Open a product page for new audio gear and you’ll usually meet a wall of numbers and units. When you’re starting out, it’s tempting to dismiss them as marketing.
fosi audio DS3 spec sheet
Those figures do have a purpose. They describe how faithfully the hardware reproduces sound and how much unwanted noise or distortion it adds. This chapter explains the common measurements, the principles behind them, and how I use them to narrow down a purchase.
Understanding DACs for Beginners: A 3-Part Series
- Part 1: What is an Audio DAC? The First Upgrade When Your Earphones Sound Dull
- Part 2: Understanding DAC Specs in 1 Minute: Bit Depth, SNR, and More
- Part 3: Which DAC Should I Buy? A Guide to Key ‘Chipsets’ for Your Sound
Bit Depth: Amplitude Resolution
Specifications such as 16-bit, 24-bit, and 32-bit describe the vertical resolution of the signal. Bit depth determines how finely amplitude can be divided into levels when sound is recorded and played back.
wood and fire studio. The difference between 8 bit and 16 bit – a much higher resolution
The orange lines in the example represent those amplitude levels. On the left, the 8-bit graph has large gaps and gives a coarse representation of the red analog curve. The 16-bit graph on the right has much closer spacing. More levels allow amplitude to be recorded with greater precision, reducing conversion error and bringing the digital representation closer to the original waveform.
Each additional 1-bit adds approximately 6.02 dB of dynamic range, the span between quiet and loud. CD-quality 16-bit audio offers about 98 dB, while 24-bit audio reaches a theoretical 146 dB. More precisely, that limit includes an additional 1.76 dB in the calculation.
The purpose of 24-bit audio isn’t simply to make the output louder. The more useful distinction is what happens to quantization noise: the tiny rounding errors produced when a continuous sound is divided into numerical levels. Greater depth pushes that error farther down beneath the signal. In other words, the extra resolution matters to the relationship between the wanted sound and the conversion noise, rather than serving as a straightforward volume control.
Imagine the noise affecting the bottommost step in the available amplitude levels. With a greater bit depth, that bottom step represents a smaller part of the total range, so the residual noise has less influence. This is another way to picture why depth is about fine resolution as well as overall scale.
In a 16-bit system, subtle instrumental details or room reverberation can be obscured by conversion noise. With 24-bit audio, that noise floor is lower, making the pauses between notes feel quieter.
What to Look For
Support for 24-bit audio is enough to start with. Much of today’s high-quality streaming and FLAC content uses 24-bit sources. A 32-bit-capable chipset provides additional processing headroom for incoming audio.
Sampling Rate: Measurements over Time
Sampling rates such as 44.1 kHz, 192 kHz, and 768 kHz describe the horizontal resolution: how often the analog waveform is measured each second. A higher rate places those measurements closer together in time, allowing a more finely spaced representation of the curve.
wood and fire studio. The sketch clearly shows the difference of a higher sample rate
The illustration shows those measurements along the horizontal time axis. In the upper 48 kHz graph, the blue sample lines are farther apart, and the depicted digital output looks stepped. The lower 96 kHz graph places samples closer together. Dividing a second into smaller intervals gives the representation more points along the original red waveform.
The Nyquist-Shannon sampling theorem says a waveform can theoretically be reconstructed up to half the sampling frequency. The CD standard of 44.1 kHz was chosen to cover the human hearing range up to 20 kHz.
Why, then, do manufacturers compete on rates such as 192 kHz and 768 kHz, far beyond the range we can hear?
One reason lies in the DAC’s digital low-pass filter. At lower sampling rates, the filter needs a steep cutoff just above 20 kHz. That can introduce phase changes or ringing at high frequencies. Higher rates allow a gentler filter slope, helping preserve a more natural texture at the top end.
What to Look For
A DAC supporting 192 kHz or above has ample technical headroom for reproducing the audible range cleanly.
Signal-to-Noise Ratio: Signal above the Noise
SNR compares the music signal with the noise generated by the device itself. Expressed in decibels, it is a useful measure of how clean an analog output is.
hollyland.
The red area at the bottom is the noise floor, including thermal noise from the random movement of electrons in powered components. It’s the background hiss you may hear when turning up an amplifier. The green line is the music signal. The gap between them represents SNR: a larger gap means the music stands farther above the noise.
What to Look For
- A direct smartphone connection typically offers 90–95 dB.
- For a dongle DAC, 115 dB or more is sufficiently quiet for outdoor use.
- For a desktop DAC, look for 120 dB or more for strong transparency.
Dynamic Range: From Quietest to Loudest
Dynamic range spans the quietest usable output above the noise floor and the loudest output before distortion. SNR compares signal and noise at a particular level; dynamic range describes the full span the device can handle.
A larger figure means more room for changes from a quiet solo to an orchestral crescendo. With 24-bit audio, the figure shows how closely the hardware approaches the theoretical 144 dB range.
Yamaha Music Blog. The difference between DNR and SNR
The diagram makes the distinction visible. Dynamic range is the entire vertical distance from the noise floor at the bottom to the “distortion ceiling” at the top, where the output starts to break up. The reference signal sits between those limits. SNR measures the gap up to that reference level, while dynamic range also includes the headroom above it. That full span is the difference between the quietest and loudest output the hardware can handle without distortion. Modern DACs often quote similar SNR and dynamic-range figures, but the measurement method still matters. Some chips can make a silent SNR test look better by muting automatically when no signal is present. AES17 measurements assess dynamic range with a signal flowing, making them a useful indication of what happens during actual playback.
What to Look For
A figure above 115–120 dB indicates enough range to convey large musical contrasts, from classical performances to film scores.
Total Harmonic Distortion (THD+N): Unwanted Changes to the Signal
Where SNR describes noise, THD+N measures unwanted distortion together with noise added as the signal passes through a device. It’s a percentage: closer to 0 is better.
audio precision.
This graph plots THD+N against output voltage. A lower trace means less distortion and noise are being added to the signal.
A pure sine wave passing through the DAC’s analog conversion stage or an OP-AMP can acquire extra harmonics because the hardware does not behave in a perfectly linear way. Those additions occur at multiples of the original frequency, such as the 2nd and 3rd, and weren’t present in the input signal. That matters because a guitar or violin already has a characteristic pattern of harmonics that contributes to its own tone. Adding another pattern through the electronics changes the texture and can make the sound less clear. THD+N combines that unwanted harmonic distortion with the device’s underlying noise. It is therefore an objective measure of how faithfully the output follows the original waveform without adding distortion and noise along the way.
What to Look For
Check the zeros after the decimal point. A rating below 0.001% suggests good preservation of the original signal. High-end equipment often quotes around 0.0001%, with three zeros before the final digit.
Amplifier Power and Impedance Matching
The amplifier stage increases the analog signal enough to move headphone diaphragms. That’s separate from conversion, although portable devices usually combine the functions. Two details are worth checking:
1. Output (Vrms and mW):
Vrms describes voltage; mW describes the power delivered into the headphone load. Most IEMs need only tens of mW, while headphones around 320 ohms can require hundreds of mW or more.
2. Impedance Matching (The 1/8th Rule):
Keep amplifier output impedance low. A common guideline places it below 1/8th of the headphone impedance to avoid phase distortion.
The details of drive capability and impedance matching deserve a discussion of their own. They belong to the amplifier side of the chain, even though a portable product often puts that stage beside the DAC in the same case. I’ll return to the relationship between output, headphones, and matching in a separate practical headphone-amplifier guide after completing this series.
Part 3: DAC Chipset Guide Preview
These measurements are the physical results of the DAC chipset and its implementation. In Part 3, I’ll compare the main chipset manufacturers and explain how their design approaches influence the finished sound.
