Getting started

Signal flow: from microphone to mixdown

7 min read

Almost every confusing problem in a home studio is a signal flow problem. The sound is not magic and it is not lost — it is being converted, gained, delayed and converted again, and each step has a job. Once you can draw the chain in your head, half the troubleshooting is done.

The chain, end to end

A voice recording in a home studio travels like this:

1. Microphone — converts sound waves into an electrical signal. A small voltage, and one that is easily unbalanced and noisy.
2. Cable and preamp — the signal arrives on a mic-level, balanced line. The preamplifier multiplies it to a usable level, and adds gain the microphone needs but cannot produce itself.
3. Analogue to digital converter — samples the voltage many thousands of times a second and turns it into numbers. This is the interface's converter, and its quality sets the ceiling on your recording.
4. Computer — the DAW stores and processes the numbers. All processing happens here from this point on.
5. Digital to analogue converter — turns the numbers back into a signal your speakers can use.
6. Monitors or headphones — the last chance to hear what you actually recorded.

Each arrow has a name and a level. Learn the names and the levels and you can ask useful questions about any piece of equipment.

The three levels you will meet

• Microphone level — millivolts. The signal leaving a microphone is tiny, and it is the level that preamplifiers exist to amplify. It is also the level most easily corrupted by noise, because the signal is so small.
• Line level — the standard "normal" audio level. Fixed-output sources such as mixers, audio interfaces and CD players output line level. It is roughly 1000 times larger than a typical microphone signal.
• Instrument level — somewhere between the two. Electric guitars output a hot, unbalanced signal that behaves like a microphone but is much louder and much noisier. Guitar amplifiers expect it.

Plugging a microphone into a line input is a real mistake with an audible result: it will be quiet, thin and noisy. Plugging a line-level source into a mic input will overload the preamp. Every interface labels its inputs for a reason.

Gain staging: the single most useful habit

Gain staging means choosing sensible levels at every step, on purpose, instead of letting each device decide for you.

The rules are simple:

• Record peaks around -12 dBFS. That is about six dB of headroom below digital zero. It is high enough that you are not wasting the converter's resolution and low enough that a louder-than-expected performance cannot clip.
• Do not gain up afterwards. If the recording is quiet, the fix is a better recording or a better mic position, not a volume knob. Raising a quiet recording also raises its noise floor with it.
• Watch the output of the preamp, not the input. Some preamps have a gain knob and a separate output knob. The number that matters is the one feeding the converter.
• Do not clip. A clipped digital waveform has lost information permanently. There is no plugin that brings it back.

If a recording is too quiet, you have a gain problem, not a volume problem. If it is too loud and distorted, you have the same problem in the other direction.

What each device in the chain is really for

Ask these four questions about anything you are considering buying, and the marketing falls away:

1. What level does it accept?
2. What level does it output?
3. Does it need power, and does that power carry signal?
4. What does it do to the timing of the signal?

That last one catches people out. Anything that converts audio introduces a small delay, and anything digital adds more. This is latency, and it is the reason you cannot play along with a recorded track in one pass unless you set things up carefully.

Balanced versus unbalanced

A balanced cable carries the signal on two wires, reversed against each other, with a ground shield around both. Any noise picked up along the way hits both wires equally and cancels out. An unbalanced cable carries the signal on one wire with a shared ground, so noise is added directly to the signal.

The rule in a home studio: use balanced connections wherever they exist. Balanced XLR from microphone to preamp, balanced TRS between equipment, and USB or FireWire between the interface and the computer — digital connections are inherently free of analog noise.

The classic unbalanced trap is a guitar plugged into a consumer interface input designed for line level. That path hums and buzzes, and no amount of editing fixes it.

Phantom power

Condenser microphones need a small amount of power — usually 48 volts, called phantom power — delivered through the same XLR cable as the signal. Interfaces provide it from a switch or button.

• Check the input before you plug in. A microphone that needs phantom power connected to an input that does not provide it will simply not work.
• Dynamic microphones do not need it, and in most cases are not damaged by it, but ribbon microphones occasionally are. Check the manual of anything unusual.
• Never apply phantom power to a microphone with a detached XLR connector or a badly made cable. The phantom voltage can arc across the pins.

Latency in the chain

Digital stages add delay. The number you feel while recording is the round trip: your playing, converted to digital, processed, converted back, and played.

Rough guide to round-trip delay at 48 kHz:

This is why interfaces have a buffer size setting, and why a slow interface feels worse to play against than a cheap one with a good driver. More on this in choosing an audio interface.

Why the chain matters more than the parts

Two setups with the same microphone and the same computer can sound completely different, because one has a clean balanced path with sensible gain and the other has a headphone splitter, an unshielded cable and a preamp pushed too hard.

When something sounds wrong, follow the signal. Where does it first go wrong? That is where the fault is. Fix that stage and only that stage.

A quick self-test

You can verify your chain end to end in about ten minutes:

1. Record ten seconds of silence in the room. Note the noise floor.
2. Speak into the microphone and watch the meter. Peaks should land around -12 dBFS for a normal speaking voice.
3. Play the recording back through the monitors and through the headphones. They should sound broadly similar in level.
4. Touch the microphone body while recording. If you hear a thump, the mount is letting structure-borne sound through — a shock mount is the fix.
5. Move the microphone 30 centimeters away and record again. If the level drops by around 6 dB, the distance law is working and your gain staging is sane.

If any of those behave strangely, the problem is in the chain, not in the mix.