Recording good, clean audio is really important. Use a good quality microphone, like the Sure 57. Try watching a TV show with less than ideal video; we can watch it although it is a bother. Try watching something with bad audio; we just can't do it.
Record into the camera and while your shooting, record ambient, or background, audio. You can use this extra audio when you put the video together in post-production.
When you are recording sound, follow these recommendations:
- Setup and test well in advance
- Always do sound check before recording
- Always use headphones
- Check for ambient sound such as wind & water, jet planes, trains, highway noise, sirens & horns, etc.
- Use blankets and wall hangings to get rid of echo or hollow sound in rooms
- Wind; use a windscreen or angle the microphone
- Listen for 60Hz Hum
- Capturing sound directly to the camera
- If off-camera recording:
- If possible, get backup recording to camera
- Using a mixer for multiple microphones
- Use a clapper to set a sync point for audio and video
- Use unidirectional microphones if possible or a shotgun microphone
- Position lavalier microphones about 15cm below head
- Tape cables and wires with duct tape
- Point shotgun microphones directly at the action
- Record several minutes of ambient background noise for post-production use
- Never compromise sound quality. Get it again if quality is doubtful
Here are the different types of microphones you can use:
Omni directional microphones pick up sounds from all around it. Its good for ambient sound: environment, outside, nature - big picture sound.
The lavalier microphone is a very small omni directional microphone that clips onto a subject. It is usually hidden in the clothing.
The Cardioid microphone picks up sound mainly in the front and some from the sides.
Unidirectional microphones pick up sound in front of them only. These microphones must be aimed at the subject to be effective.
Shotgun microphones pickup sound from exactly where they are aimed. They have a very narrow focus.
Wireless microphones are usually handheld or lavalier. They are comprised of a transmitter and receiver pair. The subject wears the microphone and transmitter and the receiver is plugged into the camera.
- Always have Extra Batteries
- Check the Transmitter Gain make sure you aren't clipping.
- Host not to rub or move fabric when shooting
Bandwidth or Data Rate
Bandwidth is a word used to describe the 'space' dedicated to transmitting a certain amount of data in a certain amount of time. Usually bandwidth is represented as bits per second or bytes per second (i.e. how much data flows through a wire, or air, in one second). As there is a limited amount of space used for transmitting sound or video, we must adjust 3 digital waveform variables to best fit our situation.
The Digital Waveform Variables
Digital sound has three variables that describe how to represent a waveform. These are:
- Sample Rate
- Sample Width
- Channel Width
In general, the higher each of these variables values are the better the sound quality and more bandwidth it consumes. There is a point, though, where the human ear cannot distinguish a difference in these variables or the equipment playing the sound can reproduce the sound any clearer. Making the values larger in this case would be a waste of time and bandwidth.
A sample rate is how many times a second the recording device looks at a waveform and converts what it sees to a number. There are different sample rates for different recording devices, each having it's own strengths and weaknesses.
Music CD's are recorded at 44Khz as are minidiscs. VHS quality tape uses 22Khz. DAT, or Digital AudioTape, uses 48Khz. High end recording studios us 96kHz equipment.
With DV, there are two common rates. 32Khz or 32,000 numbers are generated every second to represent the waveform and 48Khz. The higher the sample rate the better the waveform is represented when it is played back. This is especially true with higher frequencies.
Sample Width (or Depth)
There are 2 common sample widths that are used in video production. These are 12 bit and 16 bit. A bit is a '1' or a '0'. The recording and playback equipment will string 12 or 16 of these bits together to represent one sample. With 12 bit sampling a waveform sample can have a value of 0 to 4095. With 16 bit sampling a waveform sample can have a value of 0 to 65,535.
The 16 bit sample has more 'resolution' in which to represent a waveform, 16 times more than that of 12 bit. Therefore it can better represent a waveform more accurately, especially with sound that has high harmonic content.
The last variable we can monkey with is the number of sound channels there are. They are stereo and mono.
Sound Compression and Codecs
One of the ways to get more sound data down the bandwidth is to compress it. Different Codecs (Compressor/Decompressor) give different quality for different types of sounds. Some are better at compressing music and some are better with voice.
Sound Editing Software
Good sound editing software should be able to manipulate the 3 sound variables, give cut and paste-paste/mix capabilities and allow you to apply different filters and effects to a waveform. Our favourite is Sony's Sound Forge.