Vocal removal on documentaries

Are most documentaries easier to rip from than modern music?

I notice in many documentaries the music quiets down to accommodate narration and interviews while it keeps playing in the background.

Ok I’m a newb. I used Audacity itself plus the VoiceTrap plug-in with some fairly good results except the fact that the narration is merely distorted (maybe inverted idk) and unintelligible. It sounds like loud whispering, if that makes sense. Does this mean the narration is intertwined somewhat with the music still?

I’ve tried audacity with many other videos and audio and this has come out much better than the rest have in the past. But the staticy, loud whispering is still an obstacle. Anyway to reduce it further?

Do you want a sample of the audio?
Thanks.

Vocal Remover won’t remove stereo voice. So if the producers of the piece thought it would be a terrific idea to add a little theatrical (deep) stereo to the normally plain, centered voice, the voice will vanish, but not the stereo effect.

Also, as you noticed, unlike songs which tend to play music straight though the singing, documentaries are produced with music under. That’s what also tends to happen on radio broadcasts.

Theme.

Theme under. Open Kai. Cue!

“Hello from Los Angeles. I’m kai ryssdal and this … is Marketplace!”

Theme up. Sting!
I know you’re following all the YouTube videos that tell you how easy this is, and in fact, if you do this trick with the exact same music that the guy in the video used, yours would work, too. Success changes a great deal with the original work.

It’s also possible that they recorded the narration in stereo (pretty unusual).

The process tends to fall apart with highly compressed MP3s. Where did you get the documentary from? This works really well with original works, CDs or DVDs. Not so much from compressed downloads.

There are new tools in Audacity 2.1.1 for Vocal Isolation and Vocal Removal that don’t work the same way the older ones did.

http://manual.audacityteam.org/o/man/vocal_reduction_and_isolation.html
(new)
http://manual.audacityteam.org/o/man/vocal_remover.html
(Legacy)

Koz

Thanks for the reply. My version is up to date 2.1.1 by the way.
So straight from DVD usually works better than downloaded YouTube mps3s? Cause that’s where I get most of my stuff, downloaded directly from YouTube.

Question: Assuming I can get a hold of the DVD for the documentary, how would I be able to edit it with Audacity? Just pop it in the tray or what?

that’s where I get most of my stuff

MP3 has trouble performing post production because of the built-in compression damage. MP3 isn’t open-ended. It’s designed as an end-product. You make an MP3 > Full Stop. Not take your MP3 and make something else out of it.

Audacity doesn’t edit MP3. It imports the MP3 show in its own very high quality internal format and makes a new one when you’re done. It’s this process that can double honky, bubbly compression sound damage. If you import a 128 quality MP3 and save an edited 128 MP3, the sound quality is much closer to 64. Barely adequate for stereo.

This process also messes up delicate sound filters and tools like vocal removal. MP3 re-arranges sounds slightly to make the files smaller, so the MP3 you’re editing isn’t the same as the original show. That could be where some of your damage is coming from.


Ripping Audio CDs and Movie DVDs is done with ripping software. Sound on an Audio CD is not in formal sound files (although it seems to be). It’s in a stream plus an index system and has to be put back together for editing. The up side is the basic quality is WAV. So assuming the company didn’t make the CD from an MP3, you should be good to go. Movie DVDs can have WAV quality or compressed (or both). There’s no way to tell ahead of time.

Of course, if the production company put echo behind the voice, vocal removal still isn’t going to work.

Koz

The audio in YouTube videos is always AAC, not MP3. What your tool is doing is extracting the audio from the video then transcoding the already lossy AAC audio to the lossy MP3 format, adding another layer of quality loss and making your vocal removal job that bit harder.

Instead, download the YouTube video as the original video that is on the YouTube servers. There are numerous web browser extensions that let you do that.

Then install FFmpeg from http://manual.audacityteam.org/o/man/faq_installation_and_plug_ins.html#ffdown and drag the video file into Audacity. At that point, you now you have the audio in the same “quality” that it has in the video on YouTube.


Gale