bjork remixers
last updated: february 6, 2010
i frequently get questions on how to do remixes, but since i don't
really remix myself i'm not able to answer them. i've asked some
frequent contributors to provide some basic info and some answers to
common questions. big thanks to Petit Volcan, r-man, and dark jedi.
1 what software do you use?
Petit Volcan:
For nearly all of my remixes, I used to work mainly with Soundforge and ACID.
Soundforge is a great software to edit samples. ACID is a very simple software allowing you to loop your samples, slice them, put realtime effects, pitch, stretch and mix the whole thing together - well at least it used to be 4 years ago, so I guess now it is even better.
r-man:
Logic Platinum (in Mac and now in Windows "ambient") and Peak (in Mac)
or Goldwave (in Windows) as aid to modify the samples.
dark jedi:
I'm using Logic Express, which is a cheap Logic Audio :
very convenient, easy to learn and powerful.
But it's difficult to find free plug-ins to work with. I've chosen
this sequencer because it's Apple's one (I'm Mac user, so I try to
avoid problem by using 100% compatible stuff), and I had a lot of fun
and tracks made with a demo of Logic Audio, so I had some habits.
I'm also using an audio-editor, which is called "PEAK", a sort of
SoundForge, for editing pitch, length of loops, or edit the whole track
(for premastering i.e.).
Sometimes I create rhythmloops or basslines with reason or rebirth. Fast
& easy. And then, import them in Logic.
2 if you've tried out other software, what were the pros and cons?
Petit Volcan: I had a look at Cubase, but it was too complicated as I was just starting in this field. I suggest new remixers to focus on simple tools first.
r-man:
I've tried Cubase. But Logic is more more better (for me)!
3 so how do you extract vocals?
Petit Volcan:
That's the question ! :)
Remember most of the time, you don't get the vocals of the whole track. All you can do is pick up a few words here and there. When you know which track you want to remix, get to work :
Find a version of this track on which there is very little beats - for example, there is an a cappella version of Hidden Place on the first single release. That's a good example.
Sometimes, when you are lucky, you can simply use the track as is - all you have to do is play on it ! But it's quite rare.
So the big question : HOW TO REMOVE BACKGROUND MUSIC
To do so, I used to work with a software called Cool Edit [which is now called Adobe Audition] (but you can do it with others, I guess). On this software, there is a tool to suppress noise on a track (the *shhhhhhh* you can hear when there is no sound). What you do with this tool is :
1. opening your file
2. selecting a blank part of your track, when you only hear the noise
3. run "analyse" or something. It will determine the spectrum of the noise
4. select the whole track
5. run "process" or something. It will substract the noise spectrum from the whole track
So, to remove background music, all you have to do is to select a part of the song on which you only hear the music and not the voice ! It will calculate the spectrum from this and remove it from the whole track.
DISCLAIMER : this is *NOT* magic. The result will be really ugly, I warn you. Most of the time, instrumental and voice are overlapping on various frequencies. Therefore, when you "clean" your track, you will also delete some frenquencies used by Bjork's voice.
If you wanna hear how uggly it can be, just listen to my "Enjoy : Speed In There" mix. There are no effects on the voice : only extraction artifacts.
THE HIGH-TECH WAY
Another nifty way is to used 5.1 mixed albums. On those albums, the mix is not on two channels (stereo) but on 5 - therefore, there is usually one channel on which you hear almost only the vocals. I've never done it, but I remember someone did it with the DVD of Dancer In The Dark. During the songs, the singer's voices were on a separate channel. This way we had the film vocals to remix.
r-man:
I usually choose songs "a cappella" or arranged with few instruments.
Or, I equalize the song in "telephone" mode to cut hi and low frequences
or emphasize the Hz range of the voice.
dark jedi:
in fact I don't extract vocals. First, I look
for "acapellas" : some people have ripped the 5.1 version of the tracks
(now all bjork's album are 5.1!) and extract only the center channel,
that contains only the voices and some kickdrums or basses. You just
have to filter (highpass filter) under 200 hz and you have your
accapella !
But sometimes when I have no accapella version, I work with the whole
album track, simply adding my stuff after filtering, adding effects on
the original(vocoders, reverbs, delays, distortion...). I work with
parts of existing remixes too (for exemple, my oceania remix was made
with part of "piano & vocal version", "kellis mix", & the"album
version", plus my beats, basses, and re-arrangement).
But with this method, mixing and EQing are much harder, because you
have to work with a lot of sound-layers.
all tracks must be strictly filtered at good frequency and good
balance (this sound at left, while this other at right...)
4 do the samples you use come w/ the software or did you get them from
somewhere else? how much of the remix is people you've recorded, like
your own playing?
Petit Volcan:
There are no samples inside ACID or Soundforge. Therefore, I had to find the primary materials by myself. The good thing with sampling is that you can sample EVERYTHING. I personaly hate using premade samples - making the sample is half the work ! Plus it's really interesting. I used to sample mostly from CDs and MP3s (I've made a big use of a drum loop extracted from "Kick the PA" on Spawn's OST).
After a few years, I started buying dedicated machines to make music : samplers, sequencers ... For example, my "Hidden Place" remix is recorded live with my Korg ES-1 sampler/sequencer.
Also, sometimes you want a precise drum loop, and you can't find it anywhere. So you have to program it by yourself. There are quite a few free tools to do this on the internet. I suggest you to have a look at the good old Hammerhead Rhythm station, and also Tuareg. I guess there are many other tools now.
If you need to program a drum pattern, then you may need to program a melodic sound - to make a bassline for example. There are also great free softwares to do this. I don't have the name now.
r-man:
The software I use don't have sounds or samples archive. I take the samples
(the more simple ones also) from records. I've made my own samples archive
from my cds.
I use Itunes, RazorLame, Goldwave, Winamp or Peak for
capture or convert mp3 to audio files, and Logic Platinum or Peak for cut,
edit and make the loops.
As regards rhythm grooves, I use also some libraries ("Liquid Grooves",
"Akai", "Garage Band") that I have "stolen" in many years from some of my
friends' equipment.
I occasionally play some parts with guitar.
dark jedi:
My sounds are mostly coming from synth, they are plugins installed in
the sequencer. There are lot of sort of such "virtual instruments",
named VSTi, from basslines, old synth emulator, strings, to
drummachines, samplers ....
You can find some of them free, or there are commercial plugins, they
can be very expensive...
Then I'm adding a lot of sound fx like wooosh, claaash, fuuuuzzzz to
spice the mix, you can found them as .WAV files on the net by looking
for "samples sites". Try not to use mp3 samples, the quality is always
too bad ....
I'm using no real instruments (for now!), only my Emac and a Korg
Sampler (esx1), plus a small MIDI keyboard to enter note while
composing melodies, or for tapping beats.
5 do you have tips on matching the vocals to the beats? like
speeding up the vocals for a track w/ a faster beat, etc.?
Petit Volcan:
When you make your remix, you definitely have to keep in mind how the vocals will fit. Sometimes you only have a few words, sentences, so you can slice it in little parts and paste it on the beat. Sometimes you've got a whole vocal track and then you have to find it's BPM and make your song match it.
r-man:
I usually choose a "guide" for my remixes, and the "guide" is usually the
voice. On vocal I fit the beats and the tune. After fixing the
time and "building" the arrangement, if necessary, I adjust the phrase
length or lightly modify the tune of vocal track.
dark jedi:
I never speed-up original voice, because i hate deforming bjork's
voice. I prefer to adjust my remix on the tempo of the original song.
But with a song at 90BPM, you can add 180BPM beats,because it's double
time, it will fit perfectly to speed-up the track without changing the
voice !
To find the BPM of the original track, simply cut a loop of the track
(I mean a perfect loop, with correct start & end points, beginning with
the first kick of the measure), then play with BPM setting while loop
playback, in order to have the perfect timing, most of the time, the
BPM would be precise, 85, 90, 95 or 100 BPM. If you find 93,56BPM, your
loop may not be good.
Now, you just have to be imaginative !
bjork index
contact