Let’s say you are assigned to shoot an interview video. You set up the location, position the lights, mount the camera on the tripod and test the mic. Everything is in order. Your interview subject arrives, sits in place and as the tape rolls, the interview begins. The shoot is a success–or at least you thought it was. Once back in the studio, you discover that your normally soft-spoken subject actually got quite loud. So loud, in fact, that the audio clips most distressingly, and most frequently. Or perhaps a typically outspoken subject is overcome by emotion and whispers their way into tape hiss. Either way, your audio is not what it should be.
I’ve been burned by the too-loud or too-soft audio thing too many times. One solution is to bring along a dedicated audio guy to keep tabs on the level. Bigger crews have that luxury. However, the typically one-man-band church video dude (or dudette) has a lot to keep track of. Between lighting, composition, exposure, set dressing and making sure the tape is actually rolling (that’s another post), audio is just one more thing that is easily overlooked.
Yesterday, I was reminded of a little trick I learned a long time ago. Now, this only works when you’re interviewing one person at a time, but it just might save you from the pain and duress of bad audio. Set up your mic input so that both channels are getting the mic signal. We use a GL-2, and that is easily accomplished by using the 1/8″ to 1/8″ cable from the wireless mic receiver to the mic-in jack on the camera. Other cameras with XLRs can often be set to send one input to both channels (the JVC GY-110 oddly takes Ch. 2 and will send it to both 1 & 2).
Once you have the same signal on both channels, set one channel’s level to get pretty close to full-scale at normal talking level. Not so it’s peaking, but getting up there around -6 dB or so. Now, set the other channel to top out at -15 to -18 during normal talking.
If your person talks softly, you are probably still going to have a useable signal from the higher channel. If they are really loud, the first channel will be blown out, but the second will be fine. And if they are very dynamic (this is where this technique really shines), you can switch back and forth as needed.
This has saved me on several occasions. I normally set my main channel to have solid signal when a person is speaking normally. However, should they get really excited and get loud, the main channel will run out of bits and get grossly distorted. The second channel, set 12 dB lower, will just be coming on line. It’s a quick few slices with the digital razor blade to pick the best channel in the editor the next day. With the judicious use of compression, you can level the audio out so it still has plenty of dynamic range, yet will still be present and crystal-clear.
This technique is kind of like car insurance. You hope you never need it, but if you do, it’s nice to know it’s there.
Subscribe in RSS Reader
That’s brilliant! This is definitely going into my “save for later” part of my brain. I use Panasonic DVX-100B’s at work and this will certainly help next time we do an industrial video.
Our system is being installed as we speak, and it sounds REALLY GOOD!!! They flew the subs near the ceiling, and since I wasnt sure about that I contacted you a couple months ago, but they boosted the power to them, and hung them close to the ceiling to get the added wall bounce into it. you can see pics at http://grahamprouty.wordpress.com/2009/02/09/pictures-nearly-finished/
Thanks so much for your help and direction! It made a big difference, and we are happy with the results! I really appreciate your willingness to use your knowledge to help others! Have a great one, and thanks again!
Two (2) other great tips (although not audio related) are 1. Turning off the auto-focus, zooming all the way into your subject, focusing on your subject and zooming all the way out. The result is being able to then zooming in or out on your subject and not losing focus. 2. Keep your handheld shots steady be zooming all the way out and moving closer to your subject. The result is a steadier shot and it works great.
Sorry typo… Number 2 should state, “… by zooming…”. Not, “…be zooming…”
I would think that since -15dBFS equates to +4dBu or 0VU, it should provide enough resolution (in 24 bits) should your subject start whispering to simply raise the level in a video editor without boosting the digital quantization noise. Besides, that is what dither is for.
JB,
It might at 24 bits, but most camcorders are still recording at 16 bits. Admittedly, since we’ve gone digital with the audio of video recording, it’s less of an issue than it once was. Still it always pays to use as many bits as you can (without using too many…).
–mike