Fish(er) Tales

A student asked about my workflow approach when starting a new audio-post project. I thought about it for a few moments and then scribbled down these notes (only slightly tongue in cheek).

For me it all starts with a long cry in preparation for the heart-wrenching work ahead. Then it’s a stiff drink, perhaps a single malt scotch, and I’m ready to begin.

I prefer to watch the video nearly locked with NO interruptions so I can experience it one time as an audience member. After that, I’m too close to the material to have that perspective so I cherish and relish that initial experience.

Next, I prefer to hold a spotting session where I talk with the director (et al) about her vision for the sound and the possible overall approaches to take. Then, we go through the film together and talk specifically about certain sound moments. This step alone saves a lot of time and possible misunderstandings.

Technically, I need an OMF and a rendered file of the video. I translate the OMF using super secret, I’d-have-to-kill-ya’-software, to a format I need. Sometimes I get a drive with all the media and the NLE project files … so then I turn to an assistant who opens the files and generates the OMF(s) as I don’t do FCP or Avid.

Once translated to the DAW format I can use – usually Sony Vegas Pro, I begin the work of figuring out what the hell the picture editor did. Usually the timeline is a total freakin’ mess and it takes some time to separate dialog from temp sound effects and music that, for some reason, get mingled together. Duh!

I ALWAYS copy all of the original audio tracks, mute them, and store them at the very bottom of my timeline. This way I can always find the originals should I mess up big time. The audio from the video render file is my sync reference (along with frame rate and sample rate).

Rant time: it’s funny how many FCP editors never move beyond the initial 8 tracks that the program starts with even though you can add more audio tracks. Unfortunately, instead of adding more tracks, they comingle stuff that should be separate and on its own track.  If that’s you — put dialog on 1/2, 3/4, temp music on 5/6, and temp effects on 7/8. Please!

If you are really generous, put the room tone that goes with each scene on tracks 9/10 and mute them. You will save me hours, perhaps days, of having to find the room tone myself. And it’s easier for the picture editor to do this ‘cos they captured the footage and can find it easier.  I wrote about this here.

And please get rid of any audio effects you may have used. Let me know of a particular futzing  idea during spotting and keepa-you-hands-off-da effects (they don’t come through during the OMF anyway, thankfully). Also, replace any MP3s (shame on you!) with full quality versions of said audio as the OMF doesn’t like them and generates nonsense in its stead.

End of rant

Finally, after I’ve wasted the better part of a day getting the tracks organized, the real work begins, generally after another really good cry and more scotch.

Dialogue editing is the least creative work, so I start there. I smooth dialog via editing using  J-cuts and L-cuts (aka split edits), crossfades, and room tone fills. I also do NO noise reduction and leave that for the premix stage (iZotope RX rules). Dialogue editing requires a lot of track checkboarding to get things isolated. It’s difficult, tedious work. I can usually only do that for about four hours before I want to go Ernest Hemingway.

Therefore, I turn my attention to building the other elements that support the dialogue just finished. That means backgrounds, obvious sound effects, and Foley. Some of my ‘Foley’ comes from sound effects and other Foley sessions done in the past, so I put that in first. I do other real Foley sessions at various times throughout the project ‘cos it’s fun and breaks up the tedium of the other work. You’d be surprised how much Foley you can do with some cloth, a drill, and a box of Homies ). Thanks to Christopher Bodel for that tip (and joke!)

I always work on the first 3-5 minutes and the last 3-5 minutes first. This enables revisiting those crucial moments many times before the thing is done. If you are going to have bad sound anywhere in the film, make it the middle and not the ends. The start brings the audience into the film’s world so it has to be really good and the end sound is what people remember most.

I keep my timeline very organized using color coding and grouping/busing like sounds together. At first, I’m a bit parsimonious with tracks, and usually end up moving sounds around to more tracks later in the process. The typical film ends up between 85 to 100 tracks in the end.

I do a little bit of mixing as I go along, mostly fitting backgrounds and sound effects in, but generally it’s an ‘all faders up’ mix at this stage. I prefer clip-based automation (fades, levels, etc.) over track-based at this stage. I listen to elements on their own – such as a pass of just Foley — which always reveals missing elements and other issues. Fresh ears are invaluable!

I send versions to the director as the film comes together and that generates notes for changing things. I tend to be a ‘kitchen sink’ guy and put everything in — to give the director choices — and then strip stuff away until we find the film’s sound. It’s a variation on the ask forgiveness instead of asking permission kind of thing.

So, I continue working through the film doing a little bit of everything until it’s done (not finished). At that point I turn my attention to mixing. That usually generates a few more notes from the director.

And finally after many versions everybody is satisfied. (Except one person always hates one aspect of the soundtrack – a line, a music choice … there’s always one. Oh well. BTW: don’t be that guy!).

A note about the mythical beast known as “Picture Lock“. I prefer to start the sound work after the picture is done and locked. Sync changes are hard to conform. But directors like to fiddle. And they do. In fact, as the soundtrack comes together they often see (and hear!) the film in a new way. That gets their creative juices flowing. And so they make changes. A trim here. A shot swap here. A delete-a-line there. Oh my.

So periodically I need to regenerate the OMFs, get a new video file, and conform these changes. This usually takes a day. And it’s scary! Directors and picture editors seem to forget that changing a couple of tracks in an NLE is easy. It’s not so easy when you have 85+ audio tracks!

Hint: I never do track automation until later ‘cos it’s the one thing that often gets f-upped during a sync conform. I leave the real mix automation to the premix or more often the final mix!

Well, I hope that sheds some light on my audio-post workflow. Your comments and particular workflow tips are appreciated.