Transcription Timescales

When you need a professional transcript of your audio recordings, it is often for a particular reason. When deadlines are important, you need to make sure you get your transcript on time. In truth, there is a wide range of contributing factors that could influence the timescales involved. After all, it makes sense that a one-minute clip will be transcribed far quicker than a three-hour conversation.

Your budget will also impact turnarounds. If you can wait a week, you can often get a quality service at a discount. Finding the right balance between delivery times, cost and quality can be quite challenging. Here’s all you need to know to make the right decision for your project.

Instant Transcription Service Turnarounds: Automated Transcription Services (Speech to Text)

Automatic Speech Recognition (ASR) software can provide near-instantaneous transcription service turnaround times. The problem with automated transcription services is little to no quality control. Depending on the audio quality of your recording, the clarity of speech patterns, background noise and the number of speakers, you may need to spend a large amount of time yourself turning a transcript produced by an automated transcription service into something legible.

Read More: Best Speech to text Software: The Ultimate Guide to ARS and Transcription Services

Even when there is little background noise and the person speaks with clear diction, most automated transcription services struggle to reach accuracy rates above 80%. Whether it’s a missed word, a misspelling, or an incorrect passage, those mistakes can be very damaging indeed. It could make the piece look far less professional, and potentially lose all meaning — this outcome is arguably even worse than not having the transcript in the first place.

Automated transcription services, however, are cheap. Using an ASR service will generally cost you about £0.10 per minute of recorded audio. If you have high-quality audio, the fact that you can get your work transcribed almost immediately makes the use of ASR a good option if you need a quick and affordable solution for putting ideas into text. In most professional situations, however, opting for human transcription is the preferred course of action. You should always count on spending some time cleaning up a transcript yourself when you use automated transcription services.

Human Transcription Service: Turnaround Options and Quality Control

Human transcription services are where a real person listens to the audio before transcribing what has been said. Human transcription services deliver better quality than anything automated transcription services can offer. The human ear can still comprehend words with greater accuracy than a machine, not least because they can contextualise the dialogue.

This process, however, obviously takes longer. It takes a trained transcriptionist about an hour to transcribe 15 minutes of standard quality audio. If you have an hour-long audio file, it will take a minimum of four hours to transcribe. However, most services will charge a premium for anything faster than a 24 hour turnaround. 12 hours is generally considered a ‘fast’ job.

Of course, the more time you have, the better. If you are in a position to wait seven days or longer, for example, this can significantly reduce the costs involved without compromising on quality.

The human transcription will never match the speed of a service that uses automated speech recognition software. However, what you’ll lose in immediate turnarounds will be outweighed by the positives, which include;

• Increased accuracy, including context.
• An array of options on what type of transcript you want.
• Time savings as you won’t need to proofread it yourself.
• The ability to separate the dialogue of different people.
• Added confidence in the service provided.

If you need a transcript urgently, you cannot afford to run the risks of using an automated service. The lack of quality control with automated transcription services means you can’t count on an outcome. However, whenever using human transcription services, if you can be flexible with deadlines, you will get a better deal.

Not All Transcription Services are ‘Verbatim’

When you use automated transcription services, the software is only capable of transcribing the audio in as close to a verbatim account as the audio quality and speech patterns will allow. Human transcription services provide quality controls, but they also offer options that allow you to tailor outcomes to your needs. These allow you to maximise the value of your transcript, but these decisions can also allow you to get a transcript faster.   

The most common type of human transcription service is called ‘intelligent verbatim’ or word for word. These remove repetitions, stutters, ‘ers’ and ‘ums’. This creates an easier to read transcript that is faster to produce than a strict verbatim transcript.   

Another option is to choose a summary/notes service in which a transcriber will simply take the audio and create a transcript that focuses on the points being expressed rather than copying word for word. This is very useful if you need a transcript to simply digest the information contained in an audio file and do not need to know every detail. These are often cheaper per minute of recorded audio and can be produced even faster than intelligent verbatim transcripts. 

With human transcription services, however, you can also take the detail the other way. Verbatim human-produced transcripts can simply add back in stutters, ‘ums’ and false starts. You can also request that transcription services include comments or pauses, laughter and tone. Including these details will cost more and will take longer to produce.

Summary: Cost, Quality and Turnaround Times Are Linked, Plan Accordingly

Every transcription project is unique, which is why the timescales involved can vary accordingly. Using automated speech recognition (ASR) software will be nearly instantaneous, but will be filled with errors. By the time you’ve rectified those, it is likely that the turnaround time will look a lot more like what a human transcription service would have delivered.

Timescales are influenced by everything from the type of transcription to the quality of the original audio file. As a rule of thumb, however, you should expect it to take a trained transcriber one hour to transcribe 15 minutes of standard quality audio. If you don’t want to pay a premium, you should expect to wait 24 hours for your transcript. If you can wait longer, that often opens up the possibilities to get a reduced rate. Ultimately, the best option is the one that works for you. As long as the service provides the quality you need in a timeframe that works, you can’t go far wrong.