illuma·blogspot·com

Monday, March 29, 2010

••◊ 5D audio made simple...or is it?

Some of the major complaints I hear from "traditionalist" videographers is the form factor of the 5D mark II (or any DSLR really). They just can't wrap their heads around the lack of immediately available controls and no real world audio inputs. OK, so they have a point that the form factor is less than ideal but not totally unworkable. With the new 5D firmware upgrade that allows for manual audio levels I decided to try to use in camera audio with my lavs; which highlighted the audio input issue previously mentioned. ...the solution? ...a trip to Radio Shack for some common components.




A picture of the resulting home brew project is below along with the simple to build schematic. Nothing special right? The schematic is fairly dead brain simple and it's not like I needed XLR inputs or anything fancy for lavs. Anybody can figure this out and build it. So why blog about it? Well, this wasn't the difficult part.




The real issue to discuss here is gain structure; the arch nemesis of any audio engineer. Really you have three points of audio amplification when using a lavaliere microphone. There's amplification at the transmitter where the mic plugs in, amplification at the receiver, and amplification in the camera. At any point in the chain the audio signal can clip and you won't know where the distortion is coming from without a full triage.

I put on my scientific thinking cap and came up with the following scheme. I wanted to take 2/3rd's of the gain structure out of the equation at any one time because fiddling with buttons in the middle of the shoot doesn't give anyone a warm-fuzzy. Here's what I did...

First I determined that if the lav transmitter and the camera distort at the exact same time, then effectively that removes the lav receiver and camera as a source of distortion in the first round of triage. So how did I determine the correct gain levels for each device? I simply moved the lav microphone closer and closer to a speaker playing a 1kHz tone (earplugs highly recommended) until the lav transmitter lit up an LED showing distortion. At the same I was recording the output of the lav receiver into the camera. After importing the recorded sound into Adobe Soundbooth I could visually see when the distortion occurred (see pictures below). If the distortion appeared before the signal was clipped at 0dB, then I knew that the amplifier gain was too low in the lav receiver or camera (precedence given to the receiver, then camera). If the signal was clipped at 0dB before I saw any pre-distortion from the transmitter then I knew the gain was too high from the receiver or camera. When the transmitter and camera audio distorted at the exact same time then I knew I had the gains set correctly. The theory being if the camera was distorting at 0dB, then the lav mic transmitter was also distorting. Gain from the transmitter was arbitrary, as long as it allowed signal clipping to occur.


Camera gain too low. Receiver gain=-12dB, camera=4 clicks above mute. Note how the audio clips before the camera distorts.

Just what we want. The lav audio distorts at exactly the same time as the camera. Receiver gain=-12dB, camera=5 clicks above mute.

Too much camera gain. The camera is clipping the audio before the lav distorts. Receiver gain=-12dB, camera=6 clicks above mute.



Yeah, yeah, yeah...so what do I set the gains to, right? Well, this only works for the G3 lav system into the 5D. The process is same for other audio inputs devices as well. My gains were set as follows: receiver=-12dB, camera=5th step (counting up from mute).

Now this is just a starting point. If I start here and don't see enough audio signal the first thing I do is turn up the lav transmitter. If that's not enough then I turn up the lav receiver. ...and as a last resort I turn up the in camera amplification. Simple troubleshooting, yes? Now, if the opposite is true and you have a distorting signal then the only thing you need to do is turn down the lav transmitter - in fact that's all you can do to keep from distorting since the lav and camera signals look exactly the same time in this configuration.

So how did my interview lav audio sound? Excellent in my opinion. I'm editing it into interviews right now.

2 comments:

Mike Hall said...

I so wish that wasn't all Greek to me! It was interesting nonetheless.

Goes.2.Eleven said...

Well...ask questions. That's what I provide this information for. Doesn't do any good if I don't make it understandable. Where do you want to start?

Post a Comment