Noise Reduction in a Stereo Microphone Interface

Shielding the electronics (not lead, use steel or aluminum) won't hurt but I would imagine that it has more to do with the circuit design of the amplifiers on the circuit. You would need to buy a device that is a higher quality recording device to fix this and that would be my recommended approach for a high quality recording.

Alternatively in the past I have used recording software that will do noise cancellation post recording. What you would do is record some of that hiss noise as a sample and then apply that as a cancellation filter to your recording. It was years ago and I don't know what software I used but I found that this works quite well. I found a review of audacity for noise cancellation and it looks similar to what I used and it is free.

If you want to mess with electronics you might be able to solder in a capacitor on the input to the A to D converter but this will filter out some of the highs also and may still not solve your problem.

What value of capacitor would be suitable? Would 0.1 microfarad do?

Can't sensibly suggest a value without knowing the circuit details.

Don't honestly think it would fix it anyway, but why not try it? If you have 0.1μF to hand, give it a go. If it makes no difference, try 1μF. If the 0.1μF kills the signal (or badly degrades it), try1nF.

Maybe it's just a little rhubarb from the audience?

Ok, here's one trick I learned from a Bob Pease article that you might be able to do. Without knowing the actual circuit topology, component types and component values I cannot make any specific component repair suggestions but this may speed up your troubleshooting. With the number crunching capabilities of todays computer, FFT analysis is an easy task. [For those of you who are not electrical engineers here, a FFT analysis takes a binary sized block of discrete time signals and transforms the data into a frequency versus amplitude or power plot.] Noise is actually a signal, but most people do not know how to read the information that a noise signal brings. Do an FFT analysis of the noise signal at a known collection of gain settings. If there are any fundamental frequencies present (20dB or 30dB above noise floor) from the analysis then what you have is technically not noise but interference. The interference maybe an aliased signal so don't go hunting only for that one frequency oscillation. Hiss is typically a true noise signal but an aliased high frequency signal can sound like hiss. When this happens a broad collection of distinct frequencies will appear at consistent amplitudes. True noise will have constantly changing peaks and frequencies.

Now if instead this appears to be true noise and you are using a mathematics analysis package to generate your FFT, then you can do fancy mathematics to discern things like peak hold, V/√Hz, crest factor, logarithm and other useful forms of number crunching. But in the name of Bob, I will forgo any mathematics. Also more than likely the closest thing most people have in their computer that does an FFT is some music track mixing software that has a spectrum display with little to no numeric results capability. Now look at your noise floor of the FFT and try changing the gain of the variety of different amplification stages. If the noise floor does change with the gain then the noise is happening before or at that gain selection amplifier. You want the most dominant noise source to come from the very first amplifier stage. {Don't forget your operating system's control of any and all analog path controls like your built in microphone of a laptop. Depending on how accurately you've turned ON/OFF your sound card settings, these signals can be inadvertently get mixed in with external data card's data stream.}

Once you've identified which stage the noise is coming from, you may have simply found that you've had too much gain in the wrong stage and not enough where you should. Typically this scenario happens when the first amplifier is set too low but the following amplifiers are set too high. The best signal to noise ratio typically happens when most if not all of the gain happens in the first amplifier stage.

If the noise is coming from the very first stage and you must work at the maximum gain for that stage to get a strong enough of a signal then trace your signal path in this stage for all of the resistors in this part of the circuitry. Identify the highest resistance value resistors. The white noise hiss of a resistor increases with resistance value. Try one at a time putting a capacitor across one resistor at a time (there should not be many) that has a value of 10≈8 *10^-6 ohm farads. In some cases this will make things noisier, remove the capacitor note that resistor should not be bypassed and move on. In most cases bypassing the resistor will make no difference, remove the capacitor and note that the capacitor made no difference. If you are lucky you'll find bypassing one resistor will reduce your noise to an acceptable level. Once you've done this with all of the signal path resistors of that amplifier, more than likely you'll find one bypassed resistance that improves the noise but not enough for your liking. Now go back to the resistors that made no difference and see if now that the noise from that single resistor is lower that bypassing them can now reduce the noise further.

Thankyou for that, I have used audacity noise removal but that makes the recording sound strange. I have also tried putting capacitors over the inputs but that makes no difference. I was thinking about buying a studio microphone but I am not sure if they suffer too much from noise or not. On this microphone I can sometimes hear the mains electricity supply through it aswell.

GA - Noise is almost always a frustrating challenge due to the many possible causes. It is good that there are tools to help diagnose the cause.

Historically, it is usually a persistent, intuitive effort to determine if it is an external noise source (shield) or a noise input (filter) or a degraded performing circuit (replace noisy components) that eventually solves the problem.

The preamps used for microphone in ALL computer hardware I've seen are simply pathetic. One can use line input with an external preamp for any acceptable results. And hate to brake bad news but Bob was killed a few days ago in a car accident with his beetle, returning (?) from Jim Williams memorial. S.M.

pssst: this is a USB input job.

Does that make it an exception? S.M.

It does to the extent that both the amp and A/D functions are remote from the PC, and so don't rely on any of the computer circuitry (soundcard or whatever) for audio processing.

USB power is just as noicy as any power rail in a computer. It's a design issue to control that. But that's not what I am specificaly talking about. It's another thing to work with ~750mv low impedance capable line signal and another thing 50-60 db weaker usaly auto gain controlled. high impedance signal like a mic. It takes whole different level approach to keep noice, distortion and interference levels low on a hi gain preamp. The main reason behind the rotten mainstream approach is that cheap condenser mics used for Skype etc are quite noisy sources to start with. But if you want anything better in computer audio it's common knowlege you must work at high level like line,aux,cd-in etc input, with a decent external preamp, or a (very) expencive sound card that addresses these problems from the beginning. It's not my personal view you know.(lol). S.M.

The hiss is still there when I record without the microphones plugged into it so it is not a microphone issue.
Hmmm, it ain't necessarilly so.
An open circuit input may hiss whereas a correctly matched circuit/microphone could well silence it when plugged in.
I'd look at the matching of mic to input. Try other mics (and ones of a different type) before you jump to a conclusion.
This is a classic case of trying to solve a problem without identifying the actual problem, not that I'm saying it's easy...
Spend most time identiying the prob', then often the actual solving of it is simple.
Del

If I was to try and shield the interface with metal, would the casing have to be connected to ground?

It should prob be connected to the case of the amplifier, which in turn will prob be connected to ground.
Del