Random SplitStream Thoughts...

[Roger Merchberger]

Well, folks -- I've been perusing the SplitStream code, and I've come
to a few interesting conclusions:

1) I think that splitstream can be made to work faster by allocating
a lot more memory & increasing the file input/output buffers - right
now it only allocates memory for 1 chunk at a time, which is
128Kbytes. So, the program grabs 128K bytes in, splits & spits it out
2 seperate files (audio & video) - that seems like it'd cause a lot
of hard drive thrashing - but I won't know for sure until I change
the program to allocate more memory for the streams before I/O
happens... By my (admittedly *very* rough) calculations, there's
approx. 12,500 chunks per hour of High-quality recorded Tivo video. I
was figuring on buffering 100 chunks at a time, which would cause the
computer to utilize 12.8 Meg of RAM during execution - methinks
anyone doing video extraction should be able to afford that much
memory allocation, but I'll make sure the program is kind & gentle to
those that might be "RAM challenged." (although, if the program goes
to swap, that might actually hamper performance - but I dunno if a
command-line proggie can access swap...)

2) From what I can read in the code, there *are* bugs in SplitStream.

2a) The first is very minor: the 128K memory allocated for the stream
is never freed in the program, as far as I can see. Granted, as soon
as the program exits, it'll be returned, but if the program abends
the memory is lost until reboot. Admittedly, losing 128K is peanuts -
but I'm glad I noticed that, because if I'm not careful, after my
changes are done and the program misbehaves, losing 12M is a *bad
thing*.

2b) This one is fairly minor: When the memory is (asked to be)
allocated, it's never checked to see if the memory allocation request
actually went thru. Again, with the megamemory of today's machines,
it's "a safe bet" that 128K is available, but for some reason it
didn't go, it was never checked - and the program would run amok (or
at least die horribly). Again, I must be more careful to make sure
the memory is allocated, or else the expanded program could do
"really bad things."

I'm starting to learn a lot more about C again -- hopefully I can
finally learn it for good & keep a good chunk of it in my brain this
time... ;-)

Well, I've been plinking with allocating memory & stuff, and I've
gotten a better handle on it now, I think...

I'll be testing changes to the new zss code in a few days.

Laterz,
Roger "Merch" Merchberger

__________________________________________________
Do You Yahoo!?
Send FREE Valentine eCards with Yahoo! Greetings!
http://greetings.yahoo.com