E-guitar recording: Passive DI box or "Hi-Z" interface button?

[alexis]

Hello,

My interface (Steinberg UR28M) has a "Hi-Z" type button on two of the XLR inputs.

I've read that a DI box is good to use to get electric guitar audio into my DAW, and I do have one each of these passive DI boxes (Radial Pro D1 and D2, https://www.radialeng.com/product/prodi , https://www.radialeng.com/product/prod2 )

But I wonder if there's any disadvantage to keeping it simple by bypassing the DI, i.e., just plugging the electric guitar straight into the interface with the Hi-Z button pushed?

Thank you in advance for any thoughts!

[Wonks]

Passive DI definitely a bad idea for with electric guitars with passive pickups plugged directly in. The input impedance is too low for passive pickups.

Active DIs are better but you still need to have one with an input impedance of at least 500k ohms, and ideally 1 meg ohms.

Straight in to an instrument input on an interface is absolutely fine but then you need low latency monitoring of any amp sims being used.

You’d typically use an active DI box with a guitar/bass when you want to capture the direct electric sound before they use any pedals or play through an amp so you can then use your choice of amp sim, or reamp the signal through a real amp. In this instance, you really wouldn’t want to plug directly into the audio interface as you’d then need to send the signal out through another output on your interface. A high impedance active DI box is much easier to use here.

You can use a passive DI box after FX pedals (provided they are on or buffered) as they provide the high impedance the passive pickups need to see to stop them loosing high end.

The 1:10 rule of output to input impedance holds firm here, and guitar pickups have impedances from around 4k (e.g. a Gretsch Filtertron neck pickup) to the high 20ks for very high output humbuckers. You’re generally in the 6k to 9k for most pickups that aren’t described as ‘hot’.

But the pickup is not just seeing the input impedance of the amp or DI box as you have the volume control pot and tone control pots (tone in series with a capacitor that will pass all but the lowest guitar frequencies) in parallel with the amp or DI input impedance.

So whilst a 100k DI box input impedance on its own won’t start ‘loading’ guitar pickups with a DC resistance below 10k, it will rob more powerful pickups of high end (and they tend to be duller sounding on their own anyway). Combine that with a 250k volume pot and a 250k tone pot (which will be passing the higher frequencies we’re concerned about) wired in parallel with the a 100k DI input impedance and you get an approximate equivalence 1/R of 1/250 + 1/250 + 1/100, giving a resistance of about 55k ohms. This will start to load and remove top end from most pickups.

Make that input impedance 500k and you get an impedance of around 100k and 111k if the impedance is the amp standard of 1 meg ohm.

That’s fine for most guitars with 250k pots as these will normally be single coils, so pickups with typical DCRs of less than 8k.

Move to humbuckers where 500k pots are the standard and a 100k input impedance will result in an equivalent impedance of 71k, which is still going to load a typical ‘vintage’ style humbucker with an 8-9k DCR.

A 500k input impedance gives around a 166k equivalent, which is fine for most humbuckers up to say a Seymour Duncan JB.

1 meg here would give an equivalent 200k input impedance, so a bit better for the hotter pickups.

Which is why very hot output pickups are often fitted with 1 meg pots

1 meg pots and input impedance give an equivalent input impedance of around 333k, which should be OK for pretty much any pickup I know. The hottest I’ve had was a Kent Armstrong quad rails style with a 29k DCR, closely followed by a massively overwound P90 at 27k!

So it’s important to choose an active DI with a nice high impedance of 500k or ideally 1meg. The Radial J48, whilst often used on guitars, only has a 220k input impedance, making it ok for lower powered guitars, but it will start to load higher power pickups roughly above 8k DCR for singles coil with 250k pots, 14k DCR humbuckers/single coils with 500k pots and 15k DCR for 1 meg pots.

[Hugh Robjohns]

My interface (Steinberg UR28M) has a "Hi-Z" type button on two of the XLR inputs.

Highly unlikely! The hi-z option will apply only to the instrument input via the T(R)S socket — not the XLR connection.

I've read that a DI box is good to use to get electric guitar audio into my DAW...

Yes, when necessary, but the hi-Z option is there to replicate the functionality of a DI box so you don't need one in the signal path! Just plug the guitar straight in to the interface, select Hi-Z, and record away! (Assuming monitoring latency is low enough, of course)

I do have one each of these passive DI boxes...

They are great boxes for keyboards and anything with electronic outputs (drum machines, dj
Mixers, FX pedals, computers...etc etc), but their input impedance is really too low to work well with most electric guitars and basses, and the tone will usually suffer as a result.

Ideally, the electric guitar needs to work into an impedance well above 500k Ohms and thats why most active DI boxes, as well as mixers and interfaces with hi-z modes, present around 1Meg Ohms or more.

But I wonder if there's any disadvantage to keeping it simple by bypassing the DI, i.e., just plugging the electric guitar straight into the interface with the Hi-Z button pushed?

None at all — only positive benefits, in fact. And given the choice of passive DI or hi-z mode on an interface I'd choose the latter every time without hesitation!

[Music Wolf]

My interface (Steinberg UR28M) has a "Hi-Z" type button on two of the XLR inputs.

Highly unlikely! The hi-z option will apply only to the instrument input via the T(R)S socket — not the XLR connection..

The first two inputs are combi connections, as per my UR824. Just use a normal, TS, guitar lead.

As Wonks says, if you want to hear an amp plug-in whilst tracking you'll need to reduce your buffer size in order to get lowest possible latency.

[jimjazzdad]

So, here's a question: when a DI has a 'pass-thru' (parallel-wired) connection on the 1/4" input that is then connected to a guitar amp while the transformer-balanced signal is passed to the interface, does this affect the loading of the guitar pick-up and the tone?

[Hugh Robjohns]

So, here's a question: when a DI has a 'pass-thru' (parallel-wired) connection ... connected to a guitar amp ... does this affect the loading of the guitar pick-up and the tone?

Short answer: yes.

The input to the DI box's circuitry (whether that's a transformer or an active buffer) is effectively wired in parallel with the amplifier.

Parallel impedances sum according to the formula: 1/Rt = 1/R1 + 1/R2 + 1/R3 +...

So, if both the amp's and (active) DI box's input impedance are 1Meg Ohms each, the combined total seen by the guitar falls to 500k Ohms.

Now, that change of load probably won't result in an obvious tonal change — 500K is still pretty high.

However, switch to a passive DI where the transformer's input impedance is likely to be 150k Ohms at most. Sling that across the guitar-amp connection and the guitar now sees just 130.4k Ohms, which is getting a bit low and may well start to have an audible affect.

As a general rule, then, paralleling impedances will produce a total a little lower than the lowest value, or half of each impedance when they are the same.

Logically, then, the higher the DI box input impedance the better, as the less its affect on the total load.

For example, if the guitar amp's input impedance is lower at 500k, and the DI box is still 1Meg, the guitar sees a total of 333.3k Ohms. If the DI box had a 2Meg input impedance, the guitar would see 400k, and if the DI box was 10Meg the guitar would see 476.2k Ohms.

However, there are practical downsides to consider when it comes to raising the DI box impedance. First, a higher impedance inherently means a higher noise floor, and secondly it greatly increases the susceptibility to RF interference. Around 1 Meg Ohms is generally considered to be the sweet spot.

Instead of silly input impedances, an alternative approach to avoid the issue of a DI box pulling down the impedance seen by the guitar, is to provide a buffered output to feed the amp, rather than a simple paralleled wired link output. In that way the guitar sees only the (very high) DI box impedance, and the amp is driven from a low impedance source.... although this arrangement breaks the normal reactive interaction between guitar and amp which — in some cases — might also be considered to change the tonality!

[Wonks]

But as I said above, there's also the guitar volume and tone pots to consider in working out what the pickup 'sees' rather than what the guitar as a whole sees beyond its jack socket. The pickup doesn't care where resistances/impedances running between the signal output and ground are, just how many, their value and the resulting equivalent impedance value.

[Sam Spoons]

Yes, the guitar will see a lower impedance than it would plugged into just the DI or the amp. How significant the difference depends on the impedance of the amp input and that of the DI box, active DI boxes have an impedance sufficiently high that placing on into the circuit only makes a very small difference which may have an unnoticeable affect on the tone. As has been mentioned above passive DI's have a much lower input impedance which probably will have a noticeable affect on tone.

[Wonks]

This just reminded me that my Two Rock Studio pro 22 amp has a variable input impedance selector switch. I can't find a schematic for it, so I'm assuming that by measuring across the T and S of a guitar lead plugged into the amp, I'm basically measuring the resistance of the grid leak resistor which sets the main impedance of the amp.

The resistances I measured were (slightly lower than I expected)

600k
214k
94k
31k
7.5k

Plugging a Strat into the amp and adjusting the selector switch, obviously the 600k gave the bright sound I was expecting, but the next step down to 214k only seemed to have a very minor effect on the sound, just very slightly softer but still with a lot of definition.

Stepping down to 94k, yes, there was a bit more loss of treble and the overall volume was down slightly, but even so, it was still very useable, just slightly very less clarity.

31k and the duller 7.5k were where you really noticed the treble roll-off and associated loss of volume.

It is a quite similar effect to rolling the tone control off on a guitar, but it's a slightly less bassy sound, as it's also rolling off some of the bass rather than letting the bass through and rolling off the just the upper frequencies as a basic guitar tone circuit does. Of course the tone control on a guitar functions in two ways as not only a) where it bleeds the mids and treble off down to ground but also b) as it also lowers the overall impedance the pickups see, so there's a double high frequency roll-off effect.

My upper frequency hearing is long gone, but it's still (just) enough to hear the full range from a guitar speaker, so I might possibly be missing some of the finer detail that others with better hearing would notice, but in the really important frequencies for an electric guitar (at least with single coils), it does show that whilst having a nice high input impedance on that amp is important to get the best out of the pickups, you can actually drop the value a fair bit before it really becomes noticeable.

[Hugh Robjohns]

...which is why Radial's J48 DI box (with its 220k input impedance) isn't actually the disaster some claim of it, and why passive transformer DI boxes — which was the only game in town in the 70s and early 70s — are edging into the audibly altering but still just workable region.

Anything around 1 Meg and above is really faultless in terms of no audible loading effects.

[Wonks]

I still don’t know quite why they didn’t use a higher value though, as it will affect higher power humbuckers far more than vintage spec single coils. Though to be fair to Radial, they only really push it for bass use, where the higher frequencies are rarely as important (unless you’re doing slapping or something else that puts out a lot of treble).

[Hugh Robjohns]

Yes, it's obviously a deliberate decision as several other Radial active DI models also present a 220k impedance.

Fortunately, there are many alternatives with much higher impedance inputs.

[alexis]

Thank you again, Hugh, Wonks, Music Wolf, and Sam.

Reading all this on the run a bit, I will review, but as is usual here on SOS, it seems my question has not just been answered but also supplemented with some basic theory to help me when a similar question comes along ((as a piano player, the only pots I've ever thought of needed occasional watering! )

[alexis]

Yes, it's obviously a deliberate decision as several other Radial active DI models also present a 220k impedance.

Fortunately, there are many alternatives with much higher impedance inputs.

Not that I'm planning at this point to run out and buy any of those higher impedance active DI boxes (will try my Hi Z input first), but if I were going to do that one day, do you have any favorites you might recommend?

Thank you!

[alexis]

My plan (at this time, comments welcomed!):

1) Mic up guitarist's amp. We'll see if the lack of treatment in the oddly shaped room (which includes a wall that is part of a stairwell down) makes the recording too unpleasant to listen to or not. He'll hear this as he plays, either out of the cab speakers, or through headphones.
2) Find some "through" on his amp that sends a dry signal to my Hi Z interface input. I'm hoping I won't need to use that at all (either: live, or after the fact) to apply amp sims to the signal, because that would involve making the guitarist happy with the results of my using a tool I have zero experience with ... but in case of catastrophe with the cab mic'ing, at least there'd be something to potentially work with, if needed.

The guitarist is great, and/but has zero recording experience, so as crazy as it sounds, making everything work recording- wise will be almost completely up to me.

[Hugh Robjohns]

I was just looking at Sweetwater's catalogue to see what was easily available in the US, and my default suggestion would be the BSS AR 133. It's solidly built, affordable, sounds great, and works reliably.

Here's a (somewhat re-edited and mangled) review I wrote of it for SPL magazine (a sister title of SOS, now defunct and not available online!):

https://adn.harmanpro.com/product_revie ... iginal.pdf

Just ignore the silly XLR alternative input/output (which isn't balanced and isn't for a mic input!). It was basically included as either a locking input option, or as a link-thru output if the normal link output was re-configured as a buffered output...

In the UK the best value DI boxes are undoubtedly the various Orchid DIs (although they don't provide transformer isolation which might be a problem in some stage applications), but I don't think they are easily available (or cheap) in the USA.

https://www.soundonsound.com/reviews/or ... s-di-boxes

My personal favourite DI box is the Canford Audio Active DI (originally designed by Theatre Projects), but it's no longer made, sadly. Brilliantly versatile, though, with a 1.8M Ohm input impedance and a link output socket which can be switched from a parallel link to a buffered and transformer-isolated output (which is also balanced if connected with a TRS plug!) I can't tell you how often that has been a real problem solver... but no one else has done it in a DI box! It's a real shame they stopped making it. I scour ebay for them regularly and have snapped up a couple of extras!

https://www.canford.co.uk/Products/20-0 ... om-powered

Like microphones, I would argue that it's well worth spending decent money on a good DI box or two because they'll last a lifetime, and they determine the quality of sound you're going to get.

So while $2-300 might feel expensive for a DI box, I'd argue (for hours) that its well worth it when you consider 40 or 50 years of top quality use!

From the Sweetwater catalogue, other personal recommendations besides the BSS include the Grace Design M303 which sounds absolutely superb and is beautifully engineered, as you would expect. I reviewed it not long ago.

https://www.soundonsound.com/reviews/grace-design-m303

The various Rupert Neve Design active DI models are worthy contenders too.

https://www.soundonsound.com/reviews/ru ... gns-rndi-m
https://www.soundonsound.com/reviews/ru ... signs-rndi

The Royer Dbooster2 Is another good, if unexpected, option if its just for recording use. This unit doubles up as a superb dual mic gain-booster with dual DI functionality too — so very handy if you use low output mics occasionally, too. But it has no loop-thru outputs to feed backline amps, and no isolation transformer, so great for recording, but not much use on stage. Sounds very nice, though.

https://www.soundonsound.com/reviews/ro ... -dbooster2

The Radial J48 (and its stablemates) is very well known and highly regarded, but it has a low-ish input impedance, as discussed above. It's great for keyboards and the like, but arguably not ideal for guitars.

https://www.soundonsound.com/reviews/radial-j48-jpc

Avoid the Behringer active DIs, as they tend to be rather noisy. Fortunately, there are plenty of other marques and models around to choose from across a wide price range.

Things to look out for if choosing an active DI:

Phantom powering option saves on battery costs, but make sure it still works with the ground lift enabled. Not all do.

A ground lift switch can be a very useful problem solver.

1Meg Ohms or higher input impedance.

Output transformer for safety isolation on stage (not as important for studio recording). Cheaper boxes often omit the transformer, but while that can be good from a fidelity point of view, it's not for electrical isolation in event of a fault somewhere!

Check the input clipping level. Some clip at quite low levels. Ideally you want +9dBu or higher (without pad engaged).

Many DIs have pad switches to cope with line or speaker level sources, which is great, but their use inevitably reduces the input impedance below the 1Meg mark that might be offered without the pad. That doesn't matter for line or speaker level sources, but it's really not good if you have to engage the pad just to deal with hot pickups on a guitar — which is why the clipping headroom is important!

A buffered unbalanced output is a useful option as an alternative to a parallel loop-through/link sometimes, as it avoids unnecessary loading on the source instrument and can improve RF interference on amp feeds.

Hope that helps.

[merlyn]

Guitars lose treble because pickups are inductive. The inductance of the pickup combined with the input impedance of what it's plugged into make a low-pass filter:

L is the pickup and R is the input of an amp/DI box/whatever. The cutoff frequency is given by:

fc = R/(2*Pi*L)

As R goes up, the cutoff frequency goes up (more treble). As L goes up the cutoff frequency goes down (less treble). This explains the toan suck that bothers guitarists so much ("muh toan!") and why treble is lost when the guitar volume is turned down. A strat pickup has an inductance of around 2.5H (more treble) and a humbucker has an inductance of over 4H (less treble).

[James Perrett]

My plan (at this time, comments welcomed!):

My suggestion would be to stick with 1). If the guitarist is used to using an amp then they probably have a sound that they like and they won't be happy with a DI'd sound.

However, give yourself plenty of time to set up and experiment with mic types and positions. You'll gain a massive amount of valuable knowledge by experimenting so, if you don't have time on the day, then try to get the guitarist to come along a day or two early so that you have the time.

My top tips are to not point the mic at the centre of the speaker cone and to try using devices that you wouldn't normally consider to be microphones. One of my favourite electric guitar mics is a Sennheiser HD400 earpiece.

There are more ideas at

https://www.soundonsound.com/techniques ... ric-guitar

and

https://www.soundonsound.com/techniques ... -recording

[Wonks]

Guitars lose treble because pickups are inductive. The inductance of the pickup combined with the input impedance of what it's plugged into make a low-pass filter:

L is the pickup and R is the input of an amp/DI box/whatever. The cutoff frequency is given by:

fc = R/(2*Pi*L)

As R goes up, the cutoff frequency goes up (more treble). As L goes up the cutoff frequency goes down (less treble). This explains the toan suck that bothers guitarists so much ("muh toan!") and why treble is lost when the guitar volume is turned down. A strat pickup has an inductance of around 2.5H (more treble) and a humbucker has an inductance of over 4H (less treble).

That’s a good first approximation, but a pickup also has capacitance (as does the guitar cable) so the result is a resonant low pass filter. Depending on the pickup design there can also be eddy current losses, which is why covered pickups can sound a bit duller than their uncovered versions.

All very well described here:
https://guitarnuts2.proboards.com/threa ... ar-ltspice

[alexis]

My plan (at this time, comments welcomed!):

My suggestion would be to stick with 1). If the guitarist is used to using an amp then they probably have a sound that they like and they won't be happy with a DI'd sound.

However, give yourself plenty of time to set up and experiment with mic types and positions. You'll gain a massive amount of valuable knowledge by experimenting so, if you don't have time on the day, then try to get the guitarist to come along a day or two early so that you have the time.

My top tips are to not point the mic at the centre of the speaker cone and to try using devices that you wouldn't normally consider to be microphones. One of my favourite electric guitar mics is a Sennheiser HD400 earpiece.

There are more ideas at

https://www.soundonsound.com/techniques ... ric-guitar

and

https://www.soundonsound.com/techniques ... -recording

Cheers, James, for the advice and the references!

I'd come across the first SOS article, but somehow missed the second.

Should have all the time needed to experiment with different mic'ing positions, at least until my almost infinitely patient wife meets even her match in terms of exposure to loud, repetitive, only nominally musical-sounding noises !

[alexis]

I was just looking at Sweetwater's catalogue to see what was easily available in the US, and my default suggestion would be the BSS AR 133. It's solidly built, affordable, sounds great, and works reliably.

Just ignore ... the silly XLR alternative input (which isn't balanced and isn't for a mic input!). It was included as a locking input option... not that anyone has ever had a problem with jack plugs accidentally falling out of DI boxes....

In the UK the best value DI boxes are undoubtedly the various Orchid DIs (although they don't provide transformer isolation which might be a problem in some stage applications), but I dont thing they are easily available (or cheap) in the USA.

My personal favourite DI box is the Canford Audio Active DI (originally designed by Theatre Projects), but it's no longer made, sadly. Brilliantly versatile, though.

Like microphones, it's well worth spending decent money on a good DI box or two because they'll last a lifetime, and they determine the quality of sound you're going to get.

So while $2-300 might feel expensive for a DI box I'd argue its worth it over 40 or 50 years of top quality use!

From the Sweetwater catalogue, I can warmly recommend the Grace Design M303 which sounds absolutely superb and is beautifully engineered, as you would expect. I reviewed it not long ago. The various Rupert Neve Design active DI models are worthy contenders too.

The Royer Dbooster2 Is another good, if unexpected, option if its just for recording use. This unit doubles up as a dual mic gain booster and dual DI — so handy if you use low output mics occasionally, too. But it has no loop-thru outputs to feed backline amps, and no isolation transformer, if that matters. Sounds very nice, though, as a studio recording DI.

The Radial J48 (and its stablemates) is very well known and regarded, but it has a lowish input impedance, as discussed above. It's great for keyboards and the like, but arguably not ideal for guitars.

Avoid the Behringer active DIs, as they tend to be rather noisy, but there are plenty of others around to choose from across a wide price range.

Things to look out for if choosing an active DI:

Phantom powering saves on battery costs, but make sure it still works with the ground lift enabled. Not all do.

A ground lift switch can be a very useful problem solver.

1Meg Ohms or higher input impedance.

Output transformer for safety isolation on stage (not as important for studio recording).

Check the input clipping level. Some clip at quite low levels. Ideally you want +9dBu or higher (without pad engaged).

Many DIs have pad switches to cope with line or speaker level sources, which is fine, but their use inevitably reduces the input impedance. That doesn't matter for line or speaker level sources, but it's really not good if you have to engage the pad just to deal with hot pickups on a guitar!

A buffered unbalanced output is a useful option as an alternative to a parallel loop-through/link sometimes.

Oh my goodness thank you so much, Hugh!
.

Hope that helps.

That doesn't even begin to describe how amazing this post is.

Thank you again

[Hugh Robjohns]

That doesn't even begin to describe how amazing this post is.

I've edited it to add some review links and clarify a couple of points.

[Sam Spoons]

My plan (at this time, comments welcomed!):

1) Mic up guitarist's amp. We'll see if the lack of treatment in the oddly shaped room (which includes a wall that is part of a stairwell down) makes the recording too unpleasant to listen to or not. He'll hear this as he plays, either out of the cab speakers, or through headphones.

He won't hear (or at least probably won't notice) the dodgy room sound the as he plays, our ears and brain are exceptionally good at filtering out bad room sound and focusing on the 'wanted audio'. When we are distance miking an instrument the room sound inevitable comes back to haunt us when we listen to the recording, but, as you'll presumably be close miking the guitar speaker there may be little room sound picked up so that may be ok.

2) Find some "through" on his amp that sends a dry signal to my Hi Z interface input. I'm hoping I won't need to use that at all (either: live, or after the fact) to apply amp sims to the signal, because that would involve making the guitarist happy with the results of my using a tool I have zero experience with ... but in case of catastrophe with the cab mic'ing, at least there'd be something to potentially work with, if needed.

if his amp has a 'thru' or a DI output both will provide a buffer between the guitar and the audio interface so you won't need the HiZ input, a standard line in or mic in (for the thru and DI outputs respectively) will be fine.

The guitarist is great, and/but has zero recording experience, so as crazy as it sounds, making everything work recording- wise will be almost completely up to me.

[James Perrett]

The guitarist is great, and/but has zero recording experience, so as crazy as it sounds, making everything work recording- wise will be almost completely up to me.

I missed this the first time round...

An inexperienced recordist will try to make everything technically perfect and get really worried about making everything sound "right". This will make the guitarist feel ill at ease and lead to a poor performance.

An experienced engineer won't let the technology get in the way and will work in the way that the guitarist wants to work. All you have to do is to point a mic in vaguely the right place and record a great performance.

Anyone listening to your recording won't worry about what mic you use or where it is pointing (apart from the odd geek who's opinion isn't that important in the big scheme of things). Most listeners are going to listen to the performance and an inspiring performance is going to make them feel better about the music than a stilted, lacklustre but technically perfect one.

[Hugh Robjohns]

All you have to do is to point a mic in vaguely the right place and record a great performance.

All true... but it's a chicken and egg thing, isn't it?

To get the experience to know where the 'vaguely right place' is, and how the sound changes when pointing at different places or with different mics, you need to experiment and learn, and that requires a confident and patient guitarist who wont be fazed by the experimentation!

If you don't have one of those, then an SM57 aimed towards the edge of the cone will definitely get something workable.