What good (Windows) software is there to remove singing from songs? It's for a bloke I know who wants it for karioke, and he's not up on computers, so preferably software that's easy to use.
What good (Windows) software is there to remove singing from songs? It's for a bloke I know who wants it for karioke, and he's not up on computers, so preferably software that's easy to use.
Thanks for any answers.
er, yeah a bloke you know. he he.
i bet you could download karioke songs just like you can for the real thing.
they must come from somewhere, and i bet someone will have uploaded them to the net. you'd be able to find the popular songs that people do. but if you, erm i mean your mate wants specific songs to sing along to then i can't help you.
I know, I'd suspect the same, if someone else used the "it's for a friend of mine" line. But in this case it's true, and if it was for me, then I'd probably just download the songs I liked, from the same cough sources cough that provide any other music on the 'net, instead of messing about with conversion programs.
I suspect if such software exists, it will cost $BIGNUM.
Absolutely. However, in making remixes of other people's stuff, I have noticed that a lot of stuff produced recently seems to be fairly easy to filter. For example, one of my last remixes was of a Jamiroquai tune from their last proper album (Dynamite), and all I was using was the "AnalogX Vocal Remover" plug-in for SoundForge which worked surprisingly well. In the past, I've found this plug-in to be fairly useless, but there are certain mixes/productions that it seems to be able to filter through quite well.
But yeah, beyond that, you're looking at parting with a fair wedge of dough, unfortunately!!
and all I was using was the "AnalogX Vocal Remover" plug-in for SoundForge which worked surprisingly well. In the past, I've found this plug-in to be fairly useless, but there are certain mixes/productions that it seems to be able to filter through quite well.
These things are probably a bandstop filter, perhaps with some code to make a guess at the frequency band where the voice lies. If the frequency range of the singing doesn't overlap the instruments too much, a bandstop filter will do the job.
But for everything else where a significant amount of the instrumentation is in the same frequency range as the singing, you'll also take out a big chunk of the instrumentation, or you'll not take out all the singing. To get better than that, I suspect you're going to need AI. Or, I suspect, use the intelligence of a human, for instance, playing the sung melody on a keyboard that's connected to the computer, at the same pitch so the computer has a fighting chance of knowing what part of the sound is voice.
Comments
er, yeah a bloke you know. he he.
i bet you could download karioke songs just like you can for the real thing.
they must come from somewhere, and i bet someone will have uploaded them to the net. you'd be able to find the popular songs that people do. but if you, erm i mean your mate wants specific songs to sing along to then i can't help you.
scottie might know, he likes karioke.
:-P
I know, I'd suspect the same, if someone else used the "it's for a friend of mine" line. But in this case it's true, and if it was for me, then I'd probably just download the songs I liked, from the same cough sources cough that provide any other music on the 'net, instead of messing about with conversion programs.
http://www.youtube.com/watch?v=C_TR4po1Y9g
If you want software that does it well though, I agree with Winston.
Olanzapine will do the trick ;)
Absolutely. However, in making remixes of other people's stuff, I have noticed that a lot of stuff produced recently seems to be fairly easy to filter. For example, one of my last remixes was of a Jamiroquai tune from their last proper album (Dynamite), and all I was using was the "AnalogX Vocal Remover" plug-in for SoundForge which worked surprisingly well. In the past, I've found this plug-in to be fairly useless, but there are certain mixes/productions that it seems to be able to filter through quite well.
But yeah, beyond that, you're looking at parting with a fair wedge of dough, unfortunately!!
These things are probably a bandstop filter, perhaps with some code to make a guess at the frequency band where the voice lies. If the frequency range of the singing doesn't overlap the instruments too much, a bandstop filter will do the job.
But for everything else where a significant amount of the instrumentation is in the same frequency range as the singing, you'll also take out a big chunk of the instrumentation, or you'll not take out all the singing. To get better than that, I suspect you're going to need AI. Or, I suspect, use the intelligence of a human, for instance, playing the sung melody on a keyboard that's connected to the computer, at the same pitch so the computer has a fighting chance of knowing what part of the sound is voice.