[Libre-soc-dev] Audio and Video Codec Algorithmic analysis for instruction creation was from Re: clamping/saturation semantics

[Luke Kenneth Casson Leighton]

On Sat, Dec 12, 2020 at 8:33 PM Cole Poirier <colepoirier at gmail.com> wrote:

> and other DSP processing. Is that unreasonable?

not at all.  it's just that we've chosen a methodology (Lauri has
described it well), and, well, that's what i'm happy to follow.  if
that methodology turns out to be incomplete or insufficient, we
re-evaluate based on what is learned.

with things that are this complex i've learned (from "decades of
experience") that it's pretty useless (i.e. pointless and ineffective)
to start something by going, "right, this is complex... so uhh let's
learn everything we can from academic research for the next 2 years
before taking any action".

why?

because the "decades of research" are completely meaningless and
worthless to you because you've zero experience on which to "hook" any
kind of understanding.

by contrast, *ignoring* that "decades of research" until you've
*actually done something* you gain hands-on knowledge that actually
means you now have a comparative baseline against which the "decades
of research" can now be understood.

therefore: to *save time* we do it ourselves, first.

l.