How to Make Your Podcast Loud and Other Tips

*Update 1 - Added free tools and changed the settings and pics to reflect those tools.

I've noticed that a lot of podcasts are really quiet. If i'm sitting at home and listening on my stereo this isn't a problem, but if i'm on the train or bus or even walking down the street, some of these podcasts are impossible to hear even if i've maxed out the volume on my MP3 player. I usually run a process over certain podcasts before I load them on my player and I've decided to post that process here.

This is a quick way to make your podcast loud. It's not tips on mixing individual levels and all of that because I'm lazy plus that's something anyone could do just by listening and seeing who needs to be turned up or down. This is for when you've got the balance of everyone's voice right and the panning, which i'd suggest against unless you know what you're doing, is right. This is something you do to the final file before you upload it.

The Tools...
Audio Editor If you purchased a sound card separate from your computer, it probably came with an audio editor of some sort. Any one of them will do as long as it can load up long audio files and optionally supports VSTplugins. Some free ones such as Quick Audio and Wavosaur don't support long audio files so avoid those.

Plugins/Effect The effects I'm using are a compressor (not file compression like .zip or MP3) and and maximizer/limiter. Your audio program might use the term "Dynamics" to describe the compressor. It doesn't matter what you use, as long as it is a compressor. All of them pretty much do the same thing in the same way and differences aren't large enough to call for changes to the tips in this tutorial. Some audio editors come with effects built in and others support VST plugins, which there are thousands available for free @ KVR-VST.

The two that I will use for this example are:

- GeneComp from the BuzzComp Free Series

- George Yohng's W1 Limiter

Example Audio File Bitmob Podcast #16

The Process...

STEP 1

Load the audio file. It should look something like this picture. The red box that I drew highlights what the bulk of the audio is with a little bit of leeway. Everything outside of that is split second peaks that are louder than that. For the record, that box is around -8.5db. A different audio file will probably have a different level, just make a mental note of what it is. If you use normalizing to boost the file to a certain level, the computer takes these peaks into account and rather than boost the important part to the level you want, it will boost the peaks to the level. You may gain a little bit of volume, but overall nothing has really changed. You need a way to get rid of those peaks, which is what we'll do next.

STEP 2

Load up your compressor/dynamics processor. No matter what compressor you use or what settings it has, you only need to worry about 3 of them: attack, threshold and ratio. If your compressor has release, just leave it where it is. You may have an option called output gain or makeup, but set that to zero for now.

The threshold is the volume level that the effect kicks in at. Set your threshold to whatever the important part of the audio is (the red box) is plus -2. That was -8.5db in the example so you would set the threshold to -10.5db (-8.5 plus -2). The attack is how quickly the peaks get smashed. We want to crush them so that when we boost it up, the peaks don't throw the computer off. At the same time, we don't want to smash it to the point that it sounds horrible. Set the attack to about 1ms. The ratio of X:1 means that for every 1db over the threshold, we'll reduce it by Xdb. We want to crush the peaks so we will set this to as high as your compressor will let you.

In summary: threshold = -10.5, attack = 1 ms, and ratio = maximum. If you don't have a gain section in your compressor, run the effect and you should get something like this:

Those peaks have been greatly reduced. The blue circles represent the very loudest parts of the audio file. They are all hitting around -3db. Remember this level and move on to step 3

STEP 3

Now load up your limiter or maximizer. Like the compressor, you may have a gangload of of buttons and functions but we will be ignoring all but two of them: threshold and ceiling. Ceiling may be called "limit" or something that suggests that it's the maximum level the sound will hit at. Set your threshold to the highest part of the audio plus -4 db. For our example, this would be -7 (-3db plus -4db). Set your ceiling to -0.1 db and run the effect. Your audio file should now look like this:

As you can see, it's noticeably louder than it was before. You can actually hear this without having to be in a totally quiet room with the volume turned up to maximum.

You can smash it more by increasing the ratio and threshold and shortening the attack. Try increasing the ratio before shortening the attack though. You can make it less crushed by doing the opposite actions, but still decreasing the ratio before increasing the attack. Increasing the attack will have a pretty major effect since more of the peak slips through without getting crushed. Decreasing the ratio means that everything is crushed less, but still crushed quickly.

Some additional notes/tips

- It's critical that your podcast already sounds pretty decent. If there are any audible issues, they will be easier to hear now that the volume is louder.

- I noticed in the example that there are clicks and pops in the source file whenever someone stops talking for a short while and nobody else is talking. I think this is because of a noise gate so here is a tip for noise gates: release is your friend. Sound has a natural decay. If you chop off the sound with a quick release, there is a very noticeable effect.

- Or it's because they're editing out the silence. Cross fades are your friend and silence isn't always your enemy.

- Do a real deal sound check. I work with artists and all too often you do a sound check to make sure that their level is good and they do it at one volume. They start to sing and they've shoot up to like twice as loud. This is useless. If you're doing a sound check the whole point is to get a simulation of the volume level they'll be recording at.

- Always give yourself some space. If you've got a good sound check, don't try and crank the volume up to as high as possible. Give yourself 6db of space. You can always do an overall boost later on, but if they start distorting, you can't undo that.

- If possible, record everyone on a separate channel in your computer. For most beginners, your setup probably won't allow this. A good alternative is to at least have a mixer that has everyone on a different channel. You can adjust their volumes and do slight EQ without effecting anyone else's sound. And while we're on EQ...

- Bass cut is your friend. You can chop off any frequency below 80-120Hz. You're not trying to have your podcast beat in the trunk. Unless there is someone with a deep and sultry Barry White voice, cut those trunk rumbling frequencies off. Most mixers will come with a button to do this (usually labeled "Low Cut" or something like that) and some mics even have a switch built in that will do this.

- EQing for treble is your enemy if you're new to this. Don't boost the treble over 3 db and please don't do a treble boost by default. If you find that 3db of treble boost isn't making someone clear enough, try cutting out some middle frequencies by no more than 3db. If that still isn't working, ask the person to back off the mic and see if that helps. And while we're talking about mics...

- DO NOT DEEP THROAT THE MICROPHONE. You will destroy your sound quality. Pro-tip: you don't have to be 1 inch from the microphone in order to have your voice pick up. When you fellate the microphone, it's the same as having a person scream directly into your ear canal at the club. It hurts. It's painful. Do not do this to the mic.

- If you ever see a red light go off on your mixer or the audio level in your computer, know that something bad has just happened. There will be some occasional red moments here and there, but if this is constant, you've got something turned up way too loud.

- Mono is your friend. Don't go crazy with panning voices. I'd say don't do it at all unless you've been recording for a while and have developed a good sense of what works and what doesn't. If you must do it, don't go crazy. Try to keep everything within a "V" shape. Imagine that a V is resting horizontally on the top of your head. Keep all of the voices within that V. Don't have one person panned all the way to the left or right. That may sound cool on paper but it usually is terrible in practice.

That's all I can think of for now. Very few of these are steadfast unbreakable rules, but they should start your down a good path so that you'll know when it's OK to follow them and when it's better to break them.