Thanks for providing a repository for the audio indexing. I am interested in using your implementation as a baseline system for my research work.
Could you please provide any pointers in the code to increase the bits for storing the pair:(song ID, timestamp)? I see there is a tradeoff between bits required for the song ID and time stamp; however, I am interested in implementing the system for a larger reference database of more than 1 million files. Each file's duration could be as large as 1-2 hrs.
I wondered if it is possible to create multiple .pklz files for disjoint subsets of the reference database and merge these .pklz to a single .pklz file, or can it be possible to specify a list of .pklz files for matching a query?
Hello @dpwe,
Thanks for providing a repository for the audio indexing. I am interested in using your implementation as a baseline system for my research work.
Could you please provide any pointers in the code to increase the bits for storing the pair:(song ID, timestamp)? I see there is a tradeoff between bits required for the song ID and time stamp; however, I am interested in implementing the system for a larger reference database of more than 1 million files. Each file's duration could be as large as 1-2 hrs.
I wondered if it is possible to create multiple .pklz files for disjoint subsets of the reference database and merge these .pklz to a single .pklz file, or can it be possible to specify a list of .pklz files for matching a query?