Search for sounds from the realm of fiction
This is not the first year that I have been interested in online news. On duty, many of them are testing. I pay particular attention to the latest searches and love to test them (which I write about on my blog ). Of course, I cannot say that I know everything about the search (and who can? The question, of course, is rhetorical). So - recently I needed to solve a search problem, which, due to the peculiarities of the information sought was not only unsolvable - I can’t even imagine how to implement it.
At the concert, I heard a famous artist sing a famous English song. I even tried to reproduce the first few lines (however, I’m not sure about my English). I would like to find this song and download the corresponding mp3 file. However, how? The task is completely unrealizable, in the information field and at the level at which search engines now operate.
Let's compare a search, for example, in Google with a search, for example, in an audience where 100 connoisseurs of contemporary music are sitting. Musicians can try to sing the beginning of a song (melody), and if your voice is not completely terrible and your hearing is not completely absent, then with the composition being known enough, there is a high probability that someone will even recognize it.
With Google, all this does not work, naturally, how will you sing it?
All these essentially trivial considerations very well demonstrate the difference between algorithmic search and search in the human environment.
Nevertheless, startup authors and scientists are trying to solve this problem and launch a search as close to human as possible - a few days ago the whole Internet actively discussed the Wolfram Alpha service announcement .
However, why not dream - and at least mentally travel to the future, where the task of this kind of melody search would be solved. Imagine that search engines have learned to deal with them. But how to submit such sound information to a search engine? In the future, will Google learn to capture sound from a microphone and recognize it? Or will there be a sample library grouped by genre? What do you think?
At the concert, I heard a famous artist sing a famous English song. I even tried to reproduce the first few lines (however, I’m not sure about my English). I would like to find this song and download the corresponding mp3 file. However, how? The task is completely unrealizable, in the information field and at the level at which search engines now operate.
Let's compare a search, for example, in Google with a search, for example, in an audience where 100 connoisseurs of contemporary music are sitting. Musicians can try to sing the beginning of a song (melody), and if your voice is not completely terrible and your hearing is not completely absent, then with the composition being known enough, there is a high probability that someone will even recognize it.
With Google, all this does not work, naturally, how will you sing it?
All these essentially trivial considerations very well demonstrate the difference between algorithmic search and search in the human environment.
Nevertheless, startup authors and scientists are trying to solve this problem and launch a search as close to human as possible - a few days ago the whole Internet actively discussed the Wolfram Alpha service announcement .
However, why not dream - and at least mentally travel to the future, where the task of this kind of melody search would be solved. Imagine that search engines have learned to deal with them. But how to submit such sound information to a search engine? In the future, will Google learn to capture sound from a microphone and recognize it? Or will there be a sample library grouped by genre? What do you think?