Top 10 tips for the best audio using a DSLR

After fielding a ton of questions from the students from the Sound Advice Tour last year about how to get good sound when shooting with a DSLR camera, I thought it might be a good idea to put up a post about it - especially now that we’re nearing an Australia stop on July 9th in Sydney. Below, here are some practical-professional-proven ways to get good audio when shooting on a DSLR platform.

1. Your DSLR doesn’t record good audio.

This is the most important thing to understand about using DSLRs. They may say they can record great audio, or high bit rate, or have a wonderful microphone on board, but several issues will forever keep your audio sounding like you recorded it in a toilet if you try to use the on-board mic. 

The mic on board a DSLR is usually what’s called a “condenser” microphone which means it’s highly sensitive. That’s actually a good thing in a vacuum, but when the microphone is on a camera which makes strong high frequency whines while operating (and they all do), and can be many metres from the subject speaking – with thousands of other sounds between you and the subject – you’re likely going to get more of the ambience/room noise than the subject. In fact, sometimes these mics are positioned in such a lousy place on the camera that they can’t be used for reference audio if the subject is more than five feet from the camera. To say nothing of wind issues, shock/bump issues and high audio compression formats, it is unethical for camera manufacturers to dupe filmmakers by the hundreds into thinking their DSLR can record professional grade audio. Folks, they had to cut corners somewhere, and I assure you the biggest corners they cut were audio. Never forget, it’s a still camera first. You’ll need to record your final audio onto a separate recording unit like a Zoom F8, H6 or some other recorder with separate microphones. Get a lav and a recorder. You can do it for $500. Your on-board mic may not even be good enough in certain situations to be a sync reference.

2. Since your DSLR doesn’t record good audio, sometimes your reference audio won’t be good enough to sync sound with.

We’ve all had it happen: the subject is too far away, there was too much interfering noise, the on-camera mic is too poor or the compression setting was too crunched – and now we must manually sync all the audio clips of a scene. The automatic systems of your NLE/PluralEyes can’t analyse the reference audio to be used for syncing. It’s completely unacceptable, especially when you’ve recorded multiple discreet audio tracks which need to be synched. My heart goes out to all editors who run into this - having dealt with many projects where such audio is brought to me to fix. What situation causes this to happen most often? You guessed it: when the on-board mic of a DSLR is used as the reference. The only time I feel comfortable using a DSLR on-board mic as reference is if the camera is in studio, and the camera is fixed on sticks. Otherwise, camera bumps, wind noise, and all manner of other real-world noisy interruptions of critical synchronisation references turn an already sketchy reference recording into useless data on your camera card. It’s another reason why having a dedicated audio mix of your lavs/boom sent to your camera as reference is infinitely better than that on camera mic, and about 10X better than any external pro mic mounted on board a camera.

3. If your subject is far from the camera, you’re audio can’t be synced automatically.

Filmmaking is full of magic and the perfect process for suspending disbelief. But there is one thing which can never be suspended in production: physics. At 22ºC, sound travels through air at roughly 0.37693 m/s. Why do we care? We don’t. Unless you’re recording audio to synced video and you’re using a DSLR with a camera mounted microphone. If film rolls along at 23.976 fps, that means that a frame is roughly 42 ms long. At 0.37693 m/s, that means that sound will travel 36 feet/12 metres or so in a frame. Well? If you’re 12m or so from your subject, or at least your microphones are, your audio will be 1 frame late to picture. And when you try to sync it with PluralEyes or Adobe Premiere, sure, the audio will sync to the reference, but that reference will be 1 frame late. 72 feet is 2 frames. And of course, this isn’t a grid of latency. The further you get away from the reference mic, the further out of sync you’ll get. So, this means at 10 meters, if you pull the audio back a frame to compensate now the audio is early. And we all know that our brains are used to audio being late from tens of thousands of years of it happening naturally, but our brains are immediately taken out of your scene when audio is early even a fraction of a frame. 

42 ms doesn’t seem like much? Just try this: take one track of dialog, and copy it to another track. Then delay that audio a frame. Pretty substantial delay, no? Then apply it to picture and A/B what one frame of latency looks like - or more importantly feels like. So be careful when you have a long shot and are getting close to 10 meters away from your subject. Keep the camera logs and use those distance to-subject lists as your reference to how far you’re going to have to advance the audio of that scene to be in sync with the visuals. By the way, you’re not going to be able to sync your audio in an NLE. You’re going to have to send it out to a DAW to sync these long shots. NLEs generally don’t allow higher audio resolution than frames. And since we’re talking milliseconds, a DAW is your only course. Fortunately, at least in the case of Adobe productions (Premiere Pro/Audition), you can seamlessly exchange audio for just this purpose between NLE/DAW using their proprietary Dynamic Linking. Good luck trying to do this with any other NLE/DAW combo. It’s why we always do our assistant editing work – regardless of final edit platform – in Adobe products. 

What’s the solution? Run a mono mix of your audio from the mixer to the camera (your camera does have a mic input, doesn’t it?), and now your reference is the clean audio from the audio mixer – and it’s in sync.

4. If you must use the camera mic, use the RØDE Video Mic.

I understand. Even though gear is cheap these days, it’s not free. Moreover, I understand that sometimes you can’t operate a mixer and a boom, and operate a camera and direct. Sometimes you’re just a one-man-band and there’s nothing for it. In such a case, you’ve got to do the best you can, and in all cases, using the on-board camera mic is the worst you can. Spending a small investment in a RØDE Video Mic at least gets you a good quality hypercardioid microphone which can assist in rejecting off-axis sounds. It also gets your microphone a little ways away from the camera - and thus its noise. Having a good hypercardioid pointed at your subject is 5X better than having a badly pointed boom pole with cable rustling noise and more. Now, if you have to turn your camera away from the subject – and the microphone with it – well then you’re up a creek as you haplessly record the ground/sky/not-actors.

5. Use a boom mic, only if you know what you’re doing.

You see a lot of filmmaker’s clueless girlfriends holding boom poles in small productions. It looks and feels like a “professional production” when there’s a boom pole being held on set. Nice. But when the pole is being held statically in a scene with 3 people talking, or better yet, it’s being held “between the actors to have a good balance”. You know the only thing professional is what will happen in post when they have to ADR (automated dialogue replacement) the entire dialog track. If you don’t know what to do with a mic on the end of a boom pole or what it’s really for, just stick with lavs – or get an education about how to be the best boom op you can. Boom ‘opping’ is not for the weak or squeamish. It’s hard, and there’s non-intuitive rules and practices which must be followed in order to make it work. At the Sound Advice shows, we spend nearly an hour going over production sound and teaching attendees how to use a boom pole.

6. Never record in any format but lossless 24 bit.

So many people ask me if it’s okay to record their audio in .mp3 format. That’s like saying, “Is it okay to record my video at 650 kb/sec as long as it’s 1080p and it looks good?” Never. Ever. Period. It might be okay for dailies, but in post when edits and correction and VFX see the original footage go through generations of data loss, it’s imperative to shoot in RAW or some other near lossless format. And no, you DSLR people, even in Log, you’ll never convince me that 24mbs at 1080p is a format that should be used for anything but the most rudimentary visual post processes. Audio is no different. An .mp3 may sound good at 192kb/s when you’ve recorded it, but add EQ, compression, volume changes and it sounds like what a mouth full of sand tastes like. Hard drives are cheap. Buy more.

7. If you can’t get good audio, do wild ADR and fix with Adobe Audition.

Sometimes, it’s just not possible to get good audio. Sometimes the wind, rain, noise, cars, planes, ducks, whatever, are just too hard to deal with, and the light is going. We have to shoot. No problem. Once the takes are done, and the crew is setting up the next shot, take the actors into a quiet place and record their lines “wild” or without picture. Have them do this without moving as well. Use a good mic, and get the sound clean. They’ll be in the moment, (having just done the scene), and you’ll have your ADR without more fuss, if you need it. You might ask how one would sync such wild audio without the actors having a picture reference. Enter Adobe CC Audition. It has this amazing option call “Automatic Dialogue Synchronization.” It’s the most amazing tool I’ve ever seen for insanely accurate sync of audio to reference audio. In fact, the actor doesn’t need to be that in sync for it to work!

Doing this kind of “on-set wild ADR” also saves you from having to bring actors in to the studio in post to the tune of more cost to the production in money and time.

8. Don’t be stingy: buy a wireless lavalier system.

After you buy your DSLR and appropriate lenses, (including your primes), it’s not time to invest in lights or camera support. All of that can be rented very inexpensively in any urban area. No, what’s next is your wireless transmitter/receiver and lav microphone. Until recently, professional-grade wireless devices have been unattainable under $800, and for good reason. Professional wireless devices are highly complicated pieces of electronics. Now with the dangerous situation that North Americans face with the sell-off of wireless frequencies by the FCC, the viability of professional audio wireless is dubious at best for the independent filmmaker.

Enter RØDE microphones. They’ve developed a 2.4 GHz wireless system which is impervious to FCC rules and bandwidth sell-offs, at least for now. They sound great, are digital, and are priced under $400/channel. They even have a wireless boom mic kit for under $500. I’ve used them in the heart of New York City without a single hit – and without having to do free channel searches and the like. With the advent of affordable wireless technology, there’s no reason why every camera person/filmmaker/director doesn’t have at least 2 channels of wireless TX/RX in their camera kit, and RØDE is leading the way.

9. With Live Interviews, use a lav, and let the handheld be a dummy.

Using a hand held dynamic mic is a smart thing to do for capturing interview audio in a noisy environment – especially a wet and noisy environment. But nothing can capture high fidelity audio like a lavalier condenser microphone close to the subject’s mouth. Naturally, we don’t want to use lavs when it’s raining, but in just about every other circumstance, using lavaliers is superior for interviews and ENG-style (electronic news gathering) shooting. Some have suggested using both the interviewer’s handheld mic and a lavalier. It’s a noble idea, but foolish when one considers the awful phase cancellation issues which would develop from trying to use both signals. One should be favoured over the other, and in my book, the lav should be used exclusively, and the hand held mic should be a simple prop. It’ll look great to have a hand held going back and forth, but the hidden lav mic will do the real heavy lifting of audio information. Certainly, don’t use the on camera mic.

10. Audio is 80%.

For year, I’ve heard people saying, “Audio is one of the most important aspects of your production. It’s at least 50% of the impact that the audience experiences…” I have to avoid choking on my lunch every time I hear that. Try this: take a scene from any movie or TV show and watch it while listening to it through the worst, most distorted, broken-down speaker system. Something especially horrific to listen to. See how well you like the scene. Now watch it again through an audiophile system, but on a circa 1955 television. Preferably with a broken CRT. If it were true that audio was 50% of the experience, then we should like both examples about the same. But. That. Isn’t. How. It. Is. 

Without question people gravitate toward liking good sound/bad picture than bad sound/good picture by several times. I say 4X more. I say audio is 80% of the audience experience. It’s why we went away from silent movies and into “talkies.” It’s why independent features in film festivals have you wishing you could like them, but there was just “something that wasn’t ‘professional’ about them.” Something elusive. Something je ne sais quoi. 95% of the time it’s the audio was ill conceived and ill executed. Film schools don’t teach audio. Audio professionals refuse to let their secrets out. There are very, very few online resources to garner this knowledge. And the kicker? Mastery of sound production is not intuitive at all, because it’s centred around a human sense which isn’t primary. Now, of course, we want to have good lights and good cameras, because otherwise we’re just making a radio show. But consider this there was a radio show aired not too long ago which was so moving, so compelling that people threw themselves off bridges. No silent film ever moved the entire world to reconsider their lives to the point of ending themselves. No silent anything has ever been that powerful.

Now, I’m not suggesting media be used to hurt people. I’m suggesting the opposite. But why is audio so impactful? It’s probably because the amount of data required to trick the human brain into believing that something is “real” requires 32,000 flashes of information a second for hearing, but only 13 flashes of information a second for sight. Having a DSLR is one of the most incredible pieces of story-telling equipment a creative can have, and we all live charmed lives during a time in filmmaking where these tools are so readily available. But there is no substitute, regardless of the kind of brilliant image capturing device one has, for good audio. None. Moreover, if a filmmaker can’t create immersion for his/her audience utilising sound to its utmost, that person is a bad filmmaker. During the 8-hour Sound Advice Tour stops, we go over an intensive amount of information specifically to arm filmmakers with the same kinds of tools the Hollywood audio elite hold secretly. I hope you are able to get the kind of education and knowledge you need to insure your productions have a high audio production value. If not, I doubly hope to see you at the Sound Advice Tour stop in Sydney. Bring your DSLR.

About Mark Edward Lewis

After over 25 years of writing and producing musical scores, re-recording mixing, developing sound design, editing, writing and directing, Mark Edward Lewis has a unique perspective on how to create high production value for independent productions.

What makes Lewis an invaluable resource is his mastery of all aspects of audio production. One of the new breed of directors who approaches storytelling from an "audio first" context, his projects have a flow, pacing and emotional impact that few directors possess. Always strongly adhering to the axiom, "Post production begins in pre-production," he has righted many adrift productions with his organisation, creativity, ingenuity, and cross-discipline experience.

Lewis has worked professionally in nearly every area of both production and post production and currently works as a post production supervisor most recently on the Avenger's S.T.A.T.I.O.N. with Frank Serafine. In 2015, he was the main presenter for the Sound Advice tour in North America where he taught over 1,100 filmmakers about improving their production value with better sound.

He currently martials his deep knowledge of post production and works in Los Angeles as both a filmmaker and a post production supervisor where he can both oversee and develop creative audio content whether in sound design, mixing, music or sound editing. Along with Avenger's S.T.A.T.I.O.N., his most recent work includes, Star Trek: New Voyages Kitumba, The Holiest Thing, and directing Star Trek: Mind-Sifter. Lewis is currently post producing and directing the sci-fi series Blade of Honor and is developing a 50+ hour online resource for filmmakers called Cinema Sound.

Sound Advice

If you’re keen to expand your know-how for filmmaking, broadcasting and website audio creation, this is a must-see (and hear, naturally) event for your calendar. Just some of the topics covered include recording, editing, effects, mixing, design and inspiration. For more information about Sound Advice, follow this link for details and pricing.