| www.ethanwiner.com - since=20 1997 |
The Truth=20 About Vocal Eliminators
by Ethan = Winer
(This article first = appeared in=20 August, 1999 in ProRec, the=20 online audio magazine.)
Another fine product is the Thomson = Vocal=20 Eliminator, which is a hardware device and thus more expensive than = Vogonetm. You can learn all about the TVE, and = download demo=20 MP3 files, at the LT = Sound web=20 site.
Another option, if you're adventurous, is = Center Cut which is a public domain utility that = claims to=20 remove (or isolate) vocals in the center while maintainting the left and = right=20 stereo channels. The link above is to a page that describes the = algorithm, and=20 you can download the program in the pane at the left.
Yet another possible method is to use a = stereo=20 receiver that can decode Dolby ProLogic surround. Even inexpensive = receivers=20 have this feature, which also can remove or keep the center while = maintaining=20 the left and right channels. To use a Dolby-capable receiver you'll set = it to=20 ProLogic while playing the music, then record from the left and right = line=20 outputs to keep the music only. Or record the center channel to keep the = vocal=20 and discard the left and right sides.
Finally, the best way to obtain = sing-along tracks=20 is to buy them from a Karaoke provider. These are not the original = performances,=20 but many companies offer very high quality tracks for thousands of = popular=20 songs. Here are a few Karaoke type product sites, and Google will find = you many=20 more:
PocketSongs
For many years the back pages of audio and = recording=20 magazines have featured ads for hardware devices that claim to remove = vocal=20 tracks from a stereo recording. Lately, several audio editing programs = have also=20 claimed to offer a vocal remover feature. Is this possible? Is there = really a=20 magical way to remove the lead vocal entirely from a commercial = recording to=20 create your own instant Karaoke backing tracks? The short answer is No.=20 Sometimes a vocal can be removed almost completely, but just as often = the=20 results are disappointing. In most cases you'll be able to = reduce the vocal level, but some audible remnant of the original performance will = probably=20 remain. Further, any process that changes the vocal track is sure to = affect the=20 other instruments as well. In this article I will explain what vocal = removal is=20 all about and how it works. I'll also describe the procedure and show = how to do=20 it yourself using common audio editing tools.
HOW VOCAL REMOVAL=20 WORKS
You can reduce the level of a vocal (or other = lead=20 instrument) in a stereo recording by taking advantage of how vocals are=20 generally recorded: in mono and placed centered in the mix. Since the = vocal=20 track is present in both the left and right channels equally, you can, = in=20 theory, remove it or at least reduce its level by subtracting one = channel from=20 the other. Instruments panned away from center will not be removed, = although the=20 tone of those instruments will probably be affected. The basic procedure = is to=20 reverse the polarity of one channel, and then combine that with the = other channel. Any content that is common to both channels will thus be = canceled,=20 leaving only those parts of the stereo mix that are different in the two = channels. Reversing the polarity of an audio signal means that the parts = of the=20 waveform having a positive voltage are made negative, and vice versa. = (This is=20 often incorrectly called reversing the phase.) One important drawback = inherent=20 in vocal removal is that, by definition, it reduces a stereo mix to = mono. Since=20 you are combining the two channels to cancel the vocal, you end up with = only one=20 channel. However, there are ways to synthesize a stereo effect = afterward, and=20 that will be described later.
Important = note added=20 November 21, 2002: You cannot remove vocals = effectively if=20 your source is an MP3 file. In order to remove vocals, the vocals in the = left=20 and right channels must be exactly identical. Then when the = polarity is=20 reversed in one channel and the channels are combined, anything common = to both=20 channels - what's panned in the center - is cancelled. But MP3 encoding=20 processes the two channels separately, so they are not identical enough = to=20 cancel.
It is impossible to completely remove a vocal or = reduce=20 its level, without affecting other instruments in the mix. First, even = though=20 most vocals are placed equally in the left and right channels, stereo = reverb is=20 usually added to vocal tracks. So even if you could completely remove = the raw=20 vocal itself, some or all of the reverb is sure to remain, leaving an = eerie=20 "ghost" image. If you plan to record yourself singing over the resultant = track,=20 the new vocal can have its own reverb added, and you may be able to mix = your=20 voice loud enough to mask the ghost reverb from the original vocal = track.=20 Another limitation arises because vocals are not the only thing panned = to the center of the mix. Usually, the bass and kick drum are also smack in = the=20 middle, and those get canceled along with the vocal! However, you can = minimize=20 this problem by rolling off the lowest bass frequencies on one channel = before=20 combining it with the other. Since one channel now has less low end than = the=20 other, the low frequency instruments will not completely cancel. In = fact, of the=20 software programs I've seen that offer a vocal removal feature, none = alter the=20 low end on one channel before combining, so the bass and kick are = eliminated=20 along with the vocal.
I developed the following procedures using two = different=20 types of music. One is a tune from a friend's self-produced country = music CD;=20 the other is a cello concerto I wrote and recorded in my home studio = using live=20 classical musicians from a local orchestra. I created excerpts of these = pieces=20 in the popular MP3 format and they are available here for downloading. = This way=20 you can compare the original recordings with the processed result, to = see for=20 yourself how well vocal elimination works in practice.
STEPS FOR REMOVING=20 VOCALS
The most basic procedure is to load a stereo = Wave file of=20 the original song into an audio editor program, flip the polarity of one = channel=20 and lower the bass level somewhat, and then combine the left and right = channels=20 into a new, mono track. I use Sound Forge 4.5 from Sonic Foundry, which = includes=20 all the tools needed to manipulate audio files this way. Most other = 2-track=20 audio editors have similar capabilities, and this technique will apply = to those=20 programs as well. Sound Forge lets you load a single stereo file, = manipulate the=20 left and right channels separately, and then combine them to mono all = within one=20 edit window. But for these instructions, I split the channels into = separate=20 files to make each step easier to follow.=20
* See the notes added at the end of this=20 article.
It is possible that combining the two channels = will exceed=20 0 dB, and you will need to reduce the level of both channels a = few dB.=20 If you lower only one channel, the two channels will not combine = equally, and=20 the vocal level won't be reduced as much as possible. To roll off the = bass=20 frequencies, I used Sound Forge's Parametric EQ in the high-pass mode = set for 20=20 dB of cut starting at 200 Hz. (This filter setting affects the lows, so = why does=20 Sonic Foundry call it high-pass rather than low-cut?!) If you use Sound = Forge,=20 be sure to select the highest accuracy filter mode, since how quickly = the EQ is=20 written to the file is less important than having the filter perform = exactly as=20 you ask it to. Besides cutting the extreme low end on one channel, you = can=20 optionally reduce some of the highs too. This lets you retain strings = and=20 cymbals and other instruments that have treble content and are centered = in the=20 mix. In general, you can cut those frequencies that are outside the = vocal range=20 - for male singers you need to start the roll-off at a lower frequency = than for=20 females. Remember, the frequencies you cut from one channel are the ones = that=20 will not be canceled when you reverse the polarity and merge it = with=20 the other channel.
A BETTER = WAY
Rather than use a typical stereo audio editor = program, a=20 much better approach is to separate the left and right channels into = separate=20 files and load them into a multi-track audio recording program. The main = advantage is that you can more easily adjust the channel levels to fine = tune the=20 process for the most complete vocal cancellation. This also lets you = experiment=20 with different high and low frequency turnover points, assuming your = multi-track=20 software offers EQ for the tracks. Start with just the very lowest and = highest=20 frequencies removed, and then slide the cut-off frequencies closer to = the middle=20 until the vocal starts to leak through. Again, you are combining the two = mono=20 tracks at approximately equal levels - but with the polarity reversed, = and the extreme highs and lows rolled off on only one channel. I use SAW Plus, = which=20 has EQ and polarity reverse effects built in. These effects are = non-destructive=20 and can be adjusted in real time while the left and right channel Wave = files are=20 playing. So all I had to do was extract the Left and Right files from = the=20 original stereo Wave file, load those into separate tracks in SAW, and = add=20 polarity reverse and low-end shelf cut at 200 Hz to the left channel. = Once you=20 are satisfied that you have removed as much of the vocal as possible and = with=20 minimum damage to the rest of the track, save the mix to a new Wave file.
One useful tip is to reduce the number of = playback buffers=20 if your multi-track recorder software allows that. Normally, the more = buffers=20 you have the better because that avoids "stuttering" when playing back = many=20 tracks at once. But the trade-off is that more buffers yields a longer = time lag=20 between when you change a volume level or EQ setting and when you hear = that=20 change. So when working with only two mono tracks for removing vocals, I = set SAW=20 to use the minimum number of buffers, thus making my mix changes audible = immediately.
Earlier I mentioned that removing vocals always = yields a=20 mono sound file because the left and right channels are combined as part = of the=20 process. There are several ways you can synthesize a stereo effect to = recreate=20 some of the lost ambience. I used the BlueLine series of plug-ins by = digilogue,=20 available in a fully functional shareware version ($35 to purchase) from = the=20 author's web site at http://www.digilogue.de/. These plug-ins are = provided in the=20 universal DirectX format and also as VST versions for use with = Steinberg's Cubase. I used the BlueLine Stereo plug-in, which did a great job of = recreating=20 a stereo effect on the mono result files.
You can also create a fake stereo image using=20 equalization. Split a mono track into two identical left and right = channels, and=20 then equalize each side differently. One method is to apply a 10-band = graphic=20 equalizer to each channel, and then boost and cut alternate bands on = each=20 channel. That is, on the left channel you apply 6 dB of boost at 62 Hz, = the same=20 amount of cut at 125 Hz, boost at 250 Hz, and so forth. The right = channel is=20 then cut and boosted by the same amounts, but at the frequencies = opposite the=20 left channel: Where the left channel is boosted the right is cut, and = vice=20 versa.
Two final items are worth mentioning. First, if = your=20 multi-track software requires DirectX plug-ins for EQ and polarity = reversal, the=20 inherent delay will prevent the desired cancellation and all you'll get = is a=20 phased sound with the vocal still present. In that case you should = reverse the=20 polarity and roll off the low end in a stereo editor that writes = directly to the=20 file, and load the result back into your multi-track recorder. I'll also = mention=20 that it is possible to cancel a vocal from a stereo file while keeping = the=20 original stereo image. If you create a mono Wave file that is a simple = mix of both the left and right channels, you can reverse its polarity and mix = it with=20 the original stereo recording. This cancels the vocal and other centered = instruments, and reverses the left and right channels as a side effect. = Although=20 this should be superior to my method of reducing the mix to mono, in = practice it=20 did not work as well. More of the vocal leaked through, and the = non-centered=20 instruments were partially canceled.
THE BOTTOM=20 LINE
Does vocal removal really work? Is it worth the = effort to=20 even try? I'll leave that for you to decide. Following are two pairs of = MP3=20 clips containing Before and After versions of my attempts. The first = piece (265=20 KB for each MP3 file) is Rollin' from the CD 20 Years = Late by=20 Tom Schulz. Click here to download a 34-second MP3 clip of the original recording, and click here for the result after removing the lead vocal track. The second = selection is=20 from my Concerto for Cello and Orchestra in A minor (313 KB per = file).=20 Click here to = download a=20 38-second MP3 fragment of the original, and here for the version with the solo cello removed from the track.
Both of the After tracks were processed in SAW = Plus as=20 described previously, and then a stereo effect was synthesized using the = BlueLine Stereo plug-in. I rolled off the lows starting at 200 Hz, but = didn't=20 bother experimenting with the highs. As you can tell I was quite = successful=20 removing Tom's lead vocal, mostly because so little reverb was added to = his=20 voice. In fact, before I rolled off the low end on one channel to bring = back the=20 bass and kick, the vocal was practically inaudible. All that remains now = is a=20 muffled hint of his voice. Of course, the bass and kick have lost = definition in=20 the process, since all but the deepest components were canceled along = with the=20 vocal. With the cello recording you can clearly hear the ghost reverb, = and the=20 beginning passage also leaks through because those notes are lower than = the 200=20 Hz cut-off point. I could have lowered the EQ frequency, but that would = have=20 removed more bass content from the rest of the track.
* Added = November 14,=20 2004: I've been getting a lot of emails asking how = to=20 reverse the polarity and roll off the low end of one channel in Sound = Forge.=20 Here are the specific steps using Sound Forge version 6:
Double-click in the upper portion of the Wave = file view to=20 highlight the entire length of just the left channel. If both channels = turn dark=20 you didn't have the cursor low enough when you double-clicked. Then from = the=20 Process menu select Invert/Flip. Next, apply a low frequency shelf to = roll off=20 below 200 Hz. From the Process menu select EQ, and choose Paragraphic = from the=20 sub-menu. Check the box at the bottom labeled Enable Low Shelf, then = either=20 adjust the slider at the right until the display reads 200, or simply = type 200=20 in that field. Finally, move the smaller slider all the way to the left = until=20 the display reads -Inf.
Entire contents of this = web site=20 Copyright =A9 1997- by Ethan Winer. All rights reserved.
------=_NextPart_000_0000_01CF392E.97163E90 Content-Type: application/octet-stream Content-Transfer-Encoding: quoted-printable Content-Location: http://pagead2.googlesyndication.com/pagead/show_ads.js (function(){var p=3Dthis,aa=3Dfunction(a,b,c){return = a.call.apply(a.bind,arguments)},ba=3Dfunction(a,b,c){if(!a)throw = Error();if(2