Hi, John Hess from FilmmakerIQ.com and today
we'll try to build a foundation on the basics of recording audio for digital video. Steven Spielberg has famously said that "Sound
is Half the Picture" - but it's actually a little more complicated than that. Really
bad unlistenable sound will ruin a picture faster than bad lighting or shoddy camera
work. And in some cases like documentary or corporate industrial work, good clean sound
is actually more important than just about anything else. Fortunately, getting decent
sound isn't that hard if you understand the fundamentals of recording and implement some
solid practices. So for this discussion lets imagine recording
sound as a signal chain. At the source a microphone converts sound energy into analog electric
signals. This signal is carried down a cable and into a preamp on an audio recorder or
camera where it is converted into a digital file. Now let's dive in and look look each
piece of this signal chain starting with the end. The first thing you need to decide when recording
digital audio for video is whether to use the single or double system for recording In a single system setup, audio is fed directly
into the camera and recorded with the image. In a double system or dual system, sound is
recorded onto an independent dedicated audio recorder. Sound from the camera is still recorded
if it's available but used as a sync or scratch track Let's compare each setup. With a single system,
recording audio with video means there is no need to sync up the footage in post production.
This can be a huge time saver especially in tight turnaround situations like the news
or documentary. With a camera designed for broadcast that has a robust and professional
audio inputs and a preamp, the single system also avoids the cost purchasing additional
recorders. Video codecs have predefined settings for audio, most of them recording 16 bit,
48khz uncompressed audio - more on what that means in a second. If you want hassle-free synched audio without
having to mess with audio settings and you have a camera that has good sound inputs -the
single system may be a good choice for you. Why would anyone use the double system? First
and foremost is if you don't have a camera with professional audio inputs like many DSLRs
but there are 3rd party preamps on the market that can make any camera into a single system
setup for audio. But in my opinion there is a better reason
- The major benefit of the double system is audio quality. Digital audio recorders have
some great features that make for better recording. The first is higher sampling rate. When an analog signal is converted to digital,
the smooth analog curves of the wave signal have to be quantized - that is split up into
samples and the amplitude measured. How many times we sample the wave determines how accurately
our digital representation matches the original analog waveform. We call this sampling rate
and it's measured in kilohertz - not to be confused with the frequency or pitch of a
sound wave. At the low end we have values like 11 kHz - that's 11,000 times per second
we sample the audio. This is used for low quality internet voice transmissions. It doesn't
really sound that good but it makes for small file sizes. 44.1 kHz 44,100 samples per second
is CD quality audio. 48 kHz is the standard for digital video. This rate was chosen because
it could deliver a 22 kHz frequency response (that's refering to pitch) and work with 29.97
frames per second NTSC video - as well as 25 frame/s, 30 frame/s and 24 frame/s systems. But I really like to record audio at 96 kHz.
That's twice the sampling rate of 48 kHz and to me that extra resolution just sounds better.
I'm not sure I could pick one or the other in a blind hearing test but there just seems
like something translucent about 96 kHz that 48 doesn't have. Besides just sounding better,
having extra resolution makes post processing 96 kHz audio easier as we'll discuss in the
next lesson. Dedicated audio recorders can go up to 192
kHz - that's 4 times the standard of 48 kHz, but to my ears that's sort of overkill. Besides sampling resolution, dedicated audio
recorders can also deliver greater bit depth. Bit depth is how many different values of
amplitude each sample can be. With 16 bit audio - each sample can have one of 65,536
values - that's 2 to the 16th power. That's what most professional cameras and codecs
record but with a good dedicated audio recorder you can record at 24 bit which gives each
sample 16,777,216 possible values - This extra resolution contributes to that translucent
quality and ease of processing in post. Another reason to like recording double system
is you are no longer tethered to the camera. This is really useful in situations where
the camera needs to be moving like on a steadicam or a dolly where cables can easily snag. This
is also a consideration if you're shooting events and you don't want to run long cables
between a mix board and the camera. Dedicated audio recorders also have the ability
to compress audio wave files and record MP3 files. For recording audio for film this is
a no-no. Always record uncompressed -that's WAV files unless you have a dying need to
conserve space on your recording medium - say you need to record 8 hours continuously and
you can't get a bigger card. Compressed audio throws away a lot of useful information that
will come in handy in the post processing side and with today's memory capabilities
uncompressed audio files aren't really that big and problematic to deal with (certainly
dwarfed by the size of your video files). So for higher quality sound and freedom of
movement, I'm a big proponent of the double system recording at 96kHz 24bit uncompressed
wav files. You will have to sync the audio in post but you can use slates to line up
the audio on each shot or use sync programs so long as you record a scratch audio track
on your camera. But if speed and ease are your goal, there's nothing wrong with sacrificing
a little bit of quality to record synced sound using the single system. Whether you're shooting with the single system
or dual system sound you will be utilizing a preamp in the signal chain. Pre-amps boost
the signal of a microphone so it can be recorded. Most preamps have a switch that can be toggled
between line or microphone signal. A line signal is a strong audio signal usually coming
from a mixing console or playback device - Professional line out signals are designated as +4 dBu
which has a signal of about 1.228 Volts Root mean square. -10 dBv which is roughly 0.316
volts root mean square is the consumer level line signal.. No further amplification is
needed with line level signals A mic signal on the other hand is far weaker
typically weighing in at only 2 millivolts - that's two one thousandth of a volt. Here's
where the preamp comes into play. But when you boost the signal, you will invariably
boost noise and some preamps are noisier than others. This is another benefit of shooting
dual system sound, the preamps in dedicated audio recorders tend to be quieter than ones
found in cameras. With the mic level there may be a +48v option..
This is to supply phantom power down the line for condenser microphones that need to be
powered. Check your microphone to see it requires phantom power. If you mix up the mic and line sources - say
plug a mic level into a line level input, you will get a very weak signal if anything
at all. Plug a line level into a mic level and you will extremely distorted audio - so
make sure that you have the proper signal strength for your inputs. Now where should we keep our levels? As I've
stated 16 bit audio has 65,536 values of loudness and no more. If you go louder than those 65
thousand levels there's nothing to record and the digital file will clip - what looks
like the top of the wave being chopped right off. This is bad bad bad and it sounds ugly. So we want to avoid clipping at all costs.
and we do that by giving ourselves headroom. Now on old analog systems, 0 dBu was set to
line level and you had about 20 dB of headroom above 0 before analog systems began to clip.
In the digital world, 0 dBFS (decibel full scale) is set to the clipping point - the
maximum loudness. So to mimic the headroom of analog recorders we want our average peaks
to hit somewhere between -20 and -12 dB full scale. What I like to do is keep my average
level between -20 and -12 that way any sudden spikes and boost in loudness will top off
at -6bB - well below the clipping point. The engineers at RODE recommend keeping at
least an 18dB separation between ambient sound and the desired sound so that you can ensure
recording a clean signal. Now there is a train of thought that says
you should push it further and record hotter levels to get better signal to noise ratio
- then turn down the audio to reduce the noise - almost a Expose to the Right approach to
sound. But doing this can record thinner and flatter sound and can potentially lead to
clipped audio which is hard if not impossible to repair. If you have clean audio sources
keep your average peaks somewhere between -20 and -12 dB and your big spikes should
never rarely go above 6dB Moving up the signal chain lets look at the
audio cable that runs the sound signal from the microphone to the preamp. In the world
of audio for film and video we're most likely going to be dealing with analog cables. Analog
cables come in two varieties unbalanced and balanced. An unbalanced cable is the simplest type of
cable and therefor the cheapest. They generally have either a minijack sometimes called 1/8th
inch or 3.5mm connector, a quarter inch connector sometimes called phono plug or tip sleeve
connector, or RCA connector. And for this discussion we're assuming all cables are carry
a mono signal - that is one channel of audio even though there are stereo varieties of
the minijack and quarter inch connector. An unbalanced cable is comprised of two wires
- one serving as a ground which is the zero point for the analog signal and one serving
as the hot which is the signal itself. In a shielded low voltage cable, a metal foil
or braid is wrapped around the hot and acts as the ground protecting the signal from interference
from outside sources. Even with shielding external interference
can sneak in and create noise in the signal. Because of this unbalanced cables are mainly
used for short runs. To combat interference in longer runs we use balanced audio cables. Balanced audio cables use 3 wires and have
either an XLR connection - which is the most common type of connection with almost all
professional microphones using XLR - or TRS quarter inch connection which is tip-ring-sleeve
(this connector may also be used for stereo unbalanced signals) With balanced cables we again have a ground
wire - which is often the shield. But instead of sending the audio signal down a single
wire, we send it down two wires with the second signal reversed in polarity sometimes called
the cold signal. When the signal reaches the end of the line, the polarity of the cold
is reversed and added to hot.. Here's the neat thing that happens. If there's
any interference along the cable, it should affect both the hot and cold the same way.
At the end of the line when the polarity of the cold is reversed and combined with the
hot, the reverse interference signal will perfectly cancel itself out leaving only the
original audio signal. For this reason balanced cables are capable
of long runs without much interference and the XLR input connectors are sturdy and will
hold up to a lot of abuse on set. TRS connectors - or Tip Ring Sleeve connectors
do the same thing as XLRs except instead of having 3 pins, the different signals are handled
by the tip (hot) the ring (cold) and the sleeve (gound). Now we finally arrive at the beginning of
the audio signal chain - the microphone. Before we get into microphone placement and
selection - we need to have a brief word about impedance. Impedance isn't as critical in
the video world as it once was, but its worth keeping in mind. Without getting too technical, impediance
is a measure of opposition a device has to AC current - basically the combined effect
of capacitance, inductance and resistance. This is often designated as the letter Z and
measured in ohms or the Greek letter Omega. Low impedance microphones, sometimes labeled
Low-Z have impedance of less than 600 ohms. Medium impedance mics have between 600 and
10,000 ohms and high impedance is anything above 10,000. In our audio chain we always want to go from
low to high impedance. The microphone should be rated lower than the recorder or else you
have degraded signal. But if you stick with professional level gear
- such as any of the RODE mics we mention here or even other reputable microphones companies,
you shouldn't have an issue with impedance mismatching. It's only if you get a cheap junky mic that
you will find this issue. With that said, let's talk about microphone
placement. The first and most important thing to remember about audio recording is that
sound dissipates according to the inverse square law. Much like light, the power of a sound wave
decreases by the inverse of the square of the distance - if you double the distance
between your microphone and the sound source, you reduce the power of the sound waves to
a quarter. Triple it, and power reduces to a ninth. For this reason you want to get the microphone
as close as you can to the subject. A microphone that is sitting on top of a camera is not
doing you any favors if your subject is speaking 10 feet away - you need to get that microphone
close to the subject. But not neccesaily too close. Cardiod microphones
and other non-directional microphones exhibit something called "proximity"effect which is
the a boost of the bass frequencies when the sound source is very close to the microphone.
Radio announcers use proximity effect to fatten up the sound of their voice: K-Billy's Super Sounds of the Seventies keeps
on trucking here on FMIQ When recording audio on set you may not what
that - but you still want to get your microphone as close to the source as possible. Did I
mention you need to get the microphone as close to the source as possible? Get the mic
as close to the source as possible! For most productions this will come down to
a choice of either Boom Mic or Lav. Booming is simply putting a microphone on
a pole like this RODE boom pole and holding the boom so the microphone is just out of
the frame either from above or from below. Often times a shotgun microphone is used at
the end of the boom. Shotgun microphones are unique because they
have a particularly tight polar pattern. Shotguns like this RODE NTG2 or RODE NTG8 have a supercardioid
capsule - what gives it a tightly focused pickup pattern is the interference tube that
sits in front of the capsule. The theory behind the interference tube is
sound that travels on axis will hit the microphone capsule unimpeded. Sound that is coming from
the sides will be forced to go through slots - since sound waves will hit the slots at
different times, they will be out of phase and start canceling each other out. The longer
the interference tube, the more directional the pickup pattern. But there are some draw
backs as the real world always complicates things. Off axis Sound from moving objects
will not be filtered as well because the the wave is changing position as it is entering
the interference tube. Shotguns work best when the unwanted noise is relatively different
from the desired noise. For this reason shotguns can behave strangely in really small rooms
or in highly reverberant spaces where the off axis sound will become colored. In those
situations a basic cardiod perhaps like the RODE NT-55 may be better suited. But for booming outdoors, shotgun mics are
still a great option. Accessories like this RODE Blimp are employed to cut down on wind
noise. Further wind protection can be added with a muff, sometimes called a dead cat,
or in Rode's case - a dead wombat. The other option of recording audio on set
is using lavalier mics like this Rode Lav and Rode Pin mic. These mics are generally
attached on or near the chest. But lavs can be hidden anywhere - under a jacket, in the
actor's hair - you just have to be concerned about unwanted rustling sounds. Now for documentary work, corporate and news,
I don't personally have a problem with seeing a lav mic in the shot. In those situations
where you don't have a second take and sound is crucial, I'd rather see a small microphone
and get good sound, then try to hide the mic and run the risk of getting rustled sound.
The same cannot be said for narrative work where a visible mic can take your out of the
story. Lav mics are often used in conjunction with
wireless systems - which sort of negates our whole discussion on the audio signal chain
so far. I personally have moved away from wireless systems myself - they are expensive
(don't bother with cheap systems, they're unreliable) but they allow unparalleled freedom
of movement. But the downside for me s they are one more device that eats through batteries
and radio interference can sometimes cut into the signal and ruin a good take. Unless you
really need that freedom of movement, I prefer the security of a wired connection. From the start of the audio signal chain with
the microphone, through the cable and finally to the audio recorder, I hope we've established
a foundation of understanding of how sound is recorded for film and video. There are
a lot of nuances we didn't cover but those are things you will pick up in your filmmaking
journey. And although we have gotten technical here, there's one golden ear rule I follow,
put on a set of neutral sounding monitor and monitor your audio as you record it - if it
sounds good, it's probably good. And if there's some issues, there's some things we can fix
in post - and yes I hate saying that but we get into that in our next lesson in this audio
series. Until then, go out and make something great, I'm John Hess and I'll see you at FilmmakerIQ.com.
Thanks for sharing this!