By Rinaldi wrote:
is it possible to render an image with something like Renderman or Mantra and could one get benefits out of that?

I believe this will become increasingly feasible.

By Rinaldi wrote:
Hmmm...it got me thinking because I know PS can do a lot as long as the user knows how to do it. But packages like Combustion and Inferno are dedicated to add special effects to the overall project, such as rainfall or a stormy look. Choosing the particals option and just type in amounts or play with their sliders to have light/heavy/etc rain droplets, angle, stuff like that.

So I was thinking, instead of doing that in PS from scratch (the droplets), I load the images on a regular plane with the exact same size as the image, and add (a) layer(s) and the special effects.

Do you think this is going to be more common or will artists/wizards stay within PS?

You could use After Effects. That starts to be almost 'Photoshop with 3D, motion and particles'

SRB Photo wrote:
If the Apple UI team had been involved with Blender's development, it would be a very different story.

they could still build a wrapper on top of POV's engine

-krphoto- wrote:
AE has a true additive blending mode, which is sometimes useful for things that give off light.

What doesn't?

-krphoto- wrote:
Photoshop has Linear Dodge, which says "Add" in parenthesis next to it, because it's kind of a faux additive mode.

http://www.neilblevins.com/cg_education … toshop.htm

What is faux about it?

It's a straight arithmetic add.

NothingIsRealButTheGirl wrote:

What is faux about it?

It's a straight arithmetic add.

I use opacity all the time on layers, and always wondered why actual Additive blending looks different than Linear Dodge. The article explains why.

-krphoto- wrote:
I use opacity all the time on layers, and always wondered why actual Additive blending looks different than Linear Dodge. The article explains why.

A white circle with opacity 50% is not 'a 50% gray.' It is 50% of white plus 50% of whatever it is on top of.

A white circle with opacity 12% is not 'a 12% gray.' It is 12% of white plus 88% of whatever it is on top of.

NothingIsRealButTheGirl wrote:
A white circle with opacity 50% is not 'a 50% gray.' It is 50% of white plus 50% of whatever it is on top of.

A white circle with opacity 12% is not 'a 12% gray.' It is 12% of white plus 88% of whatever it is on top of.

It IS if it's on a black background, which is what Blevins is using for his illustrations.

I'm not trying to battle, but I must understand!

It seems to me that on a black background, if the layer mode is truly additive, a  white circle with Layer Opacity at 50% should turn completely white (rgb 255) where it overlaps a circle filled with 50% gray.

But it doesn't. In Photoshop's Linear Dodge (Add) mode, the overlapped area is 75% (rgb 192 is what I got).

There shouldn't be any complicated math, 50+50= 100.

In After Effects the overlapping area IS solid white, as expected. I just tested this.

Regardless of the math, and whether or not Neil Blevins is correct, Linear Dodge (Add) does not produce the same effect as Add does.

(apologies for straying OT)

-krphoto- wrote:

It IS if it's on a black background, which is what Blevins is using for his illustrations.

I'm not trying to battle, but I must understand!

It seems to me that on a black background, if the layer mode is truly additive, a  white circle with Layer Opacity at 50% should turn completely white (rgb 255) where it overlaps a circle filled with 50% gray.

But it doesn't. In Photoshop's Linear Dodge (Add) mode, the overlapped area is 75% (rgb 192 is what I got).

There shouldn't be any complicated math, 50+50= 100.

In After Effects the overlapping area IS solid white, as expected. I just tested this.

Regardless of the math, and whether or not Neil Blevins is correct, Linear Dodge (Add) does not produce the same effect as Add does.

(apologies for straying OT)

Have you tried his 'experiments'?  I can't replicate a single of his asserted problems - which leads me to believe there's something wrong with his setup or technique (he doesn't really go into it much) - not with PS's blend modes (which, btw, work ideally in this case).

By Rinaldi wrote:
Hmmm...it got me thinking because I know PS can do a lot as long as the user knows how to do it. But packages like Combustion and Inferno are dedicated to add special effects to the overall project, such as rainfall or a stormy look. Choosing the particals option and just type in amounts or play with their sliders to have light/heavy/etc rain droplets, angle, stuff like that.

So I was thinking, instead of doing that in PS from scratch (the droplets), I load the images on a regular plane with the exact same size as the image, and add (a) layer(s) and the special effects.

Do you think this is going to be more common or will artists/wizards stay within PS?

Yes, I do this increasingly more so every day.

I take an image from photoshop, load it into After Effects for simple things like transforming parts of an image now. Because After Effect's transform system is simply better.

It will become more common after a few tutorials are posted. People only use Photoshop because they were told to and wait for features of better dedicated programs to be included within Photoshop a few years later.

R A V E N D R I V E wrote:
It will become more common after a few tutorials are posted. People only use Photoshop because they were told to and wait for features of better dedicated programs to be included within Photoshop a few years later.

Some of us stick to PS because we can't justify the cost of AE .

SRB Photo wrote:

Some of us stick to PS because we can't justify the cost of AE .

better start looking into the bundles and save a pence

R A V E N D R I V E wrote:

better start looking into the bundles and save a pence

And buy more programs that I can't use?  No thanks.

-krphoto- wrote:
It seems to me that on a black background, if the layer mode is truly additive, a  white circle with Layer Opacity at 50% should turn completely white (rgb 255) where it overlaps a circle filled with 50% gray.

But it doesn't. In Photoshop's Linear Dodge (Add) mode, the overlapped area is 75% (rgb 192 is what I got).

There shouldn't be any complicated math, 50+50= 100.

The circle is not over black. It is over the other circle.

It's like this:

Circle 1 is white, with 50% opacity over black. Result = 50% of white plus 50% of black = a 50% gray

Circle 2 is white, with 50% opacity over a 50% gray circle. Result 50% of white plus 50% of 50% gray = a 75% gray

So a 75% gray is what you get when you put the circles together in 'Normal' blend mode. Now try all the other blend modes. Most of them appear to be ignored because Photoshop says, 'look, do you want me to Linear Dodge them? Or do you want me to opacity blend them? I can't do both. The opacity blend alone already involves a weighted summation with the background.)

SRB Photo wrote:
Have you tried his 'experiments'?  I can't replicate a single of his asserted problems - which leads me to believe there's something wrong with his setup or technique (he doesn't really go into it much) - not with PS's blend modes (which, btw, work ideally in this case).

Makes me wonder if we are all talking about the same thing if you are getting different results.

NothingIsRealButTheGirl wrote:

Makes me wonder if we are all talking about the same thing if you are getting different results.

Well at least I am happy to see all you wizards know what you all talking about, because to me it sounds like magic lol

The thread is still going and if it spawns other threads, hey I fully support that. I might even try to do what it is I have read so far, when I figured out what it is

SRB Photo wrote:
Have you tried his 'experiments'?

I didn't deem it necessary because his statements seemed consistent with my own experience.

NothingIsRealButTheGirl wrote:
The circle is not over black. It is over the other circle.

It's like this:

Circle 1 is white, with 50% opacity over black. Result = 50% of white plus 50% of black = a 50% gray

Circle 2 is white, with 50% opacity over a 50% gray circle. Result 50% of white plus 50% of 50% gray = a 75% gray

So a 75% gray is what you get when you put the circles together in 'Normal' blend mode. Now try all the other blend modes. Most of them appear to be ignored because Photoshop says, 'look, do you want me to Linear Dodge them? Or do you want me to opacity blend them? I can't do both. The opacity blend alone already involves a weighted summation with the background.)

Your explanation of what Photoshop is doing appears to be correct.  However, that's not the way other programs treat additive blending.

After Effects has Linear Dodge which produces the same results as PS's. But AE also has Add which is a separate mode altogether. Why both modes?  Because they're not the same thing.

In your example, the fact that we're "seeing through" the white layer 50% doesn't mean the underlying layer should be factored in by only 50%. "Real" Add mode simply adds 50% of white to 100% of whatever is under it, whether it's black, any shade of gray, or any other color.

-krphoto- wrote:
Your explanation of what Photoshop is doing appears to be correct.  However, that's not the way other programs treat additive blending.

Actually, it's exactly the same way that other applications handle Additive blending.  It's just a different way of handling opacity.  Would you have them break ~18 years of compatibility for the sake of doing it the way everyone else is doing it now?

-krphoto- wrote:
"Real" Add mode simply adds 50% of white to 100% of whatever is under it, whether it's black, any shade of gray, or any other color.

"Real" add mode just adds things.

When you have an upper layer at 50% you want 50% "of it" which is really a way of saying you want Photoshop to assume and blend with an invisible unmentioned layer filled with black.

Personally I'm not a fan of software assuming invisible unmentioned things, but as long as I am aware of the behavior I guess I can tolerate it.

But it's too late to change now. Plus there are adequate workarounds.

SRB Photo wrote:
Actually, it's exactly the same way that other applications handle Additive blending.  It's just a different way of handling opacity.  Would you have them break ~18 years of compatibility for the sake of doing it the way everyone else is doing it now?

Ok, technicality police. I thought it was clear from the point of the  discussion, which is about additive mode when transparency is involved. I should have said, "That is not the way other applications handle additive blending and opacity."

There's no need to break any backwards compatibility; I definitely wouldn't ask that. What I would prefer is that PS had the same Add mode that AE has. Then if we want it to behave like Linear Dodge we have that option and if we want it to behave like Add it's there, too. But according to this article from AdobePress.com it's a legal issue. [edit] Weird since they're both Adobe products.

The same article describes Add mode exactly as I described: "Add mode is every bit as simple as it sounds; the formula is

newPixel = A + B".

But it goes on to say the only difference between Add and Linear Dodge in PS is the name. Yet there is definitely a difference in their result if Opacity is used.

-krphoto- wrote:
There's no need to break any backwards compatibility; I definitely wouldn't ask that. What I would prefer is that PS had the same Add mode that AE has.

Adobe breaks compatibility all the time across their Creative Suite applications

try unchecking that "maximize compatibility" next time you save a PSD, Lightroom might not like it anymore

or the version control in Illustrator

18 years of compatibility my ass

-krphoto- wrote:
Ok, technicality police. I thought it was clear from the point of the  discussion, which is about additive mode when transparency is involved. I should have said, "That is not the way other applications handle additive blending and opacity."

The point of the discussion is whether mixed art has a place in mainstream retouching.

The point of your conversation is about misconstruing a blend mode for its blend method.  If you understand the difference, much less want to make an impact and see change, you need to put things in terms the engineers will understand.  Which means discussing it as the latter and not the former - the engineers will dismiss you out of hand otherwise.

I'm not defending the reality, only calling it for what it is.

-krphoto- wrote:
There's no need to break any backwards compatibility; I definitely wouldn't ask that. What I would prefer is that PS had the same Add mode that AE has. Then if we want it to behave like Linear Dodge we have that option and if we want it to behave like Add it's there, too. I don't know why Adobe hasn't just put it in PS by now.

Why not just clip it to a black Color Fill layer in Linear Dodge blend?  It's basically free from a file size and optimization standpoint, is (effectively for any means of output we have) lossless from a quality perspective, and above all accomplishes your desired end without breaking compatibility or adding a blend mode which is going to confuse more users than it helps.  I'm sure if you ask nicely someone around here can create an action or even write a script to make all your Linear Dodge layers behave as you'd like them to.

R A V E N D R I V E wrote:
dobe breaks compatibility all the time across their Creative Suite applications

try unchecking that "maximize compatibility" next time you save a PSD, Lightroom might not like it anymore

1.) Lightroom isn't a part of any Creative Suite.
2.) Lightroom needs 'compatibility mode' for speed, not because Adobe couldn't / wouldn't include the compositing code.  They very easily could have, but the benefit wasn't worth the cost for their alpha testers.  Are a few MB really more important to you than the time it takes for a full file load, compositing, and then previewing?  Can you imagine how slow the Develop module would be?

R A V E N D R I V E wrote:
or the version control in Illustrator

Adding new functionality while retaining the ability to interoperate with older versions isn't breaking compatibility.  It's... well, adding previously non-existing functionality.

R A V E N D R I V E wrote:
18 years of compatibility my ass

I'm referring to redefinition of the blend mode, which will quite literally break nearly as many years of files.  Entirely unlike anything mentioned above, which hasn't broken the backwards compatibility of the line.

Besides, I wasn't aware your ass had 18 yrs on it?

SRB Photo wrote:
The point of the discussion is whether mixed art has a place in mainstream retouching.

I had originally written "I thought it was clear from the point of the (off topic) discussion," but I deleted the parenthetical for brevity (the extra space is still there in fact) and it appeared everyone involved was smart enough to know what I referred and not take me to task on it. Maybe you're just giving me a hard time because I called you the technicality police. No worries either way.

SRB Photo wrote:
The point of your conversation is about misconstruing a blend mode for its blend method.  If you understand the difference, much less want to make an impact and see change, you need to put things in terms the engineers will understand.  Which means discussing it as the latter and not the former - the engineers will dismiss you out of hand otherwise.

In AE and PS they're referred to modes, and I always make an effort to be consistent to the language used in the program I'm discussing. AE's Add mode produces a different result than PS's Linear Dodge (Add) mode. The discussion revealed that the reason for the difference seems to be the blending method used. But my point was and is that Add mode doesn't produce the same result as Linear Dodge (Add).

If I happen to speak to an Adobe engineer at some point I will use whatever language they require; I didn't think that was necessary here, though as I said above, I've attempted to remain consistent in my wording and I'm mainly referring to modes. My brief experience with Adobe's engineers is they [Edit: ARE] friendly and tend to give less-technical people a little slack.

SRB Photo wrote:
Why not just clip it to a black Color Fill layer in Linear Dodge blend?

This is a good workaround that I've never considered, thanks. I'll try it when I get a second. I know how to create an action but I might be interested in a script at some point.

Again, sorry to be OT, but this tangent isn't all at my feet. I didn't expect it would need it's own thread.

-krphoto- wrote:
Maybe you're just giving me a hard time because I called you the technicality police. No worries either way.

Never! .

-krphoto- wrote:
In AE and PS they're referred to modes, and I always make an effort to be consistent to the language used in the program I'm discussing. AE's Add mode produces a different result than PS's Linear Dodge (Add) mode. The discussion revealed that the reason for the difference seems to be the blending method used. But my point was and is that Add mode doesn't produce the same result as Linear Dodge (Add).

And my point is that the blend mode produces identical results (because the math is the same), while the blending method (the opacity) interacts differently on the two programs, likely because of when they were written.

-krphoto- wrote:
If I happen to speak to an Adobe engineer at some point I will use whatever language they require; I didn't think that was necessary here, though as I said above, I've attempted to remain consistent in my wording and I'm mainly referring to modes. My brief experience with Adobe's engineers is they don't seem friendly and tend to give less-technical people a little slack.

Sometimes.  If you look into the history of the ACR vs. C1 color debate, you'll see that some arguments are lost before they begin because of how the issues are presented.

SRB Photo wrote:
Sometimes.  If you look into the history of the ACR vs. C1 color debate, you'll see that some arguments are lost before they begin because of how the issues are presented.

Opacity doesn't just darken the upper layer. It darkens the upper layer by a factor of t, then darkens the bottom by a factor of (1 - t) then sums the two.

What he is asking for (and what AE apparently does) is change the definition of opacity if a blend mode other than Normal is in effect so that the new definition of opacity becomes 'darken the upper layer by a factor of t but do nothing else'

Changing the definition of opacity midstream seems a little kludgey to me, especially since that behavior is probably not well documented and has to be intuited all touchy-feely-like. But whatev.