There is a bewildering away of recording equipment available to record your interviews, focus groups, conferences etc. for later transcription, but this article aims to inform you on the different features available for recorders, what they're for and how the will (or won't) be of use to you, as someone who is recording for transcription. There is an enormous price range from less than £50 to hundreds of pounds, and this article aims to explain indicate which features cost more and what to look out for. The different features are first listed and then explained, and this should allow you to choose a machine that's right for you and your transcriptionist. As any visit to a good website or shop selling these machines will show you, this is not a comprehensive list. It's a lost of the most important aspects from a transcription point of view.

The different features are: Recording quality Frequency response Amount of recording time Computer interface File type Editing of recorded voice e.g. rewinding and adding or deleting some words How is the recorder powered? Dimensions Microphone jack Links with a speech recognition software

Recording quality will normally be indicated as SHQ - stereo high quality, HQ - high quality, SP (short play) and LP (long play). SHQ is the best quality (and stereo of course) but will take up the most memory, so you will be able to record less at this setting. LP is the poorest quality but you can record more time. Mono is probably perfectly adequate for an interview, but if you're recording a number of different people e.g. at a meeting, conference or focus group, it might be useful to have stereo recording. To use stereo you will need a stereo microphone or more than one external microphone.

The different recording qualities relate to different frequency responses. The human ear picks up audio in the range of between approximately 20 Hz and 20 kHz. But what are the most used frequencies in speech? I'd love to know but a quick Google search gives an unbelievable range of answers! I think it's fairly safe to say it's somewhere around 250 Hz to 5 kHz, with the higher end being a high-pitched woman's voice and the lower end being a low-pitched man. So it's around those ranges that you need to be looking. There is a wide range of different frequencies available in different recording machines but frankly most will record a one-to-one or one-to-two interview to acceptable quality at SP, and possibly even at LP.

The amount of recording time will depend on which recording quality you choose, and again the range of times is enormous, so you really need to think about what your needs are. You might need a long recording time if you're going to be researching 'in the field' for significant lengths of time, with no access to your PC to download files, or if you are recording a conference, especially if you want to use a high quality recording for this, which will mean you have less time. If, however, you are recording notes or dictation, or one-to-one interviews, and can regularly download information onto your PC, there is no need to save vast amounts of data onto the machine. It is probably this factor that creates the biggest price differences, so think carefully and don't buy a machine that gives you hours more time than you need.

The computer interface is a really important consideration. If you are planning to send the recordings to a transcriptionist to be transcribed you will need to download them onto your PC first. They can then either be emailed (if small enough, see file type section), transferred by FTP, or in some cases, including my company Penguin Transcription, sent using a file-sending box on the transcriptionist's website. Even if you plan to transcribe the files yourself it is a tedious business if you have to do this directly from the 'note taker' as recorders that don't have a download interface are sometimes called. Ideally the recorder will link to your PC with a USB interface, usually requiring no extra software or drivers. Alternatively it might link through a port and come with a CD containing the relevant drivers.

The file type is another very important consideration and is perhaps the one that is most often overlooked. Examples are WAV (uncompressed, great quality but enormous), WMA (Windows Media format, compressed but reasonable quality for voice), DSS (Olympus proprietary transcription file, very compressed but specially designed for voice so good quality). These are just a few examples and you can see details of a wider range of file types in my separate article on this site 'Digital File Types for Audio Transcription'. If you plan to send your files to a transcriptionist for transcription you will ideally want files that are reasonably small and can be emailed, although there are various ways of sending much larger files over the net. WAV files will not only take a long time to send, they will take up a lot of space on your PC, and on your recorder. DSS or DVF files for instance, are much more convenient as they are small and compact, can be emailed without being blocked by the server, and are delivered quickly.

Editing of recorded voice will probably be more important if you are using the machine to take dictation rather than an interview. When dictating a letter you may want to go back and delete the last few words and replace them with something else. The recorders at the mid to upper end of the price-range usually have this feature but if it's important to you it's worth checking to make sure as they by no means all have it.

How is the recorder powered? Most recorders are battery powered but if you're gong to be using your recorder significantly it's worth checking that you can use rechargeable batteries, from an environmental and cost perspective! It's also a consideration if dimensions are important to you. In particular the batteries can add considerably to the weight. Most modern recording machines are fairly light and quite discreet though. If weight is a concern make sure that the weight listed includes the batteries.

Whether or not the recorder has a microphone jack is an important consideration, as is the quality of mike that it allows you to plug in. Really for anything but one-person dictation an external microphone is a necessity. Recording an interview with only the machine's internal microphone is likely to give disappointing results. A poor recording will lead, at worst, to a poor transcription, and at best to an expensive transcription, as it will take longer to complete! If you are recording a conference you will need more than one mike, or a stereo microphone, as already mentioned.

Some recorders link with speech recognition software. As yet this type of software is really only of use for one-voice recording e.g. dictation of notes. Even then you will need to spend some time and energy 'training' the software to recognise your voice, and then carefully check the transcript as it will doubtless be ridden with mistakes, especially on homophones (words that sound the same but are spelt differently.) However, if you are going to use one of these packages then a feature that links your recorder directly to it is a useful addition.

This list is only the most important features in my opinion for transcription of interviews, dictation, focus groups and conferences. It does not cover the recording of music which requires a much wider frequency range and a number of other specialist considerations, but I hope that it is helpful for those embarking on research that requires transcription.