1. Consider Your Audio Recording Technique
What exactly are you giving your transcriptionist to work with? Is it a cassette from your personal recorder, or is it a professionally produced audio file? Where was the microphone? There are dozens of factors that can contribute to the quality of the audio recording, and the quality of the audio recording is absolutely essential when it comes to producing a readable transcript later. A lot of people rely on personal recorders or dictation devices to record their meetings or hearings. Some of these devices record onto cassette (either standard or micro) and some record digitally, but all of them come with some risk when relied upon for recording multiple speakers. These devices are primarily intended to record a single speaker holding the machine directly in front of their lips. Placing it on a table in the middle of a meeting is no guarantee that everyone will be picked up on the microphone. If this is the route you want to go, choose a device that boasts an omni-directional microphone. For example, the Olympus LS-14 recorder has two 90-degree coverage microphones on it, which will easily cover a conference table, but is at least three times the cost of most consumer-level devices.
(Oh, and be extra wary of microcassette recorders. The format is only made possible by using thinner audiotape and a slower recording speed than standard cassettes, making them much more susceptible to damage and degrading over time. If you do use a microcassette recorder, make a backup A.S.A.P.)
If you have the budget, why not aim for a more professional recording? Multiple microphones (ideally one per speaker) with a mixer or a digital recorder with multiple inputs are now common. You have the option to adjust audio levels for each microphone separately, and you can even produce recordings where the transcriptionists can adjust these levels themselves during playback. If you are regularly producing recordings requiring transcription, a small investment in recording technology can pay you back handsomely with years of readable transcripts.
2. Be Mindful of Cross-talk and Talking Too Fast
Are your participants aware that they are being recorded for later transcription? If not, they’re probably not speaking in the direction of the microphone(s); they’re probably talking over one another, and they’re almost certainly speaking too quickly. It’s a good idea to caution the speakers, before the recording begins, that they need to speak loudly and clearly, to mind the microphone(s), and to avoid stepping on each other’s sentences. If someone acts as a moderator during the recording to periodically remind speakers of these ground rules, your transcript will benefit from it.
3. Be Prepared for Speakers with Accents
Special attention should be paid to any speakers with strong accents. Any reputable transcription company will do its best to pair your recording with a transcriptionist who has a good ear for that particular accent, but they need to be warned ahead of time. If halfway through a 2-hour meeting, there is a presentation given by a speaker from France, the unprepared transcriptionist may not have time to do anything more than give it their best shot, which could result in the loss of a lot of that presentation. And remember that the more obscure an accent is (relative to the audience), the more unlikely the transcriptionist will be able to parse it out. Urge all accented speakers to take a special effort to enunciate and make themselves understood.
4. Keep Background Noise to a Minimum
Where is your recording taking place? Is it a hot day, so you have a window open, or worse, an air conditioner running? Those sounds will compete with your speakers’ voices for prominence in the recording, so make sure that the microphones are aimed as much as possible at the people you want to hear. If the door is open to a noisy hallway, or the window to lots of street noise, close it. Ideally, do a sample recording beforehand so you can hear how the background noise competes with the speech. And don’t forget that all electronic devices emit an electromagnetic field to some degree, just by virtue of being on, and this field can show up in your recording as a hum or a buzz, often quite loud. If you have a large electronic device in the room (like an industrial copy machine), it may need to be shut off.
5. Prepare the Transcriptionist for Unique Terminology
When a transcriptionist comes across a word or phrase with which they’re unfamiliar, they’ll usually do their best to spell it phonetically. Unfortunately, this can leave the transcript looking amateurish at best (the name, “Andy Ziats” becomes “Andy Zayitts”), and misleading at worst (“SA program” becomes “essay program”), but a very easy fix is to prepare the transcriptionist beforehand with a list of keywords. This includes proper names (of both people and places), specialized terms (legal, pharmaceutical, etc.), and any acronyms to be used in the discussion. The list prepared beforehand is unlikely to be exhaustive, but it will go a long way towards reducing confusing or useless passages in your transcript.
6. Make Sure that All Speakers are Identifiable
This one comes as quite a shock to many transcription clients: when a meeting is being held with four or more participants, it is rare that the speakers’ voices are distinct enough for a complete stranger to reliably identify them throughout the transcript, especially for short “yes/no” or “uh-huh” statements. As a result, your document may eventually degenerate into a series of questions and answers attributed to “Male Voice 1” through “Male Voice 4,” and, in cases of very similar voices, even the numbering may not be possible.
Thankfully, this is another easy fix. If your schedule and budget allow, it’s a good idea to have someone sit in on the meeting solely to produce speaker ID notes. All that’s required is to number the participants, then keep a running list of the numbers in the order of which they speak, along with the first word or two of their statement. This is basic data entry and more advanced notes can be taken if time allows, but with just this bare minimum, you will see a dramatic increase in the accuracy of the IDs listed in your transcript.
7. Understand the Implications of Turnaround Time
Finally, how quickly do you need your transcript? The average professional transcriptionist takes between 2.5 and 4 hours to transcribe a single hour of audio, so if you need your 8-hour meeting which ended at 8:00 P.M. transcribed by first thing tomorrow morning, there are not enough working hours left for a single person to do the job. The task will need to be split among multiple transcriptionists, who will then lose the benefit of context when it comes to name spellings, proper terms, and even matching names to voices. A responsible transcription company will task an editor with combining the parts of the transcript and ensuring the pieces are as consistent as possible, but an editor’s practiced scan is still no substitute for the 8 hours of real-time attention that is most profitably done by a single individual. If accuracy is a higher priority than turnaround, consider giving an extra day or two for transcription time. You’ll definitely notice the difference in the final product.
No amount of preventative measures can guard against every pitfall. Eventually, you’ll have a meeting with an unexpected loud noise in the background (or a participant tapping his pen on the table the whole time), or maybe you’ll just catch a neophyte transcriptionist on a bad day. But by following the above steps, you can strongly stack the odds in your favor of receiving the best possible transcript. Feel free to email us or comment below with your experiences on this topic. Good luck!