Speaker 1: Hello everyone. Welcome back to the channel. Today I'm going to show you how to transcribe your interviews in a way that won't break the bank using the OpenAI Whisper API and then analyze them in Envivo 14. These days many subscription-based AI transcription tools can get really expensive especially if you're working with large amount of data. So with Whisper for example I was able to transcribe 18 one-hour interviews for only $7 and that's considerably lower amount of money that I spend for example on order AI. Plus a lot of these services only support English while Whisper offers multiple languages and here they are you can check them on OpenAI documentation. So if you are conducting interviews in other languages, so not in English, Whisper could be actually a great option for you. So let's dive into how you can use Whisper. Whisper is actually by itself a powerful speech recognition model developed by OpenAI. It can transcribe audio into text across several languages not only English. However one thing to keep in mind is that audio files need to be less than 25 megabytes in size. So basically if your file is larger you need to split it up into smaller segments. First thing first head over to OpenAI and sign up for API key if you don't have one yet. You'll need this key to access the Whisper transcription service. Once you have your API key you'll need to either export it as an environment variable or directly input it in your script. And in my case I already put it into the environment variable so I won't have it here. But if you don't know how to put this API key into the environment variables what you can do is in this line that says client equals OpenAI inside of the brackets you put API underscore key equals and then put your API key there. So once you have set it up you can send your interview audio files to the Whisper model for transcription. And here is a simple Python script that shows how to do it and it also requests the timestamps to make your data analysis easier. Here I chose timestamp granularities to be segment. The other option is word but I think word is too granular basically. It will make everything look like closed captioning. So you can stop this video and copy the script and the script will send your audio file to Whisper API and then return a detailed transcription with timestamps which will be useful later for syncing it with NVivo. So when you run the transcription we can expect the output to look something like this. The output includes the transcribed text then it also detects the language then the breakdown of the audio segments and each segment is given its start and end time etc etc. In this example it's just a short audio in Japanese which I prepared for this tutorial. And then next we'll going to convert these timestamps into more readable hour minute second format so to make this transition for NVivo much easier. Now that we have our transcription with timestamps we can process the data and convert the timestamps to this hour minute and second format and prepare the text to import for NVivo 14. And here you can see the Python code how to do this. You can again stop the video and copy this code to do it yourself. What this code does it processes the Whisper transcription, converts the timestamps to this hour minute second format and then saves the result as a txt file. And now this file is actually ready to be imported into NVivo as transcription. And just as an example each segment or this file looks something like this. You have a timestamp and then on the new line you have the segment itself and then you have an empty line and then this process repeats again and again. Now let's import both the transcription and the original audio into the NVivo 14 for analysis. And this is how you do it. First just open your NVivo and create a new project if you haven't done this already. Now you need to go to import and audio and choose the same audio file that you use for transcription and then put this file into your NVivo. Next go into this import tab and then files and then add transcript roles. And this is how you will actually import your transcription into NVivo together with your audio. And there you have it. So with Whisper AI it's quite easy and cheap to transcribe all your interview audios. And that's all for today. Bye.