[prev in list] [next in list] [prev in thread] [next in thread] 

List:       libmikmod-devel
Subject:    [Libmikmod-devel] More advanced mixer
From:       Simon Hosie <gumboot () clear ! net ! nz>
Date:       2000-02-28 11:23:44
[Download RAW message or body]

Right, well, I've just done the theory, and the testing, and the plots, and
...well... pretty much everything but useful code.

What I have now is significantly improved interpolation.  With eight
multiplies per sample, which I think may actually be less than high quality
mixing, it will give near ideal frequency response almost up to the nyquist
frequency of the original sample, and almost no noise at all beyond.  It
should actually offer better quality and hopefully better speed than the
current high quality mode.

There are two problems, though:

One is that it requires future and past data.  Past is easy because you just
allocate four zero entries ahead of the actual sample starting position
(this would need to be addressed in code outside the mixer).  Future is a
problem because the data beyond the end may be zeroes, copies from the loop
point, or a reflection of the end of the sample, depending on the loop mode.

The other problem is my reluctance to work on code other than the mixer in
3.2 (ie. my code).  That doesn't appear to be available for download at the
moment, so I don't know if it's meant to be worked on or not.  Still, I
think it would be a heck of a lot easier to plug a new mixer into that (and
quite a relief to remove the high quality one).

What's the situation with that code?  Is it in use, does it need to be
cleaned up more before it can be put into use, or what?


Oh, and I don't appear to be getting mail from the mailing list... or more
precisely, cowpatml@ (I don't know where this is going to be published so
I'll protect it from spam just in case) clear.net.nz isn't getting it.  Do I
smell bad or something?

[prev in list] [next in list] [prev in thread] [next in thread]