How To Do a Dance Remix

How do you do a dance remix? You’ve got these way cool ideas in your head but how do you get those ideas down within the constraints of pre-recorded material? Here’s how I approached a dance remix in a tight time crunch. I can’t say it’s the “right” way to do it, but might find some helpful info for those wanting to do a dance remix of their own.

I got a call to do a remix of Chicago performed by Frank Sinatra for the 2005 Toyota National Convention in Chicago. This was to be their big opener for the convention and would be edited into video footage showcasing their new car models.

Note: This is an educational tutorial and falls under the Fair Use copyright act for using samples in an educational setting.

I agreed to do the project and sat down to execute it with these constraints:

1) I had 24 hours to deliver a finished product.
2) Only direction given was for it to be a high energy dance remix with a little bit of a rock feel.
3) Final must be around 3:30
CHECK YOUR DETAILS
First thing I always do on any project is make absolutely sure I know what’s in the clients head. It’s MY job to figure this out, not the client’s. Fortunately, when I went to get the original Sinatra audio for Chicago I spotted that there were two similiar songs with references to Chicago. There’s “Chicago” (Chicago, Chicago that toddlin’ town….) and there’s also “My Kind of Town” (Now this could only happen to a guy like me….). I assumed it was “Chicago” I was supposed to remix but I called the producer to double check. In fact what I was supposed to remix was “My Kind of Town” which he referred to as “Chicago”, but was not. Now if I had remixed the wrong song it wouldn’t have mattered for me to come back and say “but you said CHICAGO” – bottom line is I would have messed up the job and ruined a new solid business relationship. Honestly, I think this is the main reason I have a high success rate with my work, is really checking the details and finding out what’s really in the client’s head. This is always my first step in any project.

GATHER SOURCE MATERIAL
Now armed with correct info I locate the original big band arrangement of “My Kind of Town” and listen to it several times. Problem is that the Sinatra song is only about two and a half minutes and the producer wants THREE and a half minutes. So I know right away I need to insert at least a minute of extra stuff that fits the song, has a rock feel and maintains the high energy needed for the presentation.

VISUALIZE PROCESS AND OUTCOME
I know I’m going to be working with samples, need to insert high energy elements into the mix and need a final mastered version. I am very comfortable laying down MIDI tracks in ProTools and have a pretty good MIDI setup. I felt that the MIDI tracks were essential to the mix so I could lay down dedicated bass and surrounding instrument tracks to help support any direction I wanted to go that might fight the original track a bit. General idea was to build the session in protools around the original track, edit to be 3:30, fly that template into a program like ACID to quickly insert beat elements, then fly the ACID tracks back into ProTools to add dedicated MIDI instrumental parts to fill out the final production.

I always go very slow when I first start a project, thinking of all the elements and how best to proceed. If it’s a project I’ve done before then I’m right on it, but if it’s something a little new I think it through. It’s not uncommon to see me sit in a chair for quite a while on a project. This is the most helpful thing you can do for your productions and saves you SO MUCH time once you get into it. Usually freaks clients out a bit if they’re in the same room while I start. But they “get it” as the farther we get into production the faster I go. Towards the end I am absolutely blazing fast. Think of it as a sculpture – at the beginning it’s just a big block of rock. As you get farther into it, there’s more of a direction and momentum.
BUILDING THE TEMPLATE
I uploaded the Sinatra song into Pro Tools. I have the TDM version of ProTools and it’s quite a bit more hairy to build a template than in a program like ACID where you just drag and drop. The original Sinatra audio is too slow. I altered the tempo of the source material up to around 170 (this is where it “felt” good to me, though obviously not a good dance tempo, I felt it would be best at this tempo to maintain the energy, and technically no one was going to actually dance to this anyway, it’s a presentation for car dealers). Jumping the tempo up that high ended up being a good decision. I took an educated guess that the client saying “dance remix” didn’t really mean dance remix, it meant “make it really cool so all the car dealers get pumped up at the opening of our convention”.

When I changed the tempo on the audio file, I did it to the whole complete sample. Then I could start on edits from there. If you’ve ever edited a template that’s created by ear to the audio, then changed tempos, you know how time consuming it is to revert back. Actually, for me I just start over with a new template. Why is it important the audio be to a template that locks to measures and beats? I’m going to be doing lots of MIDI tracks, and I’d like them to lock to the track. Because I’m on such a tight time restraint (24 hrs), I need to be able to move very fast in final production.

I spent probably 30 minutes listening to the audio at a wide range of tempos before I decided on the final. My benchmark for the “correct” tempo was simply what “felt good” to me. Around 150 just felt too slow, and up around 180 the song to me just fell apart, wasn’t cool anymore.

With my audio track at the tempo I like, I started to line up the audio the my MIDI grids. The opening of this particular piece is like a ballad and very loose in the timing. So my beginning of the grid I used the first bars where the drums and horns enter to establish the groove. Then I would listen through with a click for where the audio was drifting from the MIDI bars in my template. When I noticed a drift I would make a slice edit and nudge the audio to correct the timing. When lining up audio samples in ProTools, I always listen at half speed and also zoom in progressively until I am at the sample level. Use ears and use eyes to line it up. Then listen again at regular speed to see if it “feels” right. I’ve used the Beat Detective plugin on ProTools before but never really got it to work properly. I guess I’m missing the boat on how it works. Lining this all up by ear was maybe not the best approach. I had a time crunch deadline and had to work fast, so that’s how I did this project. If I have a week on the project I probably would have fully investigated the beat detective options.

After getting the main body of the song lined up in the template, I went back to the loose intro and lined it up, which required considerably more edits. In fact, it wasn’t really “lined up” because it’s ad libbed with a loose timing, I just tightened it up a bit.

Looking at the final template, the song is now only about 2:15 because of the increased tempo! I need over another minute to the song. And the intro is too long and slow. I don’t think the producer will dig a long intro like that. But it NEEDS an intro, it sets up the whole vibe of the song, the whole REASON for the song. I decided to cut the intro in half, and used the lyrics that I thought best set the song up.

Now the songs even shorter, I need another 90 seconds of music and have no idea what to do. I listen through the song in my template, and where it musically makes sense to go into a new music section, I make large edits of 8 or 16 bars to leave open space for additional music. Right now I don’t know what’s going to go there, but SOMETHING has to.

Some of the best “magical” moments in audio production happen when I leave wide open blank spaces to fill up later. When the spaces are open like that, I’m more creative to use snippets, samples and musical detours for that space – because I’m not locked into anything. As it turns out, I ended up doing a whole section around samples of Frank Sinatra to create a jazz horn section breakdown, rock guitar solos and different elements using samples of the horn section.

BUILDING DRUM TRACKS
I mixed down the audio from my ProTools template – in the mix was also a metronome click to make sure the integrity of the tempo was maintained through the silent sections. I wasn’t using a sync track so a reference was essential.

The audio templated I uploaded into Sony’s ACID program, which I have unconveniently on a seperate computer. This was easy to line up within ACID. First thing is to build the drum track – don’t recall which libraries I used, just what sounded good to me. I built a pretty beefy rhythm section using multiple layers of drum samples that would build on the different sections, and break down where it felt right.

Once rhythm was in place I added in some horn lines and some light music samples that fit well. Most of my music overdubs and cymbal crashes I’m saving for the MIDI overdubs in ProTools where I have more control and can work faster. Once the parts are in place to create the shell of the whole arrangement, I mute the original Sinatra audio and create a .wav file of the Acid overdubs and fly this back into ProTools.

Hey, now it’s starting to sound like something!

MIDI OVERDUBS
Now I can work at blazing speed. Being a keyboardist, I’m VERY fast at MIDI overdubs. I forgot to mention also, I have a pretty good ear so I’m also able to hear all the chords in my head and know what will fit where without a chart. This is a nice advantage where my early piano lessons and bar gigs come in handy. For those that aren’t players, this will slow you down a bit as you’ll have to just put in parts that seem to fit without knowing the underlying theory. If you can get some theory under your belt and develop your ear training, it will be your strongest asset when working with musical tracks.

POST PRODUCTION EDITING
Now I’ve got everything in place and do final edits to take out or add a measure here and there. I take some of Sinatra’s vocals and use them as filler samples, add in some delay effects, etc.

MIXING
I do some obvious moves to the original track – I EQ out the bottom low end to get rid of the bass so I can use my own kicks and my own MIDI bassline. Usually a hard shelf rolloff below 80hz and another gradual rolloff in the 250-300hz area. Use your ears, depends on the source material and what you want to do with overdubs. EQ the sample track to bring out Frank’s vocals a bit, which is usually in the ballpark of 5khz. I like a lot of the original elements in Sinatra’s track so I bring up the high end about 1.5db around 12hz to taste.

FINAL PRODUCT FINISHED – I THOUGHT!
The final mix is emailed to the video editor in New York. Meanwhile my mother is in the hospital in critical care and I have to leave town right away. The video editor emails me that he wants rock guitars in the mix. Totally nutty guitars all over the place to hype it up. ARGHH! Oh, and he needs it within TWO HOURS.
Because I have templates created it’s actually not to big of a deal. I fly my new edited template back into ACID, insert a cacophany of obnoxious guitar tracks. I’m guessing since the editor is younger and from New York City, he probably wants something totally crazy. I was right. He absolutely loved it and the head producer loved it. Now to my ears the guitars I inserted are really loose and a bit out of place, but when I saw the final video it all made sense and fit perfectly with the footage they were working with. Incidentally I never got to see the footage before I worked on this project.

The video editor did some edits to my music to make it fit the video, and did some edits to the video to make everything work. I got to see the finished product and it was really amazing. I’m not posting the video because I don’t know if I have permission to. Too bad, the audio itself is a little interesting, but seeing the finished product was very, very cool.

So that’s how I did it. If you have tips and tricks of your own please feel free to leave notes.

Toyota Sinatra Remix MP3 File