I'm trying to see if anyone knows of any voice isolation filters or tools. I'm trying to make the background music quieter so the voices from the anime can be used in a trailer, the music from the animes tracks gets mixed with the trailers music mix and... Well, sounds like crap.
Little help?
Voice Isolation
-
DJ_Izumi
- Joined: Wed Oct 03, 2001 8:29 am
- Location: Canada
- Contact:
-
Rozard
- Joined: Wed Oct 31, 2001 10:39 pm
Well, you could use a vocal cutting option in an audio editor, then inverting the channels and pasting it back on the original audio, thus negating everything but the voices.
RichLather: We are guests of this forum, and as such we do not make the rules.
BishounenStalker The freedom to suck is what makes the Internet rock.
-
maciko
- Joined: Thu Jul 14, 2005 12:40 am
- Contact:
What you are trying to achieve is very much outside of what is actually possible to achieve. Unfortunately extracting Voice from music is like trying to separate yellow and red paint out of orange paint. However I can perhaps give you some clues as to how to approach you problem and make it sound a bit better.
The first step is to remove as much music as possible, considering that most of vocal content will be in the center of the stereo image you can start with removing the stereo part of the sound. To do this you need to do is add the L channel to the R channel thus crating a mono file.
The next step is to roll odd any unnecessary frequencies outside of the range of human speech. So let’s assume that we will look at a bandwidth of 200Hz to about 8 kHz.
Using hi and low pass filters you would eliminate the unnecessary frequencies.
You could also gate or edit the music between the speech, where your result would have unwanted music only during the speech and not in between.
You could always try to match the pitch so that both the undesired music and the overlaid music has is in the same key or at least a friendly harmonic of one of the fundamental frequency.
By matching the frequencies you will make the interference more “musical” and therefore less intrusive. You could always consider matching time signatures and tempos.
With a bit of editing some help from natural masking you should be able to make some headway with your project.
The first step is to remove as much music as possible, considering that most of vocal content will be in the center of the stereo image you can start with removing the stereo part of the sound. To do this you need to do is add the L channel to the R channel thus crating a mono file.
The next step is to roll odd any unnecessary frequencies outside of the range of human speech. So let’s assume that we will look at a bandwidth of 200Hz to about 8 kHz.
Using hi and low pass filters you would eliminate the unnecessary frequencies.
You could also gate or edit the music between the speech, where your result would have unwanted music only during the speech and not in between.
You could always try to match the pitch so that both the undesired music and the overlaid music has is in the same key or at least a friendly harmonic of one of the fundamental frequency.
By matching the frequencies you will make the interference more “musical” and therefore less intrusive. You could always consider matching time signatures and tempos.
With a bit of editing some help from natural masking you should be able to make some headway with your project.