The system appears to be designed to fingerprint N audio files, and then identify an unclassified audio file with one of those fingerprints.
Is it possible to do a more narrow search, where you fingerprint N audio files, and then, given a specific fingerprint as well as an unclassified audio file, find the time index where that fingerprint exists in the file, if at all?
I have an application I'm testing where I split large audio files into chunks for processing. Since the files are quite large, I don't store them, as the cost would be prohibitive. However, occasionally, I want to go back and inspect the original audio source. Unfortunately, I don't control or archive the source files, so they occasionally are slightly modified, usually with different prefaces or epilogue sections that have different lengths. So I can't lookup my audio chunks with absolute time indexes. However, I could probably store a fingerprint and then search for that fingerprint within a rough time range.
The system appears to be designed to fingerprint N audio files, and then identify an unclassified audio file with one of those fingerprints.
Is it possible to do a more narrow search, where you fingerprint N audio files, and then, given a specific fingerprint as well as an unclassified audio file, find the time index where that fingerprint exists in the file, if at all?
I have an application I'm testing where I split large audio files into chunks for processing. Since the files are quite large, I don't store them, as the cost would be prohibitive. However, occasionally, I want to go back and inspect the original audio source. Unfortunately, I don't control or archive the source files, so they occasionally are slightly modified, usually with different prefaces or epilogue sections that have different lengths. So I can't lookup my audio chunks with absolute time indexes. However, I could probably store a fingerprint and then search for that fingerprint within a rough time range.
Is that something currently supported by Dejavu?