Signal to noise ratio (SNR) in audio is the difference in level between the sound you want to hear and the unwanted background noise in a system, measured in decibels (dB). A higher number means cleaner audio. A standard CD has an SNR around 96 dB, meaning the music sits roughly 96 decibels above the system’s inherent noise. That gap between “signal” and “noise” determines whether your recording sounds pristine or buried under hiss.
How SNR Is Calculated
At its core, SNR is a ratio of power. You take the power of the desired signal, divide it by the power of the noise, then convert that ratio into decibels using a logarithmic formula: 10 times the log base 10 of (signal power divided by noise power). The logarithmic scale is what gives you those familiar dB numbers. Because it’s logarithmic, every 10 dB increase represents a tenfold difference in power. A system with 80 dB SNR isn’t twice as clean as one with 40 dB; it’s ten thousand times cleaner in terms of raw power ratio.
You don’t need to do this math yourself. Every piece of audio gear lists its SNR in the spec sheet. The number tells you how far the signal rises above the equipment’s own self-generated noise when no music or voice is playing.
What Creates the Noise
The “noise” side of the ratio comes from the noise floor: the sum of all unwanted sound your equipment produces when nothing is plugged in or playing. This includes hiss from electronic components, hum from ground loops (a voltage mismatch between two connected devices), interference from power lines and fluorescent lights, and electromagnetic radiation from WiFi routers and phones. Every cable, preamp, mixer channel, and converter adds a small amount of noise to the chain.
The noise floor sets a hard limit on how quiet a useful signal can be. If your noise floor sits at, say, 30 dB SPL, any sound quieter than that gets swallowed. This is why a microphone with a low noise floor (some MEMS microphones reach as low as 29 dB SPL) can capture whisper-level detail that a noisier mic would miss entirely.
SNR vs. Dynamic Range
These two specs are related but not identical. Dynamic range is the gap between the loudest signal a system can handle without distortion and the quietest signal it can reproduce above its noise floor. SNR is the gap between a given signal level and the noise floor. The key difference: SNR equals dynamic range only when the signal is at its maximum possible level. For any softer sound, the SNR shrinks.
Consider a system with 80 dB of dynamic range. If the signal you’re recording sits 30 dB below the system’s maximum, your SNR for that signal is only 50 dB. The dynamic range predicts the best-case SNR. In practice, you rarely hit that ceiling because recording at maximum level leaves no safety margin before distortion.
SNR Across Formats
Different recording and playback formats have dramatically different SNR capabilities, and these numbers explain a lot about why formats sound the way they do.
- Vinyl records are roughly equivalent to about 11.5 bits of digital resolution. Their SNR is limited by surface noise, groove imperfections, and the mechanical nature of playback. You can hear this as the characteristic crackle and hiss between tracks.
- Analog master tape performs better than vinyl but still falls well short of digital. CD-quality audio beats analog master tape by at least 18 dB in signal-to-noise performance.
- CD (16-bit digital) offers a theoretical SNR of about 96 dB, enough to make system noise essentially inaudible during normal listening.
- 24-bit digital pushes the theoretical SNR to 146 dB. No real-world converter actually achieves this (electronic components introduce their own noise), but the extra bit depth gives engineers enormous headroom for recording and mixing.
The jump from analog to even first-generation CD players was so large that digital immediately outclassed every analog format on pure noise performance. That doesn’t mean vinyl sounds “worse” to every listener, but in measurable SNR terms, the gap is substantial.
What Good SNR Looks Like in Equipment
For microphones, SNR is calculated as the difference between a standard reference level (94 dB SPL, roughly the volume of a loud conversation) and the mic’s self-noise. A professional MEMS microphone typically specs around 65 dB SNR. Higher is better, but SNR alone doesn’t tell the whole story. Specs like dynamic range, distortion, and power supply rejection also shape overall quality.
On modern digital-to-analog converters and audio interfaces, you’ll often see a related spec called THD+N, or total harmonic distortion plus noise. This combines the noise floor with any distortion the device adds to the signal. It’s a more complete picture than SNR alone because it accounts for both types of unwanted sound. A device can have excellent SNR but still color the audio with harmonic distortion. When comparing gear, check both numbers.
How to Maximize SNR in Practice
The most effective technique audio engineers use is gain staging: setting the signal level at each point in the chain so it stays well above the noise floor without approaching the distortion ceiling. The goal is to keep your signal in the “sweet spot” at every stage, from microphone preamp through plug-in processing to final output.
A common starting point is calibrating each track in your session to around -18 dBFS (decibels relative to full scale) with all faders at their default position. Some engineers go slightly lower, around -20 dBFS, to leave extra headroom as effects and processing stack up. If you’re routing through analog hardware, you might calibrate the inputs of your summing mixer to -14 dBFS so the analog gear receives a healthy signal level.
Beyond gain staging, practical steps to lower your noise floor include:
- Shorten cable runs to reduce the antenna effect that picks up electromagnetic interference.
- Eliminate ground loops by powering connected equipment from the same circuit or using a ground lift adapter.
- Move sources of interference like phones, WiFi routers, and computer monitors away from microphones and audio cables.
- Choose balanced cables (XLR or TRS) for any run longer than a few feet, since balanced connections cancel out noise picked up along the cable.
Every decibel you shave off the noise floor is a decibel added to your SNR. In a recording chain with five or six devices in series, small improvements at each stage compound into a noticeably cleaner result. The quietest, most transparent recording isn’t about buying the single best piece of gear. It’s about making sure no single link in the chain drags the overall SNR down.