Speaker 1: Hey, everyone. How are you? Are you excited about this episode? Welcome to AI Adventures, where we explore the art, science, and tools of machine learning. I am Priyanka Vergadia, and in this episode, I'm going to show you a tool that enables you to automatically translate your video and audio content to text in real time. The world has become rather small with technology. We are all customers of services that are made in another part of the world. So we've experienced situations where the audio or video is in a language that we don't speak. And as a developer of such global apps, we clearly need to overcome the language barrier. And it's not as simple as it seems. What's so complicated about it? It's the data and the files, because they are not in the form of text. We have to first convert original data into text by using speech-to-text to get a transcription. Then we need to feed those transcriptions into the translation API to get the translated text. These are a lot of steps which cause a lot of friction and add latency along the way. And it is much harder to get a good quality translation because you might lose context while splitting sentences. So you eventually make a trade-off between quality, speed, and ease of deployment. That's where Media Translation API comes in. It simplifies this process by abstracting away the transcription and translation behind a single API call, allowing you to get real-time translation from video and audio data with high translation accuracy. There are a large number of use cases for this. You can enable real-time engagement for users by streaming translations from a microphone or pre-recorded audio file. You can power an interactive experience on your platform with video chat with translated captions or add subtitles to your videos real-time as they are played. Now let's see how it works. I'm going to say a few things here, and we will see how Media Translation API translates the microphone input into text. New Delhi is the capital of India. Washington DC is the capital of the United States. Here we see that the API is trying to understand the input from the microphone and giving us the partial and the final translation in real-time. Now how did I set this up? Let me show you. In our project, we enable Media Translation API and get the service account and key for authentication. Get the Python code sample from GitHub, and in the Media Translation folder, we see the files translate from file and translate from mic. First, we create a Python virtual environment and install our dependencies from requirements.txt. Now if we examine our translate from mic file, we see that we are creating a client for the Media Translation API service. This client includes the audio encoding, the language we will speak in to the microphone, and the language we want our translations to be in. Then with that client, we stream the audio from our microphone up to the cloud and start listening to the responses from the service. And as we listen, we are printing those responses as an output. This is also where you would incorporate this output into your application. Now while we are here, let's change our source language to Hindi and target language to English. And try this again. Namaste. Aap kaise hai? Aaj mausam bahut acha hai. Humi bahar ghoomne jaana chahiye. And it rightly says, hi, how are you today? Weather is very nice today. We should go out for a walk. Decent, right? In this episode, we learned how Media Translation API delivers high-quality, real-time speech translation to our applications directly from our microphone. You can even provide an input audio file and achieve a similar output. There's a sample in the same folder. Try it out yourself, and let me know how it goes in the comments below. Bahut bahut shukriya. Agli baar milti hun ek nai episode ke saath. Tab tak ke liye like aur subscribe karna na bhulein.