Too much dynamics or insufficent preamp? (Focusrite Scarlett 18i8)

SkyFly

New member
Cheers!

I've been making songs for some time now, but until now I haven't paid much (or more preciesly enough) attention to my recording levels. I already feel little embarrased to even ask this, maybe it's so obvious nobody even asks it, but...here it goes, cause I need an answer to this. :facepalm:

I'm recording my electric guitar straight to Focusrite Scarlett 18i8 preamp, with "Inst" switched on from MixControl. I'm using Reaper, 44.1 kHz, 24 bit. I get a good and excellent-sounding signal. Unfortunately, as I play louder parts especially from lower positions or open chords, the dynamics are all over the place! Let's say I record this riff, then I listen it, and watch what happens in "JS: schwa/audio_statistics"-plugin, or straight with meters, doesen't matter. The recorded signal hovers somewhere between -33 dB and -2 dB. Total RMS is somewhere -22 dB. Scarlett's mic preamp gain knob is at zero.

I think this dynamic range is bit too much. I think the -2 dB peaks are too hot. Why is this happening? Is my playing too rough in some places so the preamp clips? I think this is ridicolous cause there's supposed to be dynamics in playing...

What do you think? Is my playing so crappy it blows the signal all over the place or is it my guitars pickups, or is this happening because the 18i8:s preamps are not good enough to capture the dynamics or SO good I don't just understand this? :rolleyes:

I'd like to know how fellow musicians are able to record with these preamps guitar straight in if you HAVE to play softly...everything else (vocals, keyboards, enviromental samples) works and sounds like a charm when recorded, no complaints. This happens only with electric guitar. And it souldn't have hot picks, just stock Alnicos.

And, if this all is perfectly OK, I'm curious to will these brief -2 dB or so ruin the WHOLE recording, given it's "normally" at some -22 dB RMS?

Thank you in advance!
 
Nothing is "ruined" by a -2dB(FS) peak. Ideally you'd have a bit more headroom but, until you hit actual clipping, you should be fine.

Forget the numbers for a moment...how does you recording actually sound? Is it a fairly good representation of what you hear when you're playing? That's the main thing.

Depending on your playing style, electric guitar can have a very wide dynamic range. You get a transient/peak when you actually strike the keys but there's little or no resonance (like inside the body of an acoustic guitar) to sustain things "between strums". A bit of compression and limiting when you mix should even things out.

If you can whiz through the ten post restriction (there's a forum specifically for that!) maybe you should post a quick sample of your raw guitar recording so we can comment in more detail--a downloadable file sharing site would be ideal.
 
If it's not distorting, it's not a real problem. That is a pretty wide dynamic range, and you're probably going to want some form of compression on it. That would usually come from the amp, but you can get it from a sim, or a compressor or limited plugin, or whatever. Honestly, though, as long as it sounds good, it doesn't really matter what it looks like, right?

Are you playing humbuckers or single coils? The hot humbuckers in my main guitar can push the instrument ins on either of my interfaces to clip. The not-so-loud HBs on my SG get pretty close when I really dig in. I don't think any of my single coils will get quite as loud as you've described, but it's not far enough outside expectations to suspect any problem.

The thing is that the folks who design these things seem to expect "instrument level" signals to be pretty quiet compared to nominal line level, and therefore the DI inputs always have a pretty significant amount of gain on them. But "instrument level" is nowhere near a standard. Different guitars with different pickups and different players can have wildly different output levels, and some of them are just a lot hotter than "expected" therefore there's too much gain. Nevermind if you put your favorite boost pedal ahead of it!

I have found that, given a decent interface, there is no good reason that any guitar (even with weaker SCs) really needs any of that gain on the way into the machine. That is, you can run into the +4dbu line input on your interface and never have to worry about clipping even with your booster pedal cranked. You'll still be way above the noise floor of the interface, which is way below the noise floor of any passive guitar. Yes, the level ITB will then be noticeably lower, but you can add nice, clean, flat, silent digital gain after you've safely captured it, and with floating point math internal to the DAW you can add all the gain you want without fear of clipping until the very end.

...Except that the guitar really needs to see a high impedance load, and it won't get that from plugging straight into your line input. It needs a good buffer in between. Any pedal that doesn't say True Bypass on it, powered but turned off, should do the trick just fine.

But I'm not sure you need to bother with that in this particular case. You're not hitting 0dbfs. Getting close enough to make me a little nervous, and if you ever want to record a louder guitar...but right now you're ok.

On the other other hand, if you're going to run the thing into a decent amp sim with even a little bit of distortion (overdrive, crunch, dirt, fuzz) you won't be able to hear the effect of occasionally clipping the converters anyway. It's not "best practice", and sort of limits your options (pun intended), and I have gone to line ins for safety's sake, but I got perfectly usable stuff just plugging into the instrument input and letting it clip on the loudest strums since it was going into an overdriven "amp" anyway.

HTH
 
Thank you very much for your insights!

Yes, I seem to have lots of variables in this equation. One of the best points here is that in my case the signal will be overdriven anyways through an amp sim, yes. And that is actually the point that interests me most, because this is what I've been thinking, that maybe the "close-to-clipping"-signal is not a problem after all. I just want to learn the aforementioned best practices so I won't be in trouble when I start seriously learning mixing. And that's because I'm very careful with the recordings I make, they have to be precise, and, as you can image, I don't wanna record them again voluntarily. ;) Not to mention if I would have to ask someone else to record again, cause I ruined it right from the beginning...

Yes, the signal is much louder with bridge humbucker. And of course, playing style depends of the style of music. Everything easier is of course fine, and playing won't be so aggressive. Single coils and neck humbucker are fine. But me being leaning towards metal ja hard rock...well, the result is quite obvious. ;)

This is actually quite curious thing, because ages ago I used to upgrade my guitars with hotter and hotter humbuckers and now I have this great DBZ Bolero, which indeed has stock pickups. What's fascinating, is that I've never even thought about upgrading these pickups. Maybe they are hot? And in this case, too hot. That would be a shame. Because the sound indeed is excellent when recording straight to 18i8. I don't care anymore about the looks of the waveforms, I've grown up from that, so that's not an issue. I just want to know I won't ruin my recordings, if I can avoid that with for example, a compressor. But, buying a 300 € comp which is the same price as the interface, doesn't seem very clever unless I find some other use for it as well, which CAN be the case...

And I tried the "Line" mode. Unfortunately, it drained the tone. Not as good as with instrument mode. The signal's balls drop out when I do that. :rolleyes:

I find it also odd that Focusrite states in its manual that "Pad is not inteded to be used with Instrument input mode". Do you have any idea why is that? Because the level gets too low? Well, that would be a cure for me. =)

That's disapointing is the only "real" cure for this is that I have to play softer. Honestly, no one should HAVE to do that and (without any bragging, that's not the intention) my recordings are what they are because my...playing. And I would like to record that. I mean, that's the the main reason why I bought this interface.
 
Does the pad even do anything in instrument mode? Doesn't make much sense to gain it up and then attenuate back, or vice versa, does it?

When you say "drained tone" and "balls drop out"...can you be more descriptive? Plugging a passive guitar directly into a line input should kill a lot of the treble, which is often called "tone suck", but when you say "balls" I personally think of the lows and low mids, which shouldn't really change much except that the whole signal will be quieter. You have to compensate for the lower level either by adjusting the controls on the amp sim or by adding a gain plugin before the sim. Did you try through a pedal into the line input?
 
I have to try that line mode again. I remember it was considerably weaker, as it of course should be. I didn't use any pedal between, because I'm after pure, uncoloured sound. If I would go with pedals etc., I'd like them to colour the sound in desirable way. And now I don't have such a thing. :rolleyes: But I will give it another shot.

I'd like to know if anybody else has similar issue(s), or am I ranting here for nothing and should just ignore these peaks and let it burn? If it really doesn't ruin the recording.
 
The buffer in a decent pedal does the same thing as the input circuitry on your interface, only it wont add the unnecessary gain. If you are at all handy with a soldering iron (you said you've swapped pickups before) you could build a nice buffer for almost nothing, and then you'd know exactly what it was doing and choose the design that meets your expectations for purity.

But I really don't think you're going to ruin anything as is. It's not clipping anything, and it's perfectly normal for what you're doing. I kinda get the feeling that it's going to nag at you, though...

There are plenty folks around here doing the DI thing, but I think most just plug into the hole that says guitar and hit record without ever scrutinizing the result as closely as you have.
 
Just want to clarify myself, is this sort of issue common perhaps in this price range of interfaces/preamps, or with preamps more generally speaking?
 
I kinda get the feeling that it's going to nag at you, though...

And here we have a winner. :rolleyes:

It's just.....I've rolled so much different equipment in my hands that I thought it would be nice to simplify my recording setup. Cause I've had so much useless garbage in my recording chain before. But I guess I just have to use some aforementioned workarounds for this issue if it continues to nag me. But meanwhile I'd like to post example recordings as you requested tomorrow, so you can get the idea what's going on.
 
You kinda ninjad my edit there, but I didn't answer that question. AFAIK, every interface that includes an Instrument DI input applies gain to that input. You can check the specs yourself. The maximum input voltage for the DI will always be significantly less than that of the line inputs. The difference is the gain being added to the DI. I haven't actually ever looked at this on a standalone preamp with DI, but I suspect it's standard practice there too.

I think it's kinda silly, honestly. Recording at 24 bits with today's technology there is no good reason I can think of for this. The only thing I can think of is S/N concerns, but again there's just not enough noise in a decent modern preamp to touch the ocean of noise that comes with the guitar signal itself. Heck, you can get away with running mic level signals into the line ins sometimes!
 
Does the pad even do anything in instrument mode? Doesn't make much sense to gain it up and then attenuate back, or vice versa, does it?

When you say "drained tone" and "balls drop out"...can you be more descriptive? Plugging a passive guitar directly into a line input should kill a lot of the treble, which is often called "tone suck", but when you say "balls" I personally think of the lows and low mids, which shouldn't really change much except that the whole signal will be quieter. You have to compensate for the lower level either by adjusting the controls on the amp sim or by adding a gain plugin before the sim. Did you try through a pedal into the line input?

I kinda get what the OP is saying. On any proper interface (and the Focusrite is a good one) the Instrument/Line switch doesn't just change the levels; it also switches between high impedance for instrument sources and a much lower impedance for Line level. The impedance difference can make a slight change to the frequency response and, for somebody who knows the sound they want, this can be enough to "drain the tone".
 
I kinda get what the OP is saying. On any proper interface (and the Focusrite is a good one) the Instrument/Line switch doesn't just change the levels; it also switches between high impedance for instrument sources and a much lower impedance for Line level. The impedance difference can make a slight change to the frequency response and, for somebody who knows the sound they want, this can be enough to "drain the tone".
Yeah, we hit that a couple times. 10K input impedance will kill all the treble from most passive pickups, and cause some pretty severe broadband attenuation on top. It needs a buffer.
 
Allrighty then, it's a new day with new tests and conclusions!

To get new perspectives to this "issue", I took my POD HD500 to home for further testing. I actually quite gave up with this thing long time ago, and left to my band's rehearsal place for occasional recording purposes, but I never really recorded much with this. But I remember that I got good results with Spdif "Dry Input"-mode to EMU 0404 USB, which is in there still. Okay, so now I gave it another chance. Lucky me, I did so. There is quite a difference with these to, Scarlett 18i8 preamp and POD HD500 Spdif (won't say POD preamp only cause I really don't know if it does something to the signal, all I know is that it shouldn't). 1MOhm input mode on POD, I read from Mike Rivers' excellent 18i8 review that he measured 18i8 preamps to be "a bit over 200kOhm from TS when in Instrument mode", so I will trust that. My electricity meter is...somewhere. :rolleyes: Yes, this is completely different approach, and these tests are not equal. But it shows a difference. If you want to, just take a listen of these test recordings I made. Rough playing, just for demo purpose, do not concentrate on my mistakes or "not-good-enough" distorted sound blah etc. ;)

https://www.dropbox.com/sh/xwmsi546zw8q0mz/ebtC0yHt9Z

So here we have 18i8 preamp clean (guitar straight in, instrument mode), POD Spdif (guitar straight to POD, spdif from POD to 18i8), no fx or whatsoever in Reaper.

I find that the Scarlett is definately too hot. Dirty in a bad way. POD provides significantly better clean. Detailed, pleasant to my ear.

And then I tested how these two very different signals behave when using them as I'm intending to use them. This is where it gets interesting. There are Scarlett 18i8 Preamp Dirty (just Guitar Rig amp sim with keFIR cab sim) and POD HD500 Spdif Dirty (just Guitar Rig amp sim with keFIR cab sim). So they both have exactly same settings and volumes etc.

I find the Scarlett Preamp actually more punchy and roary. More aggressive. The dynamics, though overloaden, are there. POD is still more detailed, but little duller. More flat, hifi. This isn't a synonym for bad still.

I'm curious to hear what you guys think. =)
 
What a marvelous day this has been. =)

I made further testing with clean and distorted parts, and Scarlett is a winner. I recorded more amp sim overdriven parts, 18i8 makes the sound more roaring and chunky. POD sound remained almost as good, but it was duller. Neutral, flat, boring! I mean really, that bite and aggro just wasn't there. I would definately use POD for detailed hi-fi-style stuff if needed. When I think about it, I've always liked that dirtyness and powerful roar in guitar sounds, I guess that just suits my taste. :rolleyes:

Even though the signal was scarely hot in 18i8, it was better. POD makes fantastic job in clean sounds, I've always known this as I use it mainly for acoustic guitar to PA. If I would compare pure DI signals I would prefer POD. But when using that DI signal as my intention is, to process it, the choice is Scarlett.

Thanks to you guys a lot! I guess I could've perhaps come to this same conclusion on my own, but your feedback actually made me study this with a deeper focus and thought. =) Now I can concentrate on making that music, instead of staring those damn levels. ;)
 
Thank YOU for reporting back...all too often it can feel like we're posting into a vacuum!
 
What I notice recording on this interface is distorted amp sims sound overly distorted and "break up" to the point I lose any/all detail in the distortion. Have you noticed this? Have you found any solution? Does the pad help? You said FR said not to use the pad--why?
 
What I notice recording on this interface is distorted amp sims sound overly distorted and "break up" to the point I lose any/all detail in the distortion. Have you noticed this? Have you found any solution? Does the pad help? You said FR said not to use the pad--why?

I notice that with amp sims and every interface I have owned...
 
Wow, I missed SkyFly's last couple posts.

There's maybe a couple of things going on here, but I think they both come back to that extra gain on the instrument input.

First, you're just plain pushing a hotter signal into the amp sim. It's (almost) exactly the same thing as putting a booster pedal between your guitar and amp. These things are designed to run at some poorly defined nominal input level. I've generally assumed that they kind of expect that extra 10db to have been applied on the way in, but I've never seen a single one of these actually specify anything like -xdbfs = xVp2p into a real amp. That spec must exist, but for some reason none of the manufacturers think that anybody wants to know. The closest I've gotten is in the PodFarm manual it says something about how it works best with levels "around -18dbfs". So, hitting the amp harder makes more distortion at the same settings, just like in meatspace. I'd be willing to bet that if SkyFly had added 9-10db to the track he recorded through the Pod before it hit the amp sim, it would be a lot closer to the same as the one through the interface.

But then, it seems he had a relatively hot guitar to begin with, and the added gain on the instrument input was enough to push the interface itself into clipping and distortion. So now he's got a bit of extra harmonic content going into the amp, which gets further distorted and adds harmonics of those harmonics. So there would probably be some difference even if he did gain match the two tracks first. The difference between a "clean" and "dirty" booster before the amp.

So, what to do about it? Adjust the level of the input to the amp sim. All of these plugs of which I am aware have a knob - usually marked input - in the "wrapper" (as opposed to one of the control knobs on the simulated gear) which is just pure digital multiplication - completely clean and flat gain or attenuation. Turn that knob until the sim responds the way you want. Or put some other plugin before the amp sim. In SkyFly's case, turning up the input on GuitarRig might work, but if he misses that little extra distortion from the interface, he might prefer to put something like a dirty limiter, saturation, or stompbox booster model in front.

4tracker might get away with turning the input knob down, but if he is also distorting his interface, and that's not cool for whatever reason, he might have to find some way to attenuate the signal in the analog world first, or, like I talked about above, buffer the guitar and run to a Line in which doesn't add that gain.
 
"The thing is that the folks who design these things seem to expect "instrument level" signals to be pretty quiet compared to nominal line level, and therefore the DI inputs always have a pretty significant amount of gain on them. But "instrument level" is nowhere near a standard. Different guitars with different pickups and different players can have wildly different output levels, and some of them are just a lot hotter than "expected" therefore there's too much gain. Nevermind if you put your favorite boost pedal ahead of it!"

I think, Ashcat that the "poor" AI designers are between a rock and a hard place when marketing AIs these days for many of the "bedroom wannabees"? Their marketing guys think (and they may be right?) that new customers want Hot-Hot-Hot! And if they can't make their wimpy Kay's Cat Strat go into the red they won't sell the interface.

Trouble is, as you allude to Ash', if you design for "worse case" you will get clipping with a lot of guitars because AIs generally have poor headroom, especially those that are bus powered or aimed at the budget market.
The solution is pretty simple, M-Audio and others had it cracked years ago! Fit a pad switch. The venerable Fast track pro can accept +22dBu with the pad in at min' gain. Anybody got a geetar hotter than that?

But of course, switches and resistors cost money and there is only so much in the pot for any particular design...FAR be it from me to tell peeps where best to spend it!

A few years ago I made a gizmo (that was going to make me a miwyonare Rodney) that plugged into a USB port and produced 9volts at about 150mA to power a guitar pedal. I also incorporated a chip impedance converter (gain = to 1) so that the box could drive an internal (crap) soundcard. I found that a Mex Strat would only produce about -20dBFS with moderate strumming and much less on single notes. Now, those levels, even for 16bits, are totally acceptable for track building but could you sell it?
I reasoned that I should need another stage of gain, maybe a pad, maybe a pot...Starts to get messy!

Oh! BTW, I am still poor!

Dave.
 
Back
Top