DIs sound awful - why?

[Steve Conner]

Yes, you are exactly right. It is called the "impulse response", a kind of sonic fingerprint that stores all of the information about how the device alters the sound passed through it. You can then apply the impulse to another signal to colour it in the same way as if it had passed through the actual device. Doing this is called "convolution".

It has lots more applications than just cabinet modelling. You can extract the impulse response of an acoustic instrument by micing it up and hitting it. Then you can have endless fun passing signals through acoustic guitars, cymbals and so on, using Sound Forge's convolution plugin. I've used this trick to make some really freaky vocal reverbs for songs.

Convolution reverb is heavily used now. When recording classical music, you can capture the impulse response of the hall, which as far as I know is done with a gadget not unlike a blank firing pistol. (It can also be done by blasting white noise through speakers and performing a correlation.) Then you can do overdubs in a studio later and they sound as if they're in the same hall.

The main limitation is that it doesn't work for nonlinear systems. You can't capture the fingerprint of your favourite tube, for instance. I've seen attempts to get round this, but the math looks horrific, and I doubt it works very well.

[olihunt]

Aaaahhh the impedance was the problem!! I have a button on my interface that switches the input from Line to Instrument. If I have it on instrument it clips even with the gain knob on zero, so I've always used line, I assumed wrong it was just a difference in gain. I just recorded playing quietly on higher strings so that it doesnt clip on both modes, line and instrument, and the instrument mode sounds waaaay better. What the hell am I supposed to do now if it always clips?? It was a highly recommended interface :| surely it's capable of recording guitar... And sorry for the confusion about DIs, I'm used to reffering to the recording of the guitar signal as a DI and calling the device a DI box. Going to email Saffire about this, hopefully they can sort it. I have a tuner pedal but it's the korg one and it's true bypass so it does nothing

http://dl.dropbox.com/u/3059614/instrumentvsline.wav

And here is the original DI from nerve end, followed by me playing through the instrument input, with it clipping, followed by me playing through the line input. Adjusted to have similar volumes.
http://dl.dropbox.com/u/3059614/nerv...mentvsline.wav
Despite it clipping you can hear it is waaay better, moreso than in the last clip.

So as I understand it I have to buy a DI box?? What would you recommend as a cheapo one as I just spent all my money on getting my guitar set up

edit: why did my post end up here? I must have clicked the wrong reply button

[g1]

Glad to hear it worked out (sort of). Sorry if I sounded harsh or like "grammar police" lol. I just though it would be better if we all got our terminology sorted. At least I learned there is another meaning to "DI", if anyone cares to tell me what it stands for in the other context, I'm all ears.
In the focusrite info they mention that the instrument inputs allow you to go straight in with "no need for a separate DI" (box ) so I suppose if you want to use the line in you would need one. Is there no attenuator or pad for the gain? Can you back off your guitar volume so it doesn't clip the input? Your pickups must be pretty hot, or maybe still set too high? I would think that an instrument level input should be able to handle a guitar pickup without clipping, especially with the gain at it's lowest setting.

[olihunt]

Clipping to fuck - YouTube
That is me pushing every button and turning every knob on the interface. Only time it doesn't clip is if you turn it to line mode. Now to send the video to Focusrite... I've tried with turning down the guitar volume knob and it sounds good doesn't that affect the tone though? It's sure as hell better than anything I've had though. Thank you guys so much for helping me with this, I've been searching for the answer to this problem for over a year! First I found out my 6 string's tone pot had died. I replace that. Then later I find out it had a grounding problem and it's pickups were too low when I took it to a guitar tech and he sorted that out. Then I found out my 7 string pickups were too high, then too low, and then I got it set up so it's right. And now this :P This looks to be the last problem as the DIs sound good now I had tried everything before, from using different lengths and brands of cables to connect to the interface, different strings with different gauges, and had almost given hope until I asked you guys <3

Now... should I get a cheap DI box or should I save up for something like a countryman 85 or something? I don't need anything godly, just at least on par with the instrument input on my interface.

[dai h.]

If turning down the guitar volume a bit solves the problem, then I think all you need is a "pad" (just a voltage divider made up of resistors to reduce the voltage output from the guitar just enough so that it doesn't clip and you give the ugly sounding distortion). You could of course just use the guitar volume, but possibly this could be a problem (or an annoyance) if the ideal knob position is finicky and/or if you "ride" the volume knob (that is, adjust it as you play for expressive dynamics as you change from distortion to clean). I'm no expert, so I'll defer to the other guys but (vaguely) the pad (a fixed value one) would probably be two high value resistors with the output taken from the middle point (so a -6dB, or 50% signal voltage reducing pad). A couple jacks in a metal box. (The source of the problem I'm thinking is the bass strings putting out a relatively high signal and distorting the input (the way the circuitry is configured probably doesn't lend itself to sympathetic clipping/distortion like a nice guitar amp so the signal needs to be kept from clipping with the type of circuitry). Also, fixing this problem I would guess would help if you plugged in a bass as well.)

[J M Fahey]

Being a dyed in the wool minimalist, I suggest you build this instead.
It assumes 1M Interface input impedance.


Ooops! Simulposting !!

[dai h.]

hey there you go! With a nice bit of adjustability as well (and high frequency loss compensation?).

[olihunt]

Oh cool! Are there more newbie friendly instructions anywhere / somewhere I can look up what the symbols mean? I did physics for a-level but can't remember half of it xD Which would give me better sound quality, using the Saffire's instrument input with the pad, or using the line with a cheap DI box (around Ģ40)?

[J M Fahey]

Have a gifted friend build one for you.
Be generous with beer and pizza , or introduce him to that hot redhead you know and heīs been drooling over.
Canīt miss.

[dai h.]

Which would give me better sound quality, using the Saffire's instrument input with the pad, or using the line with a cheap DI box (around Ģ40)?

(Sorry if this wasn't made clear, but) the DI box won't be necessary now, since:

1) the Focusrite Saffire has a high impedance input already (this wasn't clear initially), and
2) through your experimentation, you seem to have proved where the problem lies (i.e. too much signal clipping the input amp)

As for the pad I think what JM suggests is definitely one to consider (i.e. find someone capable to build one for you). Perhaps find a piece of used gear (maybe cheap broken) that already has a couple of jacks (or holes for) plus switch so you have less work to do (drilling holes and such). Another alternative is to build a fixed one inside a 1/4 inch plug (on the end that would plug in to the Saffire high impedance input). The resistors don't need to be big (that is, large wattage ones). (If you do put the pad inside a plug, you probably want to mark it so you don't confuse it with other unaltered cables.)

[David Schwab]

So as I understand it I have to buy a DI box?? What would you recommend as a cheapo one as I just spent all my money on getting my guitar set up

What happens when you turn the gain knob down on the front? Does it still clip? I'm looking at the interface here:

http://www.focusrite.com/products/au..._6_usb/images/

And besides the gain knobs, I see a pad button as well as an instrument button. The pad button should stop the clipping if turning the gain down doesn't.

You can try running the guitar through a compressor pedal. That will allow you to turn down the output a little and keep the transients from clipping the interface.

[dai h.]

FWIW he said it was a "white faced" one (1st post), so I looked for it, and a "pro 10" plus another in a brochure appeared not to have a pad (maybe they had complaints and put them in later units--like the one above?).

(brochure showing a smaller standup type) :

http://www.focusrite.com/download/660/Saffire.pdf

(the source page for the brochure) :

Downloads the latest software and user guides for your Focusrite Product Saffire

the large (rackmount) white faced "pro 10" :

Downloads the latest software and user guides for your Focusrite Product Saffire Pro 10 I/O

[dai h.]

quick thought/idea:

if for whatever reason, making a pad isn't doable, maybe try to find a passive volume pedal (hopefully working used cheap?), then use that in a fixed position to attenuate the 7-string guitar signal a bit. It needs to be something meant for guitars (ones meant for keyboards, etc. may load down the signal too much).

[David Schwab]

FWIW he said it was a "white faced" one (1st post), so I looked for it, and a "pro 10" plus another in a brochure appeared not to have a pad (maybe they had complaints and put them in later units--like the one above?).

He needs to turn down the gain knob. They are for microphones, which need a lot of boosting. For guitar it should be set pretty low.

[J M Fahey]

Hi David S
On post #17 he states that he already

If I have it on instrument it clips even with the gain knob on zero,

so he canīt turn it any lower.
On "Line" input is does not clip, but it causes the "ugly" sound he heas been complaining about, because of the much lower impedance.
I suspect the "Pad" added in the newer ones may be low impedance too, meant for active pickups or PODs or Line Outs and such.
So heīs stuck in the no manīs land of having a very high output pickup (more than the designers thought possible) *but* also high impedance.
Maybe even more so than usual because high output in a passive pickup may mean overwinding.
You already know what that does to inductance and impedance.
I suggested him build a simple fixed high impedance attenuator or a slightly more complex one with 2 switchable settings.