New HDR Panorama feature

[P Moloney]

I haven't been able to make this work... yet!

I produce lots of 3 image panoramas from images produces from 24mm TSE II lens (Three shifts of the lens which normally merge perfectly with "merge to panorama")

Many images I shoot would be vertical shifts of architectural interiors, and a bracketed HDR Panorama would be great.

I haven't been able to make this work with any existing images - I get this message... "Unable to detect HDR exposure bracket size. Merge to non-HDR panorama instead?"

All I can think of is that the method I use for bracketing isn't correct for this feature... I normally make 3 same exposures (shift up, middle, down) then change exposure and do same again.

In the Adobe instructional video <Merge to HDR Panorama in Adobe Camera Raw - YouTube> I can see that each image has been bracketed before moving on to the next shift... is this the secret?

BTW I reordered and re numbered images so that they appear to be sequentially bracketed - that didn't work either - perhaps a slight pixel mis-match is the problem???

I am about to do an interiors job today and will try bracketing as per video... but would like a comment from someone in Adobe who might know?

Peter

[gary_sc]

Hi Dave,

Real quick because I gotta run but you said:

"Very interested in your comment that ACR only needs the +/- shots and can manage without the "0" shot....let me go try."

and:

"I have never seen anybody create an HDR without the "0" shot, using AEB its just natural to use all three images ? In googling around on this problem I watched a number of "youtubes" on the new feature and they all used all 3 images - including Julieanne Kost 🙂

Doesn't excluding the "0" shot have an impact on the final HDR - its missing a lot of the "middle" information ? Need to have a play."

If you read back you'll see I compare the light range is similar to taking a panorama with different lenses. It all has to do with the amount of data you are collecting. If you can collect more data per shot, you need fewer shots.

For a panorama, you need overlap but that also depends on the lens you are using. If you are using a 70mm lens you will need more shots than if you are using a 28mm lens to get the same spread.

Now with HDR, rather than the width of the shots, we're dealing with the full range of light within each shot. So a raw image simply has a lot more data than a jpg.

IF you are dealing with an 8 bit image, you need more shots than if you are dealing with a 12 bit image.

Ten years ago there were no raw cameras in most amateur (or even professional) hands, although some cameras did provide tif images.

Thus, you do need all three (-2, 0, +2) if you are shooting jpg but you only "need" two (-2, +2) if you are shooting raw. [Again, there are situations where they dynamic range of the scene requires more shots so just keep that in the back of your mind]

I just about always shoot all three (-2, 0, +2) because if I do not like the results from ACR or LR, OR if there is a ghosting object in the shot, I can then go over to Photomatix which does much better at deghosting and also has more tools to play around with to fine-tune the image. But for fast and efficient HDR, ACR and LR does a great job. Also FWIW, you can always try all three within ACR or LR and then also try just two and see if either has a look that you like better. Remember, it's all about what YOU like to see in your images.

But again, shooting less than two stops apart (e.g., 1.6, 1, 0.3, whatever) is a waste of time and hard drive storage.

Oh, one last comment: if you shoot 2, 3, or 20 proto-HDR images, the final size of the HDR image will essentially be the same.

HTH

[Chuck Uebele]

I did find out that the 2 stop difference was not intentional, and might be a bug. You should be able to do the HDR pano with less of a spread. What is critical is that the spread be exactly the same, so using auto bracketing should help eliminate any variance.

[DaveMascall]

Hi Chuck

I think the testing I did yesterday has covered the "spread myth". I can merge 3 sets at +/-1 and two sets at +/- 0.3. Therefore I don't think trying to merge to HDR Panorama with images taken with AEB less than +/- 2 is the issue that causes the "Unable to detect..." message. But I could be wrong 😞   Or are you saying that it is "guarenteed" to work at +/- 2 and above, but the code is "buggy" with less than +/- 2 ? In which case I refer you back to my first post on 24th Oct where I was having problems with images taken using AEB +/- 2.

To be clear ALL the testing I did yesterday (and in fact with any HDR sets I take out in the field) are with AEB set - its the right tool for the job, besides which trying to do it manually is just too messy in my opinion.

And yes, I agree using AEB should help eliminate any variance.

So.... be a hero and explain why the following three sets won't merge 🙂

All taken with constant ISO 100 and constant F4, the only thing changing is the shutter speed via AEB

Set 1 - normal 1/80th, dark 1/100th, light 1/60th

Set 2 - normal 1/80th, dark 1/100th, light 1/60th

Set 3 - normal 1/80th, dark 1/100th, light 1/60th

There does not appear to be much(!) variance there. And yet selecting all 9 shots will get thrown out by ACR with the "Unable to detect HDR bracket size" message.

And yet selecting any two of the three sets will allow ACR to detect the bracket size and start creating the HDR pano - thats just plain nuts.

Logic says ACR is looking at the EXIF data (ISO, Aperture, Shutter speed) to determine the bracket size and if all are constant then proceed .... and yet...

Further logic says if it borks on three identical sets, then it can't be using (just?) EXIF data. So if its not using EXIF, what else is it using to determine bracket size ?

Just trying to figure out what the ACR code logic is here that causes the problem, because once you understand what causes the problem when the three sets are identical, it allows you to figure out what to do when taking HDR panos out in the field with the knowledge that ACR will ALWAYS process them without throwing out the "Unable to detect..." message.

Cheers

DaveM

[Chuck Uebele]

I'm not sure why it doesn't work well when manually set. Perhaps something in the metadata when using auto bracketing. As far as the spread, I just heard that the spread isn't suppose to matter, but for some people it does. I believe Adobe is looking into this. It shouldn't be limited to 2 stops and above.

[Chuck Uebele]

Just got word that the logic behind determining the exposures has been fixed. Hopefully it will be out in the next release of ACR.

[DaveMascall]

Hi Chuck

Thanks for info, looking forward to it being available.

Any timeframe ?

Cheers

DaveM

[Chuck Uebele]

No idea when. However updates to ACR come out fairly often due to all the new cameras that need to be incorporated so that PS can read the files.

[Rod Pickett Photography]

I have been experiencing the same problem. I'm pretty sure I know what is happening.

Photo Merge -> HDR Panorama does not recognize stacks!

Photo Merge -> HDR does recognize stacks. But if I try to use collapsed stacks, HDR Panorama only selects the top image in each stack.

If I expand the stacks and select each image, HDR Panorama works.

This is clearly a bug.

Rod