An idiot's guide to audio issues
 

First, let me preface my questions with the provisio that I have read the threads in the FAQ. Sadly, while they have been helpful in many ways, I find I lack even a basic foundation that many of them seem to assume the reader will posses. Thus, I wonder if anyone can direct me to sort of an "idiot's guide" to film audio.

What I'm looking for, more or less, is a sort of checklist of things to do to or not do to minimize the amount of screw-ups I make. Now, as I only make films for fun, and I enjoy it as a learning process, I'm not looking to get everything right away. But some pointers would be nice. On to specific questions:

1) A few minutes here, on wikipedia, and elsewhere on the web has confirmed my suspicions that the abbreviations db and khz used here stand for decibel and kilohertz. I know that high frequency sounds have higher khz that lower ones, and I know decibels are a system of volume measurement. What I don't know is why the values people use matter.

I have read in other threads that some people suggest a -12db baseline for audio. Why is that? What does that mean? Are you lowering the default intake volume, or is "negative" negative in some sense other than "reality" (if reality is 0).

2) How do you use a mixer? I've rented a Shure mixer before from the local gear shop just to play around with it, but I don't really know how it works. What can a mixer control? Are there some default settings for simple dialog scenes someone can point me to? What are some of the things I should be looking for the next time I have a mixer in my hands?

3) Boom Technique: I have read elsewhere that boom operation is a complicated skill, and having tried my hand at it a few times during the last film I made, I can believe it. My new project consists mostly of different scenes with different locations with two speakers. Should I look to record each audio twice, with the boom pointed towards one each time?

Sorry for the length. Any suggestions would be much appreciated, on the three topics above, or really anything you think might help a total audio novice.

Decibels are actually a ratio that compares two levels of sound or power or energy. We can pick anything as our reference value and call that "zero dB" and everything else is either stronger or weaker than that standard. A sound that is 3dB greater than reference is twice as loud. A sound that is 3dB less than reference is half as loud, 6dB less is 1/4 as loud, etc. The negative numbers -12 and -20 etc that you're seeing comes about like this ... in a digital device sound is converted into a string of numbers with each level in the waveform corresponds to some specific number. With 16 bit recording, the norm for video and CDs, the range of numbers is from 0 through 65,535 (2^16). We arbitrarily say that the loudness represented by the number 65535 is "full scale" and the system can't handle anything louder. Since that is a brick wall that we can't go beyond we have decided to make that our reference value and call it "0dBFS." All other sounds are reference against that full scale value. So a sound that is "-12dBFS" is a sound that is 1/16 as loud as the maximum sound our system can handle and so forth.

Mixers do a lot of things - too many to go into a lot of detail here. Basically they allow you to control the levels coming from 1 or more mics and send the signals on to your recording device. They can take low level mic signals and raise them to the level devices requiring line inputs can accept. It lets you take several mics and combine their signals into a single mono or stereo signal. It also lets you monitor the levels visually with a set of meters. Many mixers also have limiters that will 'clamp off' unexpected super loud sounds before they have a chance to overload your system and filters that let you reduce or eliminate low rumbles from wind or noises from handling the mic and boom.

If you have a script your boom operator should know it as well as the actors do. As their dialog goes back and forth between them, a skilled boom operator will turn the mic to follow the conversation (often a director has the talent pause a beat or two between lines with the pauses being edited out in post). And of course you don't just shoot one take of a scene - it would be very common to shoot coverage with 5 or 6 shots or more - long two shot, close two shot, outside over-shoulder reverse on A, inside reverse on A, over-shoulder reverse on B, inside reverse on B, etc so you can edit it together into a proper dramatic flow. Each take will have its own audio and in the reverses the boom would stay with the on-camera character.

Hope this helps

I would start by reading the book "Location Audio Simplified" by Dean Miles. It has a lot of the things you are going to want to know. It is probably the best book out there that covers how to use the mixer, how to use the boom, how to wire talent with the lav. Things you need to know. You can order it from any number of audio stores (Trew, Coffey, etc.) or direct from Dean here:

http://www.locationaudiosimplified.com/

Nothing comes close to this book.

Wayne

dB is the abbreviation for decibels with the bel (named after Alexander Graham) being a unit representing a factor of 10 in power thus a 10 milliwatt signal is 1 bel (equal 10 decibels) stronger than a 1 milliwatt signal. To compute decibels one takes the logarithm of the ratio of two powers being compared and multiplies by 10. In comparing the 10 mW signal to the 1 mW signal the dB = log(10/1) = log (10) =1. If, OTOH, we are referencing to the 10 mw signal the formula is log(1/10) = log(0.1) = -1. The minus sign thus comes from the properties of logarithms. Changing from higher to lower power always gives negative decibels which should be intuitively pleasing. You may wish to play around with logarithms in Excel to gain some insight here. Note that if the comparison is made with voltages or counts out of an A/D converter the db formula uses 20 times the log because power (watts or quatloos) is proportional to the square of the volts or counts.

Hz, kHz may be easier to understand if they were called by their original names "cycles per second" or "kilocycles per second" which is the number of times the sound pressure wave, its electrical or digital representation cyles from positive to negative in one second. Heinrich Hertz was the first to transmit radio waves and so the IRE began to replace cps (cycles per second) by the unit "Hertz" some time in the 50's or 60's (if I remember correctly).

I thought 6dB was the delta where a sound would be perceived twice as loud or half as loud - am I confusing things?

- Martin

One dB is generally considered to be about the minimum discernable change in loudness. That's why we use decibels instead of bels. The ear responds logarithmically (as does the eye and probably the nose...) so while a doubling of power produces a sound which carries twice as much energy per unit time the loudness change is 10*log(2) = 3 dB which is about 3 times the minimum detectable level and the change will not seem that great. 3 dB is generally considered to be the level change which is readily discernable. If I recall from way back in my radio days the steps on the faders were 2dB - you couldn't really hear the change of one step unless listening carefully to a test signal. Today my surround system volume control displays half dB steps. Ridiculous! Well, maybe not from a marketing POV.

Go and check these video guides out mate, they are quite informative:

http://www.pixelcorps.tv/gmt1

Too many people on the web want an instant-gratification answer,
when the scope of what you're asking is WAY too broad.

You're not going to get far in high-technology hobbies by trying to
keep up with scattered web answers that you can't organize and get back to....

You need to keep a couple of good reference books in your library
(and if you don't already have a library, START ONE!)

In answering about what the minus means I completely overlooked the basic question. The optimum (in the sense of maximum NPR or signal to noise ratio) "loading" for an A/D converter when the input signal is Gaussian is acheived when the RMS voltage is approximately 12 dB (the actual number depends on the number of bits the converter is capable of) below the voltage which saturates the A/D. With digital equipment dB are often referred to this "full scale" voltage. Thus -12 dBFS is close to the ideal load (for Gaussian signals). Note that a pure tone is not Gaussian and the ideal load for it is -3 dBFS. Note also that speech is not Gaussian. The ideal load is that which never reaches 0 dB fs. Music is more Gaussian simply because there are usually several voices. A sum of signals tends to be Gaussian.

May I suggest: http://home.comcast.net/~tyreeford/AudioBootcamp.html

Regards,

Ty Ford

I second "Location Audio Simplifed". It has a lot of great info on location recording. One of the best books I have (and I have them all) is Audio in Media by Stanley R. Alten. It was one of my text books when I was going to school for broadcasting. It covers a huge amount of information and is well written. It's quite expensive but you might be able to find it used.

I'll thank you to avoid mis-characterizing a request for information as a search for an "instant-gratification answer."

As for everyone else, I appreciate the useful replies. I will definitely be looking into the resources suggested.