Aug 14, 2009

Perspectives On Some Audio Fundamentals


For the past year I have had the good fortune to work on a History Channel series titles "Greatest Tank Battles." From an audio perspective it is not overly complex but it has given me ample opportunity to ponder the fundmentals of recording a sit down interview.

In the first picture you can see my boom pole holder that will attach to any light stand and work with a very light counterbalance - in this case a Porta-Brace refillable sandbag I own into which I usually put one of my extra Cantar batteries. The light stand is also mine. I never like to borrow if I don't have to.

The audio for the series is anchored by interviews with veterans which are recorded either in front of a green screen or done "in situ" on the actual battlefields. I have always been a firm believer in double mic'ing at every opportunity. This means using both a lav and a boom mic. In the case of "Greatest Tank Battles" the technique is that of a hostless interview in which the questions being asked by the director are not used in the final edited version, only the responses. This allows me to place the boom on a stand over the subject. The interviewer is mic'ed for reference only in the edit suite. I generally split the tracks I send to the camera with the lavs on one channel and the boom on the other. All tracks are individually recorded on my Aaton Cantar X2 and I burn a DVD of this audio for the editors.

Double mic'ing makes for a much richer total sound. Lavs alone sound thin and give little sense of space. A good boom mic always sounds better than any lav I've ever heard. Lavs will only sound better in hostile audio environments due to their proximity to subject where one is using them to pick up voices. Also one can never know for sure how people behave over long interviews. I have seen some subjects fidget with their hands, cracking knuckles right across from where I've placed the mic. Too often in a fit of emotion or pique people clutch their chest with their hand producing audible crackles and pops from the lav. For the History Channel series on tanks I've also had to deal with veterans who have a chest full of medals which clink and clang whenever they shift in their chair. The lav is the most susceptible for picking up this kind of unwanted noise.

My usual documentary boom mic is the Sanken CS 3e short shotgun. I love that it is a very tightly directional supercardioid. I have friends who are rooted in drama that complain about the noise floor of the Sanken compared to their Shoeps capsules. They have a valid argument except for the fact that few docs are shot on studio sound stages. On most docs that I have worked the shooting goes on in very hostile locations for audio where the noise floor of the mic is the least of your worries. Also the Sanken has a tremendous reach which means I don't have to be literally on top of the subject to capture that rich fat audio. In other words, I don't have to crowd the frame. That said, the Sanken is TOTALLY unforgiving if you are off axis and its dynamic range is only suitable for speech. While it is clearly not the mic for every situation I feel its plusses easily overcome its minuses in many situations I'm faced with. So much so, that I own two. Actually, I bought the second when the first one went down and was sent to the factory in Japan for repair which ended up taking some six months.

While I started using the Sanken CS3e as the interview boom mic for the "Greatest Tank Battles" series, I early on wanted to switch over to the Shoeps CMIT 5u short shotgun.

The CMIT 5u is one of the sweetest sounding boom mics I own and is one I use for most sit down interviews. I haven't quite got used to using it in the rough and tumble of shooting verite. Perhaps, because it comes with push button filters on the tube of the mic which when using a full Rycote zeppelin are a pain to manipulate and because it doesn't have quite the reach of the Sanken CS3e I still use the Sanken as my dominant mic.

Earlier this year (2009) the Shoeps CMIT 5u started sounding a little off especially when I plugged it into my Cantar. I took the mic to Audio Services where the bench tech examined it and declared that there seem to be traces of moisture on the filament. Since the mic was only sparingly used we deduced the moisture must have come from condensation when packed for flight. Since I fly seemingly almost all the time the explanation seemed plausible enough. It was decided to send the mic back to the factory in Germany to bring it back up to spec. I also e-mailed a wav file I recorded of the mic to give the bench techs in Germany a sample of the digital artifacting I was hearing.

When the mic came back from Germany I immediately plugged it into my Cantar and heard that digital rumble whenever I moved my hands on the boom pole. Once again I contacted the Shoeps people with my concerns. They came back with a thought provoking e-mail:

"We listened to the... sound file, and it sounds very strange. It sounds like some digital effect but there are no digital components in the microphones (sic) signal chain. On the other hand the noise is away if the customer switches on the low cut filter. So we have another idea. Shaking the microphones (sic) the way that the customer does causes very high low frequency noises. This may overdrive the AD converter of a following recording unit. In the worst case the AD converter may produce such noises. The effect is away if the input is not overdriven any longer, for example a low cut filter prevents these signal peaks.... For mobile recording of voices we always recommend to switch on at least one of the two frequency filters."

I also passed along my concerns and the e-mail from Shoeps to the Aatechs in Grenoble, France where the Cantar is manufactured. Their response was all knowing:

"All this is coherent with the user manual p. 16 instructions about avoiding ultra-low frequencies to the preamps. We will add Shoeps CMIT 5u and the Neumann KMR 81 to the sensitive mic list. Thank you for your thorough report."

I felt like such a fool. There are days when you can't seem to see the forest because of the trees. I had become so used to working with my Sanken Cs3e's and my Sanken CS5 that don't generate ultra high low frequency handling noise to overload the AD converters in the Cantar that when I was confronted with this problem I immediately jumped to the conclusion that the fault was with the mic while conveniently forgetting to throw some filtration on the Cantar. I now know that when I use my Shoeps CMIT 5u or my Neumann KMR 81 I go to the 60 Hz -6dB setting on the Cantar input. Also, never underestimate the value of reading and then re-reading the manual.