Recording Guitar with a DI [2010 thread revived]

[chokehold]

The downside of active DI boxes is that they don't play well with high-output and especially active pickups. Shooting my Ibanez bass' active Bartolini pickups with active EQ into an active DI box results in instant clipping. The rule-of-thumb for guitars and basses is: active source - passive DI, passive source - active DI.

Good DI boxes use decent transformers that can deal with high levels of input without clipping and distorting, which is usually *not* what you find in your $150 audio interface. Or $800 audio interface. While interfaces' mic preamps are slowly becoming better, looking at the higher-voltage ones in Audient, Aphex, SPL and Lexicon models, the instrument input still seems to be an "added because there's a market" extra and not a feature that would get a lot of love from manufacturers. Or why does no available interface have a latency-free parallel output for the instrument input, which would make an additional DI box superfluous?

If the interface's mic preamps are OK, then I as a (self-) recording guitarist would 10 out of 10 times prefer a good DI box that goes to a mic preamp over any interface's instrument input. The instrument input on the Audient iD22 was pretty good tho, shame that it kept crashing OSX.

The ideal DI solution however would be to also circumvent the interface's preamps, this lets you send a line signal directly into the A/D converters of the interface without any further preamp degradation. That is, if your interface allows it - which I learned is not the case with Focusrite. Plus, it would require a decent external DI preamp. You'd connect the guitar to an external preamp that has an instrument input and line output, like a Focusrite ISA One or Chandler Germanium (type of thing), and go from the external preamp's output directly into the interface's A/D with a line cord. Best solution like evur, fur realz.

I currently use a Lehle P-Split as a first step to isolate 2 separate guitar Hi-Z signals.
One goes to the amp so I have something to chew on while recording.
The other goes into my trusty old active Monarch DIB-200 (because passive SD Pegasus/Sentient pickups) and from there on into my interface's mic preamp, which used to be the Audient but now it's the M-Audio M-Track Eight. It has decent (enough) mic preamps for my current needs, and the A/D values are at a very high level according to the spec sheet. I can't remember ever having had such clean and crisp DI tracks, a dream to reamp with. Reamp as in "real amp", but the noise in my high-gain amp sim of choice (Grind Machine) is also pretty manageable, nothing a lightly applied gate couldn't contain.

Splitting the signal in front of the DI box may seem redundant (or plain unnecessary), as the DI box offers that functionality itself, but using the P-Split instead of the Monarch's parallel out is still somewhat quieter, which means less hiss in the monitoring chain.

When I have the cash to spare, I'll add a quality DI preamp to my 500 series rack. Ditch the Monarch DI and just split the guitar signal, run one into the amp plus send the other one through the 500 preamp directly into the M-Track Eight's A/D. That should get a pretty high quality DI signal into the box and keep me happy for another couple of months.

[Compyfox]

In case of active PU into active DI's - most of them (these days) have PAD's - why not use that feature?

Saves you from having a lot of spare DI's lying around.

[chokehold]

A pad is an additional component in the chain. Leads to signal degradation, in the case of pads it's usually a loss of high frequencies, makes things "dull". I'm speaking about my own experience now and I'm generalizing, of course. There may well be some active DI box with a pad that doesn't degrade the signal. But if there is one at all, I'm guessing it won't be in the affordable price sector.

[khanyz]

You see, I just try to offer an easy-to-understand viewpoint here, while trying to evade the sometimes often involved snake oil. Especially if you say "always use a dedicated DI", but don't go more in-depth as to why.

I said that an active DI can improve things not always use one. I even mentioned an example of an interface where you don't need one.

As for Active pickups, I consider those as Passive Pickups with matched preamp (rather than a black box), so the considerations are different. They can be line matched already so plugged directly in, or with a simple isolator to get rid of noise (ground hum). However, they may need to be padded for Mic inputs as they can be too hot.

[Compyfox]

A pad is an additional component in the chain. Leads to signal degradation, in the case of pads it's usually a loss of high frequencies, makes things "dull". I'm speaking about my own experience now and I'm generalizing, of course. There may well be some active DI box with a pad that doesn't degrade the signal. But if there is one at all, I'm guessing it won't be in the affordable price sector.

Is it a "perceived" effect or can you measure it if you level match the signals?
Not trying to imply anything here, just trying to offer a different viewpoint.

I said that an active DI can improve things not always use one. I even mentioned an example of an interface where you don't need one.

So it's down to personal preferences as well as technical limitations?
I mean, if people are happy with the Hi-Z input on a ADC, there is no need for a dedicated DI unless you really(!) want to address a hardware amp at the same time, no?

As for Active pickups, I consider those as Passive Pickups with matched preamp (rather than a black box), so the considerations are different. They can be line matched already so plugged directly in, or with a simple isolator to get rid of noise (ground hum). However, they may need to be padded for Mic inputs as they can be too hot.

Interesting... So it definitely is(!) a matter of preferences as well.

And weather or not you want to convert the signal from unbalanced (Guitar) to balanced (XLR, balanced line) or keep it as is (unbalanced line).

Electronics is like lego, you just need to learn the blocks, how they combine and then how to (neatly) snap them to get what you really want. I did it as a hobby when I was a kid then specialized in it for my physics degree. Knowing it has always been handy but can quickly piss off sales assistants .

Haha, I can definitely relate to that.

Thanks for the insight so far.

[khanyz]

Matching impedances/levels is the way to get the best out of the gear you have. Whether that has acceptable quality and level of noise is a personal preference and may need improving with additional stuff.

If you consider paying £40 for a good Active DI versus what you think they spend on the Hi-Z input of an audio interface then the sound improvement would be similar. Most of the audio interface design will probably (and I speculate here) be a line level circuit to ADC/DAC. The mic/line/instrument preamps will be added to this with the mic being the star, the line in/out a direct link no brainer and the instrument an additional feature (unless you are Line 6).

Balanced/Unbalanced is a different thing and doesn't determine the signal level or impedance. A balanced signal just has the same signal going through 2 wires, that have inverted phase. This negates some transmission noise by cancelling out magnetic field effects. It can also degrade the signal slightly, as noise is random so the 2 signals are not identical and will interfere with each other. This is negligible compared to the external noise interference it prevents over long runs.

[chokehold]

Is it a "perceived" effect or can you measure it if you level match the signals?
Not trying to imply anything here, just trying to offer a different viewpoint.

I know what you mean, ears can be tricky buggers.
I have only met it as a combination of both, the degradation being audible and measurable every single time. By "audible" I mean recording samples, lowering the volume of the un-padded signal and A/B-ing between the two, and by "measurable" I mean having a fine-tuned SPAN on the master. With a pad things always sounded less thrilling than without, but YMMV.

Obviously, with less volume the perception changes. But most importantly, it also lowers the dynamic range of the signal and thereby the sound-to-noise ratio. So when you switch-in a 30dB pad and later crank the gain to +25~30dB to make up for the volume the pad loses, you'll also raise the noise floor +25~30dB higher than to where it would have been without the pad. And that does significant harm to something that'll be run through a distortion module in the next moment.

[Compyfox]

Yes, but the main thing is still to lower the signal in order to prevent clipping (in the DI), while recording(!) the signal at a healthy level (ADC).

It's not like you connect a guitar to a DI, and then the DI to an ADC and boom - instant perfect signal strength, no?
So I do assume, that even if you use the 20dB pad, that you still check your signal strength. And in this case, I do assume that the noise level is low.

I mean... you want to prevent the active PU to overdrive any other input. Or am I misunderstanding this?


Else, glad that you understood my stab at A/Bing and perceived frequency response at different signal strength.

Balanced/Unbalanced is a different thing and doesn't determine the signal level or impedance. A balanced signal just has the same signal going through 2 wires, that have inverted phase. This negates some transmission noise by cancelling out magnetic field effects. It can also degrade the signal slightly, as noise is random so the 2 signals are not identical and will interfere with each other. This is negligible compared to the external noise interference it prevents over long runs.

No, impedance should not have influence.
But shouldn't it have influence on how much you can boost the port?

At least this is what I "heard" (and read in books - also my teachers constantly bugged me with: "always go balance, else - signal loss").
[/sarcasm]

[khanyz]

I think you might be confusing terminating impedance (what's expected at each end of a lead) and transmission impedance (what's generated by the lead and is negated by balancing, thus improving signal strength).

It's getting to the point where I'd need to go technical in order to explain. It's probably better to get this info from a proper source as I'm a lousy teacher (my balanced/unbalanced description is especially bad). I once tried to explain the multi-verse concept by describing a blue pint of Guinness being within the one I was drinking. It says a lot about the state we all were in that the others understood it.

[Compyfox]

You can still try without going too technical.

I mean, there is always the "Reactive Power" example, described with a glass of beer.

[khanyz]

OK, audio-electric signals described via the medium of booze.

If you had to drink whatever was given to you, but instead of a pint of beer you got a pint of vodka, it would be a mis-match and probably messy. But if you got vodka diluted to the strength of beer, you could deal with it better. This is impedance matching.

If you got your pint of beer but others were tipping shots into it, you would probably be better off with a cup with a top (and straw) so that it couldn't be messed with. This is shielding.

If you drink with the glass only vertically upwards, then you can't get all the beer out and you've lost some. If you can invert the glass, then you get more out. This is balancing.

I'll think of a proper way to concisely explain it.

[Compyfox]

I think I'd ask for too much if you could illustrate it.


But maybe the example with the "hose" in this particular article is already helping - at least in terms of impedance conversion: Low feeds High

[khanyz]

Well, luckily I was young and dumb before smartphones so there aren't any photos .

Rather than reinvent the wheel, here's a good article that explains impedance

http://www.soundonsound.com/sos/jan03/a ... rkshop.asp

and another which explains related things

https://www.soundonsound.com/sos/jan02/ ... aq0102.asp

and one on balanced signals

http://www.aviom.com/blog/balanced-vs-unbalanced/

I hope it helps. Now to work.