Complexity in Cutting

Question

I often hear about major films using thousands of tracks for a final mix.

Do any of you have a rule of thumb while cutting effects for a particular piece of sound design such as "if I don't get what I'm looking for with up to __ elements, I'm going to start over"?

For example, you're making the effect of some sci-fi gun. You've thrown in 12 elements. It's still not working. How many elements would you try until you started from scratch? Or backed off a bit and kept the ones you liked and started throwing new ones in?

I seem to remember Erik Aadahl mentioning that he stops after 8 elements. Or was it someone else...

Anyway, I think you get what I mean.

What do you guys do?

Answer 1

There are too many possible variables to have a set rule eg - the complexity of the object generating sound - the perspectives/point of view/cutting for onscreen/offscreen panning/location - how many predubs/stems are planned etc

In the end, I stop cutting when I 'believe' the composite sound

But maybe what you elude to is this: there is nothing more frustrating than someone cutting a lot of tracks for an object or moment without having any clear idea of priorities or organisation... I've had people in the past play me heavily layered effects eg a vehicle where they have cut 64 tracks of material... and its a mush of sound. So I ask: play me the critical tracks - ONLY what has to be there for the scene to work ie none of the 'sweeteners' (i hate that word, but anyway) If they can't, then you know there is a problem

Answer 2

If I have what I feel is a lot of elements to a single sound, like lets say more than 8 for example. I go through them and mute each one at a time and listen back. If I can't tell that element is missing, I delete it.

Answer 3

sometimes I do the opposite - I'll create a special protools session just to have more creative freedom when making a sound and I'll go nuts with layers and routing and plugins.

The key is to remember that this is ingredient based cooking - so the quality of the samples that you're cutting is going to have the most direct impact on the final sound. This means lots of custom recording.

My rules of thumb:

• don't be afraid to use your tools. Break out those eqs, verbs, tracks, dopplers, etc.
• don't keep sounds in place that you don't want to hear. If you find yourself with an unwieldy layer monster, start muting stuff and see what you don't miss.
• don't add sounds for the sake of adding sounds - every layer you add needs to have a specific purpose in the overall effect.
• pay attention to your frequency spectrum. loudness does not equal tons of low end, it equals an even distribution of sound across the spectrum.
• set loud sounds up with quietness
• listen with your speakers at a good level. Where you set your speakers can influence whether you want to cut that sound in or not.

Answer 4

My only rule is that every layer of a sound has to have a reason to be there. If I have to work to justify the reason for adding a layer or keeping a layer, then it isn't needed. As I design complex sounds, I will continually go back and solo each layer. I find that it is easy to keep adding things while you are in the thick of it. Later you find that multiple layers are doing the same thing, or competing against each other, or adding noise. Before I call a sound done, I like to thin out layers that feel unneeded.

There is a quote that I love and refer to often (I heard this on the Home and Garden channel, haha):

Perfection is achieved not when there is nothing left to add, but when there is nothing left to take away.

Answer 5

I've learned through much process, head beating, migraines, etc, etc, that if the sound does not fit, then no amount of elements will make that sound fit. I personally keep a set of offline tracks in my session that I move the sound regions I am having an issue with to, then begin the process of re-building the sound from the ground up. If all else fails I have my original sound that I can revert to if needed.

As far as track count is concerned I was able to sit in on some of the sessions from Dane Davis's work on Day the Earth Stood Still. I will not go into details but the scene with the micro-robots was at least a hundred channels wide to get the swarming sound needed. So in all reality, there is a limit and I love the idea of 8 elements being a limit, but remember the best sounds are composed of broken rules!

My 2 cents, Nick

Answer 6

My personal method is that for 'garden variety' sounds (doors, etc), I try to make it happen in about 3-5 elements max. Of course there are many exceptions to this on complex sound moments, but that helps me keep the edits under control and gives me a clearly-defined boundary of knowing whether or not a sound build is working or not.

Also, I agree and follow the same ideas as Jordan and Chuck, which are great answers.

What I have found effective too is to create a set of work tracks where I build everything that I know is going to be more than 3-5 library-pulled elements. This is my sandbox where I roll up the sleeves and get into edit densities that can be 30-40 tracks wide sometimes on a design. However, when I've reached that believability and completeness tipping point that Tim mentioned, I proceed to print down the content categorically and then pull those tracks up my FX tracks. This has to be done with care though and the strategy can change with each sound situation (high mid low, versus boom hit shimmer, etc - It's all about giving the mixers the intention, while still providing them with split out prints so they can control it in the mix.

In all cases though, I religiously use region grouping on EVERYTHING - and then name the region groups in all-caps. that way, when each grouping of sound elements is easily determined on the stage. Will the stage ungroup and dig into these sometimes? Of course! But, by giving it to them this way I have found that they appreciate the forethought into organization and session cleanliness. And, it helps them easily see what's what like an interactive cue sheet (for example, by stating in the region group name whose gun shots are whose, and which tracks are ricos versus whiz bys). Sometimes I'll use the region name to drop in notes in the naming suffix for the mixer such as:

• ** SOURCE MUTED ** = these are source elements for the printed track, keep muted unless you need to dig into it

• **SWTNR FOR __ ** = the sound element was only cut in FX to sweeten or enhance something like prod FX or Foley (selling a weak O/S Foley sound), or may be only a 'partial-sound' meant to blend with Foley or Prod FX

• ** PANNED - SKIP PAST ** = already panned, save your time (only used sometimes)

• ** 3RD GRAB CHEATED ** = the third cloth grab sound has sync cheated for visual, etc
• ** MATCH GUIDE ** = this sound was created/selected to match or closely match what was heard in the guide
• ** ALT ** = alternate version of an edit or sound choices used in that edit
• ** OPTIONAL ** = optional, additional element which is non-critical but was edited; if the stage chooses to lose it, it won't affect the integrity of the primary sound

There's a ton of different ones, and these are only some specific examples I can think of. But, these can be a great roadmap for a mixer as part of that interactive cue sheet method.