Re: [Flashcoders] Problems with Gradient Banding

2008-01-18 Thread Jon Bradley 


On Jan 17, 2008, at 5:34 PM, Ed McManus wrote:

Does anyone know what causes this? It's as if for large gradients  
flash
doesn't perform enough steps making the transition from the start  
to end
colors, so on a large scale the color difference between steps is  
clearly

visible.


This is just me thinking aloud but, that's a problem with 32-bit  
color space (only 8 bits per channel, RGBA). Flash does not support  
anything greater than 32-bit color space (no 10 or 12 bit), so  
gradient banding will always be there.


Do the same in Photoshop. Choose two colors that are relatively close  
in value and fill a document the same size with any gradient type -  
banding will occur if you are in an 8-bit color space.


The only way to partially alleviate this problem is by adding a  
slight amount of noise (at least in Photoshop). It can do wonders for  
larger gradients. In Flash, I don't think this would work out too  
well (at least not in a performance-wary manner).


cheers,

jon
___
Flashcoders mailing list
Flashcoders@chattyfig.figleaf.com
http://chattyfig.figleaf.com/mailman/listinfo/flashcoders

[Flashcoders] Problems with Gradient Banding

2008-01-17 Thread Ed McManus 
Hi Guys,

After experimenting for a bit with CS3's gradients, I haven't found a way to
reproduce high-quality gradients on a large scale (matrix box size >
~1000px). I keep running into issues of gradient banding (if you've ever
seen this problem, the term should make sense) when using large radial
gradients. And if you adjust the gradient's alpha the problem only seems to
get worse.

Does anyone know what causes this? It's as if for large gradients flash
doesn't perform enough steps making the transition from the start to end
colors, so on a large scale the color difference between steps is clearly
visible.

Any tips for getting around this issue would be greatly appreciated!

Best,

Ed
___
Flashcoders mailing list
Flashcoders@chattyfig.figleaf.com
http://chattyfig.figleaf.com/mailman/listinfo/flashcoders