In this tutorial, we will create a Word Document from a YouTube Video using the automatic speech recognition service from Amazon called, AWS Transcribe.
We will upload an mp3 file to the AWS transcribe service, which will then give us a JSON file containing the text of the audio file, and then we will take that JSON file and convert it to a word document.
You will need an AWS account; you can click here to create an account for free.
You need to create an S3 bucket on AWS, look at a bucket essentially as a storage pot on S3. In essence, S3 is an AWS service and inside of that service, you can create yourself little storage pockets/folders, called buckets.
To create a bucket
The Create bucket wizard opens.
Upload your MP3 file to your bucket, give it a nice name so that when you come back to this in 6 months’ time you still know what it was about.
Now we need to transcribe the file, in essence; you will tell AWS where the file is (on S3) and then you will tell AWS what to do with the file, in this case, we want to transcribe it.
Then click on create job.
Give it a descriptive name
Leave the model type as “General model”
Click on browse S3 and choose the file you uploaded in step 2
Leave the Output data option as “Service-managed S3 Bucket”, because we will download the file immediately, we do not need AWS to keep it forever.
Now wait for the job to finish, this normally takes about 10 minutes for a 30-minute-long file.
When the file is finish, click on it
Then on this screen that appears click on “Download full transcript” it will be available in JSON format, we will convert that to a Word document in Step 5
We need to store the JSON file and our python app somewhere, so create a folder on your desktop and call it something descriptive like “Transcriptions”.
Just in case you missed it, on the AWS Transcription Jobs screen, click on the transcribed file after it completed and a new screen will appear, now on that new screen click on download full transcript and then save that JSON file in the folder you just created.
Rename that JSON file to something that you will remember, for example interview.json
Download Python by clicking on this link: Download Python 3 for windows
Now install that file, when the window pops up, click on “Add Python to PATH”
Open CMD (Inside the search field from your taskbar, enter command or cmd. Then, click or tap on the Command Prompt result.)
Now type in the following command:
pip install tscribe
This will install the library that we need. Wait for this to finish, it will look something like this:
C:\Users\erikthiart\Desktop\aws>pip install tscribe Collecting tscribe Downloading tscribe-1.3.1-py3-none-any.whl (7.2 kB) Collecting python-docx Downloading python-docx-0.8.10.tar.gz (5.5 MB) |████████████████████████████████| 5.5 MB 6.4 MB/s Collecting webvtt-py Downloading webvtt_py-0.4.5-py3-none-any.whl (16 kB) Collecting pandas Downloading pandas-1.1.0-cp38-cp38-win_amd64.whl (9.6 MB) |████████████████████████████████| 9.6 MB 6.4 MB/s Collecting matplotlib Downloading matplotlib-3.3.1-cp38-cp38-win_amd64.whl (8.5 MB) |████████████████████████████████| 8.5 MB 6.4 MB/s Requirement already satisfied: lxml>=2.3.2 in c:\users\erikthiart\appdata\local\packages\pythonsoftwarefoundation.python.3.8_qbz5n2kfra8p0\localcache\local-packages\python38\site-packages (from python-docx->tscribe) (4.5.1) Collecting docopt Downloading docopt-0.6.2.tar.gz (25 kB) Collecting pytz>=2017.2 Downloading pytz-2020.1-py2.py3-none-any.whl (510 kB) |████████████████████████████████| 510 kB 6.4 MB/s Collecting numpy>=1.15.4 Downloading numpy-1.19.1-cp38-cp38-win_amd64.whl (13.0 MB) |████████████████████████████████| 13.0 MB 13 kB/s Collecting python-dateutil>=2.7.3 Downloading python_dateutil-2.8.1-py2.py3-none-any.whl (227 kB) |████████████████████████████████| 227 kB ... Collecting pillow>=6.2.0 Downloading Pillow-7.2.0-cp38-cp38-win_amd64.whl (2.1 MB) |████████████████████████████████| 2.1 MB 3.2 MB/s Collecting cycler>=0.10 Downloading cycler-0.10.0-py2.py3-none-any.whl (6.5 kB) Collecting pyparsing!=2.0.4,!=2.1.2,!=2.1.6,>=2.0.3 Downloading pyparsing-2.4.7-py2.py3-none-any.whl (67 kB) |████████████████████████████████| 67 kB 4.8 MB/s Collecting kiwisolver>=1.0.1 Downloading kiwisolver-1.2.0-cp38-none-win_amd64.whl (58 kB) |████████████████████████████████| 58 kB ... Requirement already satisfied: certifi>=2020.06.20 in c:\users\erikthiart\appdata\local\packages\pythonsoftwarefoundation.python.3.8_qbz5n2kfra8p0\localcache\local-packages\python38\site-packages (from matplotlib->tscribe) (2020.6.20) Collecting six>=1.5 Downloading six-1.15.0-py2.py3-none-any.whl (10 kB) Installing collected packages: python-docx, docopt, webvtt-py, pytz, numpy, six, python-dateutil, pandas, pillow, cycler, pyparsing, kiwisolver, matplotlib, tscribe Running setup.py install for python-docx ... done Running setup.py install for docopt ... done Successfully installed cycler-0.10.0 docopt-0.6.2 kiwisolver-1.2.0 matplotlib-3.3.1 numpy-1.19.1 pandas-1.1.0 pillow-7.2.0 pyparsing-2.4.7 python-dateutil-2.8.1 python-docx-0.8.10 pytz-2020.1 six-1.15.0 tscribe-1.3.1 webvtt-py-0.4.5 C:\Users\erikthiart\Desktop\aws>
Create a text file in the folder on your desktop that you created in Step 4.
Call the text file app.py
Now open that file with Notepad and copy and paste the following inside:
import tscribe tscribe.write("interview.json", format="docx", save_as=" interview.docx")
Open CMD again and then paste the following command in order to get to this folder you created in Step 4.
Type it exactly like that (copy it), do not change anything, that
\%username%\ part is intentional, it will essentially resolve to whatever your username is, which I won’t know since I am sitting in Cape Town and do not know your PC name.
If you did it correctly then you will see you are successfully inside that folder.
Microsoft Windows [Version 10.0.18363.1016] (c) 2019 Microsoft Corporation. All rights reserved. C:\Users\erikthiart>cd C:\Users\%username%\Desktop\Transcriptions C:\Users\erikthiart\Desktop\Transcriptions>
Now type the following command in CMD
You will see the following after a little while.
interview.docx written in x seconds.
That means it worked! Congratulations, you can now look inside that folder and you will see a word document containing your text translation of the audio file.
Transcription of Omeshnee: Begotten from Future Cities Africa
Transcription using AWS Transcribe automatic speech recognition and the ‘tscribe’ python package.
Document produced on Sunday 16 August 2020 at 18:03:15.
|98% – 100%||1809||77.94%|
|90% – 97%||127||5.47%|
|80% – 89%||62||2.67%|
|70% – 79%||37||1.59%|
|60% – 69%||30||1.29%|
|50% – 59%||27||1.16%|
|40% – 49%||14||0.6%|
|30% – 39%||8||0.34%|
|20% – 29%||1||0.04%|
|10% – 19%||0||0.0%|
|0% – 9%||0||0.0%|
If you need any help setting this up, please follow me on Twitter (@ErikThiart) and then you can tweet me and I will help you where you get stuck.
If you found this guide to be useful and would like to make a donation, please head over to my donations page, I use the funds to pay for the hosting of this website and to buy beer on the weekends.
Google’s official WordPress plugin. Get insights about how people find and use your site, how to improve, and monetize your content, directly in your WordPress dashboard. If you have a WordPress website and would like to see how users find and use your website, Site Kit is right for you. To help you build, maintain, and grow your online presence, Site Kit gives you easy-to-understand metrics and actionable insights directly on your WordPress dashboard.Read The Article
You might have found that Hulu is limited to US viewers and the BBC iPlayer is likewise limited to UK IP addresses, however using a very easy to set up proxy server you can access BBC Iplayer, Hulu and many others from any country on all of your internet capable devices such as your Laptop, PC, Mobile Device, Kindle Fire TV, Apple TV, Google Chromecast etc. In this guide you get the ultimate viewing experience when streaming/downloading any type of regional or geo-blocked content online.Read The Article
The internet is full of restrictions, censorship, privacy invasions, security breaches, and a host of unwanted entities trafficking in your personal data. Using a VPN is a great first step toward protecting your privacy online. Privacy is fundamental to a well-functioning society because it allows norms, ethics, and laws to be safely discussed and challenged. Without privacy, a free and open society can neither flourish nor exist.Read The Article
You probably know the Sonoff brand for smart home electronics like the POW R2 or the Sonoff Basic switches, but they also released a security camera, the Sonoff GK-200MP2-B. I decided to try it out and made an in-depth review so you can make the right decision. Considering this is Sonoff, a well-known brand for robust electronics. I expect a good working, premium product that just works especially since to use the Sonoff GK200MP2-B, you need the eWeLink application which integrates very well with other Sonoff devices.Read The Article
The following mikrotik firewall rules will force all the clients on your local network like your Sony PlayStation and Google Chromecast to use your Pi-hole or your own local server as their primary DNS server, even if they have hard coded their own DNS servers we do this because many apps and devices do not use the offered DNS servers per DHCP, they are just that – an offer. Hardcoded DNS servers will still resolve and allow ads and tracking unless we use NAT rules that will redirect all DNS requests, no matter where they go, to the Pihole.Read The Article
First thing to understand is that on the hidden web, you do not know about a website unless that website is shared with you. The dark web’s content on the internet isn’t accessible through traditional browsers or standard browsing technology. Content on the dark web is designed to be hidden from search engines and from casual users – you can’t simply stumble across dark web websites by accident.Read The Article