• Most new users don't bother reading our rules. Here's the one that is ignored almost immediately upon signup: DO NOT ASK FOR FANEDIT LINKS PUBLICLY. First, read the FAQ. Seriously. What you want is there. You can also send a message to the editor. If that doesn't work THEN post in the Trade & Request forum. Anywhere else and it will be deleted and an infraction will be issued.
  • If this is your first time here please read our FAQ and Rules pages. They have some useful information that will get us all off on the right foot, especially our Own the Source rule. If you do not understand any of these rules send a private message to one of our staff for further details.
  • Please read our Rules & Guidelines

    Read BEFORE posting Trades & Request

Sound Subtraction

GelflingHand

GelflingHand

Well-known member
Faneditor
Messages
94
Reaction score
46
Trophy Points
28
LO! FOR I AM BECOME A GODDESS AMONGST FAN EDITORS! BOW DOWN BEFORE ME, MORTALS!!!

Well, I've done a little bit of coding that some of you might find useful, anyway :)

It's basically just a glorified macro, that divides the selected pair of tracks into tiny segments and runs Audacity's Noise Reduction effect on each segment. This allows it to be very specific about what constitutes noise at a given moment. The first track is used to get the noise profile, and the second has the noise reduction applied to it.

It's not flawless, and you'll have to mess around with the settings for the best results, but it's working adequately for my purposes. I've uploaded a sample of my results:

password: fanedit.org

If you'd like to try it out, you'll need to compile a custom version of Audacity (not as scary as it might sound).

Audacity source code

Sound Subtraction code - replace CMakeLists.txt in /src/, and add the two SoundSubtraction files to /src/effects/

Audacity build instructions - use -DCMAKE_BUILD_TYPE=Release

Let me know if you have any luck with it. You'll need a clean copy of the music to be subtracted,of course, eg from a soundtrack album. The two tracks need to be aligned as closely as possible - adjust the speed of the music track if necessary. NB the volume of the noise track matters - if it's quieter than the target track, it basically doesn't work, and if it's too loud the noise reduction can be overzealous. You might need to use different settings on different parts of the audio.
 
That's pretty impressive! Can you make a video that shows us how to use this?
Applying this to MI2 would be worth having to listen to Limp Bizkit one more time in order for it to cancel it out :)
 
A video seems like overkill; it's pretty straightforward if you're familiar with Audacity (and if not, I'm sure there are plenty of videos on that already).

Step 1: find a section of the soundtrack with no dialogue, so you can line up the music-only (noise) track with the movie track. If possible, zoom in right down to the sample level, but the waveforms might not be sufficiently identical for that. As long as they sound in sync when you play, it should be fine. Check that the two tracks stay in sync for the whole duration you want to process; adjust the speed/duration of the music track if not. Noise track goes on top, music+dialogue track underneath.

soundsub01.jpg


Step 2: select the part of the two tracks you want to process, then choose Sound Subtraction from the Effects menu.

soundsub02.jpg


Step 3: Choose your settings, and hit OK.

soundsub03.jpg


Step 4: If not happy with the results (don't forget to mute the noise track!), Undo and try again with different settings. Or if the music is gone from part of the track but not all of it, select a part that needs more processing and give it a second pass with new settings (mostly that means increasing sensitivity). Or you may need to use the Amplify effect on part of the noise track to make it louder or quieter. Occasionally, increasing Steps per Window helps, but it slows down processing a lot.

I haven't tested it on stereo tracks; you'll probably get better results doing the left and right channels separately. I was lucky enough to have a 5.1 version of the music available.
 
GelflingHand said:
If you'd like to try it out, you'll need to compile a custom version of Audacity (not as scary as it might sound).

Audacity source code

Sound Subtraction code - replace CMakeLists.txt in /src/, and add the two SoundSubtraction files to /src/effects/

Audacity build instructions - use -DCMAKE_BUILD_TYPE=Release
Click to expand...
For the non programmers out there, like myself, would you be able to create a step by step walk through on how to do this? I've tried github software and never can figurer out what to do. Is there anyway that you could make a downloadable custom/patched version?
 
DigModiFicaTion said:
For the non programmers out there, like myself, would you be able to create a step by step walk through on how to do this? I've tried github software and never can figurer out what to do.
Click to expand...
Are the steps at https://github.com/audacity/audacity/blob/master/BUILDING.md unclear? I can try to answer any questions you have. What operating system do you use - Windows? Do you know how to access the command line? If the pip command doesn't work, you probably need to download and run the Windows Installer for Python.
 
For me it's in a different language. I just don't have programming literacy.
 
Very kind offer. Thank you. Windows 10 64bit
 
This is a necro, but I found this thread because I'm trying my hand at a rescore of The Flash. Anyone still active here?
 
EDIT: Nevermind, please disregard. Wrong thread.

Please delete.
 
You must log in or register to reply here.
Back
Top Bottom