Automatically Transcribe YouTube Video/Audio with Google Docs.
A quick Google search for YouTube video transcription will show either paid audio transcription services like Fiverr / Rev, or other blogs offering audio transcription tools where you need to enter everything thing by hand. But luckily there is a better way.
Thanks to machine learning, computers (in this case, Google’s voice to text function) can now automatically create subtitles from any video or audio. By default, it hears your voice from the device microphone. And with some tweaking, we use that to convert any video / audio to text. (Or watch the video tutorial at the end of the article).
This workaround is free, works on both Windows and Mac (not mobile yet). Best of all, it also supports many foreign languages. Although, to be honest, it’s still not 100%. But if the sound is clear, you can easily achieve 80-90% accuracy.
Sounds interesting? So let’s see how to do this.
On the subject: How much money do youtubers make? Youtubers answer
Why copy a YouTube video
1. SEO Benefits: Unlike blog posting, YouTube cannot read your videos. Yes, there are things like title, tags, etc. that tell YouTube what your video is about. But adding subtitles to all of your videos will tell them more about the content. They can even boost your videos in search results
2. Accent: People come to YouTube from all over the world, and accent can be a big problem. For example, the English (American) accent is very different from the English spoken in India. So the signature comes in handy
3. Transcribe other videos: if you have a foreign video that has no subtitles on the Internet
4. Transcribe videos for money. If you are making money transcribing videos on Fiverr or Rev, then this workaround will help you automate 80% of your work.
5. Change the purpose of the video on your blog: if you’ve uploaded a video with unique content and want to republish it in your blog post. Or you have found some kind of video lecture on the Internet and want to transcribe it for academic purposes.
If you fall under any such scenario then this method will help.
RELATED: Difference Between Public, Private, and Private YouTube Videos
Download copies if the YouTube video already contains it
Before you get down to the hard work of creating subtitles for YouTube videos, it is best to check if they already have subtitles or not. To check, look for the “Copy” button next to it, or go to the settings and find the subtitles there.
Usually all videos uploaded to YouTube after 2014 have automatic English subtitles by default, which is pretty good if you’re a native speaker. Many professional YouTubers add signatures as well. If you see a signature, it’s pretty easy to download.
In the video description, click More Title Select Language you will see subtitles, just copy and paste them. However, for some reason, if you want to download a .srt file with timestamps, or want to do it with bulk videos, use Ccsubs or Down sub. There is also a chrome extension on GitHub that does the same.
Transcribe video / audio to text with Google Docs
There are many video to text converters online or offline, but I find Google Voice to text to be the best. It wasn’t very efficient a few years ago, but thanks to AI, this feature has changed a lot.
Google Voice to Text converts your sound to text in real time. But if you try it by playing a video on one device and recording it from another using Google Voice to text; then, unfortunately, you will not get much accuracy as most of your words will be lost in the noise.
So, the trick here is to get your computer to record system audio instead of the microphone. And then play the audio or video you want to transcribe and record it using Google docs voice to text. The calculations are done on a Google cloud server, so you’ll need an active internet for this to work as well.
Now let’s see how to do this.
Related: How to Block Specific YouTube Channels
1. Copy video / audio to text on macOS
Most computers do not allow recording audio from the computer, perhaps to avoid piracy (for example, people use it to record Spotify songs, etc.)
1. Download a third party software called soundflower; this will help us record the system sound. After that, unzip it and install.
2. Next, you need to tell MacOS to use the output audio as input. To do this, go to sound settings and set soundflower2ch on both input and output.
3. Now launch Google Chrome (yes, it only works in Chrome). Open Google Docs right-click and select Create New Document Tools Voice Input.
4. In another Chrome window, open YouTube and play any video.
5. Now go back to Google Drive, click on the Google voice icon and select your accent or language from the list, and then start recording.
And it’s all; now you have to write text on your screen.
2. Copy video / audio to text on Windows PC
Now let’s try this on Windows
1. Go to Windows Sound Settings select your recording device select Stereo Mix and set it as default. If you don’t see the Stereo Mix option, right click and enable show disabled devices.
2. Then do the same as for macOS, i.e. open Google Docs right click and select “Create New Document” “Tools” “Voice Input” 3. Play video and start recording. And that should work.
What if Stereo Mix is not available?
In many newer computers, the sound card does not support the stereo mixing option, for that you can check this article on how to record system sound without a stereo mix. Unfortunately, I haven’t tested this method, so I’m not sure if it will work.
If the Stereo Mix option is lost after upgrading your PC to Windows 10, you can install the Realtek audio driver and enable it in Windows Device Manager. Reboot the system and you will see the Stereo Mix option in your sound settings again.
How do I download a copy to YouTube?
Now that you have the subtitles in your text file, you are ready to upload it to YouTube. Here’s how to do it.
1. Go to your YouTube dashboard, click the Edit button next to your videos Subtitle / Duplicate Add New Subtitle Select Language Rewritten & Auto-sync, and then paste the text there. Synchronization takes 10-15 minutes. Don’t forget to come back in 10-15 minutes and post. Also disable automatic. I used this method to transcribe several old videos, the accuracy was always over 80%.
Update: The best speech to text app is here
A new application has appeared in the city – Voicera, which will help you decrypt any video. It’s free, works for both Android and iOS, and most importantly, it scored 95% in our conversion accuracy test.