Sept. 4: Episode 1 Post-Production

I now had all the elements we had planned to record in-house, and so was almost ready to start the post-production phase on episode 1.

The only element I still needed to get was the music for the episode. I wanted to use good music, at least as far as the quality of the recording went. This meant actually purchasing a couple of tracks from an inexpensive royalty-free music vendor. Fortunately, there a variety of these vendors available on the Web, where you can purchase a track and download it right away. Royalty-free was also important, as I thought we would probably be using the same music throughout the entire series, and certainly didn’t want to have to pay royalty fees for each use. Once I had the music tracks, I was set.

I opened up the first audio file into my audio editing software. Because the recorder we used only supported the Windows Media Audio format, I saved a copy of the file in that format. (You should always work on a copy of your recordings, never the original files, in case there’s a problem and you need to go back to the original.) However, the software warned me that WMA was a lossy format. What this meant was that every time I would make an edit and save the file, the software would compress it down a little bit more, discarding a little bit more of the file’s information to do so. This is a good tactic if all you want to do is make a copy of a file and have it take up as small a space as possible. However, it’s not a good thing to have happen when you’re editing a file and saving it repeatedly. So, I saved the file in MP3 format and started work.

Each of the different segments of the episode had been recorded as a separate audio file. This was a great benefit, as it allowed me to work on relatively small files rather than trying to edit a single large file.

First, I listened through the file and determined what edits needed to be done. In addition to playing the audio, most editing software, including what I used, displays the file as a visual waveform. Looking at the waveform while listening to the audio made it clear what needed to be cut out of the waveform, such as gaps where the speaker paused too long, or stumbled over a word.

In several cases, I was able to merge elements from multiple takes of the same segment. In one segment, everything was fine except for a short segment in which the speaker had slurred the text. Fortunately, the speaker had done fine in the other take, so I was able to cut out the good audio from one file and fit it into the other. Even though I knew were the edit had taken place, I was unable to hear the difference.

A common problem I ran into was that the different speakers in a segment often spoke at different volumes, sometimes enough to be noticeable. Fortunately, the software allowed me to select a section of the waveform and increase the volume only for that section, thus allowing me to even out the volume. On these sections, you can sometimes hear a slight elevation in the background hiss where the volume was increased, but only if you listen closely.

When all of the recorded audio files had been edited, it was time to assemble them into a single program. For this, I used the software’s multitrack function. Most commercial audio today, especially music, is recorded on multiple tracks. The recording of a guitar is placed on one trace, the recording of the drums on another, the vocals on a third and so on. The editor can then balance out the sounds to create the desired performance. Although our podcast is nowhere near as complex as a contemporary music recording, I used the same principle to piece the elements together.

I placed each segment on its own track. I then placed the music interludes on their own tracks. In those cases where I wanted to music to fade into the background behind a vocal, the software allowed me to edit what it referred to as the track’s volume envelope, lowering the volume were desired and then raising it back up when needed.

Once all the segments were in place, I exported the entire program into a single MP3 file. I expected the file to be quite large — it was nearly a half hour of audio, after all. However, I thought that at 25 MB, it was too large to be easily downloaded by some listeners. To reduce the size of the file, I altered the bitrate of the recording.

There are many factors that influence both the quality of an audio recording and the size of the final digital file. Two of the most important are megahertz and bitrate. Megahertz refers to the number of times per second that the sound is recorded when a recording is made — the higher the number, the higher the quality of the sound. The sound on an audio CD, for example, plays at about 44 megahertz, which is what our audio files had been recorded with. It is possible to reduce this value, and thus potentially reduce the size of the file, but only at the risk of reducing the quality of the sound down to unacceptable levels.

The other factor involved is bitrate. The bitrate of the file is how much information it sends to the player’s processor at one time, usually measured in kilobits per second. Just like megahertz, the higher the bitrate, the better the sound quality. However, I found that I could cut the bitrate of the episode in half, dropping it from 128 kbps to 64 kbps, with only a minimal loss of audio quality, and at the same time cut the size of my digital audio file in half, down to about 12.5 MB. I thought this was an acceptable tradeoff, especially since, as a podcast, we could get by with “good enough” audio.

Once the episode had been edited together, all that remained to do was complete the podcast’s companion website and announce the release of MOREnetworking: Podcast episode 1.

And now, on to episode 2.

hendersonl | Monday, August 6, 2007 | |


Copyright © 2007 The Curators of the University of Missouri. All rights reserved. DMCA and other copyright information.