Since the beginning, audio engineers have struggled to overcome a particularly persistent enemy – and that enemy is noise. It happens in many places and comes in many forms, but we’ll explore some of the genius technology that has been developed over the last century to keep noise at bay for the sake of clean recordings.
Thanks to Neumann for sponsoring this video and supporting audio education.
Room Noise
Unless you’re recording in a completely silent environment, the sound you’re recording will be competing against noises in the room from the moment it leaves the sound source. People have gone to extreme lengths to isolate the floors, walls, and ceilings of their rooms from outside noise. Soundproofing a room is almost always an expensive endeavor, and people will go to great lengths to reduce the noise floor in their room by only a few more decibels.
Short of renovating your home for the sake of your home studio, you can also see an improvement by moving your microphone closer to the sound source. The sound will decrease in level the further away from the sound source you are, so moving the microphone closer means the signal will be louder relative to room noise.
There’s a common myth about room noise that a dynamic microphone is better for a noisy room compared to a condenser microphone. In reality, it just depends on the polar pattern and the distance from the microphone.
The polar pattern of a microphone describes its directionality. A cardioid microphone, like the Neumann TLM 102, will pick up sound best from the front and reject sound from the rear.
Use these two principles to your advantage when recording:
- Place the microphone as close to the sound source but as far away from the noise source as possible
- Point the microphone toward the sound source and away from noise sources (according to the polar pattern of your microphone).
Microphone Self Noise
When the sound does reach the microphone, it might be safe from acoustic room noise, but there will be plenty more opportunities for noise to infiltrate our signal along the path to the recording medium.
One potential place where a lot of noise can be picked up is within the microphone. A dynamic microphone will not have self-noise in the way that a condenser microphone will because dynamic microphones are passive while condenser microphones have active circuitry within them. But dynamic mics do still exhibit some inherent noise.
The self-noise specification is often shown as “Equivalent Noise Level”. Companies like Neumann use a soundproof container to measure this spec – I saw it in person when I visited the headquarters in Berlin. However, some companies simply measure the microphone without the capsule, which is less meaningful and often gives misleading results.
If you’re going to use a condenser microphone, be sure that its self-noise is as low as possible. While a cheap microphone can have a self-noise noise of around 17 dB-A or higher, a professional-quality microphone like the Neumann TLM 102 can have a self-noise of only 12 dB-A. Its sibling, the Neumann TLM 103, has a self-noise of only 7 dB-A, which is incredibly low!
Anything below about 15 dB-A Equivalent Noise Level is very good, as that level of noise is impossible to hear in the context of a mix. When the equivalent noise level starts to exceed about 20 dB-A, you’ll start to add considerable noise to your recordings.
Similar to room noise, it depends on the level of the sound source. The self-noise will be much more noticeable when recording a quiet source compared to a loud source.
No matter what microphone we’re using, we should still be aware that the signal level at this stage of the signal chain is extremely low. Any noise that manages to infiltrate our circuit at this point could be enough to ruin it completely, so we must take measures to protect this very low-level signal.
Electromagnetic Interference
Next, the signal is headed across the room (perhaps even to a different room) over an XLR microphone cable. It’s a treacherous path because the room is riddled with electromagnetic noise at a wide range of frequencies. Plus, coming into proximity to other electrical cables or electronic devices will introduce the potential for a signal in another circuit to superimpose itself into our circuit.
This problem is almost entirely solved by using a balanced connection with a high-quality cable. An XLR cable has two wires and a shield. The signal will be sent across the cable with opposite polarity on each of the two wires. This means the signal on the black wire would be negative when the signal on the red wire is positive.
On the other end of the cable is a microphone preamp with a differential input. This input only responds to signals that are different between the red and black wire. The two equal but opposite signals now sum together as differential mode gain.
Meanwhile, the noise gathered along the path from the microphone to the preamp is now completely canceled because the noise is the same on both wires and the differential input doesn’t respond to similar signals. This is called common mode rejection.
Even if we don’t get the additional 6dB from equal but opposite signals on the two wires, we will still get nearly total rejection of the common mode noise. It’s a cool technology that helps protect the signal from noise (even over long distances). Learn more about balanced audio in this post.
If you listened to the signal now, you wouldn’t hear anything. As I mentioned, the signal level up to this point in the signal chain is extremely low. It’s a miracle it’s made it this far.
But things are about to get a whole lot worse (from a noise perspective) because we’ve now reached the microphone preamp that is responsible for boosting this very weak signal up to a line-level signal that can be recorded.
Preamp Noise
Fortunately, the microphone preamp on the Neumann MT 48 interface has an incredibly low noise specification as well… -128 dBu-A. That’s important because when the microphone signal is boosted, so is all of the room noise and electrical noise that has been acquired thus far, and we don’t want to add even more.
This is where there might be a benefit to using a low-noise condenser microphone compared to a dynamic mic because the microphone sensitivity of a condenser is typically greater than that of a dynamic one, and that means less preamp gain will be needed to achieve the appropriate signal level.
But no matter what mic is being used, the 78 dB of gain provided by the Neumann MT 48 will be more than enough.
We used to need to run levels relatively high back in the analog tape days because tape hiss was one of the biggest sources of noise in the signal chain. But there was also the potential to overdrive the input of the tape machine or saturate the magnetic tape if signal levels were too high – so it was a balance of noise or distortion.
With digital recording, tape hiss has been replaced by the much less problematic quantization noise while tape saturation has been replaced by the much more problematic digital clipping…
Quantization Noise
Now that the signal has been preamplified to line level, the analog-to-digital converter within the Neumann MT 48 audio interface needs to quantize the analog signal to a digital signal.
In any fixed-point digital recording system, we generally want to avoid exceeding 0 dB Full Scale (or dBFS) on the input meter within the DAW. This will immediately clip the waveform and cause distortion. Unlike tape saturation, digital distortion is not pleasing. On the other hand, the noise floor in a digital system is much lower than that of even the best analog tape machine.
The theoretical dynamic range of a 16-bit fixed-point recording is 96 dB. So, if you’re working with a 16-bit recording setup, it’s still relatively important to make sure you’re fully utilizing that range. Otherwise, you may start to hear quantization noise as you process the signal and decrease its dynamic range throughout the mixing process.
In a 24-bit setup, you will have a theoretical dynamic range of 144 dB which gives you a lot more room to work.
It’s important to emphasize that this 144 dB of dynamic range is only theoretical. The dynamic range of the analog-to-digital converter will always fall well below that mark. The converters in the MT 48 stand out from the crowd here, with their dynamic range of 136 dB (which results in up to four times the resolution of some other high-quality converters). Remember… 3 dB equates to a doubling of power.
In any case, I’d recommend setting your preamp gain with the goal of getting the input level on the DAW meter as high as you can while maintaining about 6 to 12 dB of headroom before clipping.
Mixing
Throughout the mixing process, the signal may be quantized again and again (resulting in additional quantization noise). This isn’t a huge problem in most DAWs though, because the dynamic range is so vast.
However, the risk is slightly higher if you intend to send the signal out of the interface, through an outboard processor, and back into the interface. The exact risk will depend on your DA converter, your interface’s inputs and outputs, the outboard gear, and your AD converter.
You can minimize the impact by limiting the number of times you go out and back and by using proper gain staging. For example, going out to an EQ, back in, and out to a compressor, will lead to more noise than going out to the EQ, then to the compressor with an analog connection, and back into the converter. But again – today’s converters are extremely high-quality so don’t worry about this too much.
Do use proper gain staging though… You will maximize the signal-to-noise ratio by sending a strong signal in and out of the outboard gear rather than a weak signal. But be sure not to clip on the way back in.
Even if you’re mixing in the box with plugins, you could still see the noise floor increase over the course of a mix.
As an example, if you send the signal into a compressor, compress the peaks, and bring up the make-up gain, the dynamic range will be compressed (and the signal-to-noise ratio will be reduced).
You do what you have to do for the mix, but just be glad that you’ve taken the necessary steps to minimize the noise floor up to this point.
Playback
Whether you’re mixing the song or just listening to a published version, the signal will have a few additional chances to pick up some noise, but luckily it’s made it through the toughest part.
In order to listen to the audio file, it will need to be converted back to analog by a DA converter and then it will need to be amplified by either a headphone amplifier or a loudspeaker amplifier.
The MT 48 has great specs for playback, as well. The headphone outputs have a dynamic range of 122 dB-A in high power mode or 117 dB in low power mode. For speakers, there are a few options for setup:
- The signal could be converted to analog by the interface, sending to the analog input on an amplifier or powered studio monitor.
- The signal could be sent from the interface to the studio monitors through a digital connection.
The Neumann MT 48 supports SPDIF and AES67, and so do some Neumann studio monitors. This means the signal won’t need to undergo an unnecessary round of conversion and it also means the signal won’t need to travel over another long analog audio cable.
In wrapping up our discussion on noise reduction in audio engineering, it’s clear that innovation and attention to detail play vital roles. With support from leaders like Neumann and others, we’ve tackled challenges such as room noise, microphone self-noise, cable noise, and preamp noise with precision. As we aim for optimal sound quality, let’s continue to harness the power of technology and expertise to achieve our goals.
One of the most important skills for reducing noise in your signal chain is gain staging. Check out this post to learn how it can make your recordings sound more professional.