Speaker 1: This is a demonstration of automated transcription using the DeepGram server with TransAna. The first time you want to use this feature, you'll need to go to the DeepGram website at DeepGram.com and sign up for a free account. Once you've signed into your account, you'll need to get your API key. Find the box labeled Create API Key. You then generate a key and give it a name. You need to remember that key to load into TransAna. The easiest way to do that is to copy it into your clipboard. We'll now switch to TransAna. I'm going to add a transcript to an existing episode. We'll call this transcript DeepGramServer to indicate where it's from, and you'll see that there's a box called Automated Transcription Methods where we'll select the DeepGram server. I now press OK, which triggers the automated transcription tool. The first page gives a confidentiality notice. It warns you not to send data to the server unless you have IRB approval, if your data requires that. If you cannot accept those terms, you should not proceed with automated transcription using the DeepGram server. If these terms are acceptable, check the box and press Next. TransAna needs to perform audio extraction on the file because we're just sending the audio to the DeepGram server, not the video, to speed things up. When this process is complete, press Next. DeepGram offers three language models, which they call tiers, that offer different levels of transcription accuracy. These tiers support transcription in different sets of languages, so it may take a moment to find the combination of tier and language that meets your analytic needs. I'll select the Nova tier here because I want the best quality transcription I can get, and the Nova tier supports the language that I need for the data file. When I look at the list of available languages, I can see that DeepGram offers multiple variations for some languages. I'll select U.S. English as that's the best match for my data. TransAna provides the option to have TransAna time codes included every sentence or between every word in the resulting transcript. For the analysis I plan on doing for this data, time codes every sentence is sufficient. Finally, if this is your first time using DeepGram, you'll need to enter your API key from the DeepGram website. In this case, I've entered it previously, so I don't need to. Once that's all in place, press the Next button, and TransAna will perform the automated transcription. Note that if you have already submitted this file for automated transcription, TransAna will not automatically resubmit the file. If you changed the settings earlier, you may want to press the Resubmit the Automated Transcription Request button to submit the file with the newer settings. You can see that there's a progress bar indicating that the automated transcription process is moving ahead and giving you an indication about how long it will take. Automated transcription on the DeepGram server tends to be very fast. When finished, press the Finish button, and TransAna will display the automated transcription. You'll notice that there is a transcription key that indicates that different colors represent different levels of certainty or confidence in the transcription. Some of the words in the transcript where the automated transcription program indicates over 90% confidence will be incorrect, while some words with lower levels of confidence will be correct. We suggest that you consider the automated transcript to be a rough first draft. You may want to go through it, playing your media file as you go, to clean it up. Alternately, some researchers only make corrections to sections of the transcript as part of the selection, coding, and categorization process that makes up analysis in TransAna. There's now a button in the Document Window toolbar that changes character formatting to black so that you can go through very quickly and reformat all of the correct text that is in color. Transcription by Trans-Expert at Fiverr.com