Audio Converters93 Product

Audio converters (AD/DA) are a key part of any recording setup: they turn analogue signals into digital data – and back again. Here you’ll learn how converters, an Audio Interface and mic preamps differ, and which specs actually matter in real-world recording.

AD/DA Converter, Audio Interface & Preamps

All three device types are building blocks of the signal chain in audio recording – and in most cases they’re combined in one unit: the Audio Interface. With Audio Interfaces from premium manufacturers such as RME, Apogee, Lynx, Antelope or Universal Audio, the AD/DA conversion and preamps can meet serious studio expectations. And even in the more affordable bracket, modern interfaces often deliver better audio quality than much of what was considered good value 15 years ago.

For complex or particularly demanding (“audiophile”) setups, it can still make sense to use separate devices. In that case, the Audio Interface connects to your computer, the converter moves signals between the analogue and digital worlds, and the preamp handles boosting low-level sources (for example mic signals). The most common reason to buy separate AD/DA converters is channel count: only a few systems offer 32+ I/O as standard, and more than 8 inputs is still uncommon on many all-in-one units – which can be tight even for a small drum recording. While many devices can be expanded via digital connections such as ADAT, going beyond 16 channels that way isn’t always straightforward.

So for higher channel counts, you’ll often add extra devices alongside your Audio Interface. Another benefit of separate converters is the option to place them in the live room rather than the control room, keeping the more error-prone analogue cable run as short as possible.

What does an audio converter do?

To digitise an analogue signal, you need a converter. In most cases it’s built into an Audio Interface and doesn’t get much attention – even though the conversion process is a crucial step in the signal path. Today, AD/DA stages in entry-level interfaces from well-known manufacturers are often surprisingly good. For high-end productions, however, older or very budget converters can still act as a bottleneck.

Analogue to digital

Before an analogue signal can be translated into a digital stream of ones and zeros, it must meet certain requirements. First, the signal needs the correct level for the converter chip – handled by suitable gain stages. Next, everything above and below the convertible frequency range is filtered out to prevent unwanted artefacts that weren’t in the source (anti-aliasing). These purely analogue circuit stages can have a major impact on the sound and are designed with great care by reputable manufacturers.

Think of the converter itself as a computer chip. Only a handful of companies worldwide produce these (for example Burr Brown, ESS, AKM, Cirrus Logic), and you’ll often find chips from the same brand used across very different price tiers. The big price differences usually come down to everything around the chip: the converter is sensitive and needs the right conditions to perform at its best. A clean power supply, proper internal signal routing and, crucially, a stable clock are essential for accurate conversion. And the analogue conditioning stage must also be top-notch if you want a faithful digital representation. It’s a bit like photography: even the best camera struggles if the lighting flickers.

Sample rate (kHz)

During conversion, the sample rate in kilohertz (kHz) determines how often per second the signal is measured and “translated into digital”. Each measurement is a snapshot (sample) taken at a regular interval – hence “sample rate”. When converting back to analogue, the process happens again in reverse: an audio curve is reconstructed, which is then smoothed by filtering into a clean waveform. A useful rule: half the sample rate equals the highest representable frequency (Nyquist). At 48 kHz, that’s 24 kHz – already beyond normal hearing. Even though some high-end devices offer extreme rates up to 768 kHz, most productions work at 44.1 or 48 kHz for a practical balance of quality and data size. Higher rates like 88.2 or 96 kHz can be used for demanding work, but only really make sense with high-quality digital processing. As always: a higher number doesn’t automatically mean a better recording.

Bit depth

Each snapshot captured at the sample rate represents the source with a certain dynamic resolution. In simple terms: the greater the bit depth, the more volume steps can be stored digitally. With low bit depth and very quiet recording levels, you can hear digital noise. That’s why in the 16-bit days it was vital to record as hot as possible. With 24-bit now standard, you typically don’t need to worry: the resolution at lower levels is high enough that noise isn’t an issue. In modern DAW workflows, signals are often comfortably levelled around -18 dBFS for healthy headroom. Some converters support 32-bit, and many DAWs work internally at 32-bit float or even 64-bit float – so bit-depth anxiety is largely a thing of the past.

Special case: DSD audio

As an alternative to the common PCM format (for example 48 kHz/24-bit), high-end audio sometimes uses DSD (Direct Stream Digital). DSD works with extremely high sample rates in the MHz range (64x to 512x 44.1 kHz CD quality) but with only 1-bit dynamic resolution. The full audio information (frequency and level) is encoded as a digital waveform with a pulse width that varies relative to the original signal. In theory, it can be played back without a traditional DA stage, but in practice it relies on noise shaping and low-pass filtering to remove quantisation noise and ultrasonic content. You’ll see DSD on formats like SACD and in some audiophile releases.