Google India has a nice activity for India’s 74th Independence Day.
Go to and try it out on your phone.
The idea is simple – you sing the Indian National Anthem. The system records it using the phone’s mike and converts it into the sound of bansuri, sarangi or shehnai (you can try all 3).
I really liked the UI design. It was very Indian and very intuitively designed, with a karaoke-style Jana Gana Mana scrolling text (only in 2 languages, though – Hindi and English) that plays as we sing.
The fact that they had added a specific note to treat this rendition with respect and add a choice (I tried both options to see what happens if I did not accept it) was extremely thoughtful! If you check the ‘I Agree’ box, you get to click the ‘Get started’ button. If you click the ‘Not for me’ button instead, you get the 3rd screen below.
The trouble with this idea is the output!
The tune of the anthem is the same and we’re all going to sing it largely the same way. The conversion to the instrumental version too would sound largely the same. For the output to be shareable, it has to have a unique imprint of us (the individual) so that we can call it our own. But because the source is the same tune, that uniqueness would obviously be missing, making it not an individualistic share/content.
It is a curious piece of engagement at best, right now. If it is not shareable, we may simply engage with its once and then move on. But if it was shareable, each iteration would lead to a lot more people trying and knowing about it (and creating their own unique digital artifact in the process).
I remember considering such an idea for social media engagement while at Flipkart. The agency had presented an idea where we could create a back-end system to convert tweets into sounds/music. The idea was to assign a musical sound (note) for each character (letters), plus special characters and use that system to convert a tweet into sound/music. We had planned to also hire a musician to even out the odd notes to make it sound interesting and musical, and not like random noise.
It was a curious idea, but interesting nonetheless. But we shot it down for a similar reason as I explained above for the Google idea. The result would simply be, for most ears, a sound… some sound. You cannot understand or reverse-engineer what the tweet was when you hear the sound. Heard without the tweet’s context, it’s just random sound. Not something I could uniquely call my own (though my tweet was my own, uniquely). The tweet was in a format (English language) that anyone else could read and comprehend too, so it becomes a shared context. The sound is just random sound with no context to anyone.
One use-case could have been a unique sound to a name. A name as a sound would be owned proudly by that name’s owner, though it would simply be a random novelty to someone else.
In Google’s case, the end result, in bansuri, shehnai or sarangi is the same, familiar tune no matter how well/badly you sang the anthem. I do understand the marvels of Artificial Intelligence and all that, but this is not a demonstration of AI that may evoke the wonder of, “Wow, how did they do that?”. The reaction is more likely to be, “Hmmm, ok!”.
But yes, special praise for the UI design. Google at least got that really right!
And I do look forward to what Google shares on 15th August 2020, as they mention : “We can’t wait to bring to you the culmination of this experience, so look out for something very special coming your way on 15th August 2020 — the 73rd anniversary of India’s Independence”. I have recorded my version and converted it into sarangi. How would the collective effort sound? I’m curious – let’s see!
UPDATE – August 15, 2020: Here’s the culmination from Google India. As I had anticipated, this hardly seems like something we all recorded individually. It sounds simply like a professionally orchestrated and recorded version of our anthem. It sounds great, but there’s nothing worth observing about AI’s prowess here on a normal listen.