Skip to main content ↑ ↓

Notice

Please note that most of the software linked on this forum is likely to be safe to use. If you are unsure, feel free to ask in the relevant topics, or send a private message to an administrator or moderator. To help curb the problems of false positives, or in the event that you do find actual malware, you can contribute through the article linked here.
Topic: Audio fingerprint of remastered songs (Read 3789 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

Audio fingerprint of remastered songs

Hi all,
I'm interested to identify different masterings of same song on quite a big library in a batch process, so to give a unique ID to each audio track depending on its detailed audio content.
Tools to set audio fingerprint are quite common but (all ?) are optimised to find song ID/tags even if content is levelled/compressed/...
My need is for a fingerprint that differs if the audio has been, even slightly, modified...
I could calculate a unique ID based on a combination of various audio parameters : time dependant amplitude, amplitude histogram, spectrum, L/R correlation,...but if there is already a usefull method, why not use it ?
Hope it is clear enough ;)
So thanks for ideas or links to existing methods !

Re: Audio fingerprint of remastered songs

Reply #2 –
Thanks, I'll have a look.
 

Re: Audio fingerprint of remastered songs

Reply #3 –
Quote from: jlohl on
My need is for a fingerprint that differs if the audio has been, even slightly, modified...

Sure you want that? Sometimes, pressing A from country/plant a and pressing B from country/plant b would differ slightly in only the least significant bit. Probably the dither. Whether that was because one did re-dither or because both got 24-bit files and dithered them down to 16 ...
You should maybe clarify what you really want. And whether you have full albums. Ripped from CD? In lossless?

My own situation was that I ripped a quite big collection with automatic metadata retrieval, not checking which pressing I had. Since I had bought people's collections on a few occasions, I had quite a few doubles. I ripped to lossless, and quite a few were bit-identical or bit-identical modulo offset. But apart from that, the "cheapest way" to get an "automated" shot at which masterings are different, would be to look at ReplayGain and/or dynamic range figures. If one version is nowhere close to peak-normalized - then it is not likely to to be the new one. And dynamic ranges? Often higher means older :-/