Computer Graphics with Photoshop - Curves behavior

Author

Curves behavior - RGB mode

ronviers@gmail.com

2007-03-28, 4:15 am

Hi,

I know I am probably going to feel stupid after asking this but here
goes.

If I grab the curve and pull it up at 45 degrees the curve gets
curvier towards the top but stays flat toward the bottom.
If I pull it down at 45 degrees the curve gets curvier towards the
bottom and stays flat at the top.
Why is that? Why the asymmetru?

How is it that the individual R, G and B channel curves can be
modified individually without changing the shape of the composite RGB
channel?

Thanks,
Ron

Mike Russell

2007-03-28, 4:15 am

<ronviers@XXXXXXXXXX> wrote in message
news:1175058499.825289.266470@n59g2000hsh.googlegroups.com...
....
> I know I am probably going to feel stupid after asking this but here
> goes.

HI Ron,

I highly doubt that. :-)

> If I grab the curve and pull it up at 45 degrees the curve gets
> curvier towards the top but stays flat toward the bottom.
> If I pull it down at 45 degrees the curve gets curvier towards the
> bottom and stays flat at the top.
> Why is that? Why the asymmetry?

Good question, and now that you mention it, I have no idea. I have a math
genius friend that I refer these things to - matter of fact the math IRC
channel would be an excellent place for your question. If you're lucky,
landon will be there. Tell him you're a friend of mine and you will get the
answer before I do.

You can get a symetrical bend by using two points, and placing them
symetrically on each side of a 45 degree line from the upper left to lower
right. Splines were originally thin pieces of wood or metal used to design
curves for boats and aircraft, and I imagine "real" wooden splines would be
symetrical even with a single point.

> How is it that the individual R, G and B channel curves can be
> modified individually without changing the shape of the composite RGB
> channel?

Many people think that the composite RGB curve represents the output from
the other three curves. It takes them as input, and the RGB curve is
applied individually to the results of the R, G, and B curves. Using the
composite RGB curve is usually not recommended because the RGB values for a
given color land on a different part of the RGB curve. If the RGB curve is
a straight line, at an angle other than 45 degrees, colors change
saturation. If the RGB curve is actually curved, colors change their hue.

For some images the RGB curve is a good thing. For example underexposed
images generally benefit from using the RGB curve because they are short on
both saturation and contrast.

BTW - the free online curves class starts this Sunday - maybe it's time :-)
http://www.curvemeister.com/support/class/index.htm
--
Mike Russell
www.curvemeister.com/forum/

ronviers@gmail.com

2007-03-28, 7:14 pm

On Mar 28, 2:50 am, "Mike Russell" <RE-MOVEm...@Curvemeister.comRE-
MOVE> wrote:

Hi Mike,

>I have a math
> genius friend that I refer these things to - matter of fact the math IRC
> channel would be an excellent place for your question. If you're lucky,
> landon will be there. Tell him you're a friend of mine and you will get the
> answer before I do.

I learned enough math to get through DC circuits and Digital, plus
some propositional stuff, so any answer I have to join a sci.math
group to get I'm not ready for. I just thought it was odd. My guess is
that it is because of a trade off - without the asymmetry the line
might not move at all - that would make for a really boring dialogue.

> You can get a symetrical bend by using two points, and placing them
> symetrically on each side of a 45 degree line from the upper left to lower
> right. Splines were originally thin pieces of wood or metal used to design
> curves for boats and aircraft, and I imagine "real" wooden splines would be
> symetrical even with a single point.

That's odd, because this is what I think of when I think of splines.

http://picasaweb.google.com/ronvier...991952135260978

What do you call those things?
I have read spline mentioned before in the context of Illustrator but
it made no sense to me - no doubt because I was trying to relate them
to my preconceptions of what a spline is.

> Many people think that the composite RGB curve represents the output from
> the other three curves. It takes them as input, and the RGB curve is
> applied individually to the results of the R, G, and B curves.

I will have to think about that one.

> Using the
> composite RGB curve is usually not recommended because the RGB values for a
> given color land on a different part of the RGB curve. If the RGB curve is
> a straight line, at an angle other than 45 degrees, colors change
> saturation. If the RGB curve is actually curved, colors change their hue.

I have noticed, and I could be way off here, that it is better to
adjust the two channels that don't need changing rather than the one
that does. It seems that going about making a change directly, say to
the blue channel for example, causes to much color shift and it is
better to increase the blues bye decreasing the red and green channel.

> BTW - the free online curves class starts this Sunday - maybe it's time

I have decided to stop being phobic about curves so I would like to
sign up but are you sure my dial-up connection will not be disruptive?
I would be amenable to any ground rules you set up that would keep it
from becoming a distraction to the other students.

> --
> Mike Russellwww.curvemeister.com/forum/

Thanks,
Ron

Mike Russell

2007-03-28, 7:14 pm

<ronviers@XXXXXXXXXX> wrote in message
news:1175096206.774555.107240@e65g2000hsc.googlegroups.com...
[re splines]
> That's odd, because this is what I think of when I think of splines.
>
> http://picasaweb.google.com/ronvier...991952135260978
> What do you call those things?

Yep. Splines too. Cousins of the other splines - the gear shaft ones were
originally done as thin slats of wood or metal, embedded in the end of the
shaft.

....
[re adjusting channels in RGB]

> I have noticed, and I could be way off here, that it is better to
> adjust the two channels that don't need changing rather than the one
> that does. It seems that going about making a change directly, say to
> the blue channel for example, causes to much color shift and it is
> better to increase the blues bye decreasing the red and green channel.

This depends on whether the image is overall too light or too dark. If you
want to add blue, and make the image darker, then decrease the amount of red
and green (aka yellow).

>
> I have decided to stop being phobic about curves so I would like to
> sign up but are you sure my dial-up connection will not be disruptive?
> I would be amenable to any ground rules you set up that would keep it
> from becoming a distraction to the other students.

It's an offline online course - if you want to download the big versions of
the images, you'll need to be patient for some of the downloads, but there
is no inconvenience for others.
--
Mike Russell
www.curvemeister.com/forum/

ronviers@gmail.com

2007-03-28, 7:14 pm

On Mar 28, 4:18 pm, "Mike Russell" <RE-MOVEm...@Curvemeister.comRE-
MOVE> wrote:

> Yep. Splines too. Cousins of the other splines - the gear shaft ones were
> originally done as thin slats of wood or metal, embedded in the end of the
> shaft.
>

So spline does not refer to the cross sectional support members in the
hull of a wooden ship but the curved template used to manufacture
them, whereas with gear splines spline refers to a specific type of
gear not a template, and finally in graphics, spline refers to any
curve used as an input to a function.
I have always liked the word spline - spline and shroud - fun to say.

> This depends on whether the image is overall too light or too dark. If you
> want to add blue, and make the image darker, then decrease the amount of red
> and green (aka yellow).

Makes sense.

> It's an offline online course - if you want to download the big versions of
> the images, you'll need to be patient for some of the downloads, but there
> is no inconvenience for others.

I will try to be prepared.

> Mike Russellwww.curvemeister.com/forum/

J. Clarke

2007-03-31, 7:14 pm

ronviers@XXXXXXXXXX wrote:
> On Mar 28, 4:18 pm, "Mike Russell" <RE-MOVEm...@Curvemeister.comRE-
> MOVE> wrote:
>
>
> So spline does not refer to the cross sectional support members in the
> hull of a wooden ship but the curved template used to manufacture
> them, whereas with gear splines spline refers to a specific type of
> gear not a template, and finally in graphics, spline refers to any
> curve used as an input to a function.
> I have always liked the word spline - spline and shroud - fun to say.
>
>
> Makes sense.
>
>
> I will try to be prepared.

FWIW, in compass and t-square drafting splines (also called "battens")
were made of wood or plastic and they came in a variety of
forms--uniform, linear taper, curved taper, tapered from the center,
tapered from the ends, etc and different stiffnesses and lengths. The
were held in place with special weights with hooks on them called
"ducks". The range of curves one could get was quite remarkable, but
the expense of a good set of battens was astronomical.
[color=darkred]

--
--
--John
to email, dial "usenet" and validate
(was jclarke at eye bee em dot net)

ronviers@gmail.com

2007-03-31, 11:14 pm

On Mar 31, 5:55 pm, "J. Clarke" <jclarke.use...@cox.net> wrote:

> FWIW, in compass and t-square drafting splines (also called "battens")
> were made of wood or plastic and they came in a variety of
> forms--uniform, linear taper, curved taper, tapered from the center,
> tapered from the ends, etc and different stiffnesses and lengths. The
> were held in place with special weights with hooks on them called
> "ducks". The range of curves one could get was quite remarkable, but
> the expense of a good set of battens was astronomical.
>
>

> --John
> to email, dial "usenet" and validate
> (was jclarke at eye bee em dot net)

On Mar 31, 5:55 pm, "J. Clarke" <jclarke.use...@cox.net> wrote:

> FWIW, in compass and t-square drafting splines (also called "battens")
> were made of wood or plastic and they came in a variety of
> forms--uniform, linear taper, curved taper, tapered from the center,
> tapered from the ends, etc and different stiffnesses and lengths. The
> were held in place with special weights with hooks on them called
> "ducks". The range of curves one could get was quite remarkable, but
> the expense of a good set of battens was astronomical.
>
>

> --John
> to email, dial "usenet" and validate
> (was jclarke at eye bee em dot net)

http://www.edsonmarine.com/newprodu...ine_weight.html

Hi John,
Expensive and apparently rare too, a quick search on ebay (titles and
descriptions) did not return a single set - I just wanted to look.
I did find these on google images:

http://www.edsonmarine.com/newprodu...ine_weight.html

pretty cool, talking about pinning a curve.

It occurs to me that the grooves inside a synchronous drive belt could
be thought of as splines too. That may not be technically correct but
they do indirectly define the shape of the curve.

Thanks for the added info,
Ron