Hi,

Alastair Robinson writes:

I've created a little patch against GIMP 1.2.5 to allow dithering of gradients, which significantly improves their appearance when printed.

If you're interested, you can find out more here: http://www.blackfiveservices.co.uk/dithergrad.shtml

How is that different from enabling "Adaptive Supersampling" in the blend tool options?

Sven

blackfive@fakenhamweb.co.uk (2003-07-17 at 1854.40 +0100):

I've created a little patch against GIMP 1.2.5 to allow dithering of gradients, which significantly improves their appearance when printed.

This is bug number 97777 (http://bugzilla.gnome.org/show_bug.cgi?id=97777).

It does not only affect printing, screen too. I had to use spread filter to fake this sometimes. You should check what is going on with CVS version (or 1.3.16), so it gets fixed there, which is where it should be.

GSR

Sven Neumann wrote:
> Alastair Robinson writes:
>>I've created a little patch against GIMP 1.2.5 to allow dithering of >>gradients, which significantly improves their appearance when printed. >>
>>If you're interested, you can find out more here: >>http://www.blackfiveservices.co.uk/dithergrad.shtml >
> How is that different from enabling "Adaptive Supersampling" > in the blend tool options?

It looks quite a lot different -- as far as I can tell from reading the patch, it does what it says on the bottle, that is, it (random-)dithers the truncated precision into the lowest bit(s) of the 8-bit precision that we support.

Conceptually I like this, and the gotchas are toggleable via the UI.

--Adam

Conceptually I like this, and the gotchas are toggleable via the UI.

I like the idea too. It should be checked in and turned on by default.

Austin

I tried the patch. It worked just fine, and IMHO should be used as a fix to the aforementioned bug.

I tried to aply adptive supersampling with maximum depth, to compare the effects with the ones from the patch: I had to kill out gimp after 20 minutes of 90% CPU use and no response.

Austin Donnelly wrote:

Conceptually I like this, and the gotchas are toggleable via the UI.

I like the idea too. It should be checked in and turned on by default.

Austin

gwidion@mpc.com.br (2003-07-18 at 1014.57 -0300):

I tried to aply adptive supersampling with maximum depth, to compare the effects with the ones from the patch: I had to kill out gimp after 20 minutes of 90% CPU use and no response.

To see supersampling at work, try doing a diagonal gradient moving the mouse one pixel in each axis, and using a custom gradient like the german one, with repeat mode triangle wave. Use two layers, one with supersampling and the other without, then toggle visibility. You will see how the supersampled version is a bit smoother, giving a orange brown looking wavy image, instead of sharply changing pixels, more like straight lines than a gradient. Work in zoom mode to build the gradients, then compare zoomed and non zoomed.

Quick conclusion: supersampling is for people working with quickly changing gradients, while dithering is for people working with slowly changing gradients. They are different things for different problems, and using the wrong one means wasting time.

GSR

Hi,

"Austin Donnelly" writes:

Conceptually I like this, and the gotchas are toggleable via the UI.

I like the idea too. It should be checked in and turned on by default.

The patch is against 1.2.5 so we cannot check it in. But we should do if we got a patch against current CVS or 1.3.16 attached to the bug-report.

Sven

Hi,

Alastair Robinson writes:

I'm working on it; I've got 1.3.16 installed and working, and it doesn't look as though the relevant code has changed too much (just been moved a bit).

It would be really nice to get a patch against 1.3.16.

BTW - are there likely to be any more releases in the 1.2 series, or is 1.3 getting 100% of the attention nowadays?

We might do another 1.2 release but I doubt that this will happen and it would surely be just be a bug-fix release with no new feature whatsoever. GIMP-1.3 is close to being released as 2.0 and support for 1.2 will be dropped then.

Sven

On 19-Jul-2003, Sven Neumann wrote:

We might do another 1.2 release but I doubt that this will happen and it would surely be just be a bug-fix release with no new feature whatsoever. GIMP-1.3 is close to being released as 2.0 and support for 1.2 will be dropped then.

Releasing the stable from 1.3 is a bad idea, and I think everyone knows it. I wrote an email a few minutes before this one, and I suggest you read it.

1.3 should become 1.4. It doesnt use gegl, and it isnt 2.0 material. Releasing 1.3 as 2.0 is possibly the worse thing any of you could ever do. You know those slashdot trolls who keep saying apple and bsd are dead? They'll say gimp is dead, and I will believe them.

On 19 July, 2003 - Patrick McFarland sent me these 1,1K bytes:

On 19-Jul-2003, Sven Neumann wrote:

We might do another 1.2 release but I doubt that this will happen and it would surely be just be a bug-fix release with no new feature whatsoever. GIMP-1.3 is close to being released as 2.0 and support for 1.2 will be dropped then.

Releasing the stable from 1.3 is a bad idea, and I think everyone knows it. I wrote an email a few minutes before this one, and I suggest you read it.

1.3 should become 1.4. It doesnt use gegl, and it isnt 2.0 material. Releasing 1.3 as 2.0 is possibly the worse thing any of you could ever do. You know those slashdot trolls who keep saying apple and bsd are dead? They'll say gimp is dead, and I will believe them.

A new version with heaps of changes is released.. that means gimp is dead? I don't follow.

/Tomas

Hi,

Patrick McFarland writes:

You know those slashdot trolls..

Yes I know those slashdot trolls. They troll and flame whatever you do. So what is your point in mentioning them? Do you believe they would not troll about 1.4? I stronly doubt that.

What is really hurting GIMP is the fact that we don't seem to be able to end a discussion and proceed with development.

Sven