[MPlayer-dev-eng] [PATCH] fix for -srate bug

[D Richard Felker III]

On Fri, Oct 29, 2004 at 01:20:13AM +0100, Ed Wildgoose wrote:
> 
> >>I'm not quite sure why you and Richard have descended to name calling?
> >>
> >>I have a 1st class honours degree in mathematics, and so I feel 
> >>reasonably comfortable with the definition of "linear".  On the other 
> >
> >apparently you are NOT! i'm not just trying to insult you or
> >something. as michael explained in his reply to this same message i'm
> >replying to, you really appear to be thinking about something very
> >different from what we're talking about!
> 
> Yeah, well, if Michael is right (and I'm sure he probably is), then it 
> looks like I need to get more involved with the resampler theory 
> (perhaps write one) to get a good grip on the basics.  I'm not going to 
> comment further I think.
> 
> However, although you are taking the opportunity to put the boot in 
> again, I don't feel that this in anyway invalidates my ability to 
> understand the concept of "linear"...

just so we're finally clear, a linear filter is (in the most general
sense) any filter that generates the output stream as a linear
transformation of the input stream (think of both streams as a vector
space under pointwise addition). however, usually when someone says
"linear filter" they mean one that comes from a convolution (i.e. the
rows of the transformation matrix are translates of the same
function), since anything else would be rather nonsensical in the
context of these applications.

> ... So I have tested libsamplerate, and the quality is good, the CPU 
> requirements are reasonable.... This is my base position...  I will also 

the cpu requirements are entirely unreasonable. such bad code that
just duplicates lavc's resampler (actually less functionality) will
not be included in mplayer!!

> I really don't understand why we are coming into conflict over this 
> stuff.

i'll explain it very blundtly. we agreed that af_resample sucks and
that something better is needed. michael added support for lavc's
resampler. yet you STILL refuse to believe that it is what you want,
and keep going on and on about this stupid libsamplerate crap!!

if you understood even the most basic mathematics involved in this and
had a reasonable argument, we would be happy to keep discussing. but
you have proven again and again that you don't understand, and you
keep MAKING UP bogus explanations (approximating sinc function with
something piecewise-linear??? i can assure you there's no such code in
any decent resampler, because it's not useful!) rather than either 1)
accepting that someone else knows more than you, or 2) doing the
proper research to learn what you're talking about before you speak
and make a fool of yourself.

i'm sorry to yell, but this just gets tedious. and it's not the first
time you've been senselessly argumentative like this on the mplayer
lists...

rich