I've just started to look at the command line version of echoprint-codegen and created an audio sample to run some tests.
The wav file is approx 5 mins long and is a continuious 4/4 beat @128bpm sampled at 11025 mono.
I used codegen to produce a fingerprint of a 20-second window but starting at different offsets into the file:
a) echoprint-codegen file.wav 0 20
b) echoprint-codegen file.wav 1 20
c) echoprint-codegen file.wav 2 20
i.e. test a) starts at the beginning of the file while b) starts at 1 second and c) at 2 seconds into the file.
All samples should be 20-seconds from the starting point.
Each code produced has a different "code count" and therefore the code lengths vary - I was expecting a fixed length code in each case (have I misunderstood?):
Why given exactly the same sample window on the same track do I get different fingerprint lengths?
If there is any documentation available to explain this please let me know where to find it :)