There are a wide variety of audio recorders on the market, and this is a rundown of the things you might want to consider when selecting a recorder starting out in wildlife sound recording, in approximate order of importance. I am assuming you will be using the recorder directly connected to the microphone (via any powering unit necessary).
This is the main thing that is different about wildlife sound recording - we have weak signals, and noisy input stages dog budget recorders, swamping the weak signals with hiss. Most people use budget recorders to record people talking or bands playing - lots of sound available, so manufacturers can get away with hissy inputs. This is a fundamental requirement - hiss added at the input can never be got rid of, and it will negate any investment you make in good microphones - the mic will be less hissy, but the hiss will be put right back from the recorder.
For the technically inclined, as a rule of thumb a noise level Ein of better than -116dBu means your recorder will be usable with good P48 capacitor mics like DPA mics or the Sennheiser MKH series. The reasoning behind this is given on the HiMD noise article. Beginners are usually using poorer quality mics that are noisier. Unfortunately, such mics are often less sensitive, so recorder input noise still matters because the wanted signal is also lower than with higher quality mics!
Manufacturers rarely specify this on low-end gear. And unfortunately it varies - early (2005-2007) CF card recorders were often poor in this respect, whereas the low-end HiMD recorders were surprisingly good.
How do you get this fundamental parameter right, if the manufacturers don't specify it? You talk to other wildlife sound recordists, for example on the WSRS forum. You can look at equipment we have tested - we usually test recorders which are up to scratch or say if there is an issue here, for example on the Zoom H2. We have only a very limited test capacity and many recorders are launched to the market every year. So there are undoubtedly many good recorders we have not listed!
You cannot rely on the manufacturer's reputation as many people seem to like to do with consumer electronics - while the Zoom H2 has issues with preamp noise, the Zoom H4 from the same manufacturer is fine.
You have to be able to set record level manually. No AGC made so far can cope with the wide variation in natural signal levels. Even the more sophisticated sound-check modes of devices like the Olympus LS14, which does a sound check for 30s before recording at a fixed level set by the sound check, isn't that suited for wildlife sounds unless your subjects make a continuous noise are can be presuaded to vocalise duirng the sound check
Lower-end recorders offer 3.5mm unbalanced inputs and no mic powering, or the poorly specified plug-in-power system. These are fine if you are going to be using self-powered mics (with a battery) as many beginners do, and modest cable lengths of up to about 10m . If you want to use P48 power. with these you either need an adapter box for about £25, or a recorder that offers P48 powering, which usually takes you to the mid-range level of recorders costing £350 or more (Aug 2008 typical prices). Your choice of microphones and recorder needs to match in terms of microphone powering and connectors.
You want to record all the sounds, including details you may not be able to hear. It is a lot of effort to get yourself and your gear to when the wildlife is, and storage is cheap enough nowadays that there is no need to throw information away to save space. Make an uncompressed recording format of at least stereo 44.1kHz sampling/16 bits a requirement. Many recorders offer both uncompressed and compressed formats which is fine. Select uncompressed for now. You may not be able to hear the difference, but some scientific analysis software can pick out more details from an uncompressed recording than a compressed one, and compression may add inaudible signals that were not originally there.
Recording format, preamp noise and electrical compatibility with your mics are the things that you have to get right. The following parameters affect the usability, or suitability if you go on long expeditions away from power and or file download capability, or you have specialised interests.
Much more important in days gone by than now. The medium is what you record onto, and is almost irrelevant in digital recording. It is nice if the memory cards are cheap, widely available and have large capacities. This seems to have settled on Compact Flash (CF) and SD(HC) cards and micro SD cards. An adapter can be used with micro SD to make these compatible with SD card slots. It can be nice if your camera and recorder use the same cards. But it does not really matter. Unlike in the days of analogue recording, the medium is a temporary storage facility until you get the recording into your computer. From then archiving is a computer file storage job which is independent of what your recorder uses. Check the price per megabyte, and the availability in your destination if recording abroad. The medium determines the maximum recording time, both in its capacity and sometimes in restrictions in maximum file size, which may be smaller than the total capacity.
HiMD was the last non-solid state medium often used for field recording, but it has been eclipsed by solid state media (DAT was another, but is at the end of its service life). There is some mechanical noise from the recorder when operating, which is a slight constraint compared to true solid state recorders which are normally silent in operation. Sony needlessly made the process of getting recordings off HiMDa little more complex than that of getting recordings off solid state recorders.
Some recorders offer a hard disk as a recording medium, which give a phenomenal recording capacity of many hours. The hard disk does, of course, introduce the old problem of mechanical noise back to the recorder, which needs to be considered in whether it suits your way of working.
This is irrelevant to the recordist who records for a day and then transfers to a PC as long at the storage does not run out over the day. For people recording on holiday or on expeditions, however, non-removeable storage is an issue because it means some method of transferring the data out has to be taken along as well. With removeable storage you simply swap for a blank medium. The advantage of nonremoveable storage is that it is always there - which will be appreciated by anyone who has got up early in the morning to record only to recall when they get on site that the memory card is still in the computer at home!
HiMD and the Nagra Ares-M2 have high gains, needing only about -65dBu for full record level, whereas many other recorders need -40dBu to -50 dBu to achieve digital full scale. Although it is the mic and mic preamp noise level that sets the noise level of the recording once it has been level adjusted in post production, if the recorder has an unusually low sensitivity, like the -41dBu for full scale of the Zoom H4 it can become hard to get enough level in the monitoring headphones to hear distant sounds well enough aim a parabolic dish correctly, particularly when tracking a moving subject.
This is not purely a function of the input sensitivity but also of the headphone amplifier gain. Higher end equipment designed for the pro market seems to have much more headphone amp gain so dish tracking is easy with a Sound Devices 702 even though it has a relatively low -46dBu sensitivity for full scale.
Proprietary batteries are always a pain, but they can make recorders cheaper/lighter. It is important for some recordists to be able to use disposable alkaline cells and rechargeable NiMH standard AA, C, D sizes. You will pay a little bit for that choice in size, weight or battery life. If you expect to be going on expeditions where AC mains is not available, you will have to think about powering options when you select the recorder. Recorders that can be USB powered can be supplied using USB power banks, external 5V battery supplies designed for externally feeding smartphones.
If you are remote from your mics, you need the recorder to be on and in record-monitor mode to hear what is going on. Many recorders will not do this for 12 hours at a stretch on one set of batteries. A recorder where the battery has to be charged in the recorder and cannot be changed in the field is an issue if you want to be out from dusk to dawn and the batteries run down by lunchtime. This type of information is also not always easy to find from spec sheets!
You get greater flexibility if it is possible to power the recorder or charge the batteries from a vehicle 12V supply.
This is a feature where on record pause the recorder is always recording to a buffer memory of a few seconds, so when you press record you have captured a few seconds before the record button was actually pressed. This takes out your reaction time in noticing that a record-worthy event is happening, which saves having to record long stretches to make sure you get the first part of a song. It can be useful with the essentially unscripted nature of wildife sound recording.
You can record at a higher sampling rate that 44.1 kHz, or at a greater bit-depth than 16 bits, or both.
You may not be able to hear the difference but many creatures can, and many vocalisations have components out of the range of human hearing. If you are interested in bats, for instance, prefer a recorder going up to 192kHz. Those interested in insects like crickets and grasshoppers may also wish to have a recorder that goes up to 96 kHz. Which you can't hear the extra frequencies directly, using software you can analyse the high frequency parts or transform them to meake the more audible. Take a look at this article on bat sounds for an example of why inaudible sounds can be of interest to the recordist. You will, of course, need microphones sensitive to higher frequencies as well.
The default bit depth is 16 bits. Most low-end gear does not offer a good enough performance to put anything other than noise in the extra 8 bits unless the input is much higher that usual wildlife sounds. It is questionable whether this is worth using with low-end kit. It can enable you to set recording levels lower and have more headroom. A 16 bit recorder with a better mic preamp will generally give you better recordings than a 24 bit recorder with poor preamps at the low signal levels found in wildlife sound...
There is great variation in the control structure on recorders. Can you easily adjust record level and mic sensitivity? Any recorder which means you have to go into a menu to adjust record level should be viewed with suspicion - wildlife sounds can vary over wide ranges, and you need to be able to track these fast to avoid clipping when faced with the unexpected. Can the screen be seen in sunlight, and at dusk? if you expect to be using handheld mics or a parabolic dish, is it easy to control the recorder with one hand?
After recording, you will want to transfer your recording into your PC for editing, and selecting the best takes. Digital recorders such as HiMD and solid state CF recorders or hard disk recorders allow you to transfer your recordings to your computer, usually via USB or Firewire. The advantage of this is your computer receives a bit-perfect copy of the sound file in the recorder, with no degradation. The transfer often takes place faster than the time it took to record the sound (ie an hour's worth of sound can be transferred in less than an hour) and you can then clear your recording media and be ready to record again.