[*]60hz buzz
[*]Low or uneven volume
[*]Intro/outtro sometimes trails on and on
[*]Room noise
[*]Coughing, mic bumping, and dog barks
[*]Lack of equalization
But I'm not here to complain, I'm here to offer solutions. To get around most of these issues I've started listening to the podcasts exclusively on my computer after doing a bunch of audio post-processing to make it listenable. I'll use podcast #110 as our case study. All screenshots and processing will be using Audacity, a free open-source audio editor.
This screenshot shows 3 different issues all at once. First, the podcast is being saved as a stereo audio file. The microphones being used are mono and there's no stereo trickery going on that requires both audio channels. If the podcast was output as a mono audio file you could A) double the bitrate and increase the overall sound quality while keeping the file size identical, or B) keep the current bitrate and now your podcasts are only half as big but would sound exactly the same.
Second, the long delay before the intro song. 5.3 seconds of room noise and whispers is kind of a long time and just feels weird to listen to. A little trimming here takes seconds and would make all the difference.
Last, the delay at the beginning isn't totally silent. Aside from the whispers you can see the 60hz buzz in the recording, especially if I zoom in. I'm pretty sure this is coming from Garthe as the Jenn-only podcast didn't have this issue. The buzz likely makes little difference on tinny laptop speakers or headphones but on a full-range system (like my car, which has enough low-end to shake your eye-balls) it's a big nuisance.
One issue you can't see in the waveform is that the intro doesn't just play, but the whole song plays during the beginning of the episode. It doesn't happen in most, but it shows that there isn't a set of pre-cut assets being used which leads to issues like this. At best it's distracting, at worst it seems like a potential, though unlikely, legal/licensing issue.
Another issue is that the volume is sometimes very quiet, then gets loud very quickly (do either to the host's change in excitement or distance to the microphone). Similarly, coughs, microphone bumps, and dog barks are extremely loud compared to the quiet talking and are very jarring, which at my computer is annoying but while driving can be dangerous. Normalizing and compressing (which I'll show in a bit) can help negate this a bit but these spikes in audio should be cut from the source audio (which will be addressed at the end).
So how do I fix back-catalog podcasts? Let's start by removing the unnecessary audio track. I'm not really sure how to explain where the menu to do this is, but look at the screenshot; it should be pretty obvious. Doing this first should cut down on the time it takes to do each other step by half since there's half as many audio tracks.
Next we're going to be compressing the audio (lowering the volume when it becomes too loud) to make the volume more uniform, then normalizing it to make sure the overall volume is loud enough but not too loud. Effects > Compressor brings up this window. These settings seem to work well enough, but I'm not an audio tech, so I'm sure someone could do much, much better.
Next we'll normalize the audio. This will re-center the peaks and valleys and set the overall volume to an appropriate... volume... Go to Effects > Normalize. These settings should be fine, but you could also set it to -2db or so.
Next we want to kill that low-frequency buzzing. Find a spot in the recording where no one is talking for a second or more (use the zoom/magnify tool to zoom in) and highlight an area where there's just buzzing. Ideally there would be no breathing or other noises.
Then go to Effects > Noise Reduction and click on Get Noise Profile. The box will go away. Click a part of the audio outside of your selection (or use CTRL+A to select all, it'll work either way) and go back to Effects > Noise Reduction. Use the settings in the screenshot (or have an actual audio tech do it the right way) and click OK.
Last thing we want to do is equalize the audio. That is, emphasizing or de-emphasizing certain frequencies to make the tone of the audio better. There's plenty of ways to do this, the EQ I used is kind of quick-and-dirty but will give you a reasonable 'radio' sound.
And finally, here's a comparison, taking audio from podcast 110 starting with the original audio and playing it again with the cleaned-up audio.
http://www.fakegamerguy.com/wp-content/ ... ompare.mp3
So this is a lot of work to do after the fact and there's still issues with the intro playing for too long, the guest audio being too low, and other misc. issues. Moving forward, the most ideal scenario would be to have each host record their own audio track (I don't know if they're in the same room, using Skype, or whatever but it should work the same either way). This way you can level-match after the fact so no one is too quiet or too loud, noise reduction is more effective, and it's easier to mute the audio track during mic bumps, coughs, dog barks, police sirens, etc. It's a few small steps that would make the podcast much, much better to listen to. And since the audio is all split up anyway adding the intro and outro tracks would be really simple and you would avoid the issue of it being 3-5x too loud like it was in... 111? 109? Something like that.
Sorry if this seems a little demanding from a newbie member. I'm not trying to poop on the podcast or anything; I really like it and just want to see it grow.
