UKC

Please give me an advice what hdr soft to use now...GIMP or affinity??
or low cost http://macphun.com/snapheal hdr tools (some mac users adviced this soft but i havent seen it before)

In reply to Caryn:

Don't bother, it's just a fad. In five years time it'll look about as classy as a starburst filter

In reply to Oldsign:

I'd argue that the fad for obvious HDR has been and gone. You just don't see as much of it around as you used to.

Probably more subtle HDR around than you would think, but then how would anyone know?

Me, I'm waiting for HDR laptop screens so I can view my pics without tone mapping and need sunglasses. Do not look into the picture of the laser with remaining eye!

In reply to davidbeynon:


Depends what you mean by 'subtle'. An image with a wide enough luminance range to necessitate HDRI will usually be fairly obvious if all those luminances are brought into the displayable range.

In this image, prior to local tone mapping, the grass in the foreground was barely visible, having been washed out by the sun.

https://www.flickr.com/photos/planetmarshall/22231906100/in/dateposted-publ...

In reply to planetmarshall:

Well, By subtle I mean no obvious gradient reversal artifacts, no stupid noise enhancement and reasonable global appearance. All the things people are talking about when they say how much they hate HDR.

I wrote a local tone mapper once that is just a simple unsharp mask type* that simply adjusts contrast enhancement in order to avoid halos. V simple approach that made a huge amount of difference.

It's slow though - need to rewrite to run on my graphics card some time.

*usually used with bilateral filter but that's just details.

In reply to davidbeynon:


Far and away the best local tone mapper I've seen is Erik Reinhard's approach.

https://www.cs.utah.edu/~reinhard/cdrom/

I wrote an implementation when I worked at Leica using separable filters, it's pretty efficient, no need for a GPU approach.

In reply to planetmarshall:

Since we can get technical...

I use the global part of that quite heavily. Didn't like the local results much though - I implemented it but never use it.

Most of the time I take a local adaptation image and just enhance in log space then pass to a global operator. It gives better results and ensures no negative numbers no matter what the <s>untrained monkey</s> user types in.

Separable filters are very fast, but I really prefer the results of a bilateral filter. You have fewer hoops to jump through to keep halos down, but the downside is that they are relatively slow. Not tried trilateral yet but my bilateral is now fast enough that I probably could probably build a sensible one if I wanted.

The system I eventually put together lets you define a directed graph of image processors that work on various channels. Nearly all of them work in CIE xyY.

The graph is fed in, modules loaded and it runs a topological sort to decide run order. Typical process is something like:

1. Pull out Y (luminance) channel.
2. perform some blur (bilateral, separable filters, segmentation etc.)
3. pass to contrast enhancement (with or without halo removal)
4. pass results to global tone mapper

I can set it up however I want of course but that's how I run it for most of my pics.

Not really updated it for a while. I do add things from time to time though. There's likely to be a few CUDA modules coming soon to replace the slow bits like bilateral & halo reduction.

Need to get some free time though.

Of course it goes without saying that you have to be a bit weird to write a command line & script based image processing tool. Never quite got around to doing a proper UI

In reply to davidbeynon:


Don't know the basis of your implementation, but some ideas for efficient implementations here, with various degrees of acceptable short cuts. http://people.csail.mit.edu/sparis/bf_course/slides/06_implementation.pdf


Nonsense.

In reply to planetmarshall:

I have two.

I developed my own algorithm, which was essentially an HDR generalisation of histogram based methods. It was fast for the time but was very much eclipsed in performance terms by the Paris & Durand approximate algorithm. I still use it though, as it's translation invariant which P&D isn't.

The other algorithm I use is basically a monte carlo approximation with importance sampling in the spatial domain. It's faster than the histogram one, and eminently suitable for GPU implementation. I just haven't got around to it yet.

I did a writeup of the "histogram" approach. Never got around to polishing it up and submitting it anywhere though.

http://www.spectral3d.co.uk/Vapourware/tonemap/bilateral.pdf

In reply to Caryn:

The main HDR tool of choice is photomatix. You can do everything from arty, to subtle controlling of highlights and shadows. It does tend to have a certain look to it tho.

A different approach, is to use things like luminocity masks in GIMP. There's a script that helps generate the masks, google for "sg-luminosity-masks.scm". Very powerful and don't have the halo problem of many HDR tools. Gimp can help you with HDR, but you'll need to go to the 2.9 branch which does multi-byte pixels for it to be worth the effort.