Mastering an Audiobook for ACX in Audacity 4 Without Macros

The big one is Macros.

Audacity 4.0 shipped without them, which means the familiar Audiobook Mastering macro can’t currently be run in the same way it could in Audacity 3.x.

For narrators who had built their whole post-production process around recording, editing, running the macro and checking the result, that is more than a minor UI change.

It removes the automated part of the mastering chain.

That does not mean you suddenly need an entirely new workflow. The useful thing is understanding what the macro was actually doing, because once you understand the chain, you have a few options.

Audacity 4 status

Audacity 4.0 does not currently include Macros. They are planned to return in some form in a later release, but there is no release date I would rely on yet.

This article is written around the current Audacity 4.0.x situation and will be updated as that changes.

What actually changed for audiobook narrators?

For ACX work, the missing Macro system is the important part.

In Audacity 3.x, Audacity’s own Audiobook Mastering workflow could be imported as a macro and run as a repeatable processing chain.

That meant you did not have to manually load the same three processes every time you finished a chapter.

Audacity 4.0 also shipped without some other workflow features used by people handling lots of chapter files, including Export Multiple and Mix and Render.

For someone recording a five-minute voiceover, that may not matter much.

For an audiobook with 30, 40 or 50 chapter files, those little workflow changes add up quickly.

There is also a separate question around ACX Check.

If you rely on ACX Check, I would verify that independently before moving a live project onto a new Audacity version.

What the old Audiobook Mastering macro was actually doing

This part is worth understanding even if Macros come back tomorrow.

Audacity’s official 3.x audiobook workflow is not doing anything mysterious.

The macro runs three main steps:

1. Filter Curve EQ using the factory preset Low roll-off for speech
2. Loudness Normalization to -20 dB RMS
3. Limiter using Soft Limit with a -3.5 dB limit

You would then run ACX Check afterwards.

That is the basic mastering chain.

It is useful, but it is also important to understand what it is not doing.

It is not listening to your recording and deciding whether the voice is too boomy.

It is not removing mouth clicks.

It is not fixing harsh sibilance.

It is not reducing room echo.

It is not choosing how much noise reduction your recording needs.

And it definitely is not proofing the audiobook.

It is essentially a repeatable technical chain designed to get an already-edited, reasonably clean recording towards the required delivery range.

That distinction matters.

If your raw recording has a bad noise floor, wet mouth noise or obvious room problems, running a mastering macro harder is not really the fix.

Why those three steps are in that order

The first stage is the low-frequency roll-off.

Speech recordings often contain energy below the useful part of the voice: desk vibration, traffic rumble, handling noise, air conditioning and other low-frequency information.

Removing some of that can make the recording cleaner before you start controlling its overall level.

But I would not treat any EQ preset as universally correct.

A voice recorded very close to a microphone with a lot of proximity effect may need more low-end cleanup than a naturally thin voice recorded on a very different setup.

That is one limitation of any fixed macro. It applies the same decision every time.

Next comes RMS normalisation.

Audacity’s official chain targets -20 dB RMS, which sits inside ACX’s required range of -23 dB to -18 dB RMS.

This is where one common problem starts.

If your narration is too quiet and your room noise is already close to the limit, raising the RMS raises the background noise too.

The meter can improve in one area while getting worse in another.

I go into that properly in the guide to why ACX rejects audiobook audio.

The final stage is limiting.

That catches peaks after the overall level has been brought up.

And there is a reason the old Audacity chain uses -3.5 dB, rather than simply limiting exactly at ACX’s -3 dB ceiling.

Why limit to -3.5 dB instead of -3 dB?

Leaving a little headroom is sensible when the final delivery file is going to be encoded to MP3.

The peak value of the encoded MP3 does not always land at exactly the same point as the WAV you started with.

A WAV sitting right on your delivery ceiling can come back slightly higher after encoding.

That is why mastering right up against a limit is rarely my preferred approach.

You want some margin.

There is also a distinction between sample peak and true peak, which gets confusing very quickly if you are new to mastering.

ACX specifies a peak requirement, but it does not state that this is specifically a true-peak measurement.

If you want the technical distinction, I have a separate guide to RMS, LUFS and peak.

A real measurement here would be much more useful than pretending the difference is always the same.

It isn’t.

So how do you master for ACX in Audacity 4?

You currently have a few realistic choices.

Do the old chain manually

If the relevant processing tools are present in your current Audacity 4 build, you can recreate the old mastering chain yourself rather than triggering it through a Macro.

The broad workflow is still:

EQ → RMS normalisation → limiting → check

But I would not publish menu instructions until they have been checked on the actual current build.

The settings from Audacity’s official 3.x workflow are:

Filter Curve EQ

• Factory preset: Low roll-off for speech

Loudness Normalization

• Normalize RMS to: -20 dB

Limiter

• Type: Soft Limit
• Input Gain Left: 0.00
• Input Gain Right: 0.00
• Limit to: -3.50 dB
• Hold: 10.00
• Apply make-up gain: No

Those are the old official macro settings. They should not be assumed to map perfectly onto a newer effect implementation without checking.

Passing the numbers is only half the job

This is where audiobook mastering gets slightly annoying.

You can hit the RMS target.

You can control the peaks.

You can get the noise floor into range.

And still have a bad-sounding chapter.

A fixed macro only knows the settings you gave it.

It does not know that your computer fan got louder halfway through the session.

It does not know that one chapter was recorded two inches further from the microphone.

It does not know that your noise gate has chopped the room tone into blocks of digital silence.

And it does not know that the de-clicker you ran before mastering has started eating consonants.

That is why I would always check the finished chapter by ear as well as by meter.

A good mastering chain should disappear into the listening experience.

You should hear the narrator, not the processing.

If you are fighting noise floor specifically, read room tone and noise floor explained and how much noise reduction is too much for audiobook narration? rather than trying to solve it by turning the mastering chain up or down.

Fixed macros have another weakness

There is a broader lesson here that applies beyond Audacity 4.

Macros are convenient, but they are brittle.

They store a chain of effects and parameters.

If an effect changes, a default changes or the underlying processing changes between versions, the same macro can produce a different result.

Audacity users have run into problems with the audiobook macro after previous updates too.

That does not mean automation is bad.

It means automated processing still needs to be checked.

The mistake is assuming that because something worked on your last book, it must produce the same result after a software update.

Whenever you change a DAW version, plugin version or mastering chain, test it on one chapter before running an entire audiobook through it.

That ten-minute check is much cheaper than discovering a problem 40 chapters later.

Your options if the macro was your whole mastering workflow

You do not necessarily need to abandon Audacity.

You have a few choices.

Stay on Audacity 3.x for the moment

If your existing setup works and you are in the middle of a paid audiobook, there is nothing inherently wrong with keeping a stable production environment until your current project is finished.

I would be cautious about changing a working production setup halfway through a book purely because a new version is available.

Recreate the mastering chain manually

This gives you almost the same logic as the old macro, just without the one-click execution.

For a single chapter, that is not especially difficult.

For 40 chapters, it gets tedious.

Use another DAW or mastering workflow

Reaper and other DAWs can handle the same basic job.

The trade-off is setup time.

If you already know one of them, that may be completely fine.

If you do not, changing DAWs purely because Audacity changed its macro system may create more work than it saves.

Move the technical pass outside the DAW

If you mainly used the macro because you wanted a repeatable mastering step after editing, you can also separate those two jobs.

Do the performance editing and proofing in your DAW, then use an automated audiobook mastering tool for the final technical pass.

Optivox does that in the browser, so it is independent of the Audacity version you use. It also handles things the old macro did not, including noise cleanup, mouth-click reduction where possible, de-essing and optional breath reduction.

You still edit and proof the chapter yourself.

The point is just to move the repetitive technical processing out of the DAW.

What about Soundscaper?

There has also been discussion around Soundscaper, a web-based project connected to an Audacity team member, and whether it can import older Audacity macros.

That may eventually be another route for people who want to keep using their existing chains.

Until that is checked properly, I would not build a production workflow around it.

Check the finished file, not just the chain

Whatever route you use, the final check is the same.

Measure the finished chapter.

Listen to it.

Check that the narration sits inside the required technical range.

Check the opening and closing room tone.

Listen for clicks, plosives and abrupt edits.

Check that a loud sentence has not suddenly slammed into the limiter.

Check that quieter passages still feel natural.

And check an actual exported delivery file rather than assuming the WAV inside your DAW tells you everything about the MP3 you are going to submit.

If ACX Check works in your current Audacity setup, it can form part of that process.

ACX Audio Lab can also help check several technical properties of a file, but neither a plugin nor a meter replaces listening.

That is probably the biggest thing I would keep from the old Audacity workflow even if the software changes completely.

Automate the repetitive part. Check the result like a human.