Messages

1

Bugs / Another Annoyance - May be a Bug

« on: January 05, 2015, 22:42:33 »

To the software designer:

When comparing two recordings of the same song where
recording #1 has a low max(abs) of around 0.39
recording #2 has a normal max(abs) of >0.99
the %precise will show <30%

I would recommend low max(abs) recordings be compared utilizing some sort of level normalization so that the %precise calculation will be more accurate.

By the way, it is a shame that word of this program has not spread.  You could sell so many more programs if people just knew about it.  The reason I believe this is that I looked specifically for a program like this many times for the last decade and finally stumbled on it by accident.  Perhaps such a common term as "Similarity" for the program name makes it less likely to surface during searches.  In any event, it is a very excellent program and I've been spreading the word as best I can.

Mike


Mike

2

Bugs / Great Program with just one Annoyance so far

« on: December 20, 2014, 12:00:20 »

First, hats off to the programmers.  I've been looking for a program like this for years and was constantly surprised that no one had used FFTs and signal processing techniques to provide a quality metric of a music file instead of just using the bit rate.  However, I do question that a 128 kbs file and even a 160 kbs file (not vbr) can be evaluated as 100% for a quality metric.  I can certainly hear a degradation in quality at 128 kbs and sometimes 160 kbs compared to the original.  It would seem that even if all the basic metrics such as clipping and spectral roll off are good, any 128 kbs file should be rated at less than 100%.  Perhaps 99% maximum for 160 kbs encoding and 98% maximum for 128 kbs encoding would be appropriate if all the other metrics are good.

Thanks,
Mike