believer in saying it how it is and being transparent. Probably the Yorkshire in me 😂
@@RabeaMassaad respect 🙏🏽
I'm really glad that you put this video up. I really like Rabea's playing and he's super talented but I felt like there was a bit of "Well it's blindingly obvious you set your input to zero, why would you think anything else, you must all be stupid" about his attitude. This video by yourself goes to show that we had been told something different all along, even by the amp sim companies and it's also quite different to how you would set levels for anything else.
Thanks for making us not all feel like idiots.
Has yer man Rabea not perhaps jumped on this to make himself a wee bit more relevant by implying that this is old news, he's been ahead of the game as hes always done this and always known about it .
@@jimsmith4611 he possibly has I'm not sure. I give him the benefit of the doubt in that maybe that's how he set his levels all along. It didn't look like he needed any more gain anyway. But it was the lack of empathy for the other side that I had issue with. Not sure why you said "yer man" though?
I like Rabea, as a guitar player, and a person. He seems like a genuinely nice guy. But I found his video, about this subject, to be pretty condescending. My reaction was about the same as yours. Also...on my Presonus Quantum 2626, if I set the instrument input gain to zero, THERE IS NO SOUND AT ALL. I MUST increase it, to hear my guitar. So Rabea's implication, that all modern interfaces "automatically" set the correct level, is false. Unless, I suppose, my 2626 is not modern enough? I believe that I am more confused now, than I was before...
And just to be clear here...I am NOT using Neural DSP. So if this conversation is JUST about THEM, then I will sit down and be quiet.
Nothing is obvious in the world of a moron. I am one and I can't point it out enough: many manuals simply suck when it comes to giving clear instructions. Ok, it made me decent at figuring out stuff myself, but that cannot be the intention of a manual.
The general consensus was ALWAYS to set the input gain as high as possible without clipping - I was doing it for many years after Misha Mansoor's guide to recording 9 years ago, and I'm sure many guitar recording guides to this day still follow these protocols. This is very typical audio recording advice that works for 99% of use cases, but now people are wisening up to the calibration stats of all digital recreations of analog gear (be it amps, or any other studio hardware). Of course, you can often compensate with the input gain in the plugins. In reality though, most users would sooner tweak the plugin itself and turn the amp's gain knob down rather than the built-in input gain controls.
ML Sound Lab's advice from the start has been to calibrate your input signal to the input light in the plugin. Grabbing your hottest pickups, chugging as hard as possible and tune your input gain so that the light just barely begins to blink. This does still calibrate it to the right levels, but it's not a realistic method to expect now that we've got a lot more exclusively-clean amps available.
We now suggest engaging the Hi-Z/instrument input and turning the gain down as low as it will go, regardless of your interface. You could calibrate further as per your interface's specs (as outlined by Ed S in his fantastic research), but these exact calibrations are too much to expect of the average player. The ballpark area is going to be good enough for most people, especially when coming from the old method of "as hot as possible without clipping". Clean tones will actually be clean, crunch tones will actually be crunchy, and high gain tones won't be a flubby, noisy mess.
I appreciate the clarification on this! I use a lot of ml plug-ins and have enjoyed them. I actually thought that we were supposed to have it lighting up green all the time. It would be really fantastic if all amp Sims could have a built-in calibration like the Joey Sturgis stuff does.
I admit this input gain thing has been a pain in the ass ever since I started trying to learn to record, and now I'm trying to switch to Reaper which has just been another headache.
I guess now I'm off to ask Dan Worrall and Adam Steel if they have an answer for my specific Reaper / pro Q question.
I've bought several ML plug-in over the years, and I've always set the input gain to just below clipping based on advice from you guys at ML, and I gotta say, I was never really happy the tones in the long run. This is why I kinda haven't used my plug-ins a whole lot, instead resorting to practice with real amps or a modeller at home. I never made the connection, that the input gain was causing my plug-ins to be all flubby. I guess that says more about me than it does the people giving the advice.
That said, I watched Rabeas video last night and I went straight to my audio interface and turned the input gain way down. Now my ML and Neural plug-ins sound great, and I feel like an idiot. 😂 I never really understood why I could get a kickass tone with a real amp or a modeller, but couldn't get anything to sound right with plug-ins.
to be fair, Misha set the input for the tone coming from his AxeFX, he didn’t dial in the gain for a DI signal
> high gain tones won't be a FLUBBY, noisy mess < So this issue affects frequency / EQ too, not just the amount of gain? I've wondered why ML and Tonex sims always seem to make my Fender bridge pickups sound like Gibson neck pickups, no spank or sparkle. Thought I was losing my mind.
@@ericolson326 kind of yes, simply boosting the input signal into an amp will make the amp respond differently.
Many high gain rock/metal players do this to archive a tighter low end response.
But it can also go too far, you really have to experiment and different amps respond differently.
I'm so glad you and Ed S. and Rabea have been digging into this. I've only got a Stomp but I've also got Helix Native and have been very frustrated that my Stomp-built presets sound super distorted when using my interface and Native.
You all are a huge asset to the digital guitar community.
I find both my stomp and helix native sound awful in my DAW, even with VERY low input gain - to the point of I stopped using it and went with Amplitube, which worked great right out the box without any crazy input gain levels.
Are you telling me 90% of plug in demos the last 5+ years are technically “wrong” tones…
I was just thinking that
Thanks for this video. I posted a similar response on NDSP's Discord after Rabea made the same "this is so obvious, how are people missing this?" comment. NDSP's Francisco was also surprised that anyone would be confused about this... despite directly contributing to the general confusion with the instructions that were previously found on their site...
I would buy a Gaslight Pedal!
I am/was one of those guys who increase the input level of the interface because I was told to do that in every single manual by every company out there since the beggining of time. The presets from any amp sim past and present always sounded bad to me. I always thought that my interface, guitar, cables, pc, monitors were crap. But then I bought a Steinberg interface, Alesis Monitors, earnie ball cables, a Dean guitar with EMG humbuckers, an I7 pc with 32GB of ram, SSD drives, etc, and the problem stayed the same. So I was lost and hopeless until now. This discovery brings me hope to my guitar tone!
Recently I put an EQ on output signal and that was the thing I needed
This is a hilarious scenario. We've always, always, _always_ heard about setting the input gain to "where it doesn't clip" and now plugin companies are all *Pikachu face* because _clearly_ you just need to put it to 0 and not fiddle with it 😂
Personally I still set my input gain on the interface higher and set the input on the plugin lower. This is because I want my recorded waveforms to appear "regular" and not require me to zoom in 1000x when trying to line things up :D
This man. If you recorded too low with this new philosophy you won't even see any waveforms
Your DAWs don't allow you to increase or decrease the visual waveform without making it louder or quieter? Have a look for this feature. It's been in ProTools for decades
@@EthanRom - You can magnify the waveforms on pretty much all DAWs without actually recording a hotter signal.
@@philmiller6655 This. Plus this whole debate is a non issue. The plugins don't care about what level you record at, they only see the level that you give them. If you trim the gain of the recorded audio or use an overdrive plugin in front, the level at which you recorded doesn't mean a thing.
It's good practice in theory because you are utilizing more of the bit depth of the recorded signal. At 24 or 32 bit audio, if you are setting your levels very low, you are only using a fraction of those available discrete values and losing dynamic range How much of a noticeable difference this makes in practice, I don't really know.
In my experience, preamps also yield better results when turned up,
I would turn up my interface and attenuate before the plugin.
Thanks for your work to uncover this and share the info. I used to add gain from my audio interface. I guess that's why I found most of Neural DSP's plugins too overboard with high gain for me when I would demo them. I'm using Helix Native now with my interface turned to zero and plugin input adjusted. I am digging the tones much more than before.
It’s stale advice from a different age. Once I learned about headroom and noise floors, I stopped maximizing the input gain on my interfaces and learned to love calibrated gear.
Thanks for calling it all out, John!
Cheers John, all, Daniel 🕊🤙🏼
Thanks for the videos! I was one of the people adding gain, getting much better tone now.
Damn, your tone and playing is always so smooth
So pure, right?!
Isn't it just... An amazing palyer ..
We've all been bamboozled!!! Ive watched so many TH-cam videos through the years of learning home recording saying to raise the input level (gain) on your interface to get " healthy signal to noise ratio", just dont clip, etc . Then time passed and people started changing this narrative a bit to= not record quite that hot (but still raise the gain on your interface so that the bulk of the signal is in about the -18 dbfs range, peaking no higher than like -6 to -9 range. Thank you to Ed, yourself and all the other people shedding light on this. My guitar sound through plugins is so much better now. Thank you
That's what I'm doing, but now this is saying I should set it to 0... which gives me a signal that PEAKS at about -19 dBFS, with a fairly hot metal humbucker (at max volume of course). That's a bit low, isn't it?
@@exscape In Ed's video (I highly recommend watching it , if you haven't already), you not only set your interface gain to zero, but then you adjust the input gain in your amp sim plugin to the value that corresponds with your specific audio interface. Ed made a whole spread sheet of the values for a bunch of interfaces.
I’ve been experimenting with guitar plug-ins since at least 2008, and I never heard the advice to not set input levels for guitar-into-amp-sim as you would for a mic. Even guitar-specific audio interfaces-once those started to appear-would give the advice to set the input level the same as you would for a microphone. I made a serious attempt to move most of my guitar processing to plug-ins several years ago, but I just couldn’t get good results. But your video on this topic has changed everything. I might end up selling a bunch of gear, and I feel like the results I’m getting even with outboard gear is better. The big question looming for me now is whether I get a six (or more) input interface and try doing stuff with my hexaphonic pickups all in-the-box too.
Great video! I just want to add, with all this discussion around recording guitar into your interface, please don't forget that the "just below clipping" advice still applies when it comes to microphones. 😊
Basically still applies to almost anything right?
@@johnnathancordy Well, certainly not bass guitar. And possibly not keys as well, since both output instrument level signals. I wonder if this has as much to do with signal impedence as level? It sure flies in the face of everything we've been told now for years.
I just found this out when I was comparing recordings from my Helix to recordings using Helix Native. The same preset sounded so different when I did the “just below clipping” method. That was when I realized that I turned my interface into an OD pedal 😂
I'm a relative newcomer to Neural DSP. I was super excited when I discovered the Tone King Suite as I've been playing TK amps for about 20 years. I was really surprised at how gain-y the plug-in was though vs the real amp. It was breaking up so quickly and didn't sound at all like a TK amp. Glad you and Ed S cleared this up for everyone!
Thanks John! I really had this idea in my head since years...now everything explains!:)
Thanks. Regarding the quality of delay and reverb ... which modeler would you choose (tone master pro, quad cortex or helix )? I saw you getting close to a bigsky with a HX ... but would you obtain same with the tmp or quad ?
Thanks for continuing the discussion! The low input approach has “fixed” the sound of my Neural DSP Toneking as well as various amps in Tonex. I’ve found that Tonex is a little all over the map though as some amps (usually the higher gain amps) sound better with the lower input while others still need the traditional level setting (or even a little hotter) to sound good. So for now, I’ll be using my ear with Tonex and will adjust on an amp by amp basis.
Thanks for this John. I could never get decent sound out of the PT 100 plug in. I gave up on it. I need to revisit now with reduced input gain.
John, thank you so much for your advocacy for normal players. Yes I have!!!
I’d be willing to bet many, if not most of us that “try” to record with digital plugins have had these gain issues. I just couldn’t get cleans correct. For example, I was trying to record on my Ibanez Jem with EMG’s and I’m using a passive DI box that has a -10db buffer before the high z input on my interface, even switched from a focusrite to ssl interface, and STILL could not get good recorded cleans out of Amplitube or Neural plugins, etc. Cue the Friedman IRX pedal with the analog front mated to a balanced digital output and holy crap, I could get correct signal representation. The way it sounded when I played it through my interface into studio monitors was the same way it sounded when I played back a recording through the same monitors. Seems simple right, but for some reason I have never had the ability to do that easily digitally and I think this was why. Kudos to your work and cheers from Texas!
Rabea did not at all shed light on anything. He just happens to have an audio interface that has the right level with the knobs set to zero. The "Problem" ist, that audio interface instrument input levels are NOT standardized. So it is guesswork. Unless you use Eds Google Document. Tried it with several interfaces and it works.
Just saying "It is obvious" does not help anyone.
Worse is the advice the companies give... when in developement they for shure use the right level, otherwise the gains would be totally off.
There needs to be some sort of gain calibration process where you can set the right level in front of the plugin. That would solve it once and for all.
And YES, I was one of those people that added gain. So greatful for Ed, you and some others for explaining it and making it popular!
Hey dude, important video thanks for making it. I feel the need to mention that your audio levels are insanely dynamic in this video. Appropriate for an action movie, but hard to watch without volume leveling enabled on my tv lol.
Didn't even know what you were gonna say, heard you playing. Liked and suscribed. holy moly
What a cluster f&*k for something so basic, especially to those creating preset packages.
Oh man, how many comments I got "Your preset sounds shit when I play it"...
I came at this from a different direction initially.
Being an old Sound Engineer by trade I usually worked to a -12dB baseline starting point for i/p levels.
This came from working with early digital 16-bit systems targeted at CD production.
BUT this was all too often a bit too lively for virtual guitar device plugins, as I soon found out.
Early experiences with Helix Presets, from a certain JNC chap, had me thinking "These are all set a bit high in the i/p or the amp gains for my liking" ...
Basically I thought, at first, that it was John's fault that his Presets were too Hot ... 🙄
Then JNC mentioned Ed's research into these matters and there it was; the disconnect between Old Best Practise and manufacturers' advice and, most importantly, the Real World.
Before Digital, we used to set i/p measured by VU metering,
The target then was to hit -6dB to 0dB VU and not get too upset by the peaks hitting above that.
The tape compression would be Very forgiving, even preferable as it turned out...
The noise floor issue came from the noise inherent in mixing console and mic preamps.
Back then, the aim was to feed the recorder with the highest manageable signal to noise ratio.
I believe these older practises are what has led to the current confusions.
back then, tape machines and consoles would be calibrated and that is no different to how things are now. Tape machines all have their own calibration (based on who sets them up), same is true of converters.
The main issue is that most people don’t have to think about this stuff for most scenarios. So generally, a lot of people don’t really understand it. There’s tons of threads on forums online that are totally muddled through around things like -18dB, VU, +22dBu, RMS, unity etc without really knowing what it relates to. So it can be quite hard to learn, because there’s a lot of misinformation online.
@@eds4754Even when one Does know what one is talking about, one still has to put one's Thinking Cap on to decipher what Others might mean innit ... 😜
P.S.
Yes Tape Machines were calibrated but to different Reference Levels and with different tolerances given the nature of the recording medium.
Most of the calibration was to do with Record Level, Frequency Response and Head Alignment.
The input level Standards were much simpler to deal with.
Pro +4dBv
Not Pro -10dBv
Of course there were differing Zeroes depending on whether it was 600 Ohms or other impedances ... but that's another story .
Thanks BBC 🙄
@@KozmykJ Ha yep. It’s essentially the same with converters and how much headroom they have. Some are 24dBu, some 22dBu, some 21, some 20, 19, 18 etc.
I'd like to disagree with your conclusion. The problem is, that people are preamplifying their instrument's signal (without realizing) to line level before "plugging" into the virtual amp which is designed (or not) to handle a normal instrument level signal but keeps this a secret
Heyy mate! Does this also applies if I plan to reamp the DI through a real amp later? Should I set the input to minimum?
Thank you for these videos! How about re-amping, should we still aim at the zero gain territory while recording with amp sims, or should we maintain "hotter" signal levels for further uses?
I'm approaching this a little differently, as I mess a lot with reamping too and mostly an using plugins as a "demo" and if the take that's there is "the take" I reamp it, if not I re-record it through my amp once I have the mic positions and what not.
But, I record as much of a signal as I can, yes adding gain from the interface, and then I use the input in the plugin to lower it. May not be correct the way to go about it, but it works for me so I have a take that I can use for reamping should I need it.
The whole thing was a game changer for me though, once I "trimmed" the input in the plugin to match our guided dbs it became way way way better. Thanks for what y'all are doing.
Hey, what is the song playing in the background of the intro jam?!
Also, I’ve been told from the beginning to strum and chug as hard as I’m going to play and dial the input knob back so that it doesn’t clip. And you do this with every different guitar you plug into your interface.
It also displays clear, big sound waves on the recorded track, versus having to zoom in so much.
Thanks for this insight! i will remember it. gain on the interface (I have UAD) set to 0. input level in amp sim (neural) also at noon (=0). However, i have quite low output pickups (single coil, set to low pickup height). So i'm coming in at -25db in my UAD. What would you do? add a boost pedal before going in the UAD? or just leave it?
I was whining a bit about the continued crappy sound from my Kemper through my monitors when recording with my DAW in the original video of yours re this topic. Tried out the settings as described in the various videos re this subject but it still sounded crappy. That said your video did make me think again about what I was doing with the various guitars I use, my Kemper settings and how it all links to my Interface and DAW. From that I realised that I hadn't set the interfaces control panel input properly in terms of whether I'm recording in stereo or mono. A very simple school boy mistake that, when set up properly, made a huge difference to the sound that I was getting. Once sorted the advice given in these videos absolutely works a treat and the Kemper is sounding as it should and as I had hoped. Am turning up the initial input ever so slightly from zero but gone is the unwanted and grotesque 'digital distortion' I was getting previously. More to lift the 'level' than anything else. Very nearly decided to sell the Kemper too had it not been for your videos on this subject making me think again about being more careful re setting up my gear.. In short thank you for this.
I also watched Rabea's video yesterday, and found his demo of the issue particularly easy to understand.
I had tried some guitar Plugins many years ago, and couldn't get anything approaching a good sound. I assumed my computer wasn't good/powerful enough or the Plugins were junk. In either case, it was a waste of time to try using them. I went to using Modelers instead.
What's the best practice for using pedals with plug-ins? I routinely use jhs preamp, an overdrive or a compressor pedal with soldano/ cory wong plug-in.
... Goodness. I had always thought my Neural DSP stuff sounded "harsh". Thanks so much!
Yep, I was a gain-adder. Since your first vids on the subject I've been using the new wisdom and it has transformed Amplitube, Helix Native and McRocklin Suite for me.
I have a question, what role does normalising audio play here? We record the DI at that low level, and set the plugin settings to much, so, if we normalise it it's going to add a ton of level and distort the plugin right? So we should not normalise guitar DI signals then?
Thank you for making this video! Definitely had to be said! As others, I was also irritated by Rabea's video, acting like there was no reason to turn up the input gain on the interface to just below clipping when it had been Neural DSP's (and others) explicit instructions to do so...
I mean it does make sense. If you have gain ahead of the plugin its alike adding a boost pedal to the signal chain.
i grew up an analog kid, so yeah, high-as-possible-without-clipping was the way I'd always done everything.
I don't quite understand what the new best practices are. What is "Minimum"? is that turning the input volume on my interface all the way down to 0, as far as it can go?
The video you linked by EdS says to turn the input on my interface all the way down as far as it can go. Say I'm going guitar>Focusrite Scarlett>Helix Native...At that point I can still hear my guitar and Native is peaking at about -36 on the input meter (set to 0db) but I'm not getting anything on the LEDs on my interface. It doesn't light up at all no matter how hard I play. Is that correct?
I've always set things to just below clipping on the interface. That would vary between whatever I was plugging in, be it mics, guitars, basses, etc. That was always the general concensus even before plugins were a thing.
One thing I like about STL Tones plug-ins, is that they have an "intelligent" input level configurator, where you tell it if you have a low/med/high output guitar pickup, and it will adjust accordingly. When I tested it, it landed at 0.2 db of the optimal setting in Eds spreadsheet. I don't know why there isn't more companies that have implemented this feature.
Well crap, thanks for shedding light on this problem i didnt know i had😂 keep it up!
Thanks for your vids on this.
I can't help but think this is one of the reasons I've now gone back to using amps. I've tried loads of Boss/Line 6 boards as well as the Kemper for a good while, but there is just something that goes missing when I'm not using an amp. And I can't really explain what it is. Fwiw I'm of the Brian May school where I use my volume knob a lot to set how much gain I get - and I've always hated how that works with every modeller I've used (I've always had to disable every noise gate as a result btw). N modeller I've ever tried does that well imo.
Fwiw I do have and use a lot of the Neural DSP plugins in the studio, but my main guitar tones are amps these days.
I was in the camp of "gain stage it like a mic" until I saw yours and Rhett's videos. I was doing this with the Neural DSP Tone King Imperial, and when I first started using it, I remember one of the most frustrating things was having to turn the "blackface" channel volume down to 1 or 2 to get it to fully clean up. But because it's emulating the actual amp... 1 or 2 didn't give it enough room to breath so it was a pretty lackluster clean sound. Really glad this all came to light! I had pretty much shelved that plugin for a while because it wasn't giving me what I was expecting, but now it's back in the rotation again.
This topic was brought up by a couple youtubers when dealing with making captures on the Tonex. People were wondering why so much gain was being introduced into their captures.
I thought I was going crazy until I tried this. I was wondering why I was getting weird noises at higher gains and not the cleans I wanted. I instinctively pulled up a fresh preset on the Tim Henson plug-in and turned my focusrite gain down to a minimum consistent signal and it’s a night and day difference. These videos have confirmed what I thought
Thx for clearing this up and for all your work!! I had also been doing it wrong.
What I'm now wondering is, what does this mean for input levels on the QC? Leave it at 0 or still boost it up?
Zero on the QC is probaby "close enough" yeh!
Thx, will try that 👍
I was definitely adding gain when I first started using NDSP on the Plini release (how was that like 5 years ago!) but I soon figured it seemed to sound better when the gain was on zero.
Thank you John. Not 'ok' how Neural DSP has handled this.
I love their plugins, but being a former QC owner (spending a lot of time on the QC Forum), I'm not overly impressed with the way they handle their interactions with their consumers - and the undelivered promises on the QC. Definitely a love/hate thing with Neural and me.
@@DrProgNerdYes.
They have fantastic top tier products with a history of repeatedly promising more than necessary only to then struggle to deliver.
Turning up to just clipping and backing off was absolutely the advice I’ve been given for gain staging for guitar, microphones and everything. I was given the advice by other musicians, college, and from TH-camrs.
I'm a humbucker player, but also noticed this problem when making presets using Helix native and then transferring them to my HX Stomp.
With Native I was having a much more saturated and gainier sound, which was odd since they should sound the same, so I came to the conclusion I was doing something wrong with my audio interface.
That was the moment I started to research online about this issue and discovered that spreadsheet made by Ed S. in some forum.
That spreadsheet completely changed my workflow and now I can confidently make presets in Native and transfer them to my HX Stomp knowing they will sound the same.
So does this advice carry over to the input gain of our modelers as well (QC, Axe FX, etc)?
only if you are using them as an interface for plugins. If you are using their on board processing, they have already optimised their input level to match their processing (because they can spec both to match).
When I was in sound engineering school, back i the dark ages! Gain gain/trim pot, was what you would use if you didn't have enough signal coming in, ie input gain. In the digital world, this is very different, noticeably when I started to use the Iridium, or now the Dream 65 and Lion. It drives me nuts, because it doesn't have the feel of an analog desk, which is what you are trying to achieve in a digital world! To the point now, that I have bought D.I box from Walrus audio with a line driver feature, to try and get a good signal in to my interface without any extra gain! So I only want to turn up the gain if i need too. Digital and analogue still have a long way to go, before it achieves a believable plus!
Rabea is flat wrong. It’s standard practice to gain stage. I’m surprised he wouldn’t know that. If a company designs in some other practice they need to be explicit about that.
Does this also change how we should set the input gain on the Quad Cortex? Ive noticed a similar phenomenon on my QC.
yes, if you're using it as an interface. check out Ed's guidelines
Recording a guitar with the Audio IF at zero, not only has an impact on SNR but also on digital resolution. If your IF or your DAW is set up to record at 16 or 24 bit, recording with the audio IF at minimum, you are only recording the lowest 3 or 4 bits. If this sounds better, it is only because these 4 bits have the right volume to hit the plugin. There are quite a few videos out there that seem to conclude that, yes , ok it sounds a bit more muffled because it is not exactly the same EQ etc., but NO, it sounds muffled because you hit the amp sim with a low resultion signal, a bit like having a bit crusher cemented in your input chain. While I can see why putting the audio IF input at 0 sounds better, the real place to address this issue is rather the input level of the plugin. And yes, you can adapt this usually (increase in Amplitube, decrease in Neural DSP, let the plugin figure it out as in Waves PRS), but this should not be the concern of us guitarists out here, it should work correctly out of the box.
You’re right, but it just highlights the issue that audio interfaces do not have a fixed standard for dBu to dBFS conversion, and they absolutely should.
The ideal would be if all interfaces used the same standard for dBu to dBFS, *and* also had auto-compensating output, so that if you turn up the gain to maximize amplitude resolution, it digitally reduces the output to maintain the same dBu/dBFS conversion.
If all of them worked like that, you could set the gain to just below clipping and everything would still be gravy. We’re way past due for audio & plugin companies to finally get smart about this, and make it consistent.
After trying this minimum gain approach, sonically it works well, much improved over the “old way”. But at this setting I can hardly see the wave form in the daw (logic) no matter how much I magnify it, which makes it difficult for punch ins, editing etc. Adding enough interface gain so it visually registers in the daw and offsetting by reducing the input gain in the plugin (helix native) seems to get to the same place sonically, any reason not to do this?
that’s fine to do
I exactly have the same issue.
i got it now, but what if you record a DI to reamp with a real amp after? How do you setup the input level?
funny thing for me is when i finally got an amp sim i plugged in with interface input as directed it didnt sound right. so i fiddled and fiddled and by chance i turned my interface gain to zero and was pleasantly surprised by how good it finally reacted and behaved, i figured i must of read something wrong in the setup guide and just left it alone and just enjoyed the sim,, then a few weeks later you put out a video on the subject ,which was great insight, and now its a thing i guess hahaha. i really appreciate your insight and Eds spreadsheet , especially now since i upgraded to a much higher end interface with a max in at +18dbu, now it really is a thing for a certain sim that seems to be all over the place with their capture process, its a bit frustrating at times to dial amps in. it really seems that their wasnt/isnt a standard base gain stage for capturing these tones.. i really hope a standard capture level is set for the industry and we wont need to get a math degree just to play some freakin guitar,, haha were musicians for cryin out loud not scholarz. but on the real, thanks for bringing attention to this little, very relevant, and now less likely overlooked vital piece of the digital puzzle. .. (i think it's easy to forget or take for granted, just how new and great these amp sims and modeling software's are. i remember only able to wish ot dream about the gear i have now with these now...) cheers thanks for the vids and great guitar playing
If I had recorded "hot" D.I. (with old advice), how I can get the right amount of input signal to send to amp sim for example? Can I measure it with a VU meter before the amp sim and use a trim plugin?
I have been following this since Rhett’s video. I use Bias Fx 2 a lot in home setting and if I set interface to 0 gain, it yields VERY low input signal on amp sim and to me doesn’t sound good at all. I have been adjusting interface input gain to bring input level on sim to -10 to 0 and it seems to sound much better. Anyone know what BiasFx2 settings ‘should’ be or am I doing this wrong? I use MOTUM4 btw. Thanks
It's gonna take me a while to get past this intro jam. And then there's a whole video after it?! Right on
I’m not sure if this was mentioned but would you set minimum gain if the guitar was plugged into a DI box?
This might be a stupid question but here goes, why isn’t the reverb included in amp with reverb like a deluxe reverb? Or a twin reverb or an ac30 ? Answers on a postcard please
Sometime early in my discovery of guitars and guitar plugins, I realized it made no sense to add any gain at all to the DI system so I got that hapbit.
However I always felt my Helix Native presets felts weak, initially throught it was "just" my guitar because it's signal is weak.
Turns out that from your video and discussions on the Line6 Helix discord, I learned that the manual of the Helix Native plugin does clarify that the input signal should reach a specific level which is green visible on the left of the plugin. I also learned that all hese plugins need some specific levels, so thanks for that.
I was getting frustrated with this on my HX stomp going into my interface. I fix this issue by recording using the USB out to my computer and it perfectly represents what I made in the patch.
It also yook me years of recording to learn this, so totally agree, it isn't so obvious, we really dont know how the gain staging of different plugins work, and in some cases like Amplitube 5, some amps are modelled with a very weird gain structure, so you need to add extra input gain to get an amp response closer to the real thing. Thanks for your videos John!. Carry on with the good work.
I think there is a missing detail here. Different pres/interfaces have different gain levels at “Instrument minimum levels.”
And if you have too low a level into your interface then you won’t get the best resolution - it can be grainy and dithered (this is with 24 and 16 bit recordings- the newer 32 bit floating point setups are much more forgiving).
So it might be useful to take the old school engineering -30dB instrument level as a guideline. Start with that as your average level and see how the plugin responds. You might be able to get away with -24dBV peaks. I haven’t messed with the various plugins but I know recording.
Anyway, some standardized ideal level is better than just “minimum”.
Or record peaks at -6dBV then attenuate into the plugin to keep good resolution without gaining out as you describe.
Most (consumer) interfaces these days have maximum input headroom of around 12-13dBu, which is roughly 3V RMS. That should be sufficient for any pickups, hotter ones will be close to clipping, lower output will still be at a good level.
The highest headroom interfaces tend to be mastering grade ones with very low noise specs (and stepped gain). It’s quite hard to record a DI that is too low in level, as the noisefloor of the interface will be well below the background noise of the pickups.
BTW big fan of your channel!
Thanks eds. I’m just concerned about the resolution as I know how lousy stuff recorded too low can sound if you bring it up in the DAW afterwards.
But you raise a good point - if you have low level stuff recorded with say a Focusrite Scarlett then you raise it in the DAW (or amp it with the plugin) you’ll be bringing the interface’s noise floor up with it.
There is some ideal sweet spot that would match all the criteria - we just need to know if this is -30 LUFs or -24 or whatever.
@@PsionicAudio Definitely wise to be aware of noise floors and any sources of problems!
The difficulty is interfaces all having different definitions of analog voltage to what digital level it is assigned. On one converter a 1V RMS sine could be a totally different level to another.
Plugins have no way of knowing what interface everyone is using, and they have to assign their own dBu->dBFS reference level. Often they don’t say what this is, and that’s why this topic is such a grey area for a lot of people.
Hopefully with the topic gaining more traction, it will become more common for manufacturers and plugin developers to share this information as standard.
I’ll be watching this as it develops. There needs to be a published standard - have any of you checked the AES white pages? There might already be a standard buried in all their stuff - assuming the plugin makers adhere to them…
@@PsionicAudio Many analog modelled plugins adhere to 22dBu=0dBFS but this isn’t really suitable for guitar pickups. Likewise, some plugins made for mix bus use something expecting a higher signal.
It’s hard to standardise because different requirements need different specs. Likewise there can be different constraints on size, power, noise, etc.
IMO 12-13dBu is sensible for guitar amp sim plugins, but it’s not standardised. There are some excellent emulations that are 1V=0dBFS that need a WAY hotter signal than you could ever record with, simply because 1V=0dBFS is the definition SPICE uses.
Thx for the great info. Can you please tell what is the song of the intro? Thx in advance
My issue is that most of us that use plugins are just guitar players, and while we may want to dabble in recording, most of us are no recording engineers; we just want to plugin, dial in, and play. We just need industry help to understand where we’re goofing up.
The advice to crank the gain so that you are as close to clipping as possible without ever reaching it is engineers best practice to reach highest possible SNR. In practice it doesn't really matter because modern AD Converters have much higher dynamic range than any guitar signal could ever reach so you can set the gain to uniy and don't bother.
I like how my FM3 does it, it has just 3 fixed input pad settings: 6, 12, 18 dB. If your signal clips, you set the pad higher until it doesn't clip anymore (Just like you heard it everywhere else) but since these values are fixed and changed in software, the decreased analog gain gets added right back digitally so your sound doesn't change. I think every guitar audio interface should handle it that way.
Yep, it would be ideal if all interfaces worked like that, but they would still have to agree and standardize on how many real-world volts = how many digital dBFS. If they all settled on one value for that, AND had an auto-compensating input like you described, we’d be golden and never have to worry about this again.
So, am I missing something or if there's too much gain going in, wouldn't you just back it off on the input or elsewhere? Each interface is going to be a little different and so is where the yellow lights are coming on. I'm guessing these were written before the HiZ inputs were common? I don't really use amp plugins since I have hardware, so that seems superfluous to me.
Set the input gain for best s/n ratio and performance before the ad converter. Then in digital post ad coverter world, adjust the input to the plugin to whatever is preferred. This way you get a healthy signal with low noise converted into 0 and 1s and the plugin will play nicely.
I'm confused, can someone help me? When I lower the gain on my Scarlett i818 my guitar audio signal is so low I can't see it (visually the audio signal). I can get my guitar to clip before seeing it with a nueral dsp plugin via input level/gain. As soon as I turn up the gain on Scarlett I can see the audio signal. I've been on this problem for a couple days now. Any thoughts?
This adjustment has greatly improved the sound of my guitar amp sims ( helix native and neural ) and my mixes are easier and better sounding. Can I assume that I can improve the sound of my old ( higher input level ) recordings by remixing them with a gain reduction plugin at the front of the fx chain to bring the level down to where it would've been if it had been recorded with the interface input level set to zero?
I've had to adjust gain depending on what I'm playing. I usually have to increase my interface gain when I'm playing leads because the signal isn't hitting as hard as when I'm riffing on the low 7th string.
Am I just not picking as hard? Maybe. But the high strings don't have as much bass. I haven't tried going into a compressor pedal first to see if that makes a difference.
It is an interesting question. I think this has come from so many interfaces being out there, and they are not 'standerdized' in any real sense. The advice discussed here is not bad advice for most audio sources, but we are talking here about using plugins and guitars using instrument and/or line level.
I think the best advice would be to 'use your ears!'... there are so many variables in this equation, (guitar, plugin, interface), that no 'one sollution' is possible. Hence the 'general' audio advice of "as much signal as possible without clipping" is given by the manufacturer here.
very interesting, if that happened with sound card and Plugins, what about pedalboards like quad cortex , helix ,etc,... why woulndn't they be affected by the same problem ? would you so intend to get your input level around - 18dbs ?
I realized this about 6 months ago when I thought the signal sounded too saturated in my neural dsp plugins, so I started turning the input volume down on my interface and guitar to get a cleaner sound. DIdn't really care what was the 'proper' way of doing it, but the lower volume did sound way better to me.
Does this apply if using Guitar Rig by Native Instruments as well?
I'm using a UAD volt 176 with my input gain knob always at zero, and it sounds super awesome
Very good video, I'm contributing with my personal take. When I had the Focusrite audio interface, it would almost clip at volume 0, so when playing clean sounds it would translate perfectly. Then I added a Palmer DI, always Gain at 0, never had any problem. I then bought an Audient interface and the same happened, volume input at 0 and everything sounded perfectly. Neural absolutely wrote something easily misunderstandable, but also the players that would find something not right, with the input level so high, I think should've used their ears and adjust to their equipment just from the fact that everyone have different guitars, pickups. interface ecc. This is my personal take, not to judge your video which I find extremely clear and useful, cheers!
I'm guilty of this. As I posted elsewhere, I did at least figure out that input gain was an issue for me. But stubbornly, I kept following what I thought was "best practice" and tried to compensate instead via other means, such as an EQ pedal between my guitar and interface input. Instead of just using my ears. The EQ is still useful mind you, but I can use it much more effectively now that I understand the signal chain end-to-end a bit better. Alas, to add the confusion I think the EQ is buffered (MXR 10 band EQ). So just going to have to what I did before all this digital stuff existed lol. Use my ears.
I'm not anti-digital by any means - I just think it requires a little investment of time to understand it and make best use of it, and some players understandably just want to plug in and play. Absolutely nothing wrong with that.
can you link where you got that sweater? I want to get one for my mom!
Yes. I added gain on the input to get just below clipping because every instruction manual, every tutorial amd every video said that was best practice. Why would i question advice from the manufacturers themselves?
💚Ed helped me immediately when I commented on his page on this. It was around 3am & I was going crazy trouble shooting. I didn’t expect him to immediately respond, but he did. Not only that, he kept responding & guiding me after every message. He never left me until I decided to perhaps sleep on his info. I was extremely impressed & appreciative. Thank you both of you. It was ironic that Bea ended up doing a vid on this. To be honest, I was surprised to hear Bea confused as to why people were having issues. And surprised to hear that he never had to think of this. I mean, come on, we all know how to set levels in a studio environment. I think that our technology is producing new “unknowns”. For example, the happy accident of the 4 cable method, never being known when these floor fx units came out…. This community is quite tight🤙🏼
I wonder if this is why my DI sounds crunchy and not smooth like everyone else's does.
I'm gonna have to test this out because I always thought it was right to set your interface right below clipping.
Yep. I read manuals & followed the bad advice, too. Consequently, I blew the lid off my amp sims without knowing it.
Thanks for the way you handled this info…
I found it helps even at higain to do the levels Ed S recommends. Yea, some "fizz" is turned down, but its more articulate and such.
One thing i can't find anywhere is what about the gain when using a D.I. Box ?? , i should still need to keep the gain at 0??! (In my casi RNDI to Apollo Twin X )
This channel has helped me get a better sound with my plugins! and I'm thankful. I only have a small issue with the word "Gaslighting". I think that every product over the last 40 years has stated that "adjust the input gain so it's not clipping". Gaslighting implies causing confusion and harm with malicious intent. I don't think companies like Neural DSP and Line 6 are intentionally trying to mess with there customers. I think a lot of people in the media today use the word gaslighting without even knowing what it means, but it's a great buzz word!
I find that using the input on the plugin vs the interface is much more noisy than the before method, is that typical?
I don't know if I am missing something or maybe my audio interface is working differently, but If i put the gain to 0, I almost have no signal at the input, so the result is almost as if my guitar volume is turn all the way down. So for me, the result is the opposite, with a plugin like NDSP Gojira, what is suppose to be a massive distorsion is more like a crunch...
I definitely added gain to just below clipping level.
After an OS update my settings got wiped for some reason, and I noticed it sounded better (for my setup) to not do that; regardless if I was using a my guitar or synth.
Oh yeah, I ran in to that too they said just below clip - what you're saying totally tracks as best I can tell. My Nashivlle Delux Tele needs a mile of gain before clipping where as the Strandberg NX 8 you basically need to go in flatlined. Cheers
Im so confused, how would you know if ur getting a good amount of signal? I used a HX going into a 2i2 into my DAW so idk how id even go about that