Behind the scenes of CZ Podcast

Plenty of CZ Podcast listeners are probably curious how we make the show and what technology we use. Here’s a look behind the scenes. If you’re planning to record a podcast, consider this a little recipe book.

I’m not sure whether Filemon came up with the idea of recording the podcast, but I’d say he did the most to get it going. Preparing to record goes roughly like this.

• One of the three of us, Roumen, Filemon or Dagi, suggests a topic, usually prompted directly by listeners. We stick to topics we have something to say about and that might interest a wider audience.
• Next we agree on when to record. That’s usually hard because each of us has plenty going on and Roumen and Filemon travel a lot. We often find ourselves in different countries and even different time zones.
• We agree on a detailed outline, or at least try to. Usually we send each other links on the subjects we plan to discuss so everyone has some context. We also arrange outside guests when needed.

We talk over Skype, setting up a conference call everyone joins. It’s an ordinary phone conversation over Skype. An hour-long podcast takes about an hour and a half to record. Just before recording we make final adjustments, decide exactly who will cover the news and the opening, and agree on how it will go.

Everyone recording is also responsible for making their own local recording with Audacity. Then we start, and everything has to come together. There’s hardly any “direction” during recording, so except on rare occasions we don’t need to make cuts. Afterward, everyone exports and shares their recording (300 MB), and processing begins.

We make the finished audio in two ways. The old way, which we used through episode eight, was to combine all the tracks, again in Audacity. The other is to use Pamela to record the Skype call directly.

Both have pros and cons. When you combine tracks you have to sync them by hand because delays make them drift out of alignment. Recording the entire call can pick up occasional lag. Someone’s speech sounds choppy. In my case that isn’t always a technical issue. I do it naturally.

Processing the audio is usually the most time-consuming part. We run the finished track through filters for noise removal, normalization, volume and so on, add the theme, and export the MP3. For example, I spent over four hours editing and polishing the ORM podcast. Filemon usually does the most work on this. He did it for all the other episodes.

The finished podcast is nearly ready to reach listeners. We listen all the way through and gather links to what we discussed. That goes into preparing the article for java.cz. We create it in the publishing system and upload the MP3 to Roumen’s site at the same time.

Downloads generate traffic in the tens of gigabytes. Over a thousand listeners downloaded episode seven, for example. At about 70 MB each, it’s easy to work out how much of your data limit it uses. There’s an amusing story about that. When CZ Podcast’s audio was still hosted on Java.cz and it went over the limit, every download cost the site operator ten crowns.

That’s about it. It may not seem like it, but from agreeing on a topic to publishing the article the whole process often takes several people more than four hours. If you have questions about what happens behind the scenes, ask in the comments. We’ll be happy to answer.

One note. I dug up a photo so you can put faces to the voices.