... If there was an all in one solution that worked better than what I'm doing currently I'd certainly be willing to pay a few bucks for it. We all see people asking questions about GIF creation in here a lot, I doubt i'd be the only one. It can be hard to get people to pay for things when there are free alternatives though, even if the paid product is vastly superior. So I don't know how worth your time it would be overall.
I mean, I think we discussed this before, but I believe that the generally poor output quality of the freeware gif encoders, is because Adobe have a ton of software patents related to perceptible quality options like dithering and quantising that they basically block anyone being able to do much in the way of improving.
I don't have full fat photoshop, but as I understand it, it offers a vastly superior quality to filesize compromise on animated gifs than any of the standalone options, and I believe thats because they all use the open source "as good as we can legally get it" encoding cores.
Its like Simplex noise is way better than Perlin noise, but is generally avoided as Simplex noise is patented and Perlin noise isn't.
I kind of feel this way as well, especially since I feel there might be questions of reverse engineering and legality if something came close to Photoshop / whatever, and to avoid all that I just stick with free and known safe stuff, plus videos if I need good quality.
I can understand the concern about any patents whatsoever. That's something
you can run into every time you flip a bit developing new stuff. But that's
something I have to deal with and not the user. So it's clear that I need to
check against some patents in the end and my come to an agreement if there is
any valid claim. Considering Photoshop, I, for a reason, haven't looked at any
of their patents, yet, (some being posted in here) to save myself from getting
influences by them, because I follow a native interest and don't want to draw
straight from any competitor. So I can safely say (to myself) that I haven't
used anything from Photoshop, which doesn't save me from violating any of
their patents in the end, but in that case we either draw from the same source
like Robert Ulichney, Digital Halftoning (1987), which is my main source, and/
or we have had the same ideas, which would be quite interesting if true. I did
a couple of stuff on my own. For example, I extended the median-cut algorithm
(free) to account for the highlights in a scene. I extended the octree
quantizer algorithm (free) in having implemented a better way to combine the
sub-nodes leading to some better shades. Etc. I would be astonished if exactly
such extenstions are patented. lol Could be the case, sure, but that's rather
unlikely. I mean Photoshop hasn't any implementation to account for the
highlights of your image. I made a test a year ago in here where I showed what
Photoshop does to your highlights and how my algorithm can deal with them.
Doesn't prove anything, indeed, but rather unlikely nevertheless.
I wouldn't even say Photoshop has the best algorithms within this regard. I
think the printing industry will hold the best of them considering doing good
photo copies. And I've heard the best of them aren't even patented to not give
any advantage to the competitor working around any such patents. Some of them
are kept a secret like the ingediences of Chanel N°5.
However, patents are just one thing. It's also about relibility and good
implementation of the status quo of what is known and free since ages. I mean,
lets have a look at what tamaster92 said last page. He was dissatisfied with
gifcam because of the green spots. He then wrote; "Found a nice little program
called licecap that does exactly what i need
". My answer follows closely
thereafter. LICEcap? It doesn't even allow you to use any dithering! It's a
pretty mess when it comes to shades, which hasn't anything to do to refrain
from any patented method. So I had a look into the code of LICEcap, no
dithering. Why? Patents? Surely not! There are a million of free patterns and
methods one can use here. Another thing I recognized within the code is that
LICEcap maps from 15 bits (555) into the octree computed color map. To do so
they simply shift 3 bits out of your 8 bits per color channel. These 3 bits
are gone for nothing. There are a couple of other things here and there, but
what I wanna say is that it is not the case that the quality of the free gif
recorders is the "best possible" before violating any patents, as LordRaptor
may have anticipated. The reality is that LICEcap is used, see tamaster92 gif
from last page which lacks in visual quality a lot. The huge bands in the gray
areas are nothing but distracting. It's here where I say this situation can be
improved, to make it more sound, more reliable. More control over the quantizer
(usually not found in any gif recorder) would also be nice to have to make for
better quality vs. size tradeoffs.
@missle, APNG may finally be
on the cusp of taking over from animated GIFs.
So for better quality but "just works" APNG may soon be viable. While for greatly reduced file size embedded video is going to beat both, leaving GIF in an odd place.
Embedded video is a different league altogether. APNG? Well, nice! Hope it
takes off this time. One more format for the new recorder. Why not. I can't
see any issues here considering APNG, because what I wrote about gif above
can be applied to APNG as well, for, 24-bit is nice, but file size will
increase. So there might be the need to cut down on the bit depth perhaps down
to index mode (palettes) to reduce file size considerably. And since my work
is mostly quantizer based it can be applied to APNG as well. And even for APNG,
it's good to have some control of the quantizer to better adapt the visual
quality with respect to file size and stuff.