home
***
CD-ROM
|
disk
|
FTP
|
other
***
search
/
Univers Mac Interactif 43
/
Univers Mac Interactif - Issue 43.iso
/
univers mac N°43
/
Reseaux
/
internet
/
internet utilitaires
/
Sparkle230 folder
/
(Docs)
/
Readme Sparkle 2.2
< prev
next >
Wrap
Text File
|
1994-11-14
|
4KB
|
70 lines
The major new features in this release are:
* PowerMac native (and very fast). A 6100/60 gives three times the
playback rate of my Quadra 610.
* Rewritten MPEG encoder to use way less memory.
* All large memory allocation is done through temp mem. Basically this
means you shouldn't ever need to increase the size of Sparkle's
partition but you can open and save files limited only by all the
memory on your mac.
* Foundations have been laid for AppleScripting. You probably can't do
any useful scripting with this release but that will change soon.
* User interface has been tweaked a little. I hope you like the changes.
* The most common problem with the 2.1 versions was that the preferences
file would become corrupted leading to random errors. The prefs file
handling has been changed in this version to make that much less likely.
If this release gives you problems, destroy the Sparkle prefs file then
try again.
* Many many bug fixes.
-----------------
The following are know small problems that I'll try to fix soon, probably
before release:
1) Scanning the directory to look for other PICTs when a PICT is opened
can take a few seconds. I might write my own code to do this rather than
using Apple's code (which scans the entire disk), or maybe I'll just add
some progress bar feedback.
2) The first time a window is opened on an 8bit color screen, or if the
color table on an 8bit color screen is altered, Sparkle takes about two
seconds (on my Q610) to calculate new color tables. I'll try to add some
sort of feedback that this is happening.
The following are smallish things that I'll add fairly soon, in between
adding major items.
* Saving one frame, or an entire movie/MPEG to PICTs.
* Support for previews in the OpenFile dialog.
* Better quality 8bit dithering, both a better algorithm (though the
present one, invented by me, is insanely fast and I don't want to lose
that speed). Also selection of an optimal palette for 8bit MPEG playing.
* I have special purpose code for playing back at 8bit color and grey
scale, 16bit color and 24bit color. I do NOT have special code for
handling 4bit greyscale, or for handling expanded images.
I might add 4bit greyscale if enough people ask.
I will certainly add special code to handle double sized, and maybe
generically expanded images. This will give much better playback
quality and speed than what we have now.
* I still would not recommend Sparkle as a default QT movie viewer.
There are a few optimizations for QT movie playback speed that I don't
yet implement, and I suspect Sparkle might die on pathological QT movies
(like those with two video track, or gaps, or no video). For this
reason, the QT movie files created by Sparkle have Simple Player as
their creator. Once I go through these QT pathologies and fix Sparkle to
cope with them, I'll switch the creator of these QT movies to Sparkle.
* There is a tacit assumption that all PICTs in a sequence are of the same
size. If your set of PICTs is of random size, they'll all be scaled to
the size of the first PICT. I don't really plan on changing this because
I I don't want to get involved with those sorts of issues. So this won't
changes unless I get some very good reason why it should.
* Any other small things you suggest.
After that I'll tackle sound. This is trickier than it may appear at
first so it'll probably appear in stages.
1) Better MPEG time-synching. I have the algorithm for this on paper
and just have to code it. As a nice by-product, this should give better
(less jerky) realtime playback of IPB frame MPEGs.
2) Actually playing audio, initially just in snd and MACE compressed form.
3) Parsing the system layer of MPEGs with audio (but tossing the audio)
4) Finally actually decoding the demuxed audio.
Obviously I'll work on these fast as I can, but I can't even guess dates
at this stage.