Re: Where do we go from here? (was: Re: Can anyone identify this card?)

[mjmahon@aol.com (Michael J. Mahon)]

Matt Jenkins wrote:

>On Wed, 10 Dec 2003 07:26:43 +0000, Michael J. Mahon wrote:

<snip>

>>I cannot escape the irony and appropriateness of this discussion in
>>a forum dedicated to one of the machines that truly ignited the world
>>of applications that we use today.  Perhaps someone involved in that
>>first ignition can point the way to another.
>
>I'm with you there. Really though, computers have always been used to
>automate already existing processes to manipulate information and media.
>At this point in time, there aren't really any forms of media left that
>aren't computer automated. The intersting thing now is how to leverage
>the digital forms of those mediums in new ways. The WWW is a first step
>along this path, but as history has shown, even when new and better ideas
>present themselves, it can sometimes be a decade or more before the value
>of them is appreciated by a large enough mass of people to bring it into
>the mainstream.

OK--here's a potential "killer app", or at least something that a very
large number of people want to do, that computers could do for them,
that is not presently available...I wish I could say "patent pending".  ;-)

Many people have home movies--either real 8mm movies or old VHS
tapes--that contain treasured memories.  Most of this footage is of
inferior quality compared to what we now expect (DVD quality).  The
prevailing attitude is that this situation cannot be fixed--"you can
copy it, but you can't put back lost quality".

This is incorrect.

Most elements of home movies are present in many frames, differing
only slightly in orientation or position.  That means that computer
processing could "track" those elements, and "stack" their representations
from multiple frames, producing a considerable enhancement in both
resolution and signal-to-noise (grain, in the case of film).  The stacked
elements could then be re-composited back into frames in their original
positions, producing a fully resolution- and SNR-enhanced sequence.
Of course, all the standard (and much less compute-intensive) corrections
to color, gamma, etc., can be performed at the same time.

The computation required to do this with high quality and minimal or
no user interaction is quite significant.  It would make motion-estimation
compression look easy.

The effect would be that, overnight, captured video of a treasured film
or videotape could be converted to a DVD with image quality 2x-4x
better than the original.

So now I've disclosed the principle of the invention.  All the elements
already exist, in scientific and forensic video post-processing, and
all the interfaces and devices required to do it are available.  All that's
needed is some programming!

Notice that the GUI component is virtually nil, since this could be
practically an appliance application (well, maybe some input is
needed to do color correction properly  ;-).  It's just a matter of
orchestrating existing algorithms--and processing trillions of
pixels worth of data.  This should easily absorb the next factor
of ten in processing power and memory capacity!  ;-)

I'd say that this little app would sell nicely for $95 or be bundled
with a high-end DVD record drive, and would make us hungry for
10GHz multiprocessors.  ;-)

It could even start a cottage industry of folks who would do the
conversion/enhancement for the people who don't want to invest
the time and money to do it themselves--with much better results
than what currently passes for conversions to DVD (but you
probably can't price it above $50/hour).

-michael

Check out amazing quality sound for 8-bit Apples on my
Home page:  http://members.aol.com/MJMahon/