Written by Gary Logsdon
There are many options for creating transcriptions, and in this article we look at paid automated transcription. In other articles, we address Free Transcriptions and Paid Manual Transcriptions.
Automated transcriptions are when a computer uses voice recognize to translate the audio with no human input. They can be a little less accurate, but can be done much more quickly.
There is a wide selection of paid automated transcription services available online, but these are my favorites. Any of these would be a reasonable choice for your transcription needs as a food blogger. The transcribing factors that are most important to you may make one or the other more attractive.
Descript is the service I ended up choosing because of its high accuracy, fast turnaround time, and competitive pricing.
Unlike most of the paid transcription services, Descript is installed on your computer rather than accessed online. I prefer this since my upload speeds are relatively slow and it was taking me more time to upload files to the service providers than it took them to do the transcription. The actual performing of the transcription is also very quick with Descript.
When I ran my test file on Descript, it's accuracy was quite good. Almost all of our audio and video has the phrase "sous vide" in it many times. Most the time Descript transcribed it as "sous-vide". This makes it much easier to do a search and replace. We often are referring to temperatures such as, "141 to 150". It also gets the temperature numbers correct more often than the other solutions.
Although Descript has good accuracy you will still need to proofread the entire transcription and make some corrections. The program has an editor included which makes this quite easy to do.
Descript has two pricing plans. The Basic plan is a pay-as-you-go plan at $0.15 per audio minute. Their Producer plan cost $20 per month and includes up to 4 hours of transcription time. If you exceed the 4 hours, additional transcription time is $0.07 per minute or $4 per hour. It also has some additional features not included in the Basic plan.
If you decide you need a specific transcription done with high accuracy, Descript has professional transcriptionists standing by with their White Glove service. It purports to provide 99% accuracy with an average turnaround of 24 hours at $1.25 per minute.
Temi is a popular automated transcription service which is quite easy to use. You simply upload your audio or video file to their site and they perform the transcription. They make it available in an online editor which allows you to make changes as necessary. The transcript can then be downloaded in a variety of formats including MS Word and PDF.
I give Temi credit since on their website they state that if the quality of your source audio is poor, their transcription will be "Mostly Unusable". With a good audio source, they suggest their accuracy will be 90-95%. Their pricing is a pay-as-you-go plan at just $0.10 per audio minute, making it the least expensive of these three services.
Temi transcribed my test file fairly accurately. Unfortunately, it struggled with transcribing the phrase "sous vide", which we use often. Temi transcribed this in a variety of forms including: suby, suvid, Sioux bead, Suvi, CVA, sue, and susie. Obviously, this makes it difficult to do a simple find and replace to fix them. You are going to need to read the entire transcription carefully in order to proof it. The other area where the transcription falters is in transcribing temperature numbers such as "150". Because of these difficulties, we decided to go with Descript rather than Temi.
Trint is another online transcription service based on AI. To use Trint you simply drop your audio or video file onto their webpage and it uploads to their site. The transcription is done quickly and is sent back to you in a variety of file formats. Their editor allows you to view the video and edit the text easily.
Trint performed well at transcribing my test file. It did the best job of transcribing "sous vide" as well as temperature numbers such as "140 - 150". The greatest downside to Trint is the cost of their pay-as-you-go plan at $0.25 per audio minute. This is higher than the other two options. There are numerous other features associated with this plan that may justify the additional cost based on your personal requirements.
These three automated transcription services will provide you with a transcript which is reasonably accurate and inexpensive. In all cases you will need to proofread the transcript and make changes where necessary. The transcription process is relatively quick and much faster than doing it yourself or using one of the free transcription methods.
We feel confident that any of these three automated transcription services would meet the needs of food bloggers who want to add transcriptions to their content portfolio. Since all three services have free trials, you may want to test them with a file or two of your audio and video to see which service meets your needs the best.
In rare cases, an automated transcription service may not be the right choice, and you will need to go to a Paid Manual Transcription Service covered in the next article.
How do you approach this? Let me know in the Makin Bacon Facebook Group or the comments below.
