I received a voiceover in WAV from a studio, where the sound engineer may have done cut and paste to construct the last part of a sentence that the VO artist missed. The result has unnatural gaps and sounds weird. I tried a few ways of cutting the gaps and fading in the pieces to make it sound passable but I can only make it sound worse. Is there any way to fix this? E.g. in the sample, the way she says 'on your own' is very halting and stands out too much.

I can't upload the whole file due to NDA.

Link to a WAV file with the problem part, a few seconds long

  • 1
    Oh dear, looks like Premier didn't respect In/Out duration while exporting. I'll change the file. Commented Aug 1, 2023 at 5:37
  • 1
    I know from experience, it's extremely difficult to "mend" something like that. When the words don't flow together, barring some dedicated newfangled granular synthesis algorithm, it can't be done. Look at how much work has been put into Alexa and Siri, and they still don't sound natural. When you imitate them, you put on an unnatural rhythm and cadence. It's just so much easier to notice the error and re-record straight away. That's why I quit as soon as I heard it lol.
    – n00dles
    Commented Aug 2, 2023 at 11:28
  • 1
    @n00dles thanks for giving it a look anyway. Appreciate the input. Commented Aug 2, 2023 at 14:52
  • 1
    I was bored, I gave it about 20 minutes. I got the flow right, but by that time it was getting very robotic on the pitch/formants, so I just gave up.
    – Tetsujin
    Commented Aug 4, 2023 at 10:17
  • 1
    Depending on how important it is and how much time you have to spend, you may need to consider finding the right phonemes in other spots in the recording and do a manual concatenative synthesis. I had to do this before with a standup comedy album where the talent messed up a really important joke in both shows that were recorded, so I had to get creative to make it sound plausible and not pull the listener out of the experience. Can the words change to something which means the same? Commented Aug 21, 2023 at 3:48

2 Answers 2


There are different tools to achieve this. One is Newtone in FL Studio. You can drag the notes of individual setences and shift them.

Here is a demo

You could even shift the words horizontally if you do not like the break

  • I have no familiarity with FL studio, so you'll have to forgive my ignorance. It looks to me like you cut up the sentence and are creating a custom pitch curve for each word. Is that correct? The end result does not sound noticeably better to me in terms of how unnatural the constructed line sounds, but I'll try and tinker with this next time I have the chance. Commented Aug 23, 2023 at 11:46
  • Newtone automatically chops each word (or syllable) into an each note that you can then drag along the piano roll. I demonstrated the cutting also in the video but realized later that it was not necessary since the word was cut automatically correct
    – Jimmy Jon
    Commented Aug 23, 2023 at 14:11

Use pitch correction software / resample (pitch shift) undesired parts and match them to rest of your audio as you wish.

  • This is a bit like a recipe for spaghetti bolognese that starts, "Get some spaghetti & some bolognese sauce…"
    – Tetsujin
    Commented Aug 1, 2023 at 17:18
  • @kubinka0505 Could you clarify/ expand on your answer? Any more detail or further explanation on how to go about it would be useful. Commented Aug 2, 2023 at 9:34

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.