UKMusic.com - View Single Post

I'm afraid to tell you that it is impossible to extract only the vocals from a track if that is what you mean to do. The only possibility you have to get them is to mail whoever recorded the song and ask for them from the original multitrack recording.

You can take a track and minimise all other noise while emphasising the vocal part but by doing so you will inevitably reduce the quality of the vocal recording. The whole thing is somewhat analogous to cooking. If you were making soup and you poured in way too much salt, you couldn't separate the salt from the soup because the salt has become part of the soup, in much the same way as the vocal has become part of the overall recording.

If you want to try and get the vocal emphasised while minimising other noises here is one way to do it, but be careful, it's quite tricky.

Part 1 (optional, but effective):

Convert the stereo recording into an MS matrix combination, using a plugin such as "Waves S1 MS". Destroy the 'side' portion of the signal and keep only the mid. This reduces all the components of the signal that are situated in the far left and right of the stereo field, the vocal is normally in the centre.

Part 2:

Take a section of the tune where there is no vocal but the same backing sound as when the vocal is playing. Remove one loop and in a multitrack editor such as 'protools' or 'cubase'. Take this loop and align it perfectly (sample accurate) with the section where you want to remove the vocal. Then invert the phase of the added part and combine the signals together, this should cancell out much of the backing whilst leaving the vocal intact. Be aware that it is important here to align the tracks perfectly and to make sure that there are no additional noises present.

Part 3 (only required if steps one or two not done or not done well enough)

Using a complex equaliser remove those frequencies unimportant to vocal detail; loww frequencies and very high frequencies. While not very effective this may help to emphhasise the effects of stages one and two.

I hope this is useful. It will take you bloody ages to get this right but good one you if you manage it.

07-02-2002, 07:00 AM	#2
DJ Fat Zed Analogue Join Date: Feb 2002 Posts: 543	accapellas I'm afraid to tell you that it is impossible to extract only the vocals from a track if that is what you mean to do. The only possibility you have to get them is to mail whoever recorded the song and ask for them from the original multitrack recording. You can take a track and minimise all other noise while emphasising the vocal part but by doing so you will inevitably reduce the quality of the vocal recording. The whole thing is somewhat analogous to cooking. If you were making soup and you poured in way too much salt, you couldn't separate the salt from the soup because the salt has become part of the soup, in much the same way as the vocal has become part of the overall recording. If you want to try and get the vocal emphasised while minimising other noises here is one way to do it, but be careful, it's quite tricky. Part 1 (optional, but effective): Convert the stereo recording into an MS matrix combination, using a plugin such as "Waves S1 MS". Destroy the 'side' portion of the signal and keep only the mid. This reduces all the components of the signal that are situated in the far left and right of the stereo field, the vocal is normally in the centre. Part 2: Take a section of the tune where there is no vocal but the same backing sound as when the vocal is playing. Remove one loop and in a multitrack editor such as 'protools' or 'cubase'. Take this loop and align it perfectly (sample accurate) with the section where you want to remove the vocal. Then invert the phase of the added part and combine the signals together, this should cancell out much of the backing whilst leaving the vocal intact. Be aware that it is important here to align the tracks perfectly and to make sure that there are no additional noises present. Part 3 (only required if steps one or two not done or not done well enough) Using a complex equaliser remove those frequencies unimportant to vocal detail; loww frequencies and very high frequencies. While not very effective this may help to emphhasise the effects of stages one and two. I hope this is useful. It will take you bloody ages to get this right but good one you if you manage it.