In order to conduct the oral history project this summer it was essential to build up a “recording” vocabulary. Historians do not typically talk in terms of digital recording, but it is important that those of us engaged in oral history understand the process where a microphone converts a sound wave into electrical energy, “generating a current that is “analagous” to the waves frequency and amplitude.” I did not write these words. Clearly. They are taken from Dr. Boyd’s work found on the Oral History Association’s website.
The Oral History Association: Learning Audio Recording Language
For those interested in acquiring the lingo of oral history audio recording I suggest visiting the “Resources” page found on the Oral History Association (OHA) website. This page offers an excellent summary of the digital recording process. The following excerpt is taken directly from Dr. Boyd’s OHA post. I only post it here in its entirety for people who follow this blog and would like to learn a little bit about the oral history language, and also for my co-workers at the Metras who will be conducting interviews in the near future and want to understand our recorder).
(From Dr. Boyd)
- Sample Rate is the number of samples or “snapshots” taken of the signal and is measured in Hz/second. The higher the sample rate (or the more samples per second) the better the digital representation will be. “CD quality “equals 44,100 samples per second or 44.1 KHz and is the minimum recommended sample rate for field recordings.
- Bit Depth refers to the number of bits used to represent a single sample. For example, 16-bit is a common sample size. While 8-bit samples take up less memory (and hard disk space), they are inherently noisier than 16- or 24-bit samples. The higher the bid depth, the better the recording, but also, the larger the file produced.
- Channels: Single channel recording is known as monaural or mono recording. Stereo recording involves the recording of 2 channels (left and right). In interviewing situations, the two channels associated with stereo recording allow the separation and isolation of a channel for the interviewer and the interviewee. This gives the recording a spatial dimension missing in mono recording and enables a listener—or transcriber—to isolate a single voice in playback. Single-point stereo microphones involve the left-right separation but do so from a single source, yielding much less sound isolation. Stereo recording doubles the data footprint—that is, the file size– of your recording. Some recorders are capable of using a stereo microphone setup but recording in mono by “mixing” the two channels together. The resulting file is half the file size of a stereo recording, but the channel isolation will be lessened. Some portable recorders will not allow you to record mono files.
- Recording Quality and Compression: Digital Recorders provide a variety of parameters to adjust regarding recording quality. Field recordings should be recorded without compression. Compressed recording formats are usually measured by “bit rate,” which is calculated by an equation involving bit depth, sample rate, and the number of channels being recorded.Lossy compression (such as mpeg compression) involves an algorithm that applies psychoacoustic principles to determine what can be taken out of an audio signal in order to make the file smaller. This involves degrading the signal. Although these algorithms are making compression more and more imperceptible, but derivative files made from your master recording will be extremely low quality and contain digital artifacts. [sample]. Compounding compression will seriously degrade audio quality. Flash and HDD storage has dramatically dropped in price, so today there is no financial need to make a compressed recording. The most common compressed format for audio is Mp3. Marantz has used Mp2 on some of their recorders.For archival purposes, uncompressed is the best way to record. Compressed audio is best utilized for creating web deliverable files, not for recording the original interview.
- File Formats
Most portable field recorders utilize a technique called Pulse Code Modulation (PCM) for digital conversion of an uncompressed analog signal. This digital data is then saved as a data file.
The standard file formats associated with uncompressed recordings are Wave (.wav), Broadcast Wave and .AIFF. Portable field recorders normally utilize Wave files. AIFF files usually are associated with Mac Computer applications. Since both are uncompressed, the quality is the same. The Wave file is the more common file format that professional and prosumer portable field recorders utilize, and unlike AIFF, enables the incorporation of metadata into the digital file itself. Many low-end consumer recorders—ones designed for dictation and usually purchased at office supply stores— record in a proprietary compressed format that can only be accessed using a proprietary software package. Because format compatibility and interoperability are of the utmost importance from an archival perspective, these low-end consumer recorders are not recommended for recording oral histories.
MPEG recording will dramatically decrease your data footprint and thus increase your recording time. The compression, however, degrades your recording quality. MPEG, however, is ideal for placing audio files on the web. Indeed, Mp3 files have become a standard uncompressed codec almost universally accepted by most computer players. Mp3 files are usually measured by “bit rate” rather than by sample rate and bit depth. Also included among more proprietary audio compression codecs are the Windows Media Audio Files, and Real Audio files, There are several “lossless” compression codecs available including FLAC (Free Lossless Audio Codec) and the Apple Lossless Audio Codec (ALE/ALAC). At this time, however, most recorders to not include these as a recording option.
- File Size
File sizes for recordings are calculated by combing bit-depth, sample rate, channels and recording time time.
16-bit/44.1 KHz/Stereo/.wav = 635.04 mb per hour
24-bit/96 KHz/Stereo/.wav = 2 gb per hourA filesize calculator can be very useful for determining the amount of storage you will need both for interviewing and for archiving your interviews to servers, external HDDs, digital tape backup, or to CDs and DVDs. There are numerous variables entailed with compressed audio file size calculation including bit rate, and the compression “codec” used to encode the file. As a general rule, choosing mono over stereo cuts file size in half. Unfortunately, the spatial dimension and separation of the original stereo recording will be lost if converted to mono.
A file size calculator for uncompressed audio is located at the website for the Louie B. Nunn Center for Oral History at the University of Kentucky.
- Recording Levels (for more detail on this topic, visit the tutorial Digital Audio Recording: Recording LevelsYou want to record as strong a signal as possible
You do not want your recording levels to clip
Higher bit-depth recording is more forgiving when boosting the levels of a low-level recording.
Locate a comfortable range for your peaks (usually between -12 and -6 decibels) at the beginning of the interview. A recording that has average peaks under -16 will normally have a greater amount of noise when the levels are boosted to optimal levels. I prefer to stay away from averaging in the -3db range because of the unpredictability of an interview. If clipping occurs, don’t panic and gently back the levels down.
- Manual level control
Manual level control involves the operator adjusting the levels by use of the input level or recording level controls. When recording with manual level control it is best to use a limiter to protect against clipping.
A limiter sets a threshold above which the signal will be gently “pushed down” in order to prevent clipping . This is preferred over ALC or AGC (see below) as it allows the operator to set optimal levels and minimizes noise while still protecting the recording from clipping. Limiters are not foolproof however, and good levels must still be determined by the operator.
- Automatic Level Control/Automatic Gain Control (ALC/ALG)
ALC and AGC are circuits in a recorder that determine an average optimal level. Use of one of these will minimize the risk of clipping but typically not produce as high a quality of recording as manual level control, because they boost quiet moments in the recording up to record level and thus boost background noise.
It is also recommended that operators utilize headphones, at least in the beginning of the interview, in order to monitor the sound quality of your recording. Levels will not indicate if there is a minor buzz, or the introduction of extra noise from a ceiling fan or an air conditioning unit, for example, that you may not have noticed during setup. A microphone can greatly exaggerate noise unnoticed by the ear, and can often be rectified by a different, typically closer, microphone placement (See microphone section). Monitoring through headphones at the beginning of a recording can greatly improve sound quality, and on occasion prevent the loss of an entire recording.
FINAL THOUGHTS ON DIGITAL RECORDING
- Use AC power if available, but keep freshly charged batteries in the recorder during recording. The accidental “unplugging” of digital recorders during an interview can result in the loss of the digital file. The batteries can serve as a backup.
- Use a good, external microphone. The better the microphone, the better the quality of the recording. An inexpensive field recorder with a good mic can produce a better recording than an expensive field recorder with a poor built-in microphone. Although some built-in mics now produce much better sound than they used to, external microphones continue to yield a better recording. Recorders with on board microphones must still be placed in between the interviewer and interviewee. The greater the distance a microphone is placed away from the sounds to be recorded the noisier the recording.
- Record as uncompressed wav files at a minimum quality setting of 16-bit/44.1KHz.
- Setting record levels manually with use of a limiter requires some practice, but once you have become comfortable with it will yield a higher quality recording than one produced with use of Automatic Level Control. Pay attention to recording levels throughout the interview. Recorders with limiters are the easiest way to protect an interview from clipping over the course of the interview.
- Turn your cell phone off during an interview. A cell phone in vibrate mode may not ring, but cellular phones periodically “handshake” with the local tower. This signal can often interfere with recorders printing unwanted noise on the recording.
- Do not be penny wise and pound foolish. Inexpensive field recorders that utilize cheap components produce poor recordings—especially when not used with an external mic. Digital recorders with inexpensive preamps will yield noisier recordings than a more professional grade recorder, which nowadays are not that much more expensive. Better recorders are getting less expensive, and now (2009) can be purchased new at prices starting at $200. OHA will continue to monitor recorder qualities and make “best practice” recommendations. Consult the “Recorders” section of the website or consult online resources for up-to-date recommendations.
- Always backup your recordings!