I could have posted this in the Paintshop, Photoshop etc. forum, but all of the image processing tools seem to have the same basic sort of blurs. What is your favorite blur to use on backgrounds to fake depth-of-field? I used to try using the "blur more" tools, but find that Gaussian blur gives me more control and, well, lets me make the background blurrier.

PhotoShop:

1. Duplicate layer
2. Desaturate image
3. Run HiPass filter (?) at value 4.0
4. Invert
5. set blend to Med
6. adj opacity to taste

sorry I cant be more specific at this time as I do not have PhotoShop in front of me. This is actually a tutorial the MStudios posted in DNA's forums a year or two ago.
It's so simple I can whip it out in less then a minute and the high-pass filter gives you much control over the blur range.

Now if just 'softening' an image, can also be used somewhat for 'field of depth' I also like this one:

1. Duplicate layer
2. Press Control+Alt+~ (that's a tildie '~')
3. step #2 will make a selection, now you Invert that selection
4. Press Delete and then deselect all
5. now you can use your eraser on a soft setting and start earasing the area(s) you want to come in clear (ie close or of 'intrest' etc)

I have at times even used both methods on a single image, or the first on the background layer and the second on my forground layer(s)

Attached Link: http://www.lunacore.com/photoshop/tutorials/tut020.htm

You just reverse what the tut is talking about and blur the background and sharpen the up close items. Not knowing how you layout looks and how many up clsoe items you have vs far aways tiems, I use this method frequently when I want to zero in on a specific area either by bluring it or the reverse not bluring it in relation to the rest of the image.

A blurry background and sharp foreground is only one example of depth of field. What about where the camera focus is mid way between fore and background?

Unfortunately I don't know of any cover all solution that works for this.

My depth of field hack jobs consist of duplicating, layering, and blending of various degrees of Gaussian blur...not ideal, but it works for me.

Quote - A blurry background and sharp foreground is only one example of depth of field. What about where the camera focus is mid way between fore and background?

Unfortunately I don't know of any cover all solution that works for this.
.

render a z-depth pass & use that as a depth map with Photoshops lens blur

Damn you Stonemason!

Making me look bad with your huge brain and your sound advice!!!

:) :) :)

Lensecare from Frischluft or Depth of Field PRo form Richard Rosenman are both excellent.

DoF Pro is technically very good. Lensecare is faster.

I prefer the Gaussian blur for the same reasons. It may not be totally realistic, but that isn't the effect I'm looking for.

Not sure if this is my "favorite" method but, I like the result :)

cyberscape: I've never thought of saving the render as a png file. Does it save the render in layers for Photoshop usage? -James

Quote - render a z-depth pass & use that as a depth map with Photoshops lens blur

Ok... how do you "render a z-depth pass" in Poser?

Dan among Den

Or, if there's no Zbuffer available, one can fake it in photoshop by featherng the selection (or working with the masking for little more advanced users).

I can't find depth map rendering option in Poser. I'd be surprized if it doesn't have it. :(

In it's absence I nay attempt to import the scene into bryce (probably via daz studio) and render a depth map there. Or some other program that does it. The tricky part would be matching the camera angle well enough.

sure you can..you can even render depth maps in poser4

courtesy of Stewer: http://www.keindesign.de/stefan/poser/dof_tutorial.html

& this thread at rdna might also offer some tips:
http://www.runtimedna.com/mod/forum/messages.php?forum_id=92&ShowMessage=274774

AWESOME!  Thank You!

In Photoshop, the "Lens Blur" option is beautiful and it will take depth map information.

cyberscape: I've never thought of saving the render as a png file. Does it save the render in layers for Photoshop usage? -James

Not as separate layers, only one. But I believe it does carry over the alpha channels.
What I do is render V3 over the default poser (gray) background and save as a PNG. This way, when I bring it into PS, the background will be 'clear'. This saves me the trouble of having to select every little seethru area at the edge of the hair. This method might change once I learn more about IBL.

Quote - cyberscape: I've never thought of saving the render as a png file. Does it save the render in layers for Photoshop usage? -James

---

Not as separate layers, only one. But I believe it does carry over the alpha channels.
What I do is render V3 over the default poser (gray) background and save as a PNG. This way, when I bring it into PS, the background will be 'clear'. This saves me the trouble of having to select every little seethru area at the edge of the hair. This method might change once I learn more about IBL.

Poser 7?
Then you now HAVE to "RenderOver -> Black" and you MUST save as PNG.
That is now the only way have your rendered Objects on their own Alpha Channel, rendering over grey or an imported background image and saving in any other format then PNG will no longer work for e-frontier feels that is wrong and does not create a proper image for working in other programs.
No longer is there a One-Render method of getting to step that was in the no deleted 'Lens Blur' link here where your subject is on its own Alpha Channel. PNG and Render Over -> BLACK is the only way now. All other methodes will leave you with an Alpha Channel that is Solid White (your entire image) and for basicaly a worthless layer you can just toss out for you will have to create one the old fashion way; Select, Lasso, Zoom In, Select, Lasso, Zoom In, Zoom Out...

Yup, this thread discusses one of my biggest grips about the way e-frontier has "fixed" Poser's render engine...but, so far, they have left us with the ability to still do it with PNG rendered over BLACK....was a lot nicer when we could do it rendering over an imported Background Image, that way we didn't have to fight those "seethru areas around the edges" for they only 'saw through' to the background anyway...

But fear not, e-frontier 'Fixed' what none of the previous versions could, so now we just have to work Harder and not Smarter.
Yup, they 'Fixed' it....

oh, to further explain for jhustead's question about 'saving layers for use in PhotoShop':

Because e-frontier has now 'Fixed' it so that PSD and TIF formats no longer save out that nice Alpha Channel but instead saves out a Solid White usless channel that you can just go ahead and delete (it serves no purpose whatsoever). You now have to render however you wish, set up your image either with background or without - all depending on your structure, myself, I like rendering my characters IN their scene so they are fully interacting, shadows and all and then 'lift them off' as described with the PNG.

Seeing as how you can no longer just use the selection tool  to Select From Image, you know have to render a Separate image, rendering your 'Alpha' section over 'Black' and save as a separate PNG, now you open both your background image and your PNG at the same time. Now there are two ways from here:

1. If your characters are already rendered over the background (the normal way - but e-f says the "wrong way") then you simply make that image active in PhotoShop, go to the menue and Select -> From Image and in the dialog window use the pull down menu to select NOT your PSD or TIF but your PNG and it will select your subject (forground item or items) and you can now Copy/Paste and voila, you have them on their own Alpha Layer - as in that Lens Blur tutorial that no longer seems to show here...

2. If your background image is void of your subject then you will just open your PNG, follow the selection steps above and Copy, then open your background and Paste into it.
   Mind you, this way means that you did NOT render a 'scene' that you are postworking, for your subject will not be 100% interacting with its suroundings and there are a lot of other stuff to prepare for and take into considerations so you get a proper final image that doesn't look like you pasted a cow into an underwater scene or something.....

igohigh: Thanks for the further explanation of my question. Now I'm not so sure about an upgrade to version 7 because “e-frontier feels that is wrong”  for me to render a certain way. Version 6 is working fine for me right now anyway. -James

Quote - igohigh: Thanks for the further explanation of my question. Now I'm not so sure about an upgrade to version 7 because “e-frontier feels that is wrong”  for me to render a certain way. Version 6 is working fine for me right now anyway. -James

Yes, Poser 6 was the last version that saves out an Alpha Channel with the 'subject' masked out from the background (much like the way Bryce saves a 'Mask' render). But in P7 you have no choice anymore but to do it from a PNG using the Render Over -> Black node ("BLACK" meaning the node 'BLACK' not the color black - took me a while to catch onto what they were saying about that).

Why anyone would want a Solid useless alpha channel that contains the entire image is beyond me? You already have the standard RBG channels, and you can create an Adjustment Layer, so what would one what the entire image's data in the Alpha Channel for? What use would it be?
Perhaps there is another use for the Alpha then I am aware of as I really only use them for Masks...?

Anyway, they are still trying to figure out the 'mysterious disconnecting BUMP node' mystery as well...another real pain Poser 6 has brought us....er, excuse me, it was all the past versions that where broken, they kept reloading those darn texture maps instead of makeing you manually reload each time you open the file!!! Duh!