Mark A. Sicoli, October 2010
I just set up a new Linguistic Anthropology lab at the University of Alaska, Fairbanks and so I recently thought about video and audio equipment and made some decisions for which I will share the reasoning and resulting equipment.
Tape or Solid State? One of the latest questions to think about is whether to use the new solid-state SD recorders or to use mini-DV tape. I consulted with the Technical Group at the MPI, and with some other linguists and decided to go with mini-DV tape (even though this means carrying more weight to and from my field sites). This was mainly because the solid-state video technology is not yet mature enough for creating media that can be archived. There are two problems with the SD video recorders that have not yet been solved: the first is that sound is compressed minimally 4x more than with mini-DV recorders that record uncompressed audio. The second is that the video coding of SD recorders is still not open, meaning that proprietary software (and sometimes hardware) is generally needed to import the video to a computer, to edit, it and convert it to another format for editing or archiving. In a few years I think both of these problems will be taken care of, but not yet. So for the lab I bought two Canon HV40 cameras, a mid line mini-DV HD recorder, which I was able to get for $650 each.
The upside to the Solid State digital recorders is the ease of transfer of files to computer which is like moving any other large computer file. It takes some time but not anywhere near the actual recoring time. Mini-DV tapes on the other have real time transfer rates meaning that it takes an hour to transfer the data of one hour of tape (though nothing has to be done during the hour except click the mouse to get it started).
With mini-DV tapes (and DAT audio tapes), you need to take the precaution of wrapping your tapes in aluminum foil before sending them through the airport x-ray devices. I bundle my tapes into either boxes of 5 or larger groupings and wrap them up together and label them videotapes. I’ve never been asked what’s inside, I think it’s still possible to see through the foil but the exposure is lessened to the point that no signal loss occurs.
One note is that many video cameras come with the hardware to charge batteries by plugging in the camera with the battery mounted. Unfortunately this means you can’t charge one battery while using another, so it is recommended to buy an external charger that can work separately from the camera. These are inexpensive and readily available.
Buy mini-DV tapes in bulk if possible. Some vendors with give a discount when buying a box of 50 tapes. You may have to ask for this, but it can save between 30-80%
Two cameras or one? While budget constraints surely dictate whether two cameras are possible, research questions and methods may necessitate two cameras. The second camera allows the opportunity for community members to record while a researcher can use the other camera simultaneously for different work. The second camera can also be used to provide different angles on a speech scene that can be useful for larger events (like town hall meetings or classrooms) or for working with multiple speakers where actions like gesture or gaze are at issue. For the lab it also allows multiple students to check out equipment for assignments or research.
External Microphone: Built-in video camera microphones do not provide the best sound quality, so an external microphone is recommended for most applications. One of the problems with the built in mics are that they can pick up machine noise from the movement of the tape and also the breathing of the camera operator. It’s best to use an external mic off the camera, or when using one on the camera, it is best to use one that is mounted in a suspension system. When buying a camera be sure that it has inputs for an external mic (you’d be surprised that some do not). These are standard 3.5mm/1/8″ stereo connectors that look like a headphone connector. We have two types of video mics in the lab: a fairly inexpensive Sony ECM-MS907 stereo microphone ($65) that can be set up off the camera on a table or on it’s own stand; and a Rode Stereo Video Mic ($250), which has a suspension mount system that can either be used on top of the camera or on its own stand. The Rode is particularly nice for capturing sound that’s farther in front of the camera and for when a camera mount is the best setup, the Sony is nice for closer and wider fields of action.
Wide-angle lenses: Video cameras tend to not have very wide fields of view, which is fine for capturing action beyond 10 feet or so but can be a problem when working inside of rooms or for interviewing contexts. There are a number of wide-angle lenses on the market that screw onto the front of the camera that can range from about $17 to $170. Wide-angle lens attachments are categorized by numbers less than one that multiplied by the lens size to which it is attached gives the adapted lens size. The two most common are 0.45x and 0.7x. If fixed to a 50mm lens, for example the 0.45x 50 = 22.5mm vs the 0.7x 50 = 35mm, so the 0.45x will give a wider field of view. The wider the lens though the more distortion may be introduced around the periphery, so the 0.7 would have less edge distortion. We are using a 0.45x from Digital Concepts that fits the 43mm lens attachment thread of the HV40 camera. It was just $17 and I’m actually very happy with it. The distortion is minimal and it gives a great field of view when working in tight areas.
Tripods: Especially for Conversation Analysis, use a tripod whenever you can. The action in the video will be most amenable to analysis if you don’t need to abstract away from the shaky camera movements. For the same reason use features like zooming and panning as rarely as possible. Gesture, proxemics, gaze, and non-verbal actions generally all depend on the physical relationships between participants and often occur in simultaneity with the speech of another. The indexical dependencies cannot be determined or are missed when panning from one speaker to another. This is another reason to use a wide-angle lens attachment: to capture all the participants in the same frame if possible.
Monitoring sound: As with audio recording, monitor the video recording’s audio track through earbuds or headphones while recording especially at the beginning of your session to make sure that your mic is turned on, the cables are fully inserted into their contacts, and basically that everything is set up right to capture the sound. The risk of using an external microphone is that if you forget to turn it on you’ll create a silent video. Regular monitoring prevents these accidents.
Importing video: Transferring HD video requires a firewire connection (when buying be sure the camera has this) and firewire cable which is generally 4-pin on the camera side and 6 or 9 pin on the computer side. Firewire connections are standard on most Macs and generally aftermarket options for Windows and Unix machines.
Video can be imported by Adobe Premier for both Windows and Mac formats; with I-Movie or Final Cut Express for Macs; and with open source editors like “Open Movie Editor” for Linux.
Transcribing, coding, and Subtitling. While transcribing and coding could have a separate blog, I’ll just mention a few of my preferences here knowing that other people have other preferences. I use ELAN for transcribing video and audio. It allows for multiple audio and video files to be linked and time-indexed to transcription files and exports in multiple formats including .srt subtitle files. The “score like” tier structure makes it fairly easy to code simultaneous features like prosody, gestures, gaze, other non-verbal actions, and simultaneous speech. It has continuous support from the Max Planck Institute and is under constant development including new modules that will semi-automize some research tasks (stay tuned and update the software regularly). Available for Mac, Windows, and Linux. http://www.lat-mpi.eu/tools/elan/download
Many video players will read .srt subtitle files and overlay them automatically as long as the .srt file has the same name as the video file. For inserting subtitled videos to Powerpoint type presentations the subtitles need to actually become part of the same file. For this I found an open source program for Mac called Submerge. Of course remember to save the file with a new name because the subtitle will be permanently burned into the video: http://submerge.en.softonic.com/mac I don’t know what’s available like this for Windows and Linux. If anyone else knows, please let us know and we’ll add the information.