HDSLR Guide Chapter 13: Audio
The audio capabilities on HDSLRs are relatively limited and in many situations will produce unusable sound. Professional users have to contend with a range of issues such as low quality built-in mics and automatic gain control. Thankfully, there are workarounds for a wide spectrum of uses and budgets. Before delving into the solutions, a breakdown of the issues and limitations of the HDSLR's audio capabilities will help determine the best solution.
Unfortunately, the built-in microphones can't be used for anything other than reference audio. Too much ambient sound gets picked up by the mic. When trying to record a person from just a few feet away, the result will be a noisy and washed out sound. The mic will also pick up the slightest handling noise as well as the motor noise caused by a lens's image stabilization. While this may not be such an issue for simple Web videos, for any professional production, a more advanced solution is required.
Many microphones produce a low voltage output, so the signal needs to be amplified in order to record a normal volume. This is typically the job of a preamplifier. The camera's built-in preamp does not meet professional standards and produces a floor of noise as the gain, or recording volume, is amplified. This issue is further complicated because some HDSLRs don't offer the ability to adjust the gain, in which case the camera automatically adjusts the gain on its own.
Automatic Gain Control (AGC)
All HDSLRs have a feature called Automatic Gain Control (AGC). AGC is designed to keep the audio levels from distorting when they're too loud. But it also raises the recording volume when the levels are too low — which introduces unwanted noise as well as picking up more surrounding sound (e.g. room sound, traffic, machine noise, etc.). The increase in gain by the camera's AGC brings out the worst in the low-quality preamps used by the camera.
In the infancy of video-enabled HDSLRs, some form of manually adjustable gain control was limited to just a few levels and could be adjusted only through the menu system. That made it difficult to adjust recording levels to prevent clipping, which can happen without the AGC turned on. As the technology advances, there are more and more HDSLRs available that are equipped with more easily adjustable input levels.
XLR Audio Inputs
Most professional microphones use an industry standard XLR connection to connect to a mixer or recording device. HDSLRs, on the other hand, only have a 3.5mm mini-jack connection (except for the GH1, which has a 2.5mm connection). Some HDSLRs have no mic input whatsoever. There are many mics available that have 3.5mm outputs; however, higher-end professional mics typically use an XLR connector.
Certain microphones need to draw external power. This is called phantom power or plug-in power. Most professional video cameras can supply this power, but HDSLRs cannot. If one of these types of mics is needed, a separate audio adapter with phantom power is needed.
Monitoring audio as it's being recorded is crucial for professional productions. Without it, any number of audio issues can arise and affect the recording without anyone knowing until it's too late. Currently, HDSLRs do not have any way of monitoring live audio recording. With the new 5D Mark II firmware update, it is possible to see the audio levels through the menu, but not while recording. The Magic Lantern firmware, on the other hand, can enable audio monitoring via the A/V output.
There are several ways of getting around the audio limitations of the HDSLR. The solutions range from simple to more complex.
1. Upgrade sound quality with a camera-mountable mic
The quickest way to achieve an upgrade in sound quality is by adding a mic that can plug directly into the camera. Since all cameras with mic inputs have 3.5mm mini-jack inputs (except for the GH1 which has a 2.5mm jack), smaller and more portable microphones need to be used. These types of mics are usually battery powered and are small enough to mount directly atop the camera or rig.
Adding a stereo mic to the mix will automatically increase sound quality as well as allow for directional pickup from intended sound sources. Using an add-on mic will also practically eliminate camera handling noise that is usually picked up with the onboard mic.
For run-and-gun situations, where rig size is an important consideration, this solution proves sufficient for recording usable sound.
2. Record audio with the camera's AGC disabled
All HDSLRs suffer from low-quality, noisy preamps and AGC, which cause extremely unclean audio recording. The 5D Mark II partially addresses these concerns by allowing the AGC to be disabled. However, for cameras that can't fully disable the AGC, using an XLR adapter with an AGC disabler can circumvent these issues and allow you to record usable audio.
Truthfully, in all cases (with or without AGC disabled on the camera) an XLR adapter is ideal because it will allow for the connection of more professional microphones, as well as the possibility to adjust audio levels during recording. Additionally, most XLR adapters offer the possibility of monitoring the sound, which is absolutely key.
For those dealing with their camera's AGC sans adapter, if the camera's volume has been lowered to limit the AGC's effects (as is possible with the Nikon D3s), the recorded sound can be too low to use. The simplest remedy would be to increase the audio's gain in post production. An editing software will typically do a better job of increasing gain than the camera, though the results will almost never be optimal.
3. Record sound using a field recorder
Instead of trying to force a still camera to record professional sounding audio, the videographer may want to simply use a dedicated audio recorder with features and inputs that suit the shoot's needs. Its audio recording capabilities will not only outdo that of the HDSLR's, but also that of a broadcast camera. With these devices, audio can be recorded at higher sample and bit rates for the most demanding situations. It also provides direct access to controls such as audio levels, which can be crucial while recording.
Several models offer XLR inputs and can feed phantom power to mics that require them. There are, however, some drawbacks to this route. First, it's a separate recording device, so each time record or stop is pressed on the camera, the same must be done on the field recorder. Second, syncing the audio in post can prove tedious, as the HDSLR does not record with a timecode. Luckily, there is a unique plug-in (PluralEyes) for Final Cut Studio that automatically syncs all of the audio tracks (more on this below).
The microphone is the first element in the chain of audio recording. Professionals will not only want to have a high quality mic, but also the particular type that works best for their applications. Most microphones operate on the principle that sound waves will cause a thin diaphragm inside the mic capsule to vibrate. Exactly how that translates into sound is what differentiates the various sorts of microphones. There are three main classes of mics to consider before getting into features and characteristics: Condenser, Electret Condenser and Dynamic.
Features to Consider
A condenser microphone produces sound when sound waves move a diaphragm in relation to an electrically charged plate. Condenser microphones are generally more sensitive to sound and can pick up a wider frequency than dynamic microphones. They're also better at recording transient sounds, or fast-changing sound levels. They generally output a higher signal level than dynamic mics. In contrast to dynamic mics, condensers require a power supply, called phantom power, which is usually 48V, though some mics operate on 12V or 24V. Many modern condensers can work with a range of phantom power voltage inputs. Keep in mind that condensers are generally more fragile than dynamic mics.
Most Electret Condenser mics do not require an external phantom power supply because it sits in a permanently electrostatic charge. Some of the better electrets, however, may require a very small internal battery, or need to be fed power through the audio cable in order to power an internal preamplifier. Electrets are generally very durable, small, and lower cost compared to other mics. Many consumer electronic devices, like video cameras and HDSLRs, utilize this type of mic. Their frequency response, self-noise, and resulting sound quality are typically not as good as that of condensers.
Dynamic microphones register sound when a diaphragm moved by sound waves produces an electric charge as it moves past a magnetic field. Dynamic microphones are generally more durable than condensers and can handle harsher production environments. They typically don't have the same frequency response and sensitivity as condenser mics. This makes them better suited for high volume situations, like live music. Unlike Condensers, they don't require phantom power, which can be more convenient.
A pick-up, or polar pattern determines the microphone's sensitivity to sound at different angles in relation to the mic capsule. An omnidirectional mic will pick up sound equally from all sides of the mic. This is represented on a polar pattern chart as a full circle. A unidirectional mic will have more sensitivity in one direction than the others. Cardioid, super-cardioid and hyper-cardioid are all unidirectional patterns.
Cardioid mics are used to "focus" on a sound source, rather than pick up all of the surrounding sounds. Super- and hyper-cardioid microphones are typically used for outdoor or noisy applications because they reject surrounding noise and pick up more of what's in front of them. In order to get the highest sound quality, cardioid mics need to be positioned correctly in relation to the intended sound source, such as a person.
Stereo microphones contain two mic capsules configured symmetrically at an angle between 90º and 135º. They are used to try to convey the natural sense of the recorded environment. It attempts to give the listener an indication of the spatial relationship of the sound source. They can be used to record a live audio performance, such as an orchestra. They also typically pick up more surrounding sound, so they may not be well suited for noisy environments. Stereo microphones produce a "wider" stereo image with less ambient sound when they're positioned closer to the sound source. When positioned farther away, the X and Y stereo image become narrower and more surrounding sound is picked up. In certain situations, using a stereo mic mounted on a moving camera may be distracting to the audience because the stereo effect will keep shifting.
Microphones have different degrees of "hearing abilities" and tend to reproduce different frequencies with more or less strength. These characteristics determine their suitability for a given application. Knowing the range of frequency of the intended sound source can help in choosing a mic that either reproduces this range "naturally" or with a "flat" response. The mic may also accentuate the particulars of a given sound. Some microphones are capable of changing these characteristics with an on-board switch. A high-pass/cut-off may be used to reduce wind noise and other ambient sound, which can produce a cleaner recording of the human voice.
Microphones are measured between 20Hz and 20kHz frequencies, which is what humans can hear. A frequency response chart for a given mic will tell how strongly the mic picks up a certain frequency, or how the end of its hearing range "rolls off" — sharply or smoothly. Most HDSLR users will be looking to record voices, which range from about 80Hz to 15kHz for adult males and females. Ideally, microphones should be matched to their intended sound source's frequency range so that more signal is given to that range.
Balanced & Unbalanced Cables
Balanced mic cables are able to withstand longer distances without the risk of interference that introduces noise to the audio signal. This is especially important at longer distances, but most professionals will use only balanced connections. If a balanced source is used, it needs to match a balanced input, or if the input is unbalanced it would need to be converted. Most consumer microphones output an unbalanced audio signal, while most professional microphones output a balanced signal. Electronic musical instruments also output an unbalanced signal.
The most common way to connect two balanced audio devices is with an XLR cable, although some may use 1/4" TRS balanced cables. The cables themselves are not the only thing that makes a balanced signal. The audio source and input both need to feed and accept a balanced audio signal.
Most current HDSLRs do not provide any power via their mic inputs; so it's important to know if a separate power source is needed, or if the microphone that will be used doesn't require one. Dynamic microphones do not require an external power source; however, most condensers and some electret condensers require power. Some may have an on-board battery compartment, while others require the power to be fed through the audio cable.
Many mixers and some XLR adapters have the ability to provide the necessary power. When plugging a mic that requires external power directly into a recording device such as a field recorder or the HDSLR itself, ensure that it can provide the necessary power, or use an adapter that has phantom power before connecting to the recording device.
XLR: Most professional microphones use this type of connection, which is typically for a balanced signal.
1/4" TRS: Many musical instruments use this type of connection. It can also be a balanced signal.
3.5mm, 1/8" TRS: Typically used on consumer or semi-professional microphones and headphones.
2.5mm TRS: Typically used by cell phones, but also found on some cameras, like the GH1. Requires a conversion adapter or cable to accept 3.5mm microphones.
Combination XLR + 1/4" TRS: Some devices, like the Zoom H4n, have combo input jacks that can accept both XLR and 1/4" TRS cables for increased flexibility and options.
Impedance is the resistance to AC (alternating current) and is measured in ohms.
High impedance microphones are usually less expensive, but the audio quality degrades when long mic cables are used, so they are usually not suitable for professional applications.
150-600 ohms — Low impedance/low Z/mic level
10,000 ohms and above – High impedance/high Z/line level
Most XLR microphones are low impedance, while 3.5mm phone jack microphones, guitars and line inputs are high impedance.
It is very important to match the impedance levels of the microphone with that of the input of the mixer or recording device, or distortion can occur. The recorded volume may also be lower than optimal. Look for the impedance level measure in ohms. The impedance of the input level should either match that of the mic, or be higher. There are impedance converters available that match incompatible microphones with a device.
The simplest, lightest and least expensive method for using a relatively higher quality mic than those found on HDSLRs is to mount a small microphone onto the camera itself. These generally have their own built-in power supply, and some form of shock mount to isolate the sounds of the camera and the operator's hands. Some come equipped with a shoemount so additional accessories aren't needed.
Most are quite small and won't weigh down the camera. A coiled audio cable will help keep the camera setup from having loose cables flying around and get ting in the way. Although they're generally a huge leap in quality compared to the HDSLR's built-in mic, they still tend to pick up on the lens' image stabilization motor noise, as well as some handling noise. A good solution for this is to mount the mics on a shoe-mount adapter that raises and increases the distance between the mic and the camera.
Shotguns, aptly named for their intimidating appearance, typically have cardioid or supercardioid polar patterns and are excellent at isolating subjects in noisy environments and the outdoors. They're widely used in filmmaking and broadcast news while mounted on a boom pole and are generally handled by a boom operator. The goal is to get the mic as close to the subject as possible, without getting in the frame of the image. Done correctly, this generally produces the most pleasing sound, as it picks up the intended sound source clearly, but also allows for a little "air" or surrounding ambient sound to be picked up.
Beyond the quality and characteristics of a microphone, the single most important factor that determines the final "quality" of the sound is proximity of the mic to the sound source. If the shoot calls for recording a person, or several people, talking, and use of a boom operator is not feasible, wireless microphones will be the best choice for getting quality sound. The voices will be fuller, and there won't be a need to pump up the gain as much. Raising the gain only picks up more noise and can eventually lead to distortion.
A lavalier, or "lav" is a small microphone that is usually clipped to a person's shirt, a few inches under the chin. Lavaliers are offered in a variety of polar patterns. A cardioid or super-cardioid lav will typically be used in a noisy environment so that only the sound coming from the mic's immediate area (usually the person's chin) will be recorded. The downside of this is that if the person turns their head while talking, the sound level will dip, which can be quite distracting to the viewers. An omnidirectional lav, on the other hand, will pick up a more natural sound and will not suffer from the same off-axis dropoff that occurs when a talking head is turned.
XLR Adapters & Mixers
In order to take advantage of professional microphones, it's necessary to adapt them for use with an HDSLR using adapters. The simplest is an XLR to 3.5mm adapter cable that can be used with either dynamic mics or a self-powered condenser. This, however, does not address the need to replace the camera's built-in amplifier, which is noisy at its best. Several options exist that can deactivate the camera's AGC (for those that won't allow manual override).
Field mixers offer the same functionality as XLR adapters, but typically have higher-quality preamps, allow for the live monitoring of audio, include high-pass filters for cutting down on wind noise, have audio level meters, and most importantly, have the capability to mix several channels of audio into a stereo output. This is done by "panning" a channel to the right or left stereo output (or somewhere in between). Done effectively, mixing can create more dynamic recordings and allow for multiple mics to be mixed down into a final stereo output. Higher-end field mixers will have additional features that offer higher-quality components and more sophisticated controls over the audio signals being recorded. However, with the added quality comes more complexity, which requires more learning and practice. Consider investing in a system that offers the features and the quality that the current project demands, while also considering future needs.
Features to Consider - XLR Adapters
The amount of input channels an XLR adapter or mixer has determines how many microphones or sound sources can be fed into it. Two channels are the bare minimum, but four or even five channels can be useful for recording several microphones at once. An example of multichannel use can be an interview with three people using wireless lavs, and a boom operator with a shotgun mic who picks up a more generalized sound.
If plans call for using the HDSLR or a field recorder that only has a 3.5mm stereo input to record the audio, make sure the mixer or XLR adapter outputs to a stereo 3.5mm cable, otherwise adapter cables will be necessary.
Most entry-level XLR adapters and mixers only have stereo outputs, even if there are four channels of inputs. The four channels need to be committed to a mix within the mixer, and sometimes this can be limiting because it means there isn't full control over each person's mic in post production. Not having this capability necessitates having a dedicated sound mixer on set because someone has to constantly adjust for the varying levels of sound coming from the different sources. Although this can be accomplished by the camera operator alone, typically the results aren't as professional. Higher-end mixers allow each incoming channel to be outputted individually, but this requires an audio recorder that records several individual channels.
Some XLR adapters and most mixers allow each incoming channel to be placed within any number of points within the stereo range. This can be used to create a final mix, where certain sounds are panned to the right, left, or middle. It can also be used to group several microphones; for example, all lavs on interviewees can be panned to one side, while the other side is used for a shotgun microphone picking up ambient sound. This way there is some level of control in post. The position of the talent on screen in relation to the stereo image also needs to be considered when mixing. For example, if a person is standing on the left side of the screen, his/her mic should be panned toward the left.
Dynamic and select condenser microphones output a tiny microphone-level signal that needs to be boosted to proper levels. Sometimes, the sound also needs to be boosted because the microphone isn't close enough to the sound source. Different preamp designs accomplish this with different methods, as well as by using varying degrees of quality parts, which all contribute to the final result. Ideally, the preamp should produce a clean, noise-free signal at a nominal volume for the recorder to capture the optimal sound.
Many video cameras and all HDSLRs employ low-quality preamplifiers that are either unable to boost the signal properly, or that introduce their own noise into the signal path during the amplification process. Some lower-end audio field recorders may also employ substandard preamplifiers. By contrast, if the camera's preamps are completely lowered, using a higher-quality preamp will produce a louder signal with less noise.
There are two forms of preamplifiers that relate to mixers and XLR adapters: passive and active. Passive preamps simply apply a volume, or attenuation change, to the incoming audio signal without applying any gain, whereas active preamps have a gain control that amplifies the audio signal's voltage level — which typically produces a cleaner sound.
Just like microphones, XLR adapters and mixers have a range in which they can hear certain frequencies better than others. Higher-end devices will be able to pick up a fuller spectrum of sound for the highest accuracy in recording — typically 20Hz to 20kHz.
Some higher-end XLR adapters and almost all field mixers have the ability to output power through the audio cable for condenser microphones that don't have an onboard power supply. It's important to have this feature if the mic that is going to be used requires it. Some mixers and adapters allow the phantom power to be applied only to all incoming channels, but others have the option to selectively turn on phantom power per channel; this conserves power and protects certain dynamic microphones from being damaged by the current.
Being able to monitor the audio during recording is crucial for obtaining professional and reliable results. Since the majority of HDSLRs do not have a way to send audio output while the camera is in standby or recording modes, it's important that the XLR adapter or mixer have this feature. Ideally, the audio being recorded would be monitored from the recording device, but since there isn't an option for this with HDSLRs, the mixer itself is used. However, some XLR adapters do not have a headphone output.
Audio Level Meter
Because most HDSLRs do not have audio level meters, it's helpful to have this capability on an XLR adapter or mixer. Audio level meters allow the operator to "see" the audio levels that are being recorded to ensure the audio level is high enough to produce a good, clean and full sound, and to be certain it does not overmodulate the signal, which causes distortion.
Almost all audio meters measure the signal in decibels (dB). Typically, it's best to keep the audio levels from going beyond 0dB on the meter. There may also be a peak indicator on the mixer that indicates when a signal has gone past a threshold and risks being distorted. Some XLR adapters only have a peak indicator instead of an audio meter.
High-pass filters help curb surrounding noise in the low end of frequencies from being recorded. Wind and air conditioners typically fall in this category. While some microphones have this feature built-in, it's good to have it for mics that don't. Most lower-end mixers will only have the option to apply this filter globally to all channels, while some higher-end models will allow the filter to be applied selectively per channel.
Some mixers have the ability to generate a tone that allows the recording levels of the device to be set accurately. Both the mixer and recording device's audio meters are "matched" so that they both indicate the same relative audio signal strength. Due to the lack of audio meters on most HDSLRs, this applies if an audio field recorder is used in conjunction with the mixer. This will ensure that the audio levels are set correctly as well as match up the recorder's meters with that of the mixer, so that when the meter indicates a certain audio level, it's accurately representing what's being recorded.
If an audio signal becomes too loud, it risks over-modulating the mixer or recording device and is almost always irreparable. Some mixers have the ability to prevent the audio levels from rising beyond a certain point where it would get over-modulated and "clip." This is the safest way to operate a mixer and allows the operator peace of mind that no matter how loud the sound source becomes, the mixer will prevent clipping. This is an especially important feature for single-person operations because it can be impossible to operate the camera and "ride" the audio levels as the sound source volumes fluctuate.
Higher-quality mixers typically have more sophisticated limiters that gradually lower the volume, because a rapid drop can be noticeable upon playback. This is also called a "curved" response.
This feature should not be confused with the infamous Auto Gain Control (AGC), which, unlike a limiter, is constantly boosting or lowering the volume
Most XLR Adapters are designed to mount underneath a video camcorder, which is why they have a small alignment, or "azimuth" pin that protrudes next to the mounting screw. HDSLRs don't have this second tripod alignment hole underneath, so mounting this type of adapter may be difficult if it is a metal pin instead of plastic, which would be easier to remove. JuicedLink's azimuth pins are plastic, while Beachtek's are not. On the other hand, Beachtek makes HDSLR-specific XLR adapters without the alignment pin. One solution is to use a quick-release adapter plate between the camera and XLR adapter.
Audio Field Recorders
Solid-state audio recorders have taken over the scene in audio field recording. They offer more durability with no moving parts and provide a much smoother workflow without having to "digitize" audio. Most of them utilize the cost-effective and widely available SD memory card format. However, be wary of recorders intended for business or personal use because some record in non-standard file formats that need to be read by proprietary software, which will slow down the workflow.
Many field recorders have some mixing features built-in as well, such as gain, limiters. Field recorders have a built-in preamp. This preamp, like on a mixer, is one of the predominant determinants of final audio quality. See mixer features (above) for details on all mixer features that may be found on a field recorder.
Consumer recorders mostly have a 3.5mm stereo input, but to connect professional XLR microphones, a field recorder must have XLR inputs. At the very least, a dedicated mixer will need to be used between the microphones and the recorder. For a simpler and more compact setup, it's best to get a field recorder with the proper connections. Some can even output phantom power for microphones that need them, but this drains the battery considerably. For the most flexibility, some recorders have dual XLR and 1/4" TRS connections within the same socket.
For HDSLR use, it's important to have both a headphone output as well as an additional aux output. The aux output can be fed into the camera's audio input to create a double-sound recording setup. True, there will be an impedance mismatch that will render the camera's recorded audio less than optimal. But the intention is to simply have a reference track that the automatic syncing software PluralEyes can analyze. It can also provide a backup of the audio, should something happen to the field recorder's file.
Most field recorders record two channels of stereo, while higher-end models offer additional channels. Having several channels provides greater control over each individual sound source in post production and doesn't require a commitment to a final stereo mix on set. If a problem is found later on with a certain mic, and there was a backup mic such as an overhead shotgun, then the problematic track can be adjusted as necessary. However, these options are more limited when recording more than two microphones to a stereo recorder.
Bit Depth & Sample Rate
When audio is recorded digitally, the level of the analog audio waveform is recorded thousands of times per second as a numeric values. A 16-bit number is capable of representing over 65,000 different values, so the waveform's level for any given sample will be recorded as the nearest one of those 65,000 values. This is a process known as quantization, turning a smoothly varying analog waveform into a series of steps, and is a principal determinant of the accuracy with which the audio can be reproduced. Higher-end systems will use 24 bits per sample, providing a much finer quantization of over 16 million levels, and consequently much smaller stairsteps.]
Theoretically, 24bit offers higher "head room" and a lower-noise floor with more bit depth. This mostly will be noticeable in high volume and dynamic scenarios, like a concert. Some audio engineers also prefer to work at higher bit depths because they offer more flexibility in post manipulation, much like the way 8 bit video files are converted to 10 bit in order to work within color grading, so there's less chance of losing information.
The sample rate is the rate at which the recorder samples the incoming audio every second. The sample rate determines the range of frequency that can be accurately reproduced. The CD quality standard sampling rate is 44.1 kHz. At this sampling rate, the highest frequency possible is about 20 kHz — which incidentally is the limit of human hearing. Many people are not able to hear the difference in audio quality beyond that sample rate, but many sound engineers maintain it's a good idea to go past that range to ensure the highest fidelity possible. Most field recorders can go up to 96 kHz, while some higher-end models go as high as 192 kHz.
There will be an issue of compatibility on the editing timeline when placing two audio tracks with varying bit depths and sample rates. One of them will have to be downsampled or upsampled in order for them to match on the timeline.
The more challenging issue to consider with higher bit depths is the additional storage it requires. However, multi-gigabyte cards are dropping in price enough to justify recording at higher bit depths.
WAV files are uncompressed and therefore will retain the maximum amount of information and quality. On the other hand, WAV files take up a lot more storage than compressed formats like MP3. Some recorders have different levels of MP3 compression that can be selected for varying degrees of quality. It comes down to how critical the sound reproduction needs to be versus the recording time needed before everything is transferred to a computer. The standard broadcast and motion picture file format is Broadcast WAV, or BWF. In addition to the actual audio, the files contain important metadata that can be read by editing software in post production to facilitate the process more efficiently.
SMPTE is the motion picture and broadcast standard for labeling each individual frame with a time stamp in order to accurately synchronize both audio and video material. Some higher-end audio recorders have the ability to record or receive an external timecode that can be applied to the recorded audio. However, HDSLRs currently do not generate or accept timecode, so this is not an important feature for now. Some lower-end field recorders can apply a time stamp to the audio, but this is not the same as SMPTE timecode.
Many small field recorders use the popular SD (Secure Digital) card format, while some higher end recorders utilize CF (CompactFlash) cards. Others may have built-in storage, in addition to a memory card. Some recorders have recording time limits.
Ideally, the recorder should allow for the creation of folders for specific projects or shooting days, and it should automatically generate scene and take information for each clip. Typically, the best file management will be found on more expensive models because they were designed for motion picture use.
If a field mixer or XLR adapter won't be used in addition to a field recorder, it's very important that there be direct access to the most important controls. The primary example of this is recording levels. Many consumer-grade recorders have the levels set through a menu system, which won't work when there's a need to adjust levels while operating the camera. Even if there is a dedicated sound mixer, not having direct control over the levels is a limitation that can affect the final quality.
Batteries generally drain quickly, especially when phantom power is being fed to the mics. It's important that the batteries be easy to swap, without having to dismount the entire recorder from the camera. Some models have back-up battery systems as well, which come in handy.
Unless there's a dedicated sound mixer responsible for recording, the camera operator will have to find a way to mount the recorder to the camera setup. This is usually done using standard shoe mount adapters or arms that have a 1/4"-20 male screw at the end which fits into the back of the field recorder. Not all field recorders have this threaded mount, however, so it's a good idea to check for one. As an alternative, the camera operator may consider wearing a small audio pouch around the shoulders or waist, but this is best used with a recorder that has audio controls on top of the unit.
Most, although not all, recorders come with a wired remote control that can operate the start/stop functions and other basic abilities. If a shoulder-support or tripod system will be used, the remote can be taped onto a handle so that the recorder can be controlled without the camera operator removing a hand from the support. This is helpful because "record" and "stop" must be pushed every time the same is done on the HDSLR. This keeps the takes organized for future synchronization.
Headphones are critical for making sure the audio is clean and problem free. Even if the audio levels indicate a strong sound that isn't clipping, there could still be poor quality sound that can only be heard by monitoring the audio. In order to monitor the audio properly for noise, echoes, distortion or unwanted sounds, the headphones need to isolate outside sounds. A closed-back design that fits around and over the ears is crucial for obtaining a good level of noise isolation.
Monitoring headphones should ideally be sensitive enough to reproduce a good audio level for reference. Some headphones are foldable for easier storage.