An Intro Guide to get Accurate Transcriptions
Whether you are getting your audio and video files transcribed the old-fashioned way (manual transcription) or through a modern service like, uh hem, Simon Says (swift, automated transcription), the quality of the audio is directly correlated with the accuracy of its transcript and this is especially true when using speech recognition.
Before we discuss recording high-quality audio, it's important to understand what constitutes poor-quality audio.
What constitutes poor-quality audio?
- Background noise suuuucks. People talking in the background, the sound of passing cars and motorcycles, and any noise other than the voice you want to transcribe is a limiter.
- People speaking over top of each other cause complications. If there's more than one person speaking at a time, it's difficult for speech recognition to differentiate between individual speakers and their words.
- And lastly, an inaudible recording is the most problematic. Whether the recording volume is too low or there's microphone interference, inaudible recordings are impossible to consistently transcribe accurately.
Poor-quality audio cause a host of avoidable transcription issues. And then it takes more time to proofread the transcript and make corrections.
What leads to high-quality audio?
Get high-quality audio by:
1. Using a directional microphone. Even cheap, directional mics are better than expensive omnidirectional ones. Though cameras and smartphones have built-in mics, they are not directional and don't provide the quality of audio you need. Lavalier mics are best for interviews. Condenser mics are only desirable in extremely quiet environments.
2. Using headphones to monitor the sound: as you record the audio, you should listen along with headphones to ensure you're capturing the quality of audio you expect, audibly. Most sound recording equipment have a headphone jack as headphones help determine if you should adjust mic placement or ask your subject to speak up.
3. Limiting the background noises: Maintaining a quiet environment is key to recording good audio. You should record in a space that is removed from street noise. Additionally when recording, phones should be off because of interference caused with audio equipment; doors shouldn't be opened or closed; any music should be turned off; and everyone other than the speakers should be silent.
(N.B. Other factors that may lead to reduced accuracy are when a speaker uses niche and industry-specific vocabulary, like oncohematology, or when a recording is of accented speakers. For example: as endearing as an Italian-accented speaker talking in English is, it can create twists for automated transcription. While you may not have options in these cases, it is helpful to know any potential limitations upfront.)
After you've finished your recording, upload it to Simon Says. We handle almost any audio/video format or codec and for most accurate results, upload the original, uncompressed audio or video files.
Poor-quality audio impedes an otherwise speedy, automated transcription process with accurate results. But, with the right equipment and recording techniques, your Simon Says transcripts will come out swiftly and accurately.