Podcasting With GarageBand
Published by James Brooks on January 6, 2020podcasting tips
I record, edit and produce all of my Happy Dev podcast using GarageBand, which is a free app from Apple. There are many other alternatives available, but I chose GarageBand for its balance between simplicity and features, which is perfect for me as Happy Dev isn't a technically complicated podcast.
Upfront warning, before Happy Dev, I had never even been a guest on a podcast, let alone produce one. I'm also no expert on audio or GarageBand, so if I have made mistakes or you have some advice on how I can improve things, let me know.
You can find more information on my setup under the Podcasting section of my uses page.
Happy Dev is made up of a few individual parts:
- The introduction theme
- Introduction narration
- Host (that's me!)
- Outro narration
- Outro theme
- Master Track
This is what a typical podcast looks like once it's finished:
Host & Guest Tracks
Separating the guest and host tracks not only allows me to individually cut out and re-position segments really easily, it also gives me the ability to overlap both sides of the conversation at once, which means that you can hear conversation and laughter from both sides. A bonus to this is that I can mute myself or the guest if I'm trying to make a more difficult edit or cut out background noise.
For myself, I use the Samson Q2U Podcasting bundle. During the interview with a guest, I record directly into GarageBand. There's no reason for this decision, other than once the interview is finished, I can get straight into editing. One problem with doing this is that when editing myself, I don't have the original track easily at hand. Note that in GarageBand, you can drag out any trimmed clip to reveal the original content.
I ask all guests to record their side of the conversation as this usually produces the highest quality recording and if our internet connection is unstable the recording doesn't capture drop outs.
Should I ever record an episode with two or more guests, I can give each additional guest their own track. At this point, I would rename Guest to the name of the person to make it easier to remember who is who. I use Host and Guest as default track names because it remains generic and I can duplicate the project without renaming a bunch of things, I also do this because should Happy Dev have a guest host (that's not at all confusing) nothing needs changing.
Out of all the processes, this is really the most time consuming. It requires listening to the conversation over and over and over again, but it's a very important process if you're producing an interview style podcast and want to add a bit of polish to it.
I start out by syncing the clips. If this isn't done, you'll hear the conversation take place at different times. One technique that makes this much easier to do is to start both sides recording and have both people clap at the same time, that will produce a peak in the track graph which you can use to visually line up against.
Above, you can see the "3, 2, 1" countdown, followed by a peak where I clap and then the conversation start.
The next step is to cut out the beginning seconds where we countdown the clap and get to it. From this point onwards, I listen to the tracks and cut out any areas of:
- Long silence. You want to be careful cutting out too much silence otherwise the conversation begins to sound unnatural.
- "Um's" and "Err's" where they sit at the front of a sentence. Cutting these out from a sentence is hard and makes the response sound choppy and again, unnatural. Sometimes though, it's good to have an "um" because it shows a period of thought. It's all down to the kind of style you want to create.
- Conversation between myself and the guest which doesn't form part of the interview. Usually this is me trying to word something a bit better or asking if the guest is okay to answer a particular question.
Although I write out a document for each guest with their introduction, notes and the questions I want to ask (in an order of sorts), sometimes the conversation naturally leads to the guest answering questions I haven't yet asked or in an order that really makes sense to the listener, or even further questions in response to their reply. When this happens, I will cut out the relevant parts from both the Host and Guest tracks and move them to a place where it flows. I find that editing like this should be kept to a minimum as I don't want to forge a conversation that didn't happen, but at the same time, I want the listeners to follow along through a story that's cohesive.
Occasionally, I will need to re-record myself asking a question. This is only the case when I have failed to properly ask the question during the interview. When this is needed, I create a duplicate Host track, align the playhead (the playhead is the line that follows the tracks as the sound plays) and re-record from there.
When editing multiple tracks in GarageBand, you can hold Shift and click the tracks you wish to modify, then ⌘T to trim the clips at the playhead. You can copy and paste multiple tracks too. As I trim down the tracks, I'm constantly moving them in time so that I can play a section or the whole thing and be able to hear it without massive gaps of nothing
For me, the hardest part of editing is deciding what to completely cut from the episode. I try to keep episodes to around 30 minutes, but I don't worry too much if it goes over. I also try to strike a balance of theme (mental health + software development) and conversation. These are all things I'm still working out myself.
Automation in GarageBand is essentally programming for sound. It allows you to control different aspects of the sound; volume, pan, echo, reverb etc automatically.
I exclusively use automation to control the volume of different tracks. For example, I want the riser clip to stay at a set volume, but then the beat to drop in volume, whilst the instrumental increases in volume and then fades out. If that sounds complicated, follow the yellow lines on the "Song Beat" and "Song Instrumental" tracks:
I try to automate as little as possible as it can be a bit cumbersome to keep in sync as you move tracks around. You can enable a Move Track Automation with Regions setting, but in my experience this can be a bit flakey, especially when moving a clip past another clip with automation.
I currently use Epidemic Sound for the intro and outro themes. I chose Epidemic as it's very easy to license the sounds and link your podcast feed to your account. If you look closely at the overview screenshot above, you'll see that I have a couple of tracks for themes:
- HD Theme Song Beat
- HD Theme Song Instrumental
- Outro HD Theme Song
Further, the "HD Theme Song Beat" track is actually made up of two clips. I combined a riser (a riser is a sound which rises in pitch) and an upbeat song to create the introduction. I split the beat and instrumental into two sounds because I wanted to be able to fade out one before the other to create an atmosphere in the sound.
I then record the intro and outro narration separately, which allows for level adjustment to make sure you can hear me. I try to time the narration with the tracks, so that I don't talk too much.
Mastering is the process of making subtle changes to the audio to balance the mix to make all of the elements sound cohesive. GarageBand provides a Master Track which you can balance your mix on. You should imagine that the Master Track is a global track that sits above all other tracks and sets the final volume etc.
After I released the first episode of Happy Dev, my friends gave it a listen and had some great feedback and ideas for me. I also received some audio help from Joe, one of my best mates who is a music teacher by day and an audio engineer by night. Joe had listened to my podcast in his car and had difficulties hearing it clearly compared to other podcasts.
He kindly explained the ideas and reasons behind mastering, and how a mix of compression + limiting will help my podcast sound louder (but importantly, not too loud) and better. If you're not familiar with these concepts, let me explain:
Compression is where you make the loud sounds quieter and the quieter sounds louder eventually making them all of a similarly equal volume. It's crucial not to simply make everything loud, because the volume of speech in conversation is very important. If the listener is hearing the podcast on a noisy train for example, I don't want them to keep adjusting the volume when our dialogue quietens. Therefore, I start off by finding a really loud and really quiet section of my podcast and looping this so that I can hear the difference, I then add a compressor and add more compression until both the quietest and loudest parts sound at a similar volume but not too loud whilst importantly still sounding natural. It's a delicate process.
Afterwards, I add a limiter. When audio gets too loud and goes into the red on my meters, it starts to sound distorted. A limiter, stops the audio from ever going into the red and distorting. Once I've added the limiter, I gradually the volume so that I can only just hear it working which will still keep my audio natural sounding but also louder than it did before.
I want my podcasts to sound both loud and clear in comparison to other professional podcasts without becoming distorted and this method of using compression and a limiter helps it sound how I like it.
This process needs to be done per-episode because the levels required will change per guest as not all guests record with the same equipment in the same environment.
When I'm finally happy with how the episode sounds, I need to export it so that I can upload it to Transistor. Originally, I was exporting to WAV which provided the best quality however, with the quality came the file size and Transistor fairly limits the maximum upload size, so I now export to MP3 by using the Export Song to Disk option under the Share menu.
I export the episode at the Highest Quality setting which provides (256 kBits/s). Honestly, I don't really understand what that means but again, when I listen on my headphones, earphones, soundbar or in the car, it sounds great, so I don't worry about not understanding that.
And that's everything I know about Podcasting with GarageBand!