PODCAST EDITING

The Complete Guide to Podcast Editing

Riley Byrne, Owner, Podigy.co

07 August 2017

A great podcast is a combination of several factors. Choices made about subject matter, tone, hosts, guests and equipment will all shape the end result in some way. However! The one thing all great podcasts have in common is ample time set aside for post production. By this I mean that somewhere along the line, before the podcast is published, someone goes through and ensures the levels are right, the flow is consistent, the audio has been scrubbed of distractions (excessive noise, long spaces of dead air, over use of filler words, etc.)  and the overall mix feels right. A good audio editor will be able to do this using audacity, but it will take a while. A great audio editor will do all of this and design a custom workflow to make it as easy as possible. By the end of this tutorial, we’re going to make you a great podcast editor who uses only 3 buttons to edit their podcast!

This guide will take you through the following:

  1. Setting up Reaper actions that combine several elements of editing into a few button presses
  2. Setting up FX Chains and templates to save time at the beginning of each new podcast. This includes using EQ, compression, gates, and other effects.
  3. Putting it all together and actually editing our podcast.

Each of our podcasts average about 1000 edits per hour. Each edit before implementing this system took between 3-9 user actions. Using this system, we make multiple edits with a single action. It might not seem like you’re saving much time over a single edit, but multiplied by a thousand, the time savings quickly add up!

Downloading our Tools

For this tutorial, we’re going to be using Reaper, a digital audio workstation (DAW) that is not only powerful and customizable, but at only $60, it is also one of the most affordable DAWs on the market.

More than that, it comes with a 60 day trial period, and you can keep using it without any loss of features after the trial ends for free! You could continue to use Reaper forever without paying for it, but with all the customizability built in, and updates rolling out every 2 weeks, I think you’ll find after 2 months that the $60 price tag is more than worth it.  Download Reaper here.

The one thing all great podcasts have in common is ample time set aside for post production.

Next, we’re going to download The SWS Reaper extension, an open source add-on that expands the functionality of Reaper considerably. In this tutorial, we’re going to use SWS to make our button presses more effective by chaining together several actions. For instance, instead of a separate button press for deleting a piece of audio, joining the two split pieces of audio together afterwards, and then crossfading them to make the edit seamless, with SWS we’ll be able to consolidate all of those actions into a single press of a key. Download SWS here.

Finally, we’re going to download TriLeveler 2, from Sonic Anomaly. This is a “set it and forget it” type of plugin that is free to boot! The Tri Leveler works by having three different compressor/expanders working in tandem to make sure your audio is always at the level you want, without it sounding overly compressed or distorted. Download the JSFX version of Trilvelever 2 here. This one is a little more complicated to install though. First, unzip the package from the website, and in Reaper click on Options > Show REAPER resource path in Explorer/Finder, and we’re going to drop our folder into EffectsMake sure you download the JSFX version.

Setting Up reaper

The greatest feature about Reaper is also the reason it is so intimidating to beginners: customizability. It is so feature-rich and flexible that you can run it across multiple computers, batch process files in a snap, sync up with others over the internet and write plugins on the fly! Consequently, there are a lot of little things we want to do right out of the box to fine tune Reaper to our specific needs. Lets first go to Reaper > Preferences > Project, and make sure that Timestamp backup is checked. Underneath, choose an interim that seems right to you, that you want Reaper to automatically backup your project file and then click Apply down in the right hand corner. Now, in the rare event that Reaper crashes, we’ll only ever have lost a few minutes of work!

Choose a length of time for the backups that suites you. I don’t mind losing 10 minutes of work in the rare instance that Reaper crashes, but you might want to backup more often.

Next, head to Media Item Defaults under the Project tab. Make sure Create automatic fade-in/fade-out for new items, length is unchecked, but that it has a timecode of 0.00.010. Similarly, uncheck Overlap and crossfade items when splitting, length, but make sure it has a timecode of 0.00.050. Reaper by default fades out the audio a bit on either side of an edit point (such as removing an “uhm”) which works well in music, but in a podcast that sudden fade in/out of audio can cause little clicks that are distracting to listeners. We’re going to make our own custom crossfades a little later using SWS, which looks at the timecodes in this menu to determine the length they should be.

Now is also a good time to go to Audio > Device and make sure Request sample rate is set to 44100. It’s a very long and complicated reason as to why this is, but sufficed to say this has the potential to save you many headaches in the future when you cannot figure out why your audio and a guest’s audio recorded remotely are slowly drifting out of sync, or why your voices sound oddly pitched down through some web players.

The block size is how often your audio will update in Reaper, and affects latency. The smaller the block size the less latency, but more strain is put on your CPU and can cause audio dropouts. 512 is a good threshold for editing, but if you’re routing your audio out of Reaper into something like Skype or Twitch, try going with 128-256. 

Finally, head to Editing Behavior > Mouse Modifiers, change the Context to “Arrange view” and “right drag”, and then double click on Default Action and select Marquee select items > Set Time Selection. It should now read Marquee select items and time. This will let us right click and drag to select time that we want to edit out with our X and Z commands now.

Make sure the grid system is disabled when you’re editing, by clicking on the grid button in the upper left hand corner of Reaper. This will allow for much more finesse when we’re trying to edit out “um”s that are in the middle sentences.

Now that we have our preferences in place, it’s time to start using some of the cool features in SWS. Head on over to Actions Show Action List. We’re now greeted with something that might look like gobbledygook to you right now, but this is actually one of the coolest features in Reaper, the ability to make our own custom actions and chain them together to make powerful shortcuts for quick podcast editing. So head to Custom Actions and click New…. This will bring up a new menu where we can begin creating our time saving editing tools.

I don’t know if arrows are helpful or dreadfully tacky but this next bit warrants some. 

First, we’re going to make a custom Ripple Edit function that automatically crossfades the resulting clips for a smooth transition between them. A Ripple Edit (or Remove Time, Shuffle Edit etc.) is when audio in a certain time frame is cut across all tracks in project, and the audio behind the edit moves up to where the cut piece of audio was before. For instance, if a podcast has a segment of dead air, instead of deleting the dead air portion and moving the resulting audio clips individually across each voice track, trying to line them up perfectly again,  with a few clicks you can do all the tracks at once and keep them perfectly in sync! Many DAWs have a Ripple Edit function, but they are somewhat clunky to implement and work against the natural flow of editing a podcast, so we’re going to make our own that suits our needs perfectly.

Ripple Edit Example: Removing Speaker 1’s interruption (Green), and syncing the two sentences of Speaker 4 (Yellow and Red) with one button press.

In the Filter bar, start typing until “Item: Select all items in current time selection” pops up and double click it to add it to our custom action. Do the same with “Time selection: Remove contents of time selection (moving later items)” and “SWS: Crossfade adjacent selected items (move later items)”. Name your action, and click OK. You’ve just created your first custom action in Reaper! If only you could talk to your past self, the one from the beginning of this guide. The things you could tell her. But I wonder if you would even recognize yourself? You’ve come a long way.

Naming your actions something you remember will be helpful in a minute.

Next, we’re going to make another custom action, this one is basically a “smart delete” button. This is going to give us a shortcut that deletes only what we’ve highlighted, 100% of the time. The farther you dive into Reaper, the more you can take advantage of settings that can change the parameters of what the regular delete key looks at, and it is frustrating (not to mention time consuming!) trying to delete a little snippet of audio only to accidentally ripple edit the entire project, or delete the whole audio track! So following the same steps as before, we’re going to add Set ripple editing off  and Edit: Cut items/tracks/envelope points (depending on focus) within time selection, if any (smart cut) and give it a custom action name.

Only you have the power to stop accidentally deleting tracks. 

Now it’s time to assign keys to our shortcuts. I personally like to use “X” and “Z” as they are close to each other, and Cmd + Z (Ctrl + Z on PC) is our undo function. This way they are nice and close to each other, and you don’t have to move your hand at all to undo an action. Finally, in the action menu, search “View: Go to play position” and assign it to “F”. This will allow us to do rolling edits, removing bits of audio with Z on the fly, and then catching up to the playhead once we have.

You will get some pop ups asking if you really want to change those assigned buttons to our new actions, and click yes to each. When you’re editing a podcast, 90% of your time will be spent cutting, ripple editing and undoing, and we’ve just reduced those actions to 3 keys that are all within reach of each other! You’ll save so much time not having to dig through menus, or looking down at the keyboard to make sure you’re hitting a complex key binding, you can easily double or even triple your efficiency at editing over other programs. But we’re not done yet…

Setting up our FX chains

One thing that is tragically underutilized in most podcast production is proper EQ and compression. Almost any decent quality mic will record your voice well enough that, with a little polish and shine in Reaper, will get you sounding like the pros over at NPR. Our goal in this section is to set up a series of plugins that will tame even the wildest of vocal performances, and save those chains as easily recallable “FX Chains” in Reaper so the perfect podcast sound is only ever two clicks away!

Right clicking in any FX box will allow you to choose FX chains that we can use to quickly get our tracks sounding great.

Let’s start with setting up our EQ. An EQ (or equalizer) shapes our audio based on frequency bands. Find out more about EQ for podcasts here. We’re going to make an EQ curve that clears out the frequencies we don’t want and leaves the ones we do relatively untouched. First, click on the “FX” button underneath the name of one of your tracks. This should open up a new window that looks fairly blank, save for “Add” and “Remove” down at the bottom. Click on “Add” and in the new window type “VST: ReaEQ (Cockos)” in the Filter text box and double click to add that effect to our track.

Arrows to arrows, you to EQ. Such is the way podcast editing goes.

First off the bat, you should see a button in the bottom left hand of ReaEQ called Add band. Click this three times, and you should see 7 little circles on the top in a straight line. This is our EQ Curve, and it takes the incoming audio and makes boosts or cuts based on the position of these dots. Right now it isn’t doing anything to our incoming audio because our curve is a straight line. Let’s change that! Click on “1”, either the button or the circle, and put in the following settings:

Enabled
Type: Band
Log-scale automated frequencies
Frequency (Hz): 87.0
Gain (dB): -1.3
Bandwidth (oct): 3.70

After putting in these settings, you should see a slight dip in the curve on the left hand of the graph, and then a straight line. Next, input the following info for the following bands.

This is what the EQ will look like when we’re done.

2
Enabled
Type: Band
Log-scale automated frequencies
Frequency (Hz): 209
Gain (dB): -2.5
Bandwidth (oct): 3.7

5
Enabled
Type: Band
Log-scale automated frequencies
Frequency (Hz): 3256.0
Gain (dB): -2.7
Bandwidth (oct): 3.7

3
Enabled
Type: Band
Log-scale automated frequencies
Frequency (Hz): 547.7
Gain (dB): -5.4
Bandwidth (oct): 3.7

6
Enabled
Type: Band
Log-scale automated frequencies
Frequency (Hz): 7400.0
Gain (dB): -1.3
Bandwidth (oct): 3.7

4
Enabled
Type: Band
Log-scale automated frequencies
Frequency (Hz): 1495.0
Gain (dB): -5.4
Bandwidth (oct): 2

7
Enabled
Type: High Pass
Log-scale automated frequencies
Frequency (Hz): 70.0
Gain (dB): 0.0
Bandwidth (oct): 1.0

This is a somewhat unconventional EQ, as most people tend to boost frequencies they like instead of subtracting ones they don’t. We’re leaving the parts of the voice we want relatively untouched, while scooping out some of the parts that can get muddy. Also, with band 7 we’re rolling off a lot of the low end rumble that can come from people speaking very close to a microphone. I like this because it gives the voice a much more neutral tone no matter where the speaker is in relation to the microphone. If you have someone who isn’t used to talking into a mic on your podcast, this rolloff will help mitigate some of the changes in tone that can happen when someone starts talking far away from the mic and then gets closer, and vice versa. Click the “+” button next to “Param” at the top, click “Save Preset as Default” and then enter in a name you’d like to save this preset as. I would suggest “Podcast Default”. To find out more about how EQ works, check out our article!

Note: If you are going for the “radio” sound on your podcast, you’ll want to turn off band 7 in the EQ. That bass-rich sound that comes from talking very close to the mic and then heavily compressing it, needs all the low end energy it can get! However, I would caution against trying to sound too much like a radio host. Podcasts are made to be listened in earphones and a very compressed, bass-heavy voice can cause ear fatigue rather quickly. The rest of this tutorial is going to focus on getting a cleaner, lighter sound that is heard on programs like Freakonomics or Radiolab.

Next, lets add “VST: ReaComp (Cockos)” to our track. Make sure it appears underneath ReaEQ. Without getting too technical, a compressor works by setting a threshold for incoming audio, and brings down the volume anytime the track goes above that threshold. The amount that it attenuates the audio above the threshold is determined by the ratio (For example: a 2:1 ratio at a threshold of -15dB will mean the compressor will allow any signal less than -15dB to pass without affecting it. However, for every 2 dB over -15dB the signal goes, the compressor will only raise in volume by 1 dB). It is always wise to use more than one gentle compressor on an individual track rather than a single harsh compressor. Find out more about how to use compressors in podcast editing here. Put the settings below into ReaComp to start getting a more consistent level from our vocal tracks. This will work in tandem with our Trileveler in a second. Save it as the default present and you’re all set.

Pre-comp: 0 ms
Attack: 10 ms
Release: 100 ms
Ratio: 2.0:1
Knee size: 5.0 dB
Threshold: -19.4

It isn’t listed here, but anything unchecked in this photo should be unchecked in your window as well

Now it’s time to crack open the Trileveler we downloaded all the way at the beginning of this guide! This is a really powerful effect that takes a lot of the thinking out of getting a consistent tone out of our vocal tracks, and it’s going to make sure each person in our podcast always sound like they are at the same level. It works by compressing and expanding the incoming signal, so that the audio sounds more consistent without sounding over compressed, and it can be set up in a few clicks! First things first, unselect “AUTO INPUT” and “AUTO THR.” unless you are a solo podcaster. These are great features, that can make for an even smoother sound, but can cause some problems in multi-person podcasts when it starts to interpret crosstalk (someone else’s voice being picked up on your mic) as audio it needs to level. Dial in the following settings:

Input Trim: 0.00 dB
Target: -19.00 LUFS
Gate Threshold: -53.00 dBfs
Gate Speed: 500.00 ms
Output Trim: 0.00 dB

Next, select “CONFIG” and bring up the “Fast Dynamics Max Attenuation” and ““Medium Dynamics Max Attenuation/Gain” settings to 10dB, and all the slow dynamics to 20dB. Also, optionally bring up the HPF to 0.10. This is a high pass filter that will ignore sound under that threshold, like bumps on the table. Save as the default preset.

Play your audio track for a bit, and adjust the Input Trim so that the three bars at the bottom of Trileveler ar as close to 0 as possible, and then adjust the Gate Threshold to the point where you can hear the audio of the speaker come through, but when they are not talking, no audio is coming through. It’s a bit tricky at first, but once you get it Trilevler will handle the rest.

The last plugin we’re going to add is a ReaXcomp, and from the pulldown menu just select “griz: female de-ess (change band 2 for your singer). This is going to help reduce all the nasty sibilance (think words with excessive “sss” sounds), which can come across really harshly in our listener’s ears. Move around the two vertical bars around band 2 until you find a nice balance of reduction in esses without it cutting into other parts of speech.

Here we’ve brought down the band 3 as well. This is a handy way to reduce some of the harshness inexpensive microphones can have in the high end. As opposed to band 2 which is bringing down our “sss” sounds, and so we are reducing incoming signal of 40 dB to one, band 3 should use a more natural sounding 10-20:1 ratio. 

Now that we have our effects in place and we’re getting a great sound, let’s save all these plugins together under one heading. Right click anywhere in the menu showing the plugin names, and go to FX Chains > Save all FX as chain. Name it how you want, and now anytime you need to add a new voice track to your podcast, you can have it set up and sounding as good as this one just by right clicking and loading this FX chain.

Finally, we want to add a little more processing on the Master track to glue everything together and make sure we limit any unforeseen spikes in volume so we don’t blow our listeners’ eardrums out. Go to View > Floating Mixer Master, and then click on the FX button for the master track. First, we’re going to add another instance of ReaXcomp, but this time we’re going to click “Delete band” and then input the following settings:

1
Active
Top Freq (Hz): 200
Threshold: -15.0
Ratio (:1): 3.2
Knee (dB): 4
Attack (ms): 30
Release (ms): 78
RMS size (ms): 5
Gain: 1.0
Log()-scale automated frequencies

2
Active
Top Freq (Hz): 5530
Threshold: -9.0
Ratio (:1): 2.0
Knee (dB): 10
Attack (ms): 20
Release (ms): 60
RMS size (ms): 5
Gain: 1.0
Log()-scale automated frequencies

3
Active
Top Freq (Hz): 24000
Threshold: -14.0
Ratio (:1): 2.0
Knee (dB): 10
Attack (ms): 20
Release (ms): 60
RMS size (ms): 5
Gain: 1.0
Log()-scale automated frequencies

Next, appropriately, we’re going to put “JS: Master Limiter” on the Master track and leave all the default settings as is, except for setting Limit (dB) to -3.0. Save these to an FX chain and we’re done fiddling with settings!

Finally, making sure there are only audio clips you need in every episode in our project (intro/outro music, recurring sponsor ads etc.) go to File > Project templates > Save project as template, name it, and you’re done! We now have a template customised to our podcast that we can call up instantly, record into and edit on with only minor adjustments to the FX chains that we’ve already put into place. It might seem like a lot to do all up front, and we haven’t even gotten into the actual editing of the podcast, but having a solid foundation that we can get going in 3 clicks (File > Project templates > [Name of Your Template]) saves so much of the time that goes into editing that we could be using to make really great show notes or coming up with shareable content for social media.

You might be wondering why we have so much preamble in a podcast editing guide. I’ve found the wall most podcasters hit after a few months has very little to do with actually recording the podcast itself, and everything to do with the amount of work in post needed to get each episode out. It might seem like overkill right now to set up so many FX chains and templates, but when you’ve left the editing of an episode to 10pm the night before it has to go up, I think you’ll be happy you made so many shortcuts today.

Finally, the Editing portion Of this editing tutorial

When we’re editing our podcast, we’re going to be looking to pull out 4 things:
1) Dead Air
2) Uhms and Ahs and Stutters
3) Crosstalk (Person A’s microphone picking up Person B’s voice)
4) Breathing, rustling, thuds and other noises made by people who forget they are recording a podcast and not having a “who can absentmindedly pick their microphone’s grill the longest” contest

And we’re going to do it all with just our X and Z buttons, and some generous portions of right clicking. Typically in DAWs the left click is used for selecting individual pieces of audio, and moving them about, or moving entire sections when Ripple Editing is on. It makes selecting a period of time in the DAW window very annoying because you always have to click just above or below actual bit of audio you want to select, making it a real pain if you have several tracks stacked on top of each other with audio. Instead, we’re going to use the right click, which in Reaper creates a time selection even if you’re clicking on the audio itself! So if you’re on a laptop you might want to consider plugging in a mouse for easy access to the right click function.

Anyways! Editing out dead air and filler words is very easy with our set up. Just right click and drag over the bit you want to remove, and hit X. That will automatically pull out the highlighted section and move everything after it up to where the edit started, and will automatically crossfade any sections that touch for a smooth transition. You see the crossfade if you zoom in close enough (Alt Scrolling, or +/), and can adjust the amount of crossfade by left-clicking on the top of the fade and dragging it to taste.

On the left we can see the default crossfade made using our custom ripple edit. By clicking and dragging the top of the crossfade, we can make it bigger in either direction. This is useful for edits mid-sentence, trying to make it as smooth as possible. 

Sometimes you’ll find it’s easier and more natural sounding to edit out an “um” that transitions into a word (for instance, when someone says “Ummmmm my personal preference is…” ) if you edit out the “uh” sound in “um” and then do a longer crossfade on the second word.  A word of warning: to make your speakers sound as natural as possible, try to edit out as little of the surrounding audio around a filler word as possible. If you start cutting words too close together, you quickly lose the natural flow of the speaker and the edits will sound more distracting than the “ums” and “ahs” ever did. With a little practice, you’ll find that you can pull out dead air and filler words in a way that sounds natural and keeps the pace of your show up!

Next, we have to deal with the crosstalk that is building up on our tracks. If you have a podcast and haven’t been gating or scrubbing your tracks to get rid of crosstalk, this is the single biggest improvement you can make. Basically, anytime someone isn’t speaking, you want make sure their track is muted some way. Without doing this, speakers can start to sound distant and the noise floor (the amount of noise on a track when no one is talking) can become distractingly high.

A lot of people will suggest setting up gates, and we do have one on Trileveler, but we don’t want to depend on this to keep out the crosstalk and keep in the good stuff, because oftentimes podcasts are recorded in less than ideal places, and in some cases the crosstalk might be louder than the speaker. You’ll run into this lots if you try recording someone loud and boisterous with someone quieter and more reserved in the same room, the loud person will oftentimes be louder on the quieter person’s track than the quiet person is!

So, to deal with this, we’re going to pull out any audio we don’t need with our trusty Z button. Highlight the audio you want to delete with a right click/drag (if doing it over multiple tracks, make sure you drag the box over  all the audio you want deleted) and press Z. It’s very simple, but can be time consuming to do, as the more people you have in an episode, the more audio you have to delete. The two best ways to do this, I’ve found, is to either pull out all the crosstalk before you start making cuts, so you just get it out of the way, or do what I call rolling edits, where you select audio and delete it while the audio is playing, and then press F to centre Reaper on the currently playing audio again. This last way is a little more advanced, but it saves you from going through the episode twice!

Finally, let’s talk about breaths. After editing hundreds of podcasts, I can say without a shadow of a doubt that every person’s breathing patterns are as unique as a fingerprint, and they are all annoying. Now, you may not share this opinion, and find the laboured wheezing of a particularly annoying braggart a comical juxtaposition that would be of benefit to your podcast, but by and large we’ll want to edit out the sharp intakes of air people take while they’re in the middle of making a point. These can be particularly sibilant and can be fatiguing on listeners’ ears. However, if we ripple edit out the entirety of every breath our guest takes, they start to sound very unnatural and almost robotic, while simply deleting the breaths can leave long, awkward gaps.

Instead, try splitting the difference. Delete the entirety of the breath, and then ripple edit out half of the resulting silence. I have found that nearly every time I employ this trick, the result sounds more natural than either edit on their own.

Putting it all Together

The main part of our podcast editing is done. However, if yours is like most podcasts, you have separately recorded intro, outro, music cues and maybe an ad to insert into your podcast once you’re done editing the main part. This is where ripple editing proper becomes very effective. Under Options select Ripple Editing Off to change it to Ripple Editing Per Track and then once more to Ripple Editing All Tracks. With this enabled, left clicking and dragging any clip will move the highlighted clip and every clip on every track after it at the same rate, allowing you to quickly make room for the intro and ads in the podcast.

Be wary though: never move tracks too quickly with ripple editing enabled, as you can sometimes move clips to different tracks and mess up the sync of the podcast. Always be ready with Cmd to undo any mishaps.

If you’re putting music underneath your voice, take an EQ and dip out some frequencies around the fundamental (80-120 Hz) and a broad swath in the high end (around 2-8kHz) to make your voice really pop out and not get lost in the music.

For most music bought from royalty-free music sites, bringing down the volume 8-12 decibels and then doing the EQ tuck gives best results. 

When you’re happy with your podcast, go to File Render and you’ll be greeted with a pop up window. Here are the settings to use if your podcast is fairly talk oriented, with the occasional music cue:

Source: Master Mix
Bounds: Entire Project
Choose your output directory
File Name: $project
“$project” gives your rendered track the same name as the project, which is a handy time saver
Sample Rate: 44100 Hz
Channels: Mono
Full-speed Offline
Use Project Sample Rate For Mixing and FX/Synth Processing
Resample Mode (if needed): Fast
Tracks with only mono media to mono files
Output format: MP3
Mode: Constant Bitrate (CBR)
Bitrate: 96 kbps

It should be noted that for rendering there is no final word on the best way to do things. However, this guide is making sure you conform to the most commonly accepted guidelines for podcasts. For instance, you may see that -16 LKFS (Loudness, K-weighted, relative to Full Scale, a way of measuring loudness) is recommended for podcasts, but we’ve set our individual tracks to -19 LKFS. However, because of panning laws, our mono file will gain 3dB of volume when going from stereo-to-mono so in the end we will be at -16 LKFS.

In the coming weeks and months we’ll be expanding out our guides to individual components (like rendering or compressors or other helpful reaper action chains), and even into the best ways to record and distribute your podcast. That being said, I encourage you to experiment with these settings and fine-tune it to the unique nature of your podcast.

Want more help podcasting?

Leave your name, email and a link to your podcast and Riley will listen to your podcast and  make you a personalized action plan to take it to the next level.

Take your podcast to the next level with our daily video tips

Get in touch with Riley, our owner, anytime

rileybyrne@podigy.co