You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I've been excitedly looking at training a 48K model of Audioseal to fingerprint some generated content, but have run into failures on a few tests with the native model.
A single watermarked file retains the fingerprint when dropped into a DAW and re-encoded during mixdown, but if I mix it with another file, say a music track or white noise, the mixdown fails the watermark check.
Recording over the air playback of a watermarked file does not appear to retain the fingerprint when checking the recorded file.
Have I missed something, or is Audioseal simply useful for validating provenance of a piece of source media, but no further than that?
The text was updated successfully, but these errors were encountered:
I've been excitedly looking at training a 48K model of Audioseal to fingerprint some generated content, but have run into failures on a few tests with the native model.
Have I missed something, or is Audioseal simply useful for validating provenance of a piece of source media, but no further than that?
The text was updated successfully, but these errors were encountered: