Hi all. Was wondering if anyone knows whether a simple resistor in the "earth" line of my cable would be enough to fix an impedance mismatch problem? If so, would it also make the audio signal very much quieter? I have a sound device that I've kitted out with a line-out. The output from the device is tame and clean, but when I sample this sound directly using the sound card, it's mixed it with white noise which isn't there at the source. I use an algorithm to remove this noise from the sample, and then EQ it, but it loses a lot of goodness this way.

Just a resistor is not enough to correct the impedance. Either an opamp or a transformer can be used as a buffer.

At what dB level is the noise exactly? The noise isn't there with any other recording? You haven't got a mixer to put inbetween?

If you listen with headphones you'll notice noise that can't be heard when using speakers. So it's not nescessarily an impedance problem.

Another option is to leave the noise there, it's part of the character. You could try to "gate" it away at silent parts.

What about going out of the Headphone out on a laptop into the Tape/aux input in a stereo reciever?

There is a mismatch but I don't think it messes with the sound.

Still, I would like to match them up.

It depends on the designs of the headphone output and the line input cirquits how well they react to a mismatch. Some inputs are very forgiving, and some headphone outputs can be used as line outputs without any issues.

Bertkoor, the noise is like pure white noise (I haven't measured dB level yet) in very short bursts. It's relatively quiet, but when multitracking it makes absolute silence impossible, and interferes quite badly with quiet parts. The odd thing is: this noise is only present when I plug the machine into the soundcard. Whether I then plug in some headphones or an amp into the soundcard makes no difference - the noise is still there. Plugging a headphone or amp directly into the sound device doesn't cause any noise to be mixed in with the sound. The sound card is OK, and there is no noise problem when I record other instruments. I could make it a part of the character of the sound, but it wouldn't be right because it's digital recording and needs a zero dB noise floor to sound clean and pro-like.

Uh no, you'd be very unhappy with a 0dB noise floor . You'd much prefer a -96 dB noise floor, I think. Have you read up on dB, dBu, dBspl, etc?

Impedance mismatch issues are pretty much a thing of the (old & transformer-coupled) past. Nowadays, the typical line outputs are very stable and have quite low effective impedance ,often on the order of 15 ohms or less, and line inputs present loads of 10000 ohms or greater.

Besides, impedance mismatch issues most often present themselves as simply level changes, not as the sudden appearance of noise.

What you're describing sounds very much like an interference or oscillation issue. Did you say you built the line output circuit yourself? ("kitted out"?). Can you tell us what opamp or output transistor? One of the most common issues with hand-built amps is that layout problems can result in the output amp oscillating at a very high frequency. The effect of this is a slight smearing of the audible sound, and it could appear as hiss on some inputs such as your soundcard. The oscillation is very easy to detect with an oscilloscope on the output device.

Another cause is some sort of interference being induced onto the signal. Again, it can affect some out/in combinations more than others. You can try lifting a audio ground or two, putting a 1 k resistor in series between the output hot and input hot, you could even try a 100 ohm resistor between the out put ground and the input ground...

Legal disclaimer: Please be careful with lifting AC grounds; poorly designed or malfunctioning equipment can possibly give you a shock. Get some advice or assistance if you're at all unsure about this. Trying some resistors in the signal path as described above is fairly safe and shouldn't hurt anything (or you)

Check for RF interference in the cable and in the device. Move it around the room a bit and see if that has an effect on the noise level.

I think kenn has put in a lot of good alternatives to check out.

Otherwise, it seems like some kind of ground loop problem almost.

Or the place where you pick you kitted out from, is badly matched. That the output signal from your sound device to kitted out input caused the problem, and it's amplified in your circuit there.

I would try breaking up the connection from sound device, and ground input of your kitted out, and see if that is silent.

I imagine you have a capacitor for signal from sound device, and the spot you pick signal might need to be picked somewhere else.

Or your kitted out, is to low impedance on input and put a load on your sound device.

Just a few more ideas into the pool.

Connect a wire from the device chassis ground,(metal case),and the other wire end to a computer case screw.

The soundcard input is coupled to the outside anolog world through a matching transformer.This circuit expects that the active input device is picking up system ground from some other place.
If the input device's impedence is very high,unlike a mic or mixer out and such,the card's input tranformer will not supply the ground.The ground is said to be "floating"...and the device can thus generate internal noise.(an open ungrounded op-amp pin on the circuit board will run away and create noise...sloppy design.)

If grounding the case doesn't work,get a microphone stepup plugin transformer from radio shack.
Something like this;
http://www.radioshack.com/product/index.jsp?produc tId=2062444&cp=2032058.2032231.2032280&allCount=33 7&fbn=Cable+type%2FAdapter&f=PAD%2FCable+Type%2FAd apter&fbc=1&parentPage=family

OR;

http://www.radioshack.com/product/index.jsp?produc tId=2062443&cp=2032058.2032231.2032280&allCount=33 7&fbn=Cable+type%2FAdapter&f=PAD%2FCable+Type%2FAd apter&fbc=1&parentPage=family

Just jumpering two of the XLR balanced connecter pins will make it unbalanced.

If your handy,radio shack will also sell a small circuit board transformer that does a similar thing for lots cheaper;
First 2 items;
http://www.radioshack.com/search/index.jsp?kw=tran sformers&f=Taxonomy%2FRSK%2F2032058&categoryId=203 2058&kwCatId=2032058

Doh! Yes I have. Trust me to think in simple terms and dB=final loudness level... You're right, -96dB noise floor i.e. zero noise is what I want.

Thanks to everyone for all the ideas here. A lot to digest all at once, and I don't desperately need this, um, "kitted out" to work right now - it's more an experiment than anything else. Reading your comments it's apparent there are several likely causes and no easy answers.