.comment-link {margin-left:.6em;}

High Definition Video for Independent Filmmakers
A How To Guide for Digital Filmmakers
Welcome all! This is my blog to share my latest research,
thoughts, etc. on utilizing HD for independent filmmaking.

YES, I am available for consulting
Contact me at mike@hdforindies.com

All content copyright 2004-2007 Mike Curtis.

Tuesday, June 26, 2007

Top Ten suggested audio workflow tips, bonus rant on good/bad calls on set 

I don't know diddly about audio production. Well, here's what I know:

1.) Get good sound.

2.) Get somebody good on set to handle it.

3.) Once burnt or echoed, It R Gonn, just like video (many many similarities to running too hot in both)

...so I don't do audio related stuff, other than to recommend don't use the built-in mike on your camcorder, odds are 99% it isn't that great (it is a common place manufactuers cheap out on cameras...because buyers don't pay much attention to it.). And get somebody good with good equipment.

OK, so that's what I know about audio. Oh - and just like video, Compression R Badd, M'kay?

That's the fun way to put, but as always the truth isn't so pithy - careful and light compression can be OK, but it limits your flexibility in post. Nothing succeeds like good source material, well balanced/exposed/noise free, with as little compression as possible in order to yield best final results and maximum room in post to have creative and technical flexibility and control. Some kinds of shots, and some kinds of post production processing, can survive more or less compression, depending. YMMV.

That's not nearly so pithy, is it?

Which leads me to this - a sound guy, Christian Dolan, came up with his own Top Ten audio related list after reading my 10 Things Not To Do.

sync.sound.cinema: The Fifteen [drops tablet]...Ten! Ten Commandments of Sound for Picture! (Part One)

sync.sound.cinema: The Fifteen [drops tablet]...Ten! Ten Commandments of Sound for Picture! (Part Two)

which references this:

An Open Letter from your Sound Department - A Production Sound Manifesto written by audio professionals

Good audio is worth having. If you're shooting on good, professional cameras (that have XLR inputs and DON'T compress the audio), I'd...hmm....not so sure if I'd say I'd ARGUE the point that it is OK to get good mikes, mix/balance, then run tethered back into the camera to record sync sound, but I'd certainly like to have a conversation about it. 

It DOES take time to sync dailies, it IS a pain in the ass. While getting the best results is extremely important, depending on the scope and nature of your production, time effectiveness does fold in there somewhere. I love production folks that blithely (or ponder seriously and then still recommend) taking steps that dramatically increase time in post. 

Sometimes these can be for the benefit of the production - like insisting on a separate, superior sound recording system. Some can be to the detriment - I cringed every time I was on set and  heard "Hey, shouldn't we fix so-and-so?" "Nah....MFIP." Which was the shorthand for Mike Fix In Post. While time on set is the most expensive, valuable, and hard to come by (getting everybody and everything in one time and place all working properly), taking 30 seconds to tuck in a greenscreen, flag a light, get the boom out of the shot, etc. can save hours to days of laborious lipstick application to a pig of a shot. 

Or, put another way, I always loathed spending time turd polishing a screwed up shot to make it viably passable rather than spending the same amount of time on the shot the way it should have been to make a decent shot look great. Or burning up my "I have X hours to spend on this shot" time fixing problems before I could get to the Make It Look Good part...and had precious little time left to make deadline.

THAT'S where/when your post dollars are spent badly - fixing broken things that weren't shot right in the first place.

It is a clear mantra of quality filmmaking to get the dollars up on screen - as in, have production value. Have the results of time/effort/money show up on screen in a GREAT looking shot, not a medicore shot that the budget got burned fixing. It is all too possible to shoot mediocre to slightly flawed footage, then burn through what COULD have been your awesome color correction or VFX budget to get fabulous looking shots, and instead that time/effort/money gets burned on making flawed shots look marginally acceptable. OR, or course, you can just go way over budget and have decent looking shots with some inescapable flaws still in there. 

For all the things that can be fixed in post at considerable time and money, it is still BEST to get it right in camera in the first place. Saying MFIP is a punt. Sometimes a wise punt balancing all the other factors, but still a punt, because you couldn't do it right on set in the first place. That's a mistake of either mishaps, lack of preparation, lack of skill, lack of proper time budgeting, whatever. But saying you'll fix it in post is admitting you couldn't get it right in the first place. BAD.

This is a great point Christian brings up about the No Luv sound gets on set - a great way to consider it is WWCD? Huh - whazzat? It is short for What Would Camera (Department) Do. He brings up the example of muddy audio from background noise. If it were people walking around in the background adding visual noise (instead of people on or off camera making unwanted, uncontrollable noise) would the camera department say "Don't worry, its OK, we'll just roto out those folks in post?" Of course not. But directors/producers all the time do that kind of thing for audio.

Remember reading that audio is at least half of making a good looking picture? This discussion is the exact moment on set when you commit to making a compromised product.

Another good example I just thought of - going back to ADR/looping is the audio equivalent of the camera department saying "We couldn't get it right on location, don't worry, we'll shoot background plates and a greenscreen production later." Time, cost, effort, complexity, and in the end...it almost never looks as good as if it had been shot there in the first place, the performances suffer a cold, distance because the actors can't relate because they aren't in the same physical and emotional space as when it was originally done or should have been originally done....etc.

OK, done ranting...for now.

Go read the three above linked articles, and ponder what it takes to make good product.

-mike


UPDATE

I'd asked why not go into the camera on set, and Christian Dolan came back with good answers:

Mike,

Most of what I wrote in the list was proselytizing, with a dash of ranting. The majority of what I do is corporate, industrial and reality, which is single system all the time, so it obviously works, but for features, I really feel that people should go the extra mile, especially considering that it may not add that much to the budget overall.

A) There is the quality issue, as measure by Jay Rose in that article that Bruce Allen linked to. I've done some of my own, non-scientific comparisons, and camera ADCs always seemed far noisier than a dedicated audio recorder.

For (non-narrative) TV, docs, and other programming with simple mixes (or whose audiences have different quality expectations), I think single-system is absolutely the way to go. It's simpler, direct, efficient.

For movies, though, things get more complicated in the mix. Generally, you will be applying some amount of signal processing, and in the case of adding gain and/or compression, you end up also raising the noise floor of the signal. With 16 bit audio (the only option in most cameras), the noise floor is higher than it is in 24 bit audio-capable recorders; throw in the self-noise of the camera's AD converters, and you're fairly limited in your range of gain before the noise floor becomes audible. When you start summing multiple tracks (as you definitely will in the mix), you'll be adding noise to noise. The more tracks you have, the more noise you'll be adding, which is why it's imperative to start with as quiet a recording as possible.

I think that the greater bit-depth argument may even carry over from HD, in that choosing 10 bit over 8 bit gives you much more latitude and wiggle room in post. (Of course, I am a audio guy, so take that one with a grain of salt:).

B) Confidence. With single-system, I set my levels with tone, tape down as many of the camera's audio switches as possible, and then cross my fingers. While I can listen to the headphone return, just about every camera I've ever listened to has had an incredibly noisy headphone output amp, which means that I have to do my critical listening off the mixer, dipping in here and there to make sure I'm still reaching camera. The noisy amp may mask subtle background noises (people talking, AC rumble,etc.), that become issues in post.

Plus, in my experience, every Sony broadcast-level camera mixed its audio tracks to mono to monitor. If I hear a crunchy buzz hit the audio, I can't tell if it's the wireless on track one, leaving the boom on two usable, or across both, rendering that take useless (unless you roll the tape back and manually switch the monitor output at the camera itself to check).

And even with the level-setting and taping off of controls, there may be a chance that a well-meaning, but very hurried cam op may bump an input setting (track one from me, track two now from the cam mic, still got audio, just sounds weird, why?) or an input level pot, giving me one more x-factor to troubleshoot.

With double-system, I have the audio recorder right in front of me. I can set,and more importantly, see, all of my audio levels. I can solo any input I want to see where that weird buzzing is coming from, and I can roll back to any take to check it without having to bother anyone else, or risk the "tape return" feature not quite doing what it's supposed to (that isn't a dig at camera at all. I've just seen it happen more than once where even veteran cam ops have rolled back to check a take, engaged the return, and ended up recording over some material).

C) Track count. As far as I know, just about every major SD/HD camera out there only accepts two line-level inputs. While two tracks can take care of most audio needs (indeed, the venerable Nagra never had more than two tracks, and countless classic films were recorded with that little Swiss tank), today's demands are changing, skewing towards more and more wireless mics. With single system, you must do a live mix on-the-fly for shots with more than two mics, which means that you end up with what you end up with. What happens if wireless #1 gives you an RF hit, spoiling the track that it was sharing with wireless #2, even though #2 was free and clear?

WWCD: Imagine a multi-camera movie shoot, where you have four cameras, but only one VTR. You have to do a "live switch" that is committed to the recording medium, without the benefit of iso-records to re-edit later. While this is all well and good for live TV, it doesn't make much sense for a narrative, where you want all the options you can have in the edit bay.

Well, as usual, I've blathered on for quite a while. I'm pretty new to tech writing; I guess economy of words will come with practice, but I'm certainly having a good time with the blog.

Thanks again for the links, and let me know if you need anything else.


Thanks Christian!

-mike

Labels: ,

Comments:
This alone made that a worthwhile read:

"5) Thou shalt not labor under the illusion that wireless mics are "magic".
(aka:
Director
"Hey, can we do anything about all that clothing rustle?"
Mixer
"Yes, either re-frame to allow a boom in, or make the actors do it naked."
Director
(beat, thinking)
"Naked, you say?"
Mixer
"Sigh...")"
 
I have always been taught that if your sound isn't good, you may as well throw crap on the screen. Or to put it another way:

Imagine the worst looking video ever with great sound. People with think you made an artistic choice.

Now imagine the best looking video with horrible sound. People will think you're an amateur.

Sound is the single most important item in a final product. I'm more a video guy, but I've watched short films where there's inconsistant sound, and it's the most distracting, annoying thing ever. At least to me.
just y 2 cents
 
This reminds me of a study about video cards. They had two computers running Doom 3 with the same video card and visual settings, One computer was set to 44khz sound, the other was set to 22khz. They asked players to play both, and say which machine had better graphics. Everyone said the machine with the 44khz audio had better graphics. Perception is a funny thing, and bleeds into itself

Dialogue must be your priority.

But never, ever, underestimate the importance of foley either... and pick up as many wilds on set as you can. Get them when you wrap a setup or shot ("Can we just get that cup hit now?"). Also, get them when you wrap shooting on a location and/or at the EOD. Get them to walk around sample all the sound they can - foot steps, clothing rustles, etc. Your post SFX will love you for it. Worst case, its fantastic reference for the acoustics. Think of it as the audio equiv of shooting reference and texture stills.
 
Mike

Even a compressed external audio recorder can often beat an uncompressed camera recorder.

Read Jay Rose's old article...
http://www.dv.com/features/features_item.php?category=Archive&articleId=23902929

For a given -35 dBU mic input, signal to noise of a Tascam DAT is 81 dB, a (compressed) Sony MiniDisc recorder is 72 dB. A high-end Sony camera is 73 dB. But a PD150 is 58 dB. And an XL-1 is 54 dB.

Why? Because firstly the DACs on the cameras often suck and secondly, there is a lot of other stuff goin' on electronically there inside the camera that is not conducive to a low noise floor.

This is why I kept saying Red should either have digital audio inputs - so that you could take audio that's been through the external DAC of a good recorder (Sound Devices, etc) and record it in sync with picture - or else just have 2 mic inputs with excellent DACs. Unfortunately, they spread their budget across 4 frikkin' analog-only inputs. Which means we might have to record double-system all over again for features because the sound might be irritatingly not-just-quite-good enough.

Bruce Allen
www.boacinema.com
 
Mike,

Thanks for the link love. You were able to explain WWCD in a far more concise way than I did. :)

Bruce,

Great article reference. If I'd had my act together, I would've linked to that one on the original post.

Thanks to all for reading.
-Christian

www.syncsoundcinema.com
 
Something people rarely touch on in these indie sound discussions is the importance of dynamic range. Just like in video it can be the lowest common denominator in good sound. For example a year or so ago I was in the market for a sound recording device I was excited about the Marantz PMD670 because of the price point and I was finally gonna get rid of mini disc and the pain of upsampling 44.1 to 48k. When you look at the specs for the PMD670 the dynamic range is only a few db better than a mini disc. The PMD is 86db versus a pro DAT which is over 100db. I ended up with a Tascam HD-P2 which has a dynamic range closer to a DAT.
So what difference will you actually hear with dynamic range? First it will allow you to a much broader range of quiet and loud things. Second it won't clip (overload) the signal on device as quickly. Just think of think of video and how dynamic range gives more variance in shadows and highlights.
On a side rant sync is not that scary when you use timecode and BFW (broadcast wave) files. Also anybody who is shooting HDV and wants serious audio of any sort better be doing second system. Mpeg layer 2 audio sucks real bad especially as an acquisition format.
Okay enough ranting I'm done now.

--Trevor
 
Post a Comment


Links to this post:

Create a Link

This page is powered by Blogger. Isn't yours?

Listed on BlogShares