Yes, audio compression is the solution. It's basically automatic level riding. Unfortunately, I'm not familiar with any Windows apps that will let you do it in one step. There are some hardware solutions (not cheap) from Dolby and such. Also you might already have it in the form of "midnight mode" or something like that in your receiver or whatever you are using to listen through.
I use compression all the time. For example on a recent video I recorded of a small concert I used compression to balance the levels between the various songs (you can see the videos
here (YouTube)). I also use it from time to time on podcasts.
Basically a compressor reduces the level of loud sounds thereby compressing the dynamic range. Then because the peaks are lower you can bring up the gain and make the whole thing louder. By controlling the threshold at which point the compression gets applied you can use a compressor to balance the levels between different segments (loud segments stay loud, but quieter parts get brought up).
The manual way would be to apply a compressor plugin in something like Audacity. Compressors usually have these controls: threshold, ratio, attack, release and output gain. Basically what you want to do is analyze what the average loudness is of the audio and then set the threshold to be just below it. Often times, the compressor plugin will have a visual indication of how many dB of compression it's applying. You want it to be around 6 dB of compression.
You set the ratio to between 3-5:1, this controls how strong the compression is. The larger the ratio the bigger the reduction (compression) is when an audio signal is about the threshold you specified. Basically, it works like this: for example, you set a compressor to -12 dBFS (digital audio is always negative in value, 0 is the loudest you can go), with a ratio of 2:1. Now you apply this to a variety of audio source with different levels:
*If the signal is at -12 dBFS or lower, then nothing happens because it's below the threshold you set (-12).
*If the signal is at -6 dBFS, then compression gets applied because it's above the threshold. How much it gets reduced is determined by the ratio. In this case, the ratio is 2:1 so anything that goes above the threshold gets reduced by half. So -6 goes in, which is 6 db above our threshold, 6/2 is 3, so what comes out of the compressor is only -9 db.
*If you send it 0 dBFS (the max), that's 12 db over the threshold, 12/2 is 6, so you get -6 dBFS coming out. Etc.
Attack and release control how fast the compressor acts. For voice something like 30ms on attack and 130ms on release is a good start.
Output gain (there's often also an "auto" button) controls how much you bring the level up after the compression. Since the compressor reduces the peaks in the audio and thereby makes a source quieter, you use the output gain to bring it back up.
There's a lot of experimenting involved and it takes time to really get the hang of compression. If you really want to get into it, you can find a bunch of tutorials online.
Hope this helps.
Btw, what podcasts are you talking about?
