Dan Ellis comments

Results 73 comments of


                                            Dan Ellis

How to Debug a Non Match

I would try a higher density; matching with 19 hashes in 10 sec is a week (although reliable) match. Try running both analysis and matching with --density 100. This will...

Can this algorithm load the historical features into memory first, so that the matching speed is improved, but I don't know how to modify your basic code

I wouldn't call the peaks "eigenvalues" (although sinusoids are the eigenfunctions of LTI systems so .. maybe?) but otherwise your description seems to cover what audfprint currently does.

License?

Thanks for your interest. I just added an MIT License on it. I hope this works for you. DAn. On Sun, Dec 28, 2014 at 5:23 PM, saskay [email protected] wrote:...

Package

I would be happy to see this done, although I don’t really understand the implications. If you want to look into it, that would be great. Thanks! DAn. On Fri,...

The "hashes dropped" proportion will increase as the database fills up. However, it's dropping hashes from every track, not entire tracks. Recognition will be mildly impacted, but everything should continue...

Max file in database

"2023" should be the total number of files you've added to fpdbase.pklz. I can't think why it wouldn't change if the %age dropped is increasing. What is the sequence of...

Max file in database

It should be fine. But you can use --list and full file paths in a text file (or even just multiple full file paths on the command line) to add...

Windows vs. LINUX

I actually do most development on MacOS, but that is much closer to Linux than Windows. I intend it to work unchanged on Linux, but I don't have access to...

what should be the format of mp3 files ? audfprint.py not working

Apparently your Windows shell doesn't expand glob patterns. In a conventional Unix shell, the filename containing a "*" would be replaced by a list of matching filenames before the program...

I found that some match matches are inaccurate. I want to know, is the time and hash generated by this audfprint continuous at the start-end time, or is it a peak value, which leads to the fact that if one song is 1 minute, another song 3 minutes, there are a lot of hashes in the previous minute in the last two minutes of the next song, which leads to the fact that the hash statistics are larger than one minute, resulting in inaccurate sorting, then the match situation is inaccurate, how to optimize this?

Hi, I'm not sure what kind of response you're after. The matching is intrinsically noisy - there will always be some spurious hash matches, which is why the ranking is...