A chapter break is where one topic ends and the next begins. Splitting a long recording into chapters turns one unwieldy file into pieces that a listener, an editor or a client can use on their own.
The job is the same whether the recording is a two hour interview, a lecture, a live set, a meeting or a day of field notes: find the moment each part should start, cut there, check the pieces begin and end cleanly, and save them as separate files.
Gleam Audio does the job with its Split tool. It runs offline on macOS, Windows and Linux, and nothing is uploaded. Version 1.0.0 rebuilt the tool around what you get out, so the window now shades and numbers each part before you commit to anything. The steps below are from version 1.1.1.
What you need
- The long recording in a common format. Gleam Audio opens wav, mp3, m4a, flac, ogg, opus and many more.
- Gleam Audio installed. Everything it needs comes inside the app, so there is nothing else to install.
- An idea of where the boundaries are: a running order, an agenda, a setlist or your own notes. Without one, listen through once and note the times.
Clean and level the recording before you split it. Split first and the noise and the loudness have to be fixed on every part separately. The complete guide to cleaning up spoken audio puts that order together.
1. Add the recording and pick it
Drop the file on the window, or click Open Files…. Then pick it in the list.
That step matters, because Split works on one file at a time. It is not like Remove noise, Enhance and Convert, which take the whole list in one run. Pick the file you are splitting, mark it, and run.
2. Find the boundaries on the picture of the sound
Pick Split in the sidebar. The window shows the recording as a picture of the sound, which is where the quiet gaps between sections are easiest to see.
Click anywhere on the picture to move the playhead there and listen from that point. Then land on the boundary exactly:
- Space plays and pauses.
- The arrow keys move the playhead a second at a time, or a tenth of a second with Shift.
Click the picture first, or the keys do nothing.
3. Mark each split
With the playhead on the boundary, press Split here. The button sits right on the playhead. The S key does the same thing.
Work through the recording and mark every boundary the same way. Two things to know while you do it:
- The picture shades and numbers each part as you go, so you can see the shape of the result before you save anything.
- Drag a split by its round handle to move it. You do not have to land it perfectly the first time.
4. Check every part before you save
The list under the picture shows every part with its start, end and length. That list is the real check on the job.
- Read the lengths. A part that lasts four seconds is usually a slip, not a chapter.
- Play just one part. Listen to its first and last few seconds. The opening word should not be clipped, and the ending should not run into the next topic.
- Join a part with the next one when the boundary between them was in the wrong place.
Do this before you run, not after. Splitting again means running again and clearing up extra files.
5. Use Split evenly into when there are no boundaries
Not every recording has natural breaks. A live set, a language course, or a long recording going to a platform with a length limit often needs equal pieces and nothing clever.
Split evenly into cuts a long recording into equal parts. Give it the number of parts you need and the lengths come out the same. Play any part to check it before you save, the same as with marked splits.
6. Choose where the parts go and what they are called
Two settings in the bottom bar decide the output.
- Save to puts each part next to the original by default, or in a folder you pick. The folder is remembered, along with the last five you used.
- Name sets the pattern the parts are named from. The braces button adds tokens like
{name},{tool},{n}or{date}, and the bar shows an example as you type. Set it before the run so a chapter folder does not fill with names you have to fix by hand.
7. Run the split
Click Run at the top right. The parts are written to the folder you chose, and your original recording is never changed.
When the run ends, a bar under the header says what happened, with a button to open the folder. The Activity button, at the right of the bottom bar, lists everything you made since you opened the app, grouped by run, and any file still on disk can be played again or shown in Finder.
Then check the parts against the pre-publish checks before they go anywhere. Loudness, the noise floor and the format matter just as much on a five minute chapter as on a two hour file.
When splitting is the wrong tool
- Chapter markers inside one file. Some players and hosts read a chapter list from the metadata inside a single file. Split writes separate files, one per part. If the destination wants markers in one file, keep the source and add the markers in a tool that writes them.
- A recording with no natural breaks. A conversation that wanders is easier to leave whole and trim with Cut, which marks a range and keeps only what is between the two handles. That need not wait until the end: Cut works on one file at a time too.
- Noise and level problems. Splitting cleans nothing. Fix the noise and the loudness on the whole recording first, or you repeat the work on every part.
- A damaged source. A clipped or echo-heavy take stays that way in each of its parts. No split repairs it.
Takeaways
- Split works on one file at a time. Remove noise, Enhance and Convert take the whole list.
- Find each boundary on the picture of the sound, then press Split here or the S key.
- Drag a split by its round handle when it lands a moment late.
- The list under the picture shows every part with its start, end and length. Play one before you run.
- Split evenly into is the answer when the recording has no natural boundaries.
- Set the save folder and the name pattern in the bottom bar before the run, not after.