User talk:PeterSampson/Repairing bad takes

From Audacity Wiki
Jump to: navigation, search
Peter 26Sep12: Text copied here from Manual "old page" > Tutorial - Common Editing Tasks - Part 2 - Cleaning Vocal Tracks
  • Gale: 03Sep12: Could be useful for a FAQ on redoing fluffs / replacing bad bits.
    • Peter 3Sep12: Looks a bit to complex for a simple FAQ - how about making a tutorial based on this material. If you agree I'm prepared to have a stab at a first draft based on this material. But note that I'm no expert at this, it's not what I normally use audacity for :-)
  • Gale 09Sep12: I see Steve plans on to have a section on Punch-In, which is a major part of "fluff repair", but cross-fade is not necessarily essential for punch-ins even for music; I often manage even with music by overdubbing into the same selection in a new track then replacing the old selection with the new by cut and paste. I think fluff repair demands a FAQ so people have increased chance of finding it (and it must mention Append Record), but quite likely the FAQ would be little more than a pointer to one or more other pages. I'm fine with Peter and/or others working on it, though it could I think end up as Wiki Tutorial instead of Manual Tutorial.
  • Gale 26Sep12: Perhaps a better working title may be "Repairing bad takes"?
    • Peter 26Sep12: Certainly a better working-title than "Fluff repair" - therefore implemented.
      I am on holiday shortly the week after next so I won't have time to work on this properly before then - so it's earmarked for a post-vacation project.

The Situation

You have:

   Vocals, Speech or Wallas (fx made with voices)

You want to:

  • remove inappropriate breathers, coughing, bad takes, parts you didn't like and bad noises
  • use only the good parts of takes
  • construct sequences from parts of different takes

You can combine three ways of handling this:

  1. Silence stuff you don't want.
  2. Substitute it with,
    • a piece of ambience to prevent a sudden hole in the sound texture
    • more fitting breather, that works better, if you're cutting a breather away.
  3. Use the Envelope tool on the Tools Toolbar to create volume automation to pull down those parts, so they won't sound as loud in the end.

When do I use what ?

The first option is a tool, but is rarely used by itself.

In audio books, radio interviews and sometimes in dialog for a film, the vocal tracks will stand alone in many parts of your project.

Therefore the second option is general practice. Many times, the third option is used as well (Envelope tool) ///'ADD SMALL IMAGE NOT LINK??/// The holes you create with it are easier to control and change at later times, but still need to be filled up with replacement material.

When using the Envelope Tool, replacement material is usually overlayed by placing the filler material on another track right at the area in the timeline as the hole is. For this reason only <a href="menu_edit.html#silence">silence</a> unwanted audio if it's too obtrusive.

Bad takes are of course not usable, so usually you'll have to cut that stuff away.

Just remember, that <a href="menu_edit.html#cut">cut</a> and <a href="menu_edit.html#cut">delete</a> (cut without copying to the clipboard) work like cutting tape away and sticking the remaining pieces together. If you're doing things that require critical timing, you need to keep this in mind. Use the Silence function instead.

Continuous sound is the key in those cases and having the ambience, even if it's just a bit of quiet dark noise, drop away in many parts by silencing them isn't going to sound as good as keeping a steady sound texture.

Therefore filling those holes with material on another track and fading the edges to make it all sound continuous is the preferred way to do it.

There are situations, such as vocals in a song, that you'd like to sound as clean as possible.

The best way is to have a very quiet recording room or location.

The second best way is to use volume automation to get rid of any unwanted stuff. It allows you to change or take back what you did after you have made those changes to the volume curve with the Envelope Tool.

With volume automation you can mute sections of the audio without actually making any edits to the audio data.

The Three Options - a quick How To


  • Select an unwanted piece of audio.
  • Listen to it a couple of times, adjust your selection if necessary, and listen again.
  • Select Silence from the edit menu or hit CTRL+L on your keyboard.
  • Fade the edges of the audio beyond your selection to smooth it out.
  • Listen to it. If it doesn't sound good, hit Undo or press CTRL+Z on your keyboard, and try again at fading those edges.
  • Repeat this procedure for every part you don't want in your audio.


  • Silence all unwanted bits as described in the section above. Substitution expands and improves that option. Be sure to only silence the bits you don't want and don't forget to fade the edges around the silenced material. Then come back here.
  • Now that you've silenced all unwanted, create a new audio track.
  • Now find a piece of audio that'll work as a substitute for those deleted parts. This is usually a piece of ambience somewhere else in the audio track. The speaker might have made a longer pause, so look in those areas first. Select that bit and duplicate it.
  • Using the Time Shift Tool, move the duplicated ambience directly beneath the first silenced part in your vocal track.
  • Now select the entire track by clicking on Track Control Panel just beneath the Mute and Solo buttons, and duplicate it. Mute that new track. This will be our filler for the next gap in the original track.
  • Turn your attention back to the first bit of ambience, that is now right underneath the bit that we silenced in our vocal track.

Using the Envelope Tool

  • Switch to the Envelope Tool on the Tools Toolbar
  • Left-Click in to an area to create a new automation point or left-click on an existing automation point and drag to change the location of the point.
  • The automation you have written remains bolted to the audio, so it will move along with the audio when you use the Time Shift Tool.
  • An example:
    An example of the Envelope Tool in use
    Using the envelope tool - an example

From: Manual>Splitting and Submixes

Manual Cross Fade

In our example, we have a small sentence of speech, where the speaker made a pause after the first word. We'd like to eliminate that pause.

The part after the pause is selected

Track selected after the pause

Then the Split New function is used to pop the selected audio to a new track

The Track Split into two

The Time Shift tool Splitting and Submixes Time Shift Tool.png is selected and the audio on the lower track is moved left.

The Split track is moved to reduce the pause
Move Audio and select for fading

Now, it's a good idea to listen to the two tracks individually for breathing sounds for example.

Use the solo button of the tracks for this. Then listen them both in the mix. Again, you can use the solo buttons for this.

If you have a lot of other tracks playing at the same time, press the solo buttons on both tracks. There should be no over lapping or cut-off breathing sounds.

When you're satisfied, fade out the last two thirds of the overlapping upper part of the track, and fade in the first two thirds of the lower overlapping audio.

Two thirds, and not the whole overlapping audio, are chosen to keep the level of audio constant. If the whole overlapping parts were faded, you would get a level drop of 3 dB in the middle of those fades.

You can check this out by taking a piece of music, duplicating it, and then fading the tracks, one fading out, the second fading in. In the middle of those fades, the level of the mix will drop audibly. Do a fade over last two thirds for the fade out and first two thirds for the fade in, and you probably won't notice any change in level.

Two thirds is a guideline, but not the law, so you may have to experiment a little.

=== Mixing===
Splitting and Submixes Mix.png
Bits and Pieces spread all over the screen
First four tracks selected for quick mixing

Select the tracks you want to mix together by SHIFT+clicking on the Track Control Panels. In the graphics above, the first four tracks are selected.

Then select Mix and Render. In this example I have quick mixed everything down to two tracks :

Splitting and Submixes Rendered.png
Bits and Pieces, quick mixed down to two tracks