# How to transcribe audio fragments in Bengali Hi there! Thank you for joining us. Here are instructions on transcribing audio fragments via our special web platform. Please, read this tutorial carefully and keep it open while working. All the examples here are in English but the same rules of transcription apply to Bengali as well. ## What do you need for start - Personal computer of any kind. Unfortunately, our platform doesn't work correctly on tablets or smartphones. - Google Chrome browser. Our platform doesn't work properly in other browsers. If you don't have Chrome installed on your PC, you may download it from the official website: https://www.google.com/chrome/. - Headphones or a quiet working environment. ## What should you do The task consists of two main parts: 1. To listen to audio fragments and transcribe them into text. 2. To leave a note whether there is background noise on the audio or not. There are specific requirements for both steps that are described in more detail below in this instruction. ## Let's meet the interface To make your work easier, we created a special platform. Let's explore its interface. This is the page that you will see after login. Please make sure that the translation into English is turned off in your browser. Otherwise, some elements of the interface might be displayed in a foreign language. ![](https://hackmd.io/_uploads/ByRrnpCy6.png) <br> | Element|What's that | | -------- | -------- | | **0** | The name of the task. | | **1** | This is a vizualisation of sound wave with the whole audio. | | **2** | Blue segment is an indicator of selected (cut out) fragment that you need to transcribe. | | **3** | Two sliders that help to regulate where the fragment should start and end. | | **4** | The field where you should type in the text from the selected (blue) fragment. | | **5** | 'Lose (play) the whole track'. <br>This is the play and stop button that plays the **whole audio** file. | | **6** | 'Lose (play) a segment' <br> This is the Play and stop button that plays from the beginning of the **selected (blue)** segment only. | | **7** | 'First second'<br>The play button for playing the first second of the selected (blue) fragment. | | **8** | 'Last second' <br>The play button for playing the last second of the selected (blue) fragment.| | **9** | 'Ignore (skip)'<br> This is the button that skips current audio file. Use it if you're hesitating about what to do with this audio file. | | **10** | 'Accept' <br>Use this button every time after you typed the transcribed text.| | **11** | 'Reject' <br>Use the Reject button if there's no understandable speech on the audio, or the background music is too lound, or the speaker speaks not in Bengali.| | **12** | 'Back' <br> This button redirects you to the prevous audio file. | | **13** | 'Increase the zoom' <br>'Decrease the zoom' <br> These buttons help to zoom in or out the sound wave so that it's easier to cut out the fragemt. | | **14** | 'Speed'<br> 'normal'<br> '50% faster' <br> '50% slower'<br> These buttons help to regulate the speed of the audio if it's needed.| | **15** | ID of the current audio. If you have questions about the audio, you may write down this ID, send it to Lisa, and she'll help you.|<br> ## How to transcribe the text 1. Tap on the 'Lose/play the segment' button to play only the blue segment (6 on the image) and listen to the audio fragment. 2. Write down in the field (number 4 on the image) the text that you hear on the blue segment. **Important!** You need to write down only what you hear on the blue segment (2 on the image). You do not need to write down the text from the whole audio. Special requirements for transcription are mentioned below. 3. After you typed all the text, please, put two slashes (//), and after that leave the relevant number: - 0 if the background is silent; - 1 if there's any kind of background noise (music, sound effects, clapping, sounds of nature, etc.). More info on that you'll find below. 4. After that, tap on the blue button 'Accept' (number 10). ## Rules of the transcription 1. Write down ONLY and EXACTLY what you hear. For example, if the speaker says *'I go went to the hospital'* despite it's grammatically incorrect, write down it as is: *'I go went to the hospital'*. 2. Please, write down ANY numbers that the speaker says as words, not numbers (even if it’s a part of brand name). For instance, ‘8’ should be written down as ‘eight’, ‘Iphone 11’ as ‘Iphone eleven’, ‘store45’ as ‘store forty-five’ and so on. Large numbers, eg. 2023, should be written down the same way as the speaker pronounces them. If the speaker says ‘twenty twenty-three’, then you should transcribe in this way. If the speaker says ‘two thousand and twenty-three’, then in this way. 3. Sometimes the audio segment may end in mid-sentence. In this case, you need to write down only what you hear. For example, the speaker says: *'I liked these cookies. Honestly, I like the swee'* Logically, we understand that the speaker probably says *'sweets'*. However, in this case, we have to write down what we hear only: *'I liked these cookies. Honestly, I like the swee.'* 4. If you cannot understand some sentences/words at all, please do not guess. Instead, just reject the audio file (tap on the button 'Reject'. It's number 11 on the image). 6. Please do the same (reject the audio) if the speaker speaks NOT Bengali. Even if you know the language that the speaker speaks, you do not need to transcribe any other language speech. ## Rules of scoring '0' or '1' After each piece of transcribed text, please make a space, enter two slashes (//) and leave: - 0 if there's no backgound sound/noise; - 1 if there's any background noise (music, clapping, people coughing, sound effects, rain, and other sounds of nature, etc.) ## Examples of transcription <audio controls="controls" src="https://robotvera.ru/media/en_mrbeast_c8VcUnz3nVc_5299229.0_5311149.0.b765c7f9-7cf.short.mp3"> </audio> Transcribed text and score: `but at the same time, you’re limited by personality like. // 0` -- <audio controls="controls" src="https://robotvera.ru/media/en_mrbeast_c8VcUnz3nVc_1199230.0_1211070.0.77ac9f27-a87.short.mp3"> </audio> Transcribed text and score: `you have to be, you have to work on multiple videos at a time, because most of our videos take months to produce. // 0` -- <audio controls="controls" src="https://robotvera.ru/media/en_demo_OxGsU8oIWjY_330000_362000.49c9f654-61e.short.mp3"> </audio> Transcribed text and score: `this is mind blowing enough, but what's even crazier? // 1` ## How to check your progress To check your statistics, click on the word 'Statistics' at the top of the web page. You'll see this info: ![](https://hackmd.io/_uploads/Hy2T5pRJ6.png) ------- ## Video tutorial Here's the short video tutorial. Please watch it before starting the work too: <iframe width="560" height="315" src="https://www.youtube.com/embed/e5jcs1VM0e0?si=3qD2CDNlMhmwbNaE" title="YouTube video player" frameborder="0" allow="accelerometer; autoplay; clipboard-write; encrypted-media; gyroscope; picture-in-picture; web-share" allowfullscreen></iframe> --- If you need more examples of have any questions, please do not hesitate to contact to Liza on Upwork. Thank you and take care!