What Is Dither in Audio? A Complete Guide for Mixing and Mastering
TL;DR: Dither is a very small amount of noise added to a digital audio file right before you reduce its bit depth — for example, when you bounce a 24-bit mix down to 16-bit for a CD. Without dither, that reduction creates a harsh, gritty distortion called quantization error, especially in quiet passages like reverb tails and fade-outs. Dither replaces that distortion with a much softer, more natural hiss you barely notice. You only need it when you are lowering bit depth, you only apply it once, and it always goes as the very last step, after every other process is finished.
What Is Dither, in Plain Terms?
Dither is a tiny, deliberate amount of noise added to an audio file at the exact moment you reduce its bit depth. That is the whole idea. Adding noise on purpose sounds strange, but that small amount of noise is exactly what keeps a bit depth reduction sounding clean instead of harsh.
Think of bit depth as the number of steps on a staircase. A higher bit depth, like 24-bit, gives you a huge number of very small steps, so the staircase feels almost like a smooth ramp. A lower bit depth, like 16-bit, has far fewer steps. When you convert from a high bit depth to a low one, every sample in your audio has to land on the nearest available step. That rounding is called quantization error, and left untreated, it creates a rough, digital-sounding distortion that is most noticeable in the quiet parts of a song. Dither breaks up that rounding pattern before it happens, turning a harsh, correlated distortion into a soft, random hiss that is far less noticeable and does not follow the music.
🎯 Free Download: Mixing & Mastering Cheat Sheet. Grab my free cheat sheet for quick reference points on bit depth, export settings, and mastering chain order. You will also get my free Complete Audio Mastering guide in the same download. Keep both next to your workstation. Download the Free Mixing & Mastering Cheat Sheet →
Why Does Quantization Error Happen?
Every digital audio sample is stored as a number, and bit depth decides how many possible values that number can take. At 24-bit, you have more than 16 million possible values for every single sample, so the sound is represented with extreme precision. At 16-bit, you still have plenty of values — 65,536 of them — but far fewer than 24-bit.
When you convert a file from a higher bit depth to a lower one, every sample value has to land on one of the lower bit depth’s available steps. Any value that does not land exactly on a step gets rounded to the closest one. On a loud, busy passage, this rounding is completely inaudible, because the music itself masks it. On a quiet passage — a reverb tail slowly fading out, a soft piano note, a vocal ad-lib trailing off — that rounding has nothing to hide behind. It shows up as a faint, gritty, almost crunchy quality instead of a smooth fade into silence.
The most skipped step in mastering, explained in one picture.How Dither Actually Fixes This
Dither adds a very low level of noise to the signal before the rounding happens. This might sound like it should make things worse, but the opposite is true. That small amount of noise randomizes the rounding pattern, so instead of a harsh, structured distortion your ear picks up as unwanted grit, you get an extremely quiet, random hiss. The human ear is very good at ignoring random noise, and very good at noticing structured, repeating distortion. Dither trades one for the other, and the trade is almost always worth it.
Some dither tools also offer noise shaping. This pushes the added noise toward the very high end of the frequency spectrum, an area your ears are naturally less sensitive to, which makes the noise even harder to perceive. Noise shaping is a nice bonus, but the basic idea of dither works well even without it.
Do You Actually Need Dither?
Here is the simple test. If you are reducing bit depth, you need dither. If you are not reducing bit depth, you do not.
- Bouncing a session from 32-bit float down to 24-bit or 16-bit: you need dither.
- Bouncing a 24-bit mix down to 16-bit for a CD or a platform that requires it: you need dither.
- Exporting a file at the same bit depth you have been working in the whole time: you do not need dither.
- Bouncing stems or a rough mix that will go through further processing later: you do not need dither yet. Wait until the final bit depth reduction.
That last point matters more than most guides mention. Dither should only ever be applied once, at the very last bit depth reduction in the entire process. If you dither a file and then reduce its bit depth again later, you are adding noise on top of noise for no benefit.
Dither in Mixing vs Dither in Mastering
During Mixing: Almost Never
Do not dither individual tracks, buses, or a rough mix bounce. During mixing, you are almost always working at 24-bit or higher, and you are going to process that audio further — whether through more mixing revisions or by sending it off for mastering. Adding dither at this stage does nothing useful and only introduces noise that gets carried through every step after it.
During Mastering: At the Very End, Once
Mastering is where dither actually belongs, and even then, only at the final step. The correct order is simple: finish every piece of mastering processing first — EQ, compression, saturation, limiting, all of it. Only then apply dither, immediately before the bit depth reduction that happens on your final bounce. I walk through the full signal chain, including exactly where dither sits relative to your limiter, in my guide on what order to put plugins in your mastering chain.
Dither and 32-Bit Float Recording
If you record in 32-bit float, dither becomes relevant later than you might expect, not earlier. Recording in 32-bit float does not mean you need to dither right away. It just means you are working with a huge amount of headroom while you record, mix, and master. Dither only comes into play at the very end, when you finally reduce that file down to a delivery bit depth like 24-bit or 16-bit. I cover the recording side of this format in more detail in my guide on 32-bit float recording.
Dither for Different Delivery Scenarios
Streaming Releases
Most streaming platforms today accept 24-bit files, so you often will not need to reduce bit depth at all for a streaming delivery, which means no dither is needed either. If a specific platform or distributor requires a 16-bit file, apply dither as the very last step before that final bounce.
CD Mastering
The CD standard is fixed at 16-bit. If you are mastering at 24-bit or higher, which almost everyone does, you will need to reduce to 16-bit for the CD master, and that means dither is essentially mandatory here. Skipping it on a CD master is one of the more common ways quiet passages end up sounding rougher than they should.
Vinyl Mastering
Vinyl cutting engineers almost always want a high-resolution file, 24-bit or higher, so you typically will not dither the file you send for a vinyl cut. If you are also preparing a separate 16-bit digital release from the same master, that is where dither applies. Apply it on that specific bounce, not on the file headed to the cutting engineer.
Stem Delivery
If you are sending stems to another mixing or mastering engineer, keep them at 24-bit or higher and skip dither entirely. Stems are meant for further processing, and dithering them early just adds unnecessary noise before the work is even finished.
How to Apply Dither Correctly
The process is simple once you know where it belongs in the chain.
- Finish all of your mastering processing first. EQ, compression, saturation, and limiting should all be completely done before dither ever comes into the picture.
- Confirm your final delivery bit depth. Know whether you are bouncing to 24-bit or 16-bit before you touch dither at all.
- If you are staying at the same bit depth you have been working in, skip dither entirely.
- If you are reducing bit depth, add a dither plugin as the very last item in your signal chain, after the limiter.
- Bounce your file once. Do not reopen that bounced file and reduce its bit depth again later, since that would mean dithering twice.
Common Dither Mistakes
- Dithering a rough mix or individual tracks that still have more processing ahead of them.
- Applying dither before your limiter instead of after it, which defeats the purpose — the limiter changes the signal after dither was already applied.
- Dithering the same file more than once across multiple bounces, which just stacks unnecessary noise.
- Skipping dither entirely on a CD master, then wondering why quiet passages sound slightly rough or gritty.
- Adding dither when you are not actually reducing bit depth at all, which adds noise for no reason.
None of these mistakes are dramatic on their own. But they are an easy way to leave a small amount of avoidable distortion, or unnecessary noise, sitting in an otherwise clean master. If you want to see where this fits inside the bigger picture of getting a mix ready for mastering, my guide on how to prepare your mix for mastering covers the full export checklist.
Key Takeaways
- Dither is a small, deliberate amount of noise added right before you reduce bit depth, and it replaces harsh quantization distortion with a soft, unnoticeable hiss.
- You only need dither when you are lowering bit depth. If the bit depth stays the same, skip it.
- Apply dither once, as the very last step, after all other mastering processing including the limiter.
- Do not dither mixes, stems, or files that still have more processing ahead of them.
- CD masters almost always need dither. Streaming and vinyl often do not, depending on the final delivery format.
Related Articles
- Sample Rate and Bit Depth Explained. The foundation dither builds on, explained in plain terms.
- 32-Bit Float Recording. Why dither matters later than you might think if you record in this format.
- How to Use a Limiter in Mastering. Where your limiter sits in the chain relative to dither.
- What Order to Put Plugins in Mastering Chain. The full signal chain, including exactly where dither belongs.
- How to Prepare Your Mix for Mastering. The complete export checklist for handing off a mix.
Ready to Go Deeper?
Dither is a small step, but getting the final bounce wrong can leave avoidable distortion in an otherwise great master. If you want my complete mastering workflow, including exactly where dither, limiting, and every other stage of processing belongs, the Pro Audio Mastering Course walks you through it. It covers the entire chain I use on every release, from the first EQ move to the final export.
Learn more about the Pro Audio Mastering Course →
📥 Your Free Resources: Do not forget to download your free Mixing & Mastering Cheat Sheet. The same download also includes my free Complete Audio Mastering guide, covering the export settings, bit depth decisions, and mastering chain order this post builds on. Download both here →
About the Author
Adrian Milea is a Mixing and Mastering Engineer, Music Producer, and Educator with over 25 years of experience. He has releases on Armada Music, features in DJ Mag, and teaches producers worldwide through his courses at adrianmilea.com.
Until next time, Stay Cool and Make Great Music,
Adrian Milea
Last Updated: July, 2026
Time needed: 5 minutes
How to Apply Dither Correctly
- Finish all mastering processing.
Complete EQ, compression, saturation, and limiting before dither enters the chain at all.
- Confirm your final bit depth.
Know whether your final bounce is going to 24-bit or 16-bit before touching dither.
- Skip dither if bit depth is unchanged.
If your export stays at the same bit depth you worked in, dither is not needed
- Add dither as the last step.
If you are reducing bit depth, place a dither plugin as the final item in the chain, after the limiter.
- Bounce once.
Export the file a single time. Do not reduce its bit depth again later, which would mean dithering twice.
Dither is a small, deliberate amount of noise added to a digital audio file right before its bit depth is reduced, for example when bouncing a 24-bit mix down to 16-bit. It prevents a harsh distortion called quantization error and replaces it with a soft, mostly unnoticeable hiss.
No. You only need dither when you are reducing the bit depth of a file. If your export stays at the same bit depth you have been working in, dither is not necessary.
No. Mixes and stems sent for further processing should stay at 24-bit or higher without dither. Dither belongs at the very final bit depth reduction, not before further work is going to happen on the file.
Dither goes at the very end of the mastering chain, after every other process including the limiter, and immediately before the final bit depth reduction on your last bounce.
Often not. Many streaming platforms accept 24-bit files, so no bit depth reduction happens at all. If a platform specifically requires a 16-bit file, apply dither as the final step before that bounce.
Not really. Recording in 32-bit float just gives you more headroom while working. Dither still only matters at the final bit depth reduction, whenever you eventually bounce down to 24-bit or 16-bit for delivery.
Yes. Dither should be applied once, at the last bit depth reduction. Dithering an already-dithered file adds unnecessary noise on top of noise that is already there, with no benefit.
Almost always. The CD standard is fixed at 16-bit, and most mastering happens at 24-bit or higher, so a bit depth reduction is required, and dither should be applied at that final step.