Re: Animated Gifs
rja.ca...@excite.com
> <rja.ca...@excite.com> wrote in message
> news:1138226138.3...@z14g2000cwz.googlegroups.com...
> > It seems that Internet Explorer now(?) imposes a minimum interval in
> > GIF animation of 10 centiseconds.
>
> Some kind of anti-epilepsy feature, perhaps?
Not mentioned as such in the Microsoft knowledge base, that I can find.
And while I'm no expert at all, I think that speed would still set off
an epileptic who was sensitive to it - there are various triggers.
I'll cross-post this to the Opera group, touch wood.
Alister wrote:
> I would be grateful if as many as possible could look at:
>
> http://homepages.tesco.net/~acg.gcs/tmp/test.html
>
> with as many different browsers as you have.
I think the question is: is the following true, and if so then why?
1. That Microsoft Internet Explorer for Windows imposed (but not in its
first version) a speed limit on animated CompuServe GIF images: the
time interval between one image frame and the next is specified in x
1/100 seconds, but it is always taken to be at least 10/100 seconds,
i.e. no less than 0.1.
2. That some other Web browsers and other tools follow this rule. Some
do not.
3. That Opera version 5 did not(?) Opera 6 did. Opera 7 didn't.
Opera 8.51 does. We're waiting for 9...
At this point we seem to be stuck with the behaviour anyway in a MS
Active Desktop setup, we're basically curious (yes?)...
Perhaps this is also what CompuServe did, which therefore would be
correct(?)
Gates's Law also was mentioned ("The speed of software halves every 18
months") but surely not Opera??? ;-)
Jernej Simončič
> 3. That Opera version 5 did not(?) Opera 6 did. Opera 7 didn't.
> Opera 8.51 does. We're waiting for 9...
It seems to me that Opera 9tp1 plays the animations even slower than IE6...
--
begin .sig
< Jernej Simončič ><>◊<>< jernej simoncic at isg si >
end
Richard Grevers
> Paul Harman wrote:
>> <rja.ca...@excite.com> wrote in message
>> news:1138226138.3...@z14g2000cwz.googlegroups.com...
>> > It seems that Internet Explorer now(?) imposes a minimum interval in
>> > GIF animation of 10 centiseconds.
>>
>> Some kind of anti-epilepsy feature, perhaps?
>
> Not mentioned as such in the Microsoft knowledge base, that I can find.
> And while I'm no expert at all, I think that speed would still set off
> an epileptic who was sensitive to it - there are various triggers.
>
> I'll cross-post this to the Opera group, touch wood.
>
I studied this in some depth back then Robert, as it affected some 24 fps
animations I made from 16mm film. I also posted bugs to Opera each time
they broke it :-)
1) Microsoft configured IE so that any frame with a zero interval defined
rendered 0.1 seconds after the previous frame. As far as I can tell,
animations with nonzero intervals less than 0.1 second are rendered as
designed - my 0.4s gifs certainly did. This might have been because slow
CPU speeds meant that intervals much shorter than 0.1 seconds weren't
actually possible in the early 1990's
2) Working on the "If it works in MSIE it must be ok" principle, one or
more gif generating applications output gifs with a default (when the
author didn't override it) interval of 0. The apps probably used some
Windows components for rendering so no one noticed.
3) Opera 3 built its own gif rendering widget and followed the spec
exactly (did you know that Opera supported the text functionality of
gif89a?). Hence Opera got a ton of bug reports that some gifs rendered
much too fast. (An interval of zero means render the next frame as fast as
you can, and I have sometimes used a zero).
4) In a subsequent version, Opera "fixed" the bug by setting a minimum
refresh rate of 0.1 seconds. This made the previously "broken" gifs render
the same as IE, but broke mine by making them play below half speed. I
grumbled.
5) Opera reverted the previous "fix"
6) At some point (O6) they actually managed to get it right. (zero
defaults to 0.1, 0.01 - 0.09 rendered faithfully. Since then I've lost
interest and haven't kept track.
However, some large gifs do seem to run very slow in Opera (cf irfanview).
I'd guess that someone stopped a memory leak in such a way that O needs
too much CPU.
since I'm using news.opera.com I'll leave you to relay this back to
alt.fan.pratchett
--
Using M2, Opera's revolutionary e-mail client: http://www.opera.com/m2/