I’m sure you want to get millions of downloads and that costs money. You also want high audio quality and that is good, but you are wasting a few megabytes on the episode’s audio quality.
I checked the 6th episode.
The ogg file is targeted at 160 kbps, which means -q5. This is way overkill. Default oggenc on Ubuntu with q3 (between ~112 and ~128 kbps) and probably q2 (between ~96 and ~112 kbps) is more than sufficient since 99% of the show is human voice and not music. You can save 40% ogg file size with no noticeable audio quality changes.
As for MP3 is similar, but not exactly the same. You are encoding at 160 kbps which is also way overkill for voice. If you are using LAME (which you should) you should try -V6 (~115 kbps) or -V5 (~130 kbps) which will give you a substantially smaller mp3 file and the same audio quality.
As you probably know the recommended setting for LAME (mp3)
http://wiki.hydrogenaudio.org/index.php?title=LAME#Recommended_encoder_settings
and for ogg
http://wiki.hydrogenaudio.org/index.php?title=Recommended_Ogg_Vorbis#Recommended_Encoder_Settings
Are based on compressing CD quality music. You probably have a CD quality source for your voices, but voices carry very little information compared to music so you can compress it a lot more and still get transparent compression results.
I usually hear podcasts (radiolab) complain about bandwidth costs, but they still have huge files for no reason.
