Stacked and HDR'd Maybug

[rjlittlefield]

Since your setup is apparently quite stable and repeatable, there is a completely different approach that you might like to try.

Run the middle set of exposures through ZS, allowing it to adjust scale. Save the project, creating a .ZSJ project file. The ZSJ file contains an XML-format description of the project, complete with easily identified file names. Edit the file names to reference another set of exposures, delete the chunks of XML having to do with cached images, and save the result to a new ZSJ file in a new directory tree, thus creating a new project that references different image files but has exactly the same alignment. Open that project in ZS and run Stack > Align & Stack All, which then should skip the alignment because it has already been done. You can confirm that alignment is being skipped by comparing the logs via Options > View Console Log. Repeat for the third exposure stack.

It should be fairly simple to do this process if you first rename your input files so that corresponding images have the same number, with only a different constant prefix for each of your stacks. There are quite a few bulk renaming utilities that will do this job. I use Rename Master from JoeJoeSoft, for no more reason than that's the first one I ran across back in 2005 that did what I wanted.

--Rik

[AndrewC]

You can touch up the HDR image by feeding it back into ZS as an input image. It may be a bit fiddly to get the alignment correct, however, since Photomatix has now altered that.

--Rik

I hadn't thought of that, I was going to try making a "short stack" of the antenna in each exposure stack and use those to remove the halo before HDR ..... but I've used up my buried_in_the_study time for this week

For future reference, if I try this again I'll do everything possible to maintain alignment between focus and exposure stacks that way I should be able to run with no alignment in either package.

Andrew

[lothman]

another option might be the new pentax K7 which can do incamera hdr from three shots, saving one jpg.

[AndrewC]

another option might be the new pentax K7 which can do incamera hdr from three shots, saving one jpg.

Another one is the Rocih CX-1 which can automatically do HDR and focus but i'm not sure it it can combine the two

[rjlittlefield]

Another one is the Rocih CX-1 which can automatically do HDR and focus

Regarding focus on the Ricoh CX-1, all I find is "Multi-target AF can shoot seven consecutive images at seven pre-defined focal distances. After shooting, you can select the image with the preferred focus. The consecutively shot images are recorded as a single file."

Is there more capability, or is this it?

--Rik

[lauriek]

With all the electronics in modern DSLRs I wonder if this could actually be achieved by a 'simple' firmware hack.

Perhaps one of us should try to get in touch with those Russian chaps who came up with the Canon P&S hacked firmware (CHKD or something like that?)?!

It's surely just a question of writing a simple 'program', almost like a time lapse, which some cams do already have:

Shoot.
Adjust focus forward by minimum step.
(optionally) Wait for flash recharge.
Repeat above.

And then attaching this program to a menu option or button...

But when I say 'simple' - I imagine the manufacturers do what they can to ensure this is not so simple. Which is a darn shame really, I for one, and many others I'm sure would be happy with a more 'open' camera...

[lothman]

With all the electronics in modern DSLRs I wonder if this could actually be achieved by a 'simple' firmware hack.

I had a demonstration of a Canon 450D controlled by software from IMAGIC which do automated stacking (by using fine steps of the focal distance). They said it was difficult to linearize the exif data provided by the cam, but it worked perfect on a tripod selecting upper and under limit in life view and the rest was done automatically by the software/camera.

Also some HDR funcionality is implementated.

[AndrewC]

Another one is the Rocih CX-1 which can automatically do HDR and focus

Regarding focus on the Ricoh CX-1, all I find is "Multi-target AF can shoot seven consecutive images at seven pre-defined focal distances. After shooting, you can select the image with the preferred focus. The consecutively shot images are recorded as a single file."

Is there more capability, or is this it?

--Rik

Here's a hands-on review

http://www.photographyblog.com/reviews/ ... x1_review/

[AndrewC]

With all the electronics in modern DSLRs I wonder if this could actually be achieved by a 'simple' firmware hack.

Perhaps one of us should try to get in touch with those Russian chaps who came up with the Canon P&S hacked firmware (CHKD or something like that?)?!

It's surely just a question of writing a simple 'program', almost like a time lapse, which some cams do already have:

Shoot.
Adjust focus forward by minimum step.
(optionally) Wait for flash recharge.
Repeat above.

And then attaching this program to a menu option or button...

But when I say 'simple' - I imagine the manufacturers do what they can to ensure this is not so simple. Which is a darn shame really, I for one, and many others I'm sure would be happy with a more 'open' camera...

... and then getting someone to write some software which automatically ingests the files and in realtime performs the HDR combination and builds up a focussed stack as you sit back sipping the cocktail of your choice

Andrew

[rjlittlefield]

Here's a hands-on review

http://www.photographyblog.com/reviews/ ... x1_review/

Thanks for the link. Sounds like what I'd politely call an "evolving feature".

You can potentially also use the Multi-target Auto Focusing mode to shoot the 7 images, then combine them together in Photoshop or a similar application to create a single image with wider focus than a single image allows. In practice, however, the CX1 tends to always focus on a particular part of the scene and doesn't differentiate enough between the 7 shots to really make this technique effective

But it's getting there, and for focus stepping it's a small step between "works at all" and "gets tuned to work well".

What's not so small is for users to then understand that the scene really does have to be static and the camera can't move around very much.

There's a useful analogy with panorama stitching. It's easy to shoot a distant landscape -- just be sure you have enough overlap and the clouds aren't moving too fast. But it's surprisingly difficult to shoot a living room even if no people are in it. Miss the rotation point by more than a few mm and parallax will mess up the seams. The same thing happens in focus stacking, but there are seams all over the place.

I fully expect to see camera support for focus stacking. I'll just be interested to see how well it sells, and what the manufacturers tune it to do.

Laurie, I share your thoughts about an "open" camera. The current situation puzzles me. I can understand why manufacturers would want to make their interfaces opaque for say lenses and flashes -- sell more hardware that way. But I don't understand why they do the same thing with their control interfaces. It's not like they're making a zillion bucks selling control software. Wouldn't they do better to open up those interfaces and sell more cameras?

--Rik

[AndrewC]

Control interfaces: the trouble is you "open it up" and then change something and everyone starts complaining about you. Either you have a common interface protocol (like GEM or SECS in equipment automation) or you just keep it to yourself and if someone hacks it, good for them.

Andrew

[rjlittlefield]

Control interfaces: the trouble is you "open it up" and then change something and everyone starts complaining about you. Either you have a common interface protocol (like GEM or SECS in equipment automation) or you just keep it to yourself and if someone hacks it, good for them.

There are other options. Consider these guys: http://www.iqeye.com/megapixel-network- ... amera.html . You plug the camera onto a network and assign an IP address. The camera embeds an HTTP server with some standard pages for common control functions. It also embeds a shell that accepts a bunch of appropriately arcane but human-readable command lines for setting everything that's settable.

Do the commands vary between cameras? Probably. Certainly the allowable parameters do. Does anybody give them grief about it? I don't know. But they seem to be doing quite well just by making cameras that do pretty good image acquisition and are easy to talk to.

The reason I know about IQeye is that a couple of years ago, we had a job at work that needed a couple of remote-controllable cameras to do a 360-degree high-res inspection of a hazardous space. My first thought was to use one of the SLR-like point-and-shoot cameras with its nice integrated super-zoom lens. But too bad -- every manufacturer we looked at had either never provided sufficient control of their camera, or had provided it at some point, but then removed it from current models. I talked with one of the developers of remote control packages, thinking I had overlooked something, and he said nope, that was just the way it was, and his product looked to be doomed. So we shopped around some more. Found these IQeye guys. Stuck on an external controller for lights and lens control, and ended up with a nice little package whose interface consisted of power and a network cable. Piece of cake. The image quality was OK for the job. But it would have been so much better to have the image quality of a modern super-zoom point-and-shoot. I still get ticked off thinking about it.

--Rik

[lauriek]

I think open stuff generally is taking off - obviously open source software is now quite mature and now we're /starting/ to see open source hardware [e.g. http://en.wikipedia.org/wiki/Arduino]. Some large manufacturers are testing the water at least [http://www.dd-wrt.com/dd-wrtv3/dd-wrt/about.html]. I think within a year or three we may see a decent open camera.

I can see no way that making a camera more versatile would impact sales in a negative manner...

[AndrewC]

There are actually a few bits of software that let you control exposure and some other camera settings. I think they work through commands available in Nikon's SDK. Of course there is Nikon's own Camera Control Pro as well. However, I haven't seen any that let you control the camera focus. Possibly because the focus algorithms have a lot of IP ?

I have been toying with the idea of exploring it, combining exposure bracketing through a camera interface and focus through stage movements but it is way down on my priority list.

Andrew

[lauriek]

What were we saying?

http://news.stanford.edu/news/2009/augu ... 90109.html