Untangling Audio: Navigating Reformatting Challenges

Lauren Katz Records Management

As a Reformatting Specialist, Leif Johnson has come across his fair share of challenging projects. Recently, he was presented with a seriously tricky audio reformatting issue that required his expertise and problem-solving abilities. Below, Leif explains how his technical skills saved the day!

“We had court audio records that were recorded at a slower speed to capture more data. This means they play back much faster and must be slowed down in post. Oddly enough, Side B of these audiocassettes sounded like it was the same audio but backwards. This makes no sense! If they were trying to save space on their cassettes by recording at a slower speed, why would they have duplicate audio recorded to side B?

So I did some research and scrutinized the recordings. First I noticed that the 2 stereo channels on side A were clearly from different microphones. The waveforms and spectrograms were completely different. On one channel, the judge might be very loud, but the subject is quiet. On the other channel, the subject is loud and the judge is quiet. I determined they had two separate microphones feeding unique mono signals into the left and right channels.

From there, I compared Side A’s stereo channels to Side B’s stereo channels. It was easy to see the waveforms were slightly different, yet the audio was from the same court session. This meant that the audio wasn’t recorded in stereo, it was recorded in quadraphonic. This is a format usually reserved for music. I believe that in order to capture different areas of the courtroom, they had 3 to 4 microphones placed in different locations. Each microphone fed a mono-signal into its own channel. They must have had a quadraphonic cassette recorder (not very common), and all four channels on the cassette were recorded forward simultaneously.

Because we don’t have a cassette player that plays quadraphonic sound, this means that when we flip the cassette and playback side B, it comes out backwards. In order to properly capture the full audio, I capture side A in stereo, then flip the tape and capture side B backwards and in stereo. In post-processing, I reverse side B to play forwards. Then I split all of them into four mono tracks, which I slow down, lower the pitch and clean up a little. Because each track was from a different mic, they all have slightly different noise profiles. Once I’m done with post, I mixdown all four tracks into one audio file to get the accurate sound recording from the original court hearing.”