Earlier today, Eric Nagel sent me a message on Twitter (here I am embedding it using the new Twitter embed function in WordPress 3.4. Spiffy!):
— Eric Nagel (@ericnagel) June 15, 2012
There’s about a half-a-meg-per-minute difference in the examples there, which is pretty significant when you’re talking about podcasts that can last anywhere from 30 to 60 minutes or more. I have two bits of advice about this. Pun totally intended.
Before I get into the lengthier part about bitrate, I’ll mention images.
Check Your Image
Your mp3 file carries metadata that includes a bit of text: title, artist, copyright, album and a lot more. It also carries an image—your show’s artwork. Your art can be anywhere from 128×128 pixels on up to 1400×1400 pixels (no need to go higher than the iTunes recommended size). Optimization for your mp3’s art is just as important as optimization for your website. Run your image through an editor with an eye toward reducing the file size before adding it to your episode’s metadata. There’s no reason to be carrying a 2MB image when you can reduce it to a few kilobytes with no noticeable loss of quality.
Check Your Bitrate
MP3 files are encoded with a bitrate setting: how many bits of data, on a per-minute basis, are you using? Your options will vary according to whatever editor you’re using, but typically you will have at least the following:
- 32 kbps – just about rock bottom voice quality. Think AM radio.
- 64 kbps – approximately what you might hear on FM radio.
- 96 kbps – better than FM radio, this is my personal sweet spot.
- 128 kbps – low-quality music downloads use this one.
- 256 kbps – higher-quality music downloads might use this one.
- 320 kbps – very nearly CD quality
That’s kilobits-per-second. Using a little math, you can approximate the file size, but your show recording/editing software should do that for you. Compare the following screenshots. I use Adobe Audition, which uses a different encoder than a program like Audacity, but the results in either program (or whatever you’re using) will be approximately the same. The file being exported in these examples is a recent episode of Geek Dads Weekly. It’s a .WAV file with a length of 67 minutes.
You can see there that at 128 kbps, the file is approximately one megabyte per minute (we saw that in Eric’s example above), while at 64 kbps, the file is approximately half-a-megabyte per minute (we saw that in Eric’s example, too). Those two bitrates are, I believe, the most commonly found in podcast mp3’s. Certainly, most shows that I’ve ever downloaded have been one or the other.
Optimizing Your Podcast Episode
Why do some podcasters encode at 128 kbps while others encode at 64 kbps? I have a few theories. One is that 128 kbps is the default export setting in many programs, so podcasters just hit ‘OK’ and go with it. Another is that many podcasters feel that they need the 128 kbps setting because it makes them sound really good – after all, that’s the bitrate sometimes used for music, so it must be awesome for voice work.
The thing to remember though, is that it is just voice work. You don’t need 128 kbps to sound great. Your voice doesn’t have nearly the level of intricacy that music has. You need 128 kilobits every second to capture all the sounds in a piece of music. Your voice needs far, far less.
Is 64 kbps the right answer then? Lots of podcasters swear by it, and the file size does get very small. I would certainly not recommend going any lower or your quality will suffer. 64 kbps seems to be the sweet spot, but…
I like 96 kbps. I split the difference between 64 and 128 because I have music in my shows. Remember what I said about music – you need more kilobits-per-second to capture everything. I have found that while some music will survive the encoding at 64 kbps, most will sound a little worse than I prefer. Encoding at 96 kbps will result in just about any music sounding good (not great, not awesome, but podcast-worthy).
My bottom line then, is this:
If you don’t have any music in your show, encode your mp3 at 64 kbps. If you do have music in your show, try 64, but maybe plan on using 96 kbps and living with the slightly larger file size. There is no technical reason to encode any higher than 96 kbps in my opinion.
Production masters can massage the files, work with the music in subtle ways, and figure out how to get great sounding music into a 64 kbps file. For everyone who doesn’t want to spend a long time in post production just to save a few megabytes… this is how I roll.
Just for a laugh…
At full near-CD quality, a 67-minute podcast would be…
There isn’t a podcaster on the planet that needs to sound that good.
Hope that helps! Comments are open for people that have other ideas or want to give me high fives.