Recording audio and video

One of the most frequently asked questions about field recording is whether good quality audio can be captured with a video camera. In fact, I that the number of people interested in doing simultaneous high-quality audio and video capture (at least in linguistics and related disciplines) far exceeds the number of people who want to use either medium alone.

The use of both audio and video (A/V) technologies poses a significant methodological, logistical, and financial challenge. I am going to assume that most readers are interested in capturing reliable data, but keeping the cost and hassle to an acceptable minimum. I am going to keep this idea in mind throughout this article.

SCENARIO #1: You need equally good audio and video signals

A/V signal routing

Simple

By this method, you route your microphone feed(s) to an audio recorder only and use the in-camera audio track for reference only (Figure 1). This is a simple solution and it has the advantage of freeing up the video camera for untethered use. The SMPTE time code is provided by the video camera. You can easily replace the in-camera audio track with the one captured with the audio recorder in post production.

Figure 1. Audio signal routing to an audio recorder only; the in-camera audio for reference only

Complex and flexible

By this method, you route your microphone audio feed(s) simultaneously to your audio recorder(s) and the video camera (Figure 2). Make sure your field pre-amplifier/mixer has an extra (in addition to the L + R "mains") stereo output (usually a 1/8-inch stereo out connection). This solution offers you extra flexibility by providing high-quality in-camera audio field mixes for quick editing and distribution (e.g., Web podcasts, intranet use, preliminary analysis, etc.). In addition, you can add an extra audio recorder to provide redundancy in case one of the devices fails during field production. Synchronization can be further simplified by means of a time clock slate (see more below).

Figure X. Multiple, simultaneous routing of audio feeds

A/V synchronization

SMPTE time code synchronization

The general idea is to record audio simultaneously with a SMPTE time code-capable digital recorder (such as Fostex FR2, Tascam HD-P2, etc.) and a digital video camera, so that audio and video can be easily synchronized in post production. Time code allows the synchronization of several digital devices and has standard practice in the motion picture industry for years.Note that SMPTE time code uses frame-level synchronization. This may be a new concept to those of you who are used to audio editing because in the digital audio world, we usually work with sample-level synchronization. Frame-level synchronization is the default editing mode in most professional digital video editing software and it helps "snap" audio and video tracks together in perfect sync.

Time code can be generated by one of your recorders or, better still, by a dedicated digital time code slate (e.g., those made by Denecke). You must keep in mind, however, that time code slates are very expensive. Unless you have a big budget and plan to use video capture for most of your field recording projects, the cost of the slate might be prohibitive. You may, however, look into the possibility of renting one, if you happen to live in a major metropolitan area. Rental costs are reasonable, so if you need to use a slate sporadically or if you need one to learn the skills for future projects, renting might be a very good option. Alternatively, you can quite easily substitute an ordinary clapperboard for a time code slate and use it to create a audible time "marker" in the recording which will help you sync up your footage in post production.

Clapperboard synchronization without SMPTE time code

1. Make sure your video tape is pre-striped (N/A if you use a tapeless camera)
2. Start recording on the video camera
3. Start recording on the audio recorder
4. Use an old-fashioned clapperboard slate (or two 20-inch wood 2x4 pieces of wood) to make a sharp "clap" noise
5. Import both audio and video feeds into your video editing software
6. Identify the clap sound by locating the characteristics spikes in the waveforms
7. Drag the audio-only track to visually match the spike of the audio-video track as closely as you can
8. Your software will probably force your timeline to "snap to frame" which should be perfectly sufficient for ordinary audio/video synchronization
9. Now you can mute the audio track of the video footage and use the audio-only footage instead
10. Your video suddenly has acquired a great audio sound!

SCENARIO #2: Digital camcorder as the only recording device

Some people believe that they can minimize cost, reduce clutter, remove unnecessary complexity, and simplify the A/V recording methodology by getting rid of the audio recorder altogether. Also, in some markets, digital camcorders are more readily available than portable audio recorders. They tend to be less expensive, as well. While it is certainly possible to capture decent audio signals with a camcorder, it is somewhat of a compromise. Consumer-grade camcorders are designed primarily as video recording devices and even though their digital audio specifications are impressive (e.g., 16-bit, 48,000 Hz stereo PCM audio in the DV format), their audio recording capabilities are rather limited. This is true of both Standard Definition (SD) and High Definition (HD) consumer-grade camcorders.

Connecting a microphone to a consumer camcorder is not as straightforward as it might seem. In theory, you should be able to follow the same guidelines as those included in the last section of my post on matching microphones with portable recorders ("Starving student") because camcorders typically use the plug-in power microphone interface. However, your biggest problem is going to be controlling recording levels (the so-called "Auto Gain Control"). The camcorder will try to adjust levels automatically and, if it is not perfectly matched with the external microphone, it will cause a lot of noise in the presence of quiet sounds and distortion in the pr esence of loud sounds. For example, this problem has plagued modern DLSRs, which have trouble working with external micrphones and microphone pre-amplifiers. You should consult your camera maker's technical support documentation before buying an external microphone.

Sony, for example, makes its own line of stereo camcorder microphones which do an excellent job. I have been using the inexpensive Sony ECMMSD1 stereo microphone with my Sony DV camcorder and have been very pleased. While the sound is nice and rich, I would not use it for any serious work because it lacks the acoustic and electrical performance I need. You may try connecting the Audio-Technica ATR3350 or Azden ECZ-990 to your camcorder but I cannot guarantee that it will work.

Of course, with this approach, you do not need to worry about time code. After you have finished recording, all you need to do is transfer your footage to a personal computer for further editing and/or audio track extraction. There are two easy ways to extract a WAVE file from a DV or HDV file (I assume those will be the most commonly used file formats in this case). You can do this while capturing the video onto your PC via FireWire. Simply instruct your capture software (e.g., Adobe Premiere, Apple Final Cut Pro, or equivalent) to capture the WAVE file only, or capture the entire A/V signal and then export the WAVE file out of your video editing software. Similar methods will work with video formats other than DV or HDV. Then you will have to find a way to deal with resulting the stereo WAVE file, as all your speakers' voices will have been automatically mixed down into a stereo file, so separating the voices out for analysis will be rather difficult.

What do buy?

Video camera

I often receive questions through this website about which video camera to buy. For low-budget projects, any of the current consumer camcorders in the $1,000 range will do, provided it has a microphone input. However, recommending a high-end video camera is difficult without knowing more about one's budget, geographical location, tech support needs, etc. Besides, I do not really follow the A/V market closely enough to be familiar with the latest product portfolios. All I can recommend is that you buy a camera with the following features:

• a 3-CMOS sensor (or equivalent)
• a robust and stable video format, e.g. HD 1080/60i, 30p, 24p
• dual XLR mic/line audio interface (important!)
• SMPTE time code input and output
• optional manual focus, exposure, and white balance
• long battery life
• headphone monitoring

Audio recorder

It would be nice if the audio recorder had the following features:

• dual XLR mic/line audio interface
• SMPTE time code
• 24-bit/96,000 Hz PCM audio format
• headphone monitoring
• synchronization with SMPTE time code
• digital audio I/O
• long battery life

Concluding remarks

There is no doubt that simultaneous A + V recording for the purposes of speech research has grown in popularity over the years. While most consumer-grade video equipment is relatively inexpensive, the cost of professional video production can be prohibitively expensive. In addition to high cost, video production requires extra equipment, such as lights, light stands, light modifiers, reflectors, and battery packs. The use of video technology in the field can, therefore, be rather cumbersome; so much so that working without an assistant is virtually impossible. However, the extra effort can often result in impressive work. Walt Wolfram's series of documentaries on dialect diversity is a great example of skillful use of audio and video technology in field linguistic research.

What about wireless audio?

Wireless audio recording, particularly in video production, is ubiquitous these days. However, I typically do not recommend using wireless technologies. The acoustic performance of some of the high-end systems has significantly improved over the years, though the same cannot be said of the lower end of the product line. There's the real risk of interference, signal drop-out, fluctuating frequency response and dynamic range. But, to me, the most important disadvantage of wireless audio is the extra layer of complexity. If you are out in the field, working alone, far away from your office or lab, interviewing subjects in their homes, places of employment, public places, etc., the last thing you want is to wrestle with too much technology. Besides, wireless audio does not really offer any real advantage over wired recording, does it?