The Agency Record Blog

EXAMPLE: Optimal Recording Levels

A client of mine in Atlanta, GA recently recorded a phone conversation on her computer. She used an inexpensive phone bridge device, and recorded the signal using Audacity, a very useful free software recording application. The same call was also recorded using an online conference call service. She sent me both of the resulting files, asking me which one would be better to use for its intended purpose (a downloadable MP3 file for her customers). Since I would be the one editing together the file, adding intro/outro clips that help to professionally brand her products, she wanted me to have the benefit of choosing the best source recording.

I loaded up both files in my favorite professional audio editing suite, and examined the waveforms. Before I ever pressed play, I knew which recording I would be using (see illustration below, followed by an explanation).

waveformcomparison2.gif

The top waveform (A) was made using Audacity, and the levels were too high for my client’s microphone and voice, resulting in the distortion of the audio signal. Waveform B was automatically recorded by the web conferencing site, and while the audio is much more consistent across the entire recording, the levels were too low.

So which one did I know I wanted to use?

Recording B, unequivocally. The levels were low, but raising the overall volume is much easier and results in a much more coherent audio track than trying to take a distorted loud track and remove the distortion. Once distortion is introduced into the recording, it is very hard to remove it without creating other nasty side effects. Signal level distortion can happen either because the microphone input levels were set too high or the equipment being used to interface with the phone system was mismatched or unable to handle the dynamic ranges of both ends of the conversation (there is usually a disparity between the level of your voice and the voice of the other party when recording).

So what steps can be taken to find the ideal middle ground, the perfect recording that isn’t so loud it causes distortion, but not so soft that the volume has to be artificially raised or “normalized” (a process that usually also raises the volume of other extraneous noises unnoticed before in the recording)?

  1. Do several sample recordings, and closely watch the recording level indicators. If they jump into the red area of the scale, lower the recording level in the software you use.
  2. It is just as important to listen to the recording, too. Just because your level indicators aren’t peaking into the red area, the phone bridge device you are using might be distorting the signal before it ever leaves the unit on iots way to the recording software. If the device is adding distortion (either when you speak or the other party speaks or both) you will need to see if the device has a way to adjust the signal strength. Some do ($$$$), but most don’t ($).
  3. Another option, and one I would strongly encourage you to explore, is to get off the phone company’s lines and use your high-speed internet connection to make the voice call. This is not an expensive option when you use Skype to place the call, and a simple recording tool such as PowerGramo to record the call. I will cover these two tools in more depth in a later post (let me know in the comments if you would find such info helpful).

Obviously, there is quite a bit more technical information that could be offered here, but I like to keep things simple on the blog. If you have any questions, please let me know!

No comments yet.

Leave a Reply