Actions

[bruce1951]

Just a short question for the technical gurus.
I'm using Filter Forge as a plugin. My version, (because I couldn't justify the extra expense), only allows 8 bit and a max of 3000 pixels in any one direction.
Due to the nature of my work 16 bit isn't critical but the 3000 is. I use a min of 6000 x 4000 images. 99% of the time the 6000 is the horizontal side.

So to get around this I created an action.
I open a file, 6000 x 4000. Duplicate and save as a PL file.
Select the top layer.
Start the action, Layer/Convert layer Type/ then from 16 bit to 8 bit. Then the action goes to Layer/Scale Layer and 3000px wide.
Easy. Then once finished in Filter Forge another action reverses the action. IE back to 16 bit and 6000px wide.

So my technical question is - is it important to image/file quality if resizing or bit depth is performed first or second?

bruce

[Hoogo]

I would resize first.
Going from 16 to 8 bit is like rounding or cutting digits after the decimal point. Usually you get better more accurate results when you first do all the calculations, and then round the result.

Maybe your action could cut the picture into 4 smaller ones?

[bruce1951]

Thanks Hoogo. Cutting the picture into 4 wouldn't work due to the nature of the filters.

bruce

[russellcottrell]

Once you reduce the image to 8 bit, you can’t truly go back to 16 bit; the information is lost. It is like a staircase that goes from small steps, to large steps, then to large steps that are drawn with a lot of small line segments; you don’t get the small steps back.

It sounds like what you are trying to do justifies the extra expense . . . .

[bruce1951]

I understand the loss of quality/information. I've never been able to unscramble an egg either!!! I'm simply trying to minimise the amount that is lost.
My work involves a lot of 'artistic' affects. Such as water colour. By its very nature detail is lost. But the less I lose the better.

bruce