Google’s artificial intelligence is now able to identify samples of tracks under a second long with Google Assistant. Yes, we no longer need to wait for someone to tirelessly figure it out and post it on YouTube. We’ve seen the Daft Punk ‘Discovery’ samples, Mobb Deep, Nujabes, and Madlib. But now, anyone can find unknown samples through Google Assistant’s song recognition technology.
How It Works
We’re all familiar with Shazam. It uses similar audio fingerprinting methods and tries to match the tempo and structure near perfect to obtain a result. But, you either need good data to figure out that song at someone’s festival set or you get a completely random suggestion from a different genre. So, it doesn’t always work. Google Assistant, though, uses deep neural networks so their song recognition is more advanced. This means that it can detect samples chopped or time-stretched, while also being under a second long.
“If the sample you are trying to find is part of a much longer drawn-out chord or texture, you can time-stretch or crossfade a loop to make the AI think it’s longer. You can also repitch the sample to form a chord progression if you guess it correctly.”
DJPasta
This technology is fairly new but scratches the surface of its potential: sample spotting. The tech will get more reliable over time, as advancements in AI continue. Tracklib reports that this new method “puts control into the hands of anyone looking for certain samples, as opposed to relying on a fingerprinting system fully controlled by a tech giant.” Plus, original owners can access multitracks, and not have to manually edit the audio to isolate any samples.
Will Shazam evolve, become extinct, or continue to be a game show? The notion of recognizing a sample under a second is almost inconceivable..almost. Read more on Tracklib’s blog regarding Google Assistant’s AI here.
