How I Make a Podcast

There are a lot of ways to make a podcast, but after producing hundreds of episodes of Software Defined Talk, here's what I've settled on — what works best for me, and maybe it'll work for you too.

1. Record

I record almost everything remotely, using Riverside. I've found it the most reliable of the virtual options out there — Zoom, Restream, and the rest. I keep it simple: the basic studio for recording, and I'll sometimes stream an episode to YouTube live as we record.

Once we're done, I export each person's track as raw audio only, with no processing applied, and download the transcript too — that comes in handy later for show notes and pulling quotes.

2. Edit

With the recording finished, I move into PodEditor and import every track — mine, my co-hosts', and any guests. Since everyone recorded in the same session, the tracks come in already synced.

The first thing I do is make sure every track is the same length. Riverside will often leave a second or two of extra time at the end of each person's individual track, so I find the shortest track and trim every other track to match it. It's usually a few seconds off a few tracks, never any actual content — but keeping every track the same length keeps the whole session tidy.

From there I link all the tracks together — so a cut, trim, or move on one carries across all of them — and start editing. Most of it is editing for content: cutting segments, tightening things up, getting rid of side conversations that don't need to be there. I add ads into the timeline as needed, and I add chapters as I go rather than leaving them for the end.

The intro, outro, and ad reads I use every episode live in Favorites, so I'm dragging them into the timeline rather than hunting for the file each time.

3. Mix

Before I send anything off, I'll run PodEditor's Auto Level over the session to get every track roughly balanced — a quick pass, not a final one.

I'm not an audio engineer — I have a working knowledge of limiters, EQ, and the rest — but for the real polish I rely on Auphonic almost entirely. It's quick, it's easy to use, and the episodes always come out sounding great.

You can send Auphonic a full stereo mix of the episode, or upload the individual tracks — uploading them separately gives Auphonic a little more control over how everything gets balanced. Either way, use high-quality source audio (WAV or FLAC, not MP3): you tell Auphonic separately exactly how to encode the distribution file — usually mono MP3 at 64 kbps — so source quality is the only thing actually at stake.

That's what PodEditor's Export Multitrack Stems is built for — it exports every track as its own file, along with the chapter markers and their timecodes, ready to upload straight into Auphonic for the final mix.

Once Auphonic is done, I'll sometimes make one or two small tweaks in Auphonic's own web editor — mostly just checking that a stinger or bumper didn't get leveled out along with everything else. Then I download the finished audio, and the episode's ready to publish.

4. Publish

For hosting, I've been using Fireside.fm for years — I upload the episode and show notes there and publish straight to the feed from their site. I've also had a good experience with Buzzsprout. Beyond that, my honest recommendation is just to find a host whose workflow you like, and learn it well — that matters more than which one you pick.

It's not the most complicated workflow, but it works for me. Have a better way of doing it? Email us and tell us what you're doing.