How to Listen To Your Audio And Know When It Sounds Unnatural

By Holly J. McDede

Listening For “Bad Edits” 

When you’ve finished your mix, you shouldn’t be able to listen back and hear the edits you made along the way. In other words, there shouldn’t be anything in your audio file that sounds “unnatural.” When you do listen for quality control, here’s your checklist of what you’re listening for. 

If the answers to any of these questions is YES, you’ve made some bad edits. 

  1. Do I hear “double breaths” or does it sound like someone is missing a breath? 
  2. Are the levels consistent? 
  3. Are there parts that sound jarring, like an edit?
  4. Are there parts where the words or speakers sound unnaturally close together or abrupt? 
  5. For music beds underneath speakers, is it difficult to hear and focus on what the speaker is saying? 
  6. For music, are my fades too sharp or abrupt? 

Once you’ve completely finished your edits, you should listen back to your piece twice. 

First time around: Content 

Second time around: Tiny Technicalities

And then, you should listen again to see if the piece sounds okay based on the things-to-watch-out-for listed above. For this second listen, absolutely DO NOT think about the content and the words being said. You’re just trying to hear if anything sounds bad. Listen to the whole piece, writing down time codes (the minute and seconds in the segment) whenever you hear anything that sounds bad. When you’re done, go back and clean up those edits. And then listen again. 

Order of Operations 

  1. Edit for content 
  2. Get the order of your interview or story where you want it to be 
  3. Then, work on making cuts/edits to the interview or story 
  4. Listen for content 
  5. Listen for bad edits 
  6. Make bad edits better
  7. Listen again
  8. Repeat 

Breaths 

GROUND RULES: 

Don’t cut off people’s breaths, because people breathe! So if you take away their breath they won’t sound like a (normal) person 

First we hear some sentences with the breath kept in and then with it removed:

Breath Example 1

Avoid hearing two breaths next to each other 

This is an example of two clips put together. The first clips keeps its breath at the end and the second clip keeps its initial breath. This is too many! Just use whichever one sounds best and get rid of the other.

Breath Example 2

  • When making cuts, try to do so after a breath or after transition words like ‘but’, ‘and’, and ‘so’ 

Ummmms

Ums are similar to breaths in that, unless the um’s are super distracting, you should try to avoid cutting them. People say “um” when they speak. It’s natural. So don’t obsess over getting rid of all of them. 

That said, there are some instances where um’s can be distracting! If there are multiple in a row, or a sentence has 30% um’s, it might be time to remove one or two “easy um’s” (um’s with silence on either side).

There are um’s you can’t make sound better, or that are more or less attached to another word. You can try, but just leave it if you don’t succeed. Um’s can help us stay true to the actual pacing and sound of someone’s speech, so don’t think they are something to always be erased!

Upcuts 

  • You want to avoid what are called ‘upcuts’. Upcuts are audio that has been clipped, meaning is missing the beginning or end of a sound element. Often this happens when you cut a speaker mid-sentence.  It’s called an upcut because the person’s voice is going up in tone.
  • To avoid upcuts, try making your edit at the end of a phrase or sentence, which is when the speaker’s tone is likely to be going down

Upcut Example 1: 

Upcuts can also happen when you leave off the final sound of a word. This is common with final consonants. Listen to how the “k” got cut off of “back” and the “s” got cut off of “edits.”

Upcut Example 2: 

Sometimes it’s easier to see this visually. You can see how the K and the S are separate from the rest of the word and might get chopped off: 

Room Tone 

Room tone makes it way easier to make certain edits. Room tone is basically the sound of the room or area where the interview has taken place. You use room tone to smooth out transitions, particularly if the interviewee and the interviewer are in different rooms.

Room Tone Example 

Levels 


Above all, be consistent with your levels. You can see a level meter at the bottom of the session. Depending on which version you. have of Hindenburg, you’ll want your levels between -15 to -12 decibels.

Transom outlines the rules like this:

  • Traditional public radio: -24 LUFS
  • Podcasts: -18 LUFS

 



This guide was created by Holly J. McDede based training at KALW public radio, with voicing and fact checking assistance from Kristin McCandless. Holly is a reporter and producer at KALW public radio and KQED.