Slashdot Mirror
GIMP 2.4 Released
Enselic writes "After almost three years since the release of GIMP 2.2, the GIMP developers have just announced the release of GIMP 2.4. The release notes speak of scalable bitmap brushes, redesigned rectangle/ellipse selection tools, redesigned crop tool, a new foreground selection tool, a new align tool, reorganized menu layouts, improved zoomed in/zoomed out image display quality, improved printing and color management support and a new perspective clone tool."
CMYK colors and other functionality that keeps it from being able to replace Photoshop completely? Not to understate all the effort that has been put into it, but something like that does seem pretty basic for three years of development.
2.3 was the devel branch leading up to 2.4
ART on dA
I hope they moved the gui closer to that of Paintshop. I can't tell you how many times I've been unable to edit an image for one reason or another, or the expected behavior is what happens. I know a lot of people love GIMP and its scripting abilities, but seriously, when they're trying to enter the market dominated by a few programs with that same gui and behavior, they should replicate it.
The stable GIMP releases have even numbers. The last stable release before 2.4.x was GIMP 2.2.x, starting with 2.2.0 released in December 2004. So that was almost three years ago. There were several bug-fix releases in the meantime, up to 2.2.17.
The unstable 2.3.x releases ended with the last versions becoming release candidates for 2.4.
-Raphaël
Maybe 5 of the posts will have something actually illuminating. The rest of them will be GIMP and Photoshop fanbois going at each other. Let me save everybody the trouble.
GIMP has an unprofessional name! Waaaaaaaaaaah!
GIMP only does 8-bit color! Waaaaaaaaaaah!
GIMP isn't UI identical to PhotoShop on every menu 3 levels deep! Waaaaaaaaaaaah!
GIMP manages windows sucky! Waaaaaaaaaaaaaaah!
Does not! Does too! Does not!.................
Adjustment layers. If you're not familiar with the adjustment layers that Photoshop 5 software introduced, they're layers that copy pixels from layers below them and run a filter on them, and they automatically update when the layers below them are changed. It's been said that GIMP is one of the best Photoshop 3/4 clones around.
A lot of the key algorithms, particularly for color space conversion, are patented. Guess who holds a bunch of those patents?
Despite your lame attempt at humor, you make a very good point. Photoshop is a tool, and a very versatile one at that, which is used by professionals to get the job done right. It's expensive and complicated, and for good reason.
The GIMP, on the other hand, is a comparatively simple tool, though still very useful and quite versatile in its own right. It is what us amateurs use because the pro tools are overkill and/or too expensive. It also happens to be free, in more than one sense of the word, which makes it ideal for its target audience. For example, I do web graphics sometimes. Why in the world would I spend close to US$500 for something that is rarely used and would be overkill to boot? I'd rather use my free image program with more tools in its toolkit than I would ever need for that task.
This is why I will never understand the PS vs. GIMP debate. GIMP will never be a Photoshop killer because there is no need for a Photoshop killer. Those who need the power of Photoshop will buy it (or steal it), those who don't will use GIMP or another simple tool.
GIMP was NEVER intended to replace, duplicate or mimic photoshop. Neither was it created to draw users from photoshop. Unfamiliar doesn't mean bad or uncomfortable. They go their own way. Some like it, some don't. You are free to use gimpshop if you like to. I really see no points in this interface discussion.
Respecting your software freedom to share and modify the program has never been an option with Photoshop, no matter how much you pay. Freedom has always been a part of the GIMP.
Why stress software freedom? I want the social solidarity that you only get in freedom; I want to be independent from masters and make sure my computer only obeys me. I'd rather have less functional or powerful free software than a more powerful or reliable proprietary program because I can hire people to improve the free program or I can ask the community to help me improve the free program. I can't free Photoshop. The catch here is that most people haven't been taught to value their software freedom, so they don't know to look for it and they haven't been taught to think of the consequences when their freedom is absent. I aim to change this by teaching people to value freedom for its own sake. I hope you will too.
Digital Citizen
The tool in question is improving with each iteration. Eventually it will get to a level when it is usable by professional people, as it is it is good enough for many people.
We had *nothing* 10 years ago.
Some people simply don't understand the dynamics of open software and how the cumulative improvements are not lost and will eventually get you where you need to be.
IANAL but write like a drunk one.
You seem to think CMYK is somehow 32 bit (8 bits of each?) and that because 8-bit rgb has 24 bits it can't represent it, but because 16-bit rgb has 48 bits it can.
This is wrong. CMYK has FOUR dimensions. It is completely impossible to represent it in a 3 dimensional space. You claim is like saying that if I put finer graduations on a ruler, it can suddenly measure 2 dimensions rather than one!
The converters you talk about (and incidentally are in Gimp already, and in printer drivers when you send them rgb colors) map the 3-D space into the 4-D space. But they cannot fill the 4-D space, any more than you could fill a room with a piece of paper (while keeping the paper's shape a non-fractal). Thus there are CMYK colors that are not output. This has NOTHING to do with color resolution. No useful RGB->CMYK converter will produce both CMY=0,K=1 and CMY=1,K=1 output. Even if the CMYK device was 1 bit per ink and thus only capable of printing 16 different colors, you could not represent all those 16 possibilities with 24, or even 48, or 96 bits, or an infinite number of bits of rgb!
In reality the highest quality CMYK printing devices available have much less than 8 bit resolution in how much ink they lay down (once you take into account errors in ink delivery and spread). The resolution is so low that the volume represented by the RGB->CMYK conversion is over-sampled by many times when the source is 8 bit rgb. So actually 16 bits does not help one tiny bit in the area you are asking for.
The reason for more than 8 bits is for processing in the digital realm. For instance if your picture is 1/4 as bright as you want it, and you multiply by 4, then you lose two bits of resolution (as the bottom 2 will be zero). If your screen shows 8 bits and the original was 8 bits, you have effectively reduced your screen to 6 bits. If the original was 16 bits (and your screen was showing the top 8 bits) then after the multiply your screen is still showing an 8 bit image (the top 8 bits of the remaining 14). (that is not real accurate, a correct program with knowledge of sRGB would do something more complex and you would lose more than 2 bits at the bright end, less at the dark end).
Also more than 8 bits should absolutely use 16 bit half float data. 16 bit integers is a total waste of effort. Float data has the advantage that it is not clamped (this eliminates gamut limitations), and that a vastly larger range of useful data. Even 16 bit data would start to lose resolution on an 8 bit screen if multiplied by more than 256 (actually somewhat larger if sRGB is correctly followed). But 16-bit float would allow a multplication by 65540 or so before there would be loss. The only reason for 16-bit integers was that older computers could not do float fast enough, but this is not a problem now, modern graphics cards even take half-float data directly.