I use kdenlive/audacity for editing my own videos (mostly mountain bike & family events). While I understand the mechanics of audio/video editing, the sound in my videos is mediocre at best. Here is a sample of my work, with music, voice and ambient sound:

Some problems I can't solve: what's the best sound levels when mixing music, voice and ambient sound? How I make a neat transition between music and voice? How I make a home-recorded voice sounds like as an open-field recorded voice (to replace part of a in-location recording)?

I searched YouTube, but most videos are about foley and/or how-to use specific programs, while I'm looking for general ideas and principles for a great sound treatement.

So, what resources you recommend?

  • Buy a separate mic. You can connect it to your main camera so you don't have to sync it manually later. Avoid the camera mic if possible. Make your own "dead cat" to cut wind noise.
    – LordNeckbeard
    Commented Jan 22, 2013 at 18:43

2 Answers 2


Get a decent quality sound recorder and directional mics with good windscreens if you want to get good audio on location. Balancing music and vocals is going to be a trick you'll have to play by ear. You can apply things like normalization and compression if the audio for voices tends to get softer and louder. Either of these approaches will shrink the dynamic range (the difference between the quiet and loud parts) and make it so that music is not going to be too soft in some parts and too loud in others. You might also have to normalize the audio, though most music you get on CD is already normalized and compressed during mastering.

Really, looking for information on mastering in general would probably be helpful for you.


Your hardest question is: "How I make a home-recorded voice sounds like as an open-field recorded voice?"

Below is also a workflow on cleaning audio recordings. But first to the home/field recording question:

Before decreasing the quality of your home-recorded voice, try by all means to increase the quality of your field recording (for example by using a small shotgun mic with a windshield pointing from your bike handlebar to yourself). Here and here are some comparisons of shotgun mics.

There will still be a noticeable difference between home and open field recordings. To deal with this I have two practical ideas:

  • Could you ask a woman to speak the home recorded commentary? This would make your story more appealing, and nobody would be irritated about the different recording qualities, because they're natural.
  • If you have to speak everything on your own, I wouldn't do these recordings at home, but on a separate bike tour. Then you have the same sound quality for free.

Now, for your question about editing your audio recordings, I would like to expand the answer given by AJ Henderson.

This is my personal workflow that I use in Audacity:

1. Preparation: If you look at the waveform of your recording, you will see a couple of clippings (amplitude at the maximum). This can be the wind, but also a truck passing by. If this is a part that you will not use later, select it roughly and choose Generate/Silence in the menu (don't delete the selected part, as this would make your video out of sync). If this is a part that should be included in the final video, take more care in selecting the part, and then choose Effect/Amplify. There, enter a negative value so that the marked part will be at the level or below of the voice.

2. Noise reduction: If the noise is disturbing, select some noise without voice as reference (1 second or more). Then choose Effect/Noise Removal/Get Noise Profile. Now mark a part with voice, and select Effect/Noise Removal/Preview. If the result of the preview is okay, remember the values in the "Step 2" section, press Cancel, now select the whole track, and do the Effect/Noise Removal again, enter the remembered values, and now click OK. Attention: Sometimes this effect worsens the quality. Then I just skip the noise reduction.

3. Compression: If applied well, the compressor brings a loud voice and a silent voice to the same level, while leaving the noise untouched. You know that in every sentence there are loud and less loud words. This can irritate the listener, especially when we turn our heads away form the mic. To do the compression, select again a representative part of the track, then click Effect/Compressor. When the preview sounds okay, remember the values, press cancel, select the whole track, go to Effect/Compressor again, and now click okay. The option "Make-up gain for 0db" should be enabled.

4. Normalization: Now your track will have the maximum level of 0db, which is too much. Perhaps later-on you will like to amplify certain words or passages and need some headroom for this. As for my projects, I decided to generally bring all my voice recordings to -3db. Some TV recommendations demand -9db, whereas every YouTube user chooses his own, going up until 0db. That's why we have to regulate the volume so often while switching from one YouTube video to the other. So, in short, -3db is somewhat in the middle. To normalize, choose Effect/Normalize, and enter -3db in the effect dialog.

(Important: Please save your project after every step into a new file. If you do mistakes, you will only recognize them in the video editor, and then you will be happy if you don't need to redo everything.)

Finally, you can export this cleaned track as a WAV file. When you now join different recordings to one new clip in your video editor, they should better fit together.

Your Answer

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