I just learned a valuable lesson in life. I was helping my dad troubleshoot a recording problem. (He’s running sound at a church now – the tree doesn’t fall very far from the apple, eh?). His church is providing podcasts of their sermons, and up until two weeks ago, they sounded decent. But the last two sermons sounded like the Devil himself had infiltrated their computer recording system. After some triaging, we figured it was because he, in an attempt to vastly reduce the file size, encoded the file to 16kbps. But not just randomly. He had example audio bytes that were mp3 encoded at 16kbps that did indeed sound decent. So why did those last two podcasts sound like junk?
As it turns out, when encoding to that low of a bitrate, keeping the signal in stereo does bad things to the quality. Technically, it’s splitting the samplerate apart to use on each channel, and 16kpbs encoding just doesn’t leave enough data for both channels. So, in conclusion, always mixdown to mono before encoding that low-bitrate mp3 podcast:
That will certainly be helpful to know now that Pantano is moving towards sermon and newsletter “E-News” podcasts. Here are the apparent “money/awesome” specs for a decent sounding, ridiculously small mp3 podcast:
Mpeg Type 2.5, Layer III
With that you can fit 55 minutes into 6MB! And, it doesn’t sound too bad. It sounds about like the best AM radio you ever heard, but if it’s just speech, than that should be fine in most cases. Just don’t do it in stereo! Otherwise it will sound like a horde of breeding buffaloes.