Re: Apple II Music Video - Welcome to the Jungle!!

["Frank M." <frank_o_rama@hotmail.com>]

Thanks! i think it looks pretty cool, too! thanks for the code,
michael. i was so focused on getting a proof-of-concept that my code
was a bit sloppy. and i actually didn't know you could use offsets like
that in ProDOS (i was always more of a dos3.3 guy since that's how the
bootleg of Hard Hat Mack showed up and also because hard drives were
something we never used). it will be even better when/if we can get it
running at 25 frames per second. that's when the dithering effects will
really stand out, i think. i could write a quick program to load larger
chunks of frame data into high memory and then just rapidly swap banks
of memory into the GR screen. you could probably do hundreds of frames
a second that way. problem is that you can only fit so many frames into
RAM before you have to do another disk access. maybe it could be fast
enough to fit into the delay between swapping frames.

i'm rendering a couple more movies today, so there'll be other things
to work with. i realize the dithering method i'm using isn't perfect.
i'm not as much a programmer as i am an imaging nerd (photoshop is my
'day job'). since i'm currently using photoshop to do the downsampling,
i'm sort of constrained by what i can automate. but i can see how it is
possible to write an application that you run a quicktime movie through
and it produces these 13,707 block files formatted and ready to go. and
on top of that, 40x48 pixels isn't really a whole lot of information to
work with, so it only follows that you'll lose some information.

you realize the next step is running a sample using DAC522 during the
delay between frames, right?  ;)   even though there was never much of
a demo-scene for the A2, it's never too late to start.....


=f=


Michael J. Mahon wrote:
> Frank M. wrote:
> > Here's the code for rendering the lores photoshop TIFFs onto the A2 GR
> > screen.....
>
> Frank, it's pretty awesome to watch the results of your rendering
> on an Apple II!
>
> I notice lots of sequences where there are regions containing
> very little detail.  For example, outlines of figures are
> often visible, but any detail in the figure is missing (which
> would be expected if the figure is only a silhouette in the
> original).
>
> Have you considered using dynamic thresholds between colors that
> track "median" colors of regions, so that details would tend to be
> rendered instead of "falling into the same color bucket"?
>
> If a little noise were introduced into these thresholds, it would
> also add time-domain color dither, which would have the effect of
> increasing the palette.
>
> I realize that I'm talking about an adaptive algorithm that may
> have to be "custom" instead of a pre-built image filter, but maybe
> this has already been done to handle the "posterization" operation...
>
> I think that a little increase in color distortion would be a small
> price to pay for increased visible picture detail.
>
> -michael
>
> NadaNet networking for Apple II computers!
> Home page:  http://members.aol.com/MJMahon/
>
> "The wastebasket is our most important design
> tool--and it's seriously underused."