search

dpwe / audfprint

Landmark-based audio fingerprinting
MIT License
536 stars 121 forks source link

How search works? #52

Open loretoparisi opened 5 years ago

loretoparisi commented 5 years ago

Could you please depict how this landmarks based fingerprint search works in your implementation? Also, assumed I want to embed the audio representation in a different way (let's say spectrum or correlogram instead of fingerprint), would the landmark based representation work? Thank you.

dpwe commented 5 years ago

You could try reading Avery Wang's papers e.g. https://www.ee.columbia.edu/~dpwe/papers/Wang03-shazam.pdf The point about the spectral-peak-pair landmarks is that some of them remain invariant to channel and additive noise, giving the representation robustness. You can try to use any other kinds of derived features, but they will likely have different properties in terms of robustness and discriminability.

DAn.

On Fri, Dec 21, 2018 at 5:51 AM Loreto Parisi notifications@github.com wrote:

Could you please depict how fingerprint search works? Also, assumed I want to embed the audio representation in a different way (let's say spectrum or correlogram instead of fingerprint), would the landmark based representation work? Thank you.

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/dpwe/audfprint/issues/52, or mute the thread https://github.com/notifications/unsubscribe-auth/AAhs0fQD3shUFUr0OUOGxUC--8iQ9ohyks5u7L0bgaJpZM4ZdyhT .