Great song. Great video. Very cool

The weird interlacing works here, you're right. It gives it this quirky, ironic quality. I love these 80s-influenced synth-pop styles when they're done well, like here, and so you have the slickness of the widescreen and cinematic tones with that video-y layer to it, enhanced by the step-printing.

"Proper music video"? Dude, don't change a thing!

@B3Guy, are you singing all the vocals, not just the vocoded/processed voice? I was perhaps unclear with my first post, the singer sounds like the voice of a young person, or are you saying that is you and processed that also? I really like sub sound of the music at the start, then the blend of melodic waves that follow later.

edit: B3, do you mind explaining the 'step-print' method you applied, as @BurnetRhoades mentions above? Thanks

Yup. All the vocals are me, the chid and the really deep one are EQed, compressed, lightly auto-tuned, vocal transformed (pitch and timbre shifting), and then of course side chain compressed to the drum track. The child vocal also has a fairly heavy tape delay effect on it to mask some of the telltale signs of pitch manipulation. The deep voice is sent to a bus in order to double it and that second part is shifted down another octave. A little reverb and there you have it.

Neither are vocoded, but there is a very airy voice part that is vocoded synth. It sounds like a synth, but run through various vowel sounds, "Oh, Eh, Ah Uh", etc.

I learned a lot from making this song, and I know I will do better next time around.

The "step-print" is just slowing down the footage by halving/quartering(etc) the speed w/o any frame blending or new frame creation. Technically, most of this is 24fps footage slowed down by 25%. Effectively it holds every frame of the file for 4 frames in the 24fps timeline, so it plays back at 6fps. The only thing that really keeps this from being a 6fps video is the flicker effect, which of course flickers every frame at 24fps (that and the few clips I left full speed). I personally prefer this choppy slomo look compared to overcranking (shooting higher FPS and conforming it down). It is not for every situation, but I like making grittier films, and this is a more hack/gritty way of slowing down footage.

I've only seen one Woo film (Hard Boiled), but I loved it. Need to watch more of his HK stuff. Quentin does it some, the epic walk in Reservoir Dogs comes to mind.

Surely you know EQed . . . equalizer? If you learn one post-audio skill, it should be how to EQ a human voice and how to make pockets in the mix for that voice. Just that will make everything sound so much better. It is just as important to do this as it is to point the mic in the right direction. Audio folks are the unsung heroes of good films, and audio is the most commonly neglected element in low/no budget films. You'd be surprised how cinematic 30fps 720p 4:3 camcorder footage starts feeling with a good sound mix attached.

Auto-tune is just a tool for keeping things on pitch. You tell it what notes you are using (or what scale/key the music is in), and it corrects the pitch of the recording to the nearest note in the scale at any given moment in the audio. The most drastic example of which would be the "T-Pain" sound, in which auto-tune is used very aggressively, and the audio snaps abruptly from one note to the next. Auto-tune gets a bad reputation because of this, but it is actually a very handy tool and can be used in undetectable ways to keep vocals on pitch. Very few people actually sing exactly on pitch, after all.

Vocal transform/pitch shift is just a fancy name for tuning something up or down, making a voice higher or lower (in this case, my baritone voice is tuned up to sound like a little girl and also down for the other part to sound deeper.)

A delay effect does just that . . . it delays the sound, and usually also echoes it. In the olden days, this delay used to be achieved by recording live signal onto a magnetic tape and then playing that sample back at regular intervals, creating an echo. In this case, I turned the echo off so that the tape recorded sound is only played back once. That way, I get the weaving warping sound of playing back an old cassette. Of course, I don't actually have tape, this is a digital emulation but it sounds pretty similar. This is a commonly used effect in the new 80s-influenced stuff to give it some mojo.

Side-chaning compression is widely used in modern music mixing. Essentially, the compression filter on one track is "side chained" to the output of another track. The compressor on track A then only compresses (becomes active) when there is sound from track B. Side-chaning to the drum track, especially the kick is very important to give the drums that big punch. Whenever the drum beats, the side-chanied track gets compressed, making in sound quieter and less full so that the drum sounds big and powerful in comparison. In this song, I believe everything is side-chanied to the drums.

TL;DR I just explained some music fx that make things in my song sound more awesome. Also learn to EQ a voice.

My face when you mentioned Drive :-D

Love that film, it is what got me on to this music genre in the first place.

And yes, it will sound best with headphones.

Equalizing a voice ABSOLUTELY applies to movies. The basics are:

1. Apply a multi-band parametric equalizer
2. Pick one band, and turn its dB all the way up and its Q-factor all the way up (usually 10, it will look like a sharp spike, not a gentle hump)
3. Play the track and sweep the frequency of this band back and forth over the spectrum. You will hear it making a small slice of the audio much louder than the rest.
4. Now the IMPORTANT PART. You aren't just sweeping the band frequency up and down willy-nilly for kicks and giggles. You're actually trying to FIND something: the plosives. Plosives are those hard sounds in a person's speech: "P"s, "T"s, "S"s, "K"s, etc. Sweep the frequency around until you find the spot where it makes the plosives very loud, then leave the frequency for that band right there.
5. Now that you have found the exact frequency of the plosives with your surgical audio spike, you can turn the Q and dB back down a bit so that it makes a nice little mound instead of a drastic spike.
6. Pick a different band and make another spike as before, except this time, find the vowels. Be careful, just below the vowels is a muted, boomy bass end to a person's voice, you don't want that, you want the clearer vowel area just a bit higher.
7. Soften the second band to a nice mound like you did the first.
8. And that's the tricky part, finding the vowels and plosives, and making little hills to increase their volume. After that, it's really an adjust-to-taste thing. You may find it helps to take a third band and make a little valley in between your two nice little hills. You may find that one hill needs to be taller, or shorter, or wider. You may want to use the high cut band to chop off the top of the EQ after the plosives band, and/or it may be good to use the low cut band to chop off some boomy bass sounds below the vowel band.

Maybe I'll make a tutorial sometime.

Edit: and concerning "Sonic Pockets" for the voice in the mix,, I'm not sure I can divulge that juicy little tidbit, but maybe you can guess where the pockets go and what they are for after that vocal EQ description ;-)

@B3Guy "...Surely you know EQed . . . equalizer?"

Yeah, but most of that other terminology or trying to visualize what you mean with "bus" or what "tape delay" really does for you, "side chain compressed"...that's all magic to me. I still don't really picture what it's doing in my head where most complex visual methods for picture manipulation are pretty easy to pick up by comparison. Just how I'm wired I guess.

So, just know I was being complimentary because I regard audio and music a much more difficult subject to master. Going through some of the online ColorGHear docs at one point Shian references what's necessary to restore "true black" to DSLR footage like the GH2. The precise offset in Levels of your black point by -.07 undoes that broadcast convention baked-in that's largely an anachronistic hold-over from broadcast NTSC and kinda stupid if you ask me.

But that's info I'm like, "oh, of course, that makes sense" where so many other grading tutorials feature people just arbitrarily pushing their stuff around by eye (if they acknowledge the incoming broadcast range in the first place). That's the kinda stuff that just clicks moving forward but I rarely have those same levels of "ah, I see, good to know," sorta moments with anything audio related. I'm betting you can hear a piece of raw audio and don't have to just play with it and poke around to get it to sound good. There's specific techniques and part of the recorded audio spectrum that need to be manipulated in some way to account for how humans hear to make it more pleasing or to impart some kind of emotional register that aren't accidental, the same way there are known techniques involving angle, color, lighting direction, scale, etc.

Demonstrating ability in both of these areas really sorta pisses me off ;)

There isn't any part of the process, from a visual, technical or even logistical standpoint to movie making that intimidates me like audio. The skills someone has to master to take raw audio and make it sound "like a movie" is really cool.