Welcome to our community

Be a part of something great, join today!

  • Hey all, just changed over the backend after 15 years I figured time to give it a bit of an update, its probably gonna be a bit weird for most of you and i am sure there is a few bugs to work out but it should kinda work the same as before... hopefully :)

Using Adobe "Fill" in Camera?

Tom Lowe

Well-known member
Joined
Jan 5, 2007
Messages
8,520
Reaction score
1
Points
0
The HDR thread reminded me that I wanted to post something about the possibility of using an algorithm like Adobe's "Fill" inside the camera, as a look? This way, one can expose for the sky, for example, and still hold more foreground objects than would otherwise be possible.

Here is an example of the expanded "dynamic range" this software produces, using the same RAW frame with and without the effect:

1j7m1y.jpg


2nhevew.jpg


I have no idea whether software like this could be included into firmware on a camera, but I thought I would toss the idea out there anyway. I know that Jim, Jarred, Graeme and the gang don't have enough on their plate already, and are always looking for new hairbrained schemes to add to their workload. :patriot:

And all you smartasses don't bother asking me about "where was the grad nd..." I didn't have one! :hand:
 
It's shadow / highlight tone mapping - brightening up dark regions, while leaving local tonality intact. It's a great grading tool - I use the equivalent in Photoshop or Aperture all the time. (Got to be careful as it can be "overdone" though)

However, for in camera..... It would be a very "expensive" effect in terms of hardware requirements as you need large scale spatial filtering.

Graeme
 
Tom, do you ever shoot with your shirt on? :)

Seriously, thats a really nice effect, and ND's are overused as well, it appears to be a simple and fast solution for correcting dusk shots.

Kenneth, that appears to be a compliment to me ;)
 
Graeme, you think this would require too much processing juice for the upcoming cameras?
 
Perhaps it might change the way you composed, shot and exposed if you had an idea what you were actually able to achieve with software like this in camera?
 
But you're not going to be watching the image in the field on a D.I. level viewing system (on a theater screen), so any decisions you made after seeing a gamma curve applied like this will still be a guess as to the final look.

Since you can't get more out of RAW anyway, you're just lifting the shadow end independently from the highlights, sort of a black gamma stretch -- and usually noise is a result of using that trick. So personally, I'd rather shoot as if I couldn't easily lift the shadows rather than be fooled into thinking I can in the field by seeing this shadow lift, only to find on the big screen later that I have a lot more noise in the shadows than I wanted.

Besides, when you shoot in HD, you generally can imagine what you can and can't do with the image in post later, through experience, so I'm not sure of the need to spend time in the field playing around with shadow levels, etc. If RAW is the widest dynamic range you can record anyway, it's not like you're getting something extra with this trick, and if you want to see everything being recorded, you can switch to RAW VIEW.

The one thing you learn over and over again with digital is that there is no free lunch -- any trick that looks too good to be true usually is. The contrast problem of the shot, with the dark ground and the hot sky, hasn't gone away. Playing around with gammas on the set to see if you can "save" high-contrast shots usually just gives you a false sense of security that everything is OK because you aren't looking at the images on a big screen where you can really spot noise problems.
 
Graeme, you think this would require too much processing juice for the upcoming cameras?

It would need quite a bit but really this kind of thing is best done in post where it can be done properly without realtime limitations. It's also something that you don't really want to do to the raw so it's usefulness is mainly as a monitoring preview.
 
I see what you are saying, guys.
 
If Epic offers RGB 1080p, doesn't it already need to possess the ability to debayer on the fly? Or am I wrong?

Sorry for throwing out yet another hair-brain idea. I appreciate that you guys humor me on this stuff.
 
Yes, although I think the processing needed to get 1080P RGB in-camera already reduces your frame rate selection, so there may be limits to the amount of image manipulation in-camera. On the other hand, usually an RGB camera allows different gamma tables to be used, so one with the shadow lift may be one option.
 
If Epic offers RGB 1080p, doesn't it already need to possess the ability to debayer on the fly? Or am I wrong?

Sorry for throwing out yet another hair-brain idea. I appreciate that you guys humor me on this stuff.

It isn't a cheap operation since it requires a pretty large convolution size I believe. It has to inspect a pretty large area to determine if it's a an area in shadow or just a darker pixel. So that means big caches and lots and lots of processing power. (Every pixel requires its own LUT based on the pixels around it.)

But like David said you could at least get the general feel with a custom LUT. Then just know that when you do the real tone mapping later it'll look much better.
 
Back
Top