Yeah, super-resolution is my favorite feature of MAnalyzer! And I don't mind the 'inaccuracy' when it's enabled. But as the first post in the thread noted, i'm also talking about the case where it is off.
With super-res off, and time resolution at 100% as you suggested in the email chain, I can get an infinite-average graph that trends with only 1 or 2dB slope to it. It's not super flat on the small scale (i.e. a bit bumpy), but it's pretty close. Higher FFT sizes seem to have more bumpiness.
And that's totally cool, if that's just the way the analyzer is designed for the reasons you've described; I just want to understand the tool I'm working with. Thanks for the attention to it, and thanks for the work on the plugins.
MAnalyzer bug or user error?
-
- KVRer
- Topic Starter
- 14 posts since 26 Oct, 2014
-
- KVRian
- 1094 posts since 23 Sep, 2006
I still want a super resolution mode for the actual audio signal, for the spectral effects.
According to Vojtech it's very heavy on cpu and would be complicated to build, but it could avoid the time accuracy vs frequency accuracy compromise.
According to Vojtech it's very heavy on cpu and would be complicated to build, but it could avoid the time accuracy vs frequency accuracy compromise.
-
MeldaProduction MeldaProduction https://www.kvraudio.com/forum/memberlist.php?mode=viewprofile&u=176122 - KVRAF
- 14339 posts since 15 Mar, 2008 from Czech republic
That's actually quite the problem - it is basically processing several parallel differently sampled FFTs and merging them together. Unfortunately it's rather problematic to make the transformation lossless and even worse reversible... Plus I was thinking about the resynthesis and I think it would end up with completely f*** up phase.
