AI and voice recognition are doing a good job
The operation of subtitling is a kind of hell for every videomaker. Each of us has behind us sleepless nights to hunt in long speeches, sentence after sentence, cell after cell. From the research I’ve done in these months, however, I can begin to tell you that the era of hours lost behind the subtitles of your videos could be a memory. The merit is of a developer just over forty in Tel Aviv called Eli Leshem who has developed a miraculous app called “Autocap” that adds the so-called “caption”, subtitles, in fact, on the finished video.
The tech behind
I had a long chat with Eli and I discovered a world. Behind Autocap there are servers that do voice recognition, passing the entire video and translating the sentences said in sentences written with a cure of 80-90%. I had never encountered similar applications, until I discovered it thanks to a good tip of @smartfilming, aka Florian Reichart, German wizard of the android world. Instead, I had seen speech to text applications that, however, never showed me the accuracy in the transcript that shows Autocap, on which the need for intervention never passes 10 percent of the words said.
Behind this app there are plenty of servers that do voice recognition and that, among other things, learn as you feed them videos. Precise algorithms equipped with artificial intelligence that makes self – learning, wrap the subtitles in long videos, on which, thanks to Autocap, you can also intervene with new fonts and different colors. To discover all the secrets, in any case, I leave you to the chat I did a few days ago with Eli who is in English and, of course, is not subtitled :-). Last note: Autocap, for now, is only for Android, but will soon be developed for iOS.