---
title: "Text to WAV on Mac: Lossless Export for Editing"
description: "Export text to WAV on a Mac for editing. See how sample rate, bit depth, and channels change file size, and when a lossless WAV master beats a compressed file."
canonical: "https://www.murmurtts.com/blog/text-to-wav-mac-lossless"
---
[Murmur](https://www.murmurtts.com/)/[Blog](https://www.murmurtts.com/blog)/Guide

Guide

# Text to WAV on Mac: Lossless Export for Editing

Export text to WAV on a Mac for editing. See how sample rate, bit depth, and channels change file size, and when a lossless WAV master beats a compressed file.

![](https://www.murmurtts.com/murmur-icon.webp)Murmur·November 3, 2026·11 min read[Buy $49](https://murmur-licenses.tarunyadav9761.workers.dev/checkout)[Download free for Apple Silicon (3-day Pro trial)](https://murmur-updates.tarunyadav9761.workers.dev/download/latest)Intel Mac? [Download here](https://murmur-updates.tarunyadav9761.workers.dev/download/intel).Not sure? Apple menu > About This Mac shows your chip.

Free plan with Kokoro voices, scripts up to 1,500 characters · 3-day Pro trial, then $49 once

On this page

What "lossless WAV" actually meansSample rate: match the source instead of inflating itBit depth: 16-bit PCM is the practical editing masterHow the settings change file sizeHow to export text to WAV on a MacWhen WAV is the wrong choice, and when Murmur is not the toolFrequently asked questions

[Back to the journal](https://www.murmurtts.com/blog)

To turn text into WAV on a Mac, generate the speech in a local text to speech app and export the result as WAV. A WAV file that holds uncompressed linear PCM is lossless: it stores the samples the engine produced instead of throwing detail away the way a compressed format does. For an editing master the three settings that matter are sample rate (how many snapshots per second), bit depth (how much detail each snapshot carries), and channel count (mono or stereo), because together they set both the fidelity and the file size you will drag into an editor.

This guide is about the export decision, not the writing. It covers what lossless means for a text to speech file, how to choose sample rate and bit depth, the exact size math for common settings, and the steps to export a clean WAV on macOS. It also says where WAV is the wrong choice so you do not ship a 90 MB file to a web page that needed a 3 MB one.

## What "lossless WAV" actually means

WAV is not an encoding, it is a container. The [Library of Congress format description for WAVE](https://www.loc.gov/preservation/digital/formats/fdd/fdd000001.shtml) calls it a wrapper that can hold an audio bitstream plus other data chunks, and notes that the common bitstream is LPCM (linear pulse code modulation). LPCM is the part that is lossless: every stored sample is a direct measurement of the waveform at that instant, so nothing is discarded on playback. When the container holds LPCM with the standard PCM format code, that is the lossless WAV most editors expect.

The same LoC page lists the file signature as RIFF at the start and WAVE at byte eight, which is why nearly every editor, video tool, and archive accepts a .wav file without complaint. The [McGill MMSP WAVE specification](https://www.mmsp.ece.mcgill.ca/Documents/AudioFormats/WAVE/WAVE.html) describes the classic layout in more detail: a 44-byte header for simple PCM, with fields for sample rate and bits per sample, followed by the sample data in little-endian order. That header is exactly what a local engine writes when it saves a mono PCM WAV.

"Lossless" describes how the samples are stored, not how many exist. If the speech engine produced mono audio at 24,000 samples per second, a lossless WAV at 24,000 Hz keeps that audio exactly. Exporting the same audio as a 48,000 Hz WAV does not invent new detail; it just stores roughly twice as many samples describing the same waveform. Understanding that one point prevents most sample-rate mistakes.

## Sample rate: match the source instead of inflating it

Sample rate is samples per second, written in hertz. Speech models commonly synthesize at 22,050 or 24,000 Hz, because that range captures the frequencies a human voice uses. Murmur writes its WAV files at the rate the model produced them, and its default export rate is 24,000 Hz mono. If you upsample a 24,000 Hz generation to 44,100 or 48,000 Hz, the file gets bigger and the numbers look more "pro," but the audible content does not improve. You are paying in megabytes for interpolation.

There are two honest reasons to change the rate anyway. First, a target platform may demand it: some video editors and course platforms standardize on 48,000 Hz, and a couple of delivery specs name 44,100 Hz. Second, you may be mixing the narration with music or effects that already live at 48,000 Hz, and keeping one project rate avoids repeated conversion. In those cases, resample once, deliberately, and keep the original file as the master. Do not resample a file three times because three tools each guessed a different rate.

- Voice narration, audiobooks, voice notes: keep the model's native rate (often 24,000 Hz) as the master.
- Video projects and many course platforms: convert once to 48,000 Hz if the editing spec requires it.
- Podcast delivery: 44,100 Hz is common, but check the host's current spec before converting.
- Music or sound-effect beds in the same timeline: match the project rate rather than the voice file's rate.

## Bit depth: 16-bit PCM is the practical editing master

Bit depth is how many bits describe each sample's amplitude. Standard PCM WAV supports 8-, 16-, 24-, and 32-bit samples; PCM data is stored as two's complement integers, and the bits-per-sample field in the header tells the player which to expect (per the McGill specification). More bits raise the theoretical dynamic range by roughly 6 dB per bit and lower the quantization noise floor, so 24-bit gives more headroom for heavy processing than 16-bit.

In practice, for text to speech of clean narration, 16-bit PCM is the right master. The synthesized signal is already a quiet, controlled range, there is no microphone noise to preserve, and 16-bit integers give about 96 dB of theoretical dynamic range, far more than spoken word needs. Murmur writes 16-bit PCM WAV by default, which is the same thing most editors, audiobook toolchains, and video apps expect. Reaching for 24-bit or 32-bit float only helps if your downstream chain applies large gain changes, aggressive EQ, or many stacked effects, and even then the gain is marginal for voice.

Channel count is the third lever and the easiest one to get wrong. Speech is almost always captured or synthesized as mono. Duplicating a mono file into stereo doubles the size and adds zero information. Keep narration mono through editing, and go to stereo only when you are placing the voice in a stereo mix with music and effects. If you want the wider picture of masters versus delivery files, the [TTS voiceover export guide](https://www.murmurtts.com/blog/wav-m4a-mp3-tts-voiceover-export) walks through the split between them.

## How the settings change file size

Uncompressed PCM size is arithmetic, not a mystery. Use this formula: size in bytes equals sample rate times (bit depth divided by 8) times channels times duration in seconds. The table below applies it to common export settings for a single mono voice file, in decimal megabytes (1 MB is 1,000,000 bytes). Your editor may report slightly different numbers because it counts the header and may use binary megabytes.

| Settings | Bytes per second | Size per minute | Roughly 10 minutes |
| --- | --- | --- | --- |
| 16-bit, 24,000 Hz, mono | 48,000 | 2.88 MB | 28.8 MB |
| 16-bit, 44,100 Hz, mono | 88,200 | 5.29 MB | 52.9 MB |
| 16-bit, 48,000 Hz, mono | 96,000 | 5.76 MB | 57.6 MB |
| 24-bit, 48,000 Hz, mono | 144,000 | 8.64 MB | 86.4 MB |
| 16-bit, 48,000 Hz, stereo | 192,000 | 11.52 MB | 115.2 MB |

Read the first and rows together. A 10-minute narration at the engine's native 24,000 Hz mono is about 28.8 MB, roughly a quarter of the same audio at 48,000 Hz stereo. Nothing about the speech is worse in the smaller file. The last row is four times the first for what is, for a single speaker, the same sound.

One ceiling is worth knowing. The classic RIFF WAV header uses 32-bit sizes, so a single file is capped just under 4 GB; the LoC entry explains that the WAVE_FORMAT_EXTENSIBLE variant and the RF64 extension exist to go beyond that. At 24,000 Hz mono 16-bit, 4 GB is around 24 hours of audio, so a normal narration, audiobook chapter, or long document will never hit it. Murmur's own exporter checks the limit and stops with a clear error rather than writing a broken file, which is the behavior you want from a master export.

## How to export text to WAV on a Mac

The workflow is short because the hard part happened in the engine. If you are still choosing an app or wiring a script, the [Mac text to speech workflow](https://www.murmurtts.com/blog/mac-text-to-speech-workflow) covers the local setup. Once you have a generation you like, the export itself is a few decisions.

1. Generate and listen. Play the full text before exporting; fixing a mispronounced word after export wastes the whole render.
2. Confirm the source rate. In Murmur the file is written at the model's rate (24,000 Hz by default), so decide now whether the destination needs a different one.
3. Choose WAV, not the compressed option, for the master. WAV is the editing copy; M4A or another compressed format is for delivery.
4. Leave channels mono for a single voice and leave bit depth at 16-bit unless you have a specific reason to go higher.
5. Export to a dedicated masters folder with a descriptive name, not the project root, so the lossless copy is never confused with the delivery copy.
6. Import the WAV into the editor and place it at the start of the timeline, then trim silence at the head and tail rather than re-rendering.
7. Keep the untouched WAV. Export the finished mix as a separate compressed file for distribution and archive both.

Murmur is a local voice studio for Apple Silicon and Intel Macs running macOS 15 or later. Its current export options are WAV and M4A, so the lossless master and the compressed delivery file come from the same project. It runs the models on your machine rather than uploading your text, which matters when the source text is a private document. If that is your use case, the guide on [offline text to speech for private documents](https://www.murmurtts.com/blog/offline-text-to-speech-mac-private-documents) is the closer match.

### Worked example: a five-minute narration

Say you paste a 750-word script, generate it, and get 5 minutes of speech. At 24,000 Hz mono 16-bit that is 48,000 bytes per second times 300 seconds, or about 14.4 MB. The same audio at 48,000 Hz stereo 16-bit is 192,000 times 300, about 57.6 MB, four times larger for identical speech. The 14.4 MB WAV opens instantly in any editor, leaves a small footprint in version control or a backup, and still carries every sample the engine produced. That is the whole case for keeping the native rate and mono.

Now say the destination is a video editor that insists on 48,000 Hz. Resample the 14.4 MB master once to 48,000 Hz mono, which lands near 28.8 MB, and edit from that. The doubled size is the real cost of the platform requirement, and you can see it before you commit. Do not also convert to stereo unless the mix needs it.

## When WAV is the wrong choice, and when Murmur is not the tool

WAV is a master and an editing format, not a distribution format. For a podcast feed, a course player, a web page, or a YouTube upload, upload the compressed result instead. A 10-minute WAV near 28.8 MB is four to five times the size of a reasonable compressed version with no audible gain for speech once it is encoded well. Export WAV, edit, then encode the final mix to your platform's format. Keep the WAV so a future revision starts from the clean copy.

Text to speech is also not the only reason to reach for a WAV. Reads of long manuscripts, documentation, and audiobook material benefit from the same lossless master, and if you are producing hours of narration, the [audiobook narration guide](https://www.murmurtts.com/blog/text-to-speech-for-audiobook-narration) covers the batch and chapter workflow that keeps the file sizes manageable.

Be clear about the boundaries. Murmur is a macOS app, so it is not the answer if you are on Windows, iOS, or Android, and it does not do live or real-time streaming speech. It is not a cloud service, so it will not help a team that needs a shared browser project with collaboration built in. If your work is a live voice agent or a Windows pipeline, a different tool fits, and it is better to know that before buying. For a Mac creator who wants a lossless local master in a few clicks, the WAV path above is the point.

## Frequently asked questions

## Sources

- [Library of Congress: WAVE Audio File Format (fdd000001)](https://www.loc.gov/preservation/digital/formats/fdd/fdd000001.shtml)Accessed 2026-10-14
- [McGill University MMSP Lab: WAVE or RIFF WAVE sound file specification](https://www.mmsp.ece.mcgill.ca/Documents/AudioFormats/WAVE/WAVE.html)Accessed 2026-10-14
- [Apple Developer: AVAudioFile](https://developer.apple.com/documentation/avfaudio/avaudiofile)Accessed 2026-10-14

## Export text to WAV on your Mac with Murmur

Murmur is a local voice studio for Apple Silicon and Intel Macs running macOS 15 or later, with WAV and M4A export from the same project. The website edition costs $49 one-time, with a free plan and a 3-day Pro trial, and a 7-day refund policy. Listen to the public samples before buying.

[Buy Murmur · $49](https://murmur-licenses.tarunyadav9761.workers.dev/checkout)[Download free for Apple Silicon (3-day Pro trial)](https://murmur-updates.tarunyadav9761.workers.dev/download/latest)Intel Mac? [Download here](https://murmur-updates.tarunyadav9761.workers.dev/download/intel).Not sure? Apple menu > About This Mac shows your chip.

macOS 15+ · Apple Silicon and Intel Macs · Windows beta · 7-day refund policy

Made for your Mac

## Less reading. More making.

[Buy $49](https://murmur-licenses.tarunyadav9761.workers.dev/checkout)[Download free for Apple Silicon (3-day Pro trial)](https://murmur-updates.tarunyadav9761.workers.dev/download/latest)Intel Mac? [Download here](https://murmur-updates.tarunyadav9761.workers.dev/download/intel).Not sure? Apple menu > About This Mac shows your chip.[Explore Murmur](https://www.murmurtts.com/#product)