What Is Mastering? | Audio Mastering Explained
Audio mastering is the process of finalizing music by balancing art and science. It relies on empirical data collected through technical analysis, and on emotional insight steered by intuition and a deep connection to the music.
Generally speaking, mastering is composed of three discrete processes:
Quality control, fine-tuning, and deliverable assembly: how mastering works
Mastering is three things, in this order: quality control, fine-tuning, and deliverable assembly.
Quality control means listening to the mix critically, in an accurate room, for the first time. That last part matters. A mastering engineer has no history with the record. There is no memory of what the kick sounded like before the compression was added, no attachment to any of the decisions that accumulated over weeks of mixing. There is only what is actually there, heard fresh, in a room designed to tell the truth about it.
What shows up in that first listen is what the process is built around. Clicks, pops, distortion, phase issues, imbalances that were masked by room modes in the mix environment. These are not failures of the mix engineer. They are what happens when audio moves from one acoustic environment to a different one, particularly when the second environment is calibrated specifically to be accurate. Technical evaluation also includes true peak and inter-sample peak monitoring, dynamic range and crest factor analysis, spectral balance, stereo imaging, and comparison against reference material.
Fine-tuning is where the adjustments happen. These are not large moves. Mastering is not a place for dramatic changes. It is a place for small, precise decisions that improve clarity, balance, impact, and translation. Equalization to address tonal balance. Compression used not to control peaks but to add cohesion. A limiter at the end, set to an appropriate level for the music, not for a number on a meter.
Every adjustment is made with one question in mind: does this translate? Does it hold up on a different system, in a different room, at a different playback level? That is the standard.
Sometimes the right answer is nothing. A mix that is already balanced, already translating, already at the right level does not need to be touched. Knowing when to leave it alone is part of the job.
Deliverable assembly is the technical preparation for release. Different formats have different requirements, and they are not interchangeable. A master for streaming is not the same as a master for vinyl, CD, or high-resolution digital. Each format places specific demands on level, frequency content, sequencing, and file structure. This work includes properly formatted files, embedded metadata, ISRC codes, track markers, PQ information, sample rate conversion, dithering, and compliance verification. It is not creative work. It is critical work. A technically flawed deliverable can undo everything that came before it.
Streaming. The major platforms normalize loudness, but normalization is not a substitute for a master that translates. Each platform handles loudness differently, and a master optimized for one is not automatically optimized for all. The deliverable is a 24-bit WAV at the native sample rate, sequenced, trimmed, and verified against true peak and integrated loudness targets. For Apple Digital Masters, there are additional encoding requirements and a certification process that ensures the file holds up through AAC conversion at the consumer end.
Vinyl. Pre-mastering for vinyl is its own discipline. Low-frequency content that works in digital can cause a stylus to jump a groove or prevent a cut from being made at all. Sibilance that is manageable on streaming becomes a real problem on lacquer. Side length, level, and the balance between channels all factor into what is physically possible to cut. A vinyl pre-master is not the digital master with the level turned down. It is a separate pass, made with the cutting lathe in mind.
CD.The CD master is a 16-bit, 44.1kHz DDP 2.0 image file, and every detail matters: track markers, ISRC codes, PQ data, crossfades, gaps, and the precise sample-accurate placement of every edit. Dithering is applied at this stage, not before. A DDP delivered with an error in the PQ subcodes or a miscoded ISRC is a problem that surfaces at the replication plant, not at the mastering desk, and fixing it after manufacturing has begun is expensive.
Hi-res digital. High-resolution masters are delivered at the native sample rate of the session, typically 88.2kHz or higher, at 24-bit. The goal is preserving the full resolution of the recording for download platforms and archival use. What matters here is that no unnecessary conversion happens upstream. A hi-res master that was sample-rate converted and then converted back is not actually hi-res. The chain has to be clean from session to delivery.
Cassette. Mastering for cassette means working with a format that has real physical constraints: limited high-frequency response, elevated noise floor, and a bias characteristic that varies between tape stocks. A cassette master is typically louder and brighter than the digital master, with careful attention to high-frequency content that will compress on tape. It is not a novelty format to be treated as an afterthought. If someone is pressing cassettes, they care about the format, and the master should reflect that.
The format is not a detail. It is the destination. Understanding what each one demands, and preparing for it accordingly, is what separates a master from a file.
Mastering vs. automated tools: what human judgment does that software cannot
Automated and preset-based tools can adjust level and tonal balance. That is not nothing, but it is also not mastering. What they cannot do is listen. They cannot hear that the chorus feels smaller than it should, that the low end is masking the kick in a way that will translate badly on a phone speaker, or that the dynamic contrast in the bridge is actually the best thing about the record and should not be touched. They process. They do not evaluate.
Mastering involves judgment: about what the music is trying to do, what the listener is going to feel, and what the specific playback context demands. That judgment comes from experience with real rooms, real formats, and real records across real genres. A tool optimized on a dataset does not have a relationship with the music. It has a model. Those are not the same thing, and the difference is audible.
Loudness, dynamics, and LUFS: why the number is only part of the story
Streaming platforms normalize loudness. The number on the meter is only part of the story.
Normalization means a master that is slammed to the ceiling gets turned down, but it does not get its dynamics back. What the listener hears is a record that has been compressed to a uniform density and then attenuated. It does not sound loud. It sounds flat. Fatiguing. Like something is missing, even if they cannot name what.
A master that preserves dynamics, tonal balance, and headroom does not just survive normalization. It benefits from it. The contrast between quiet and loud is still there. The low end has room to breathe. The limiter was used as a tool, not a ceiling to press against. That is what carries across formats, across playback systems, and across the increasing variety of ways people listen to music.
The integrated LUFS number tells you where a master lands on average. It does not tell you whether it sounds good. Short-term loudness, crest factor, and true peak all contribute to how a master actually behaves on a given platform and a given system. Understanding what each one means, and how they interact, is part of the job.
Translation is the standard. Not the target.
For a deeper look at what actually happens in the room, see What Does a Mastering Engineer Actually Do.
Common questions
What does mastering do? Mastering is three things: quality control to identify and correct errors, fine-tuning to optimize the audio, and deliverable assembly to prepare the required files for distribution. Every part of the process serves one of those three functions.
What does a mastering engineer do? A mastering engineer listens to the mix critically in an accurate room, makes small precise adjustments to level, tone, and dynamics, and assembles the final deliverables for every format the release requires.
What is the difference between mixing and mastering? Mixing combines and balances individual tracks within a session. Mastering takes the finished stereo mix and prepares it for release. They are separate disciplines, done in separate rooms, by engineers with different tools and different listening environments.
