[kh1234567890]

Idiots!

Some interesting repetitions in the EXIF Makernotes suggest to me a bit of a kludge in a hurry to get it out of the door ...

[Quicksand]

They've used DNG in a way that Adobe's own products can't handle properly! Lightroom! ACR in Photoshop!
Idiots!

At this early stage it's not at all clear to me whether they are (a) subverting the DNG format in a way that fails to meet or goes beyond the spec, or (b) just generating a perfectly compliant DNG that current software packages (including the validator) just don't know how to handle yet. They may have even had a dialog with Adobe on this. Who knows?

There are some advantages to including four separate uncombined images in a single raw file, in particular enabling the raw processing software tool (or its user) to decide how to handle artifacts. If the four images are combined in-camera that opportunity is lost.

[MJSfoto1956]

There are some advantages to including four separate uncombined images in a single raw file, in particular enabling the raw processing software tool (or its user) to decide how to handle artifacts. If the four images are combined in-camera that opportunity is lost.

+1

Wise advice. Let's see how it plays out in the coming months.

[JohnBee]

Just putting out there, that RAW Therapee has no issues with either the DNG or PEF files.

[rawr]

RAW Therapee has no issues with either the DNG or PEF files

In which case, the 'idiots' may be the Adobe LR and ACR developers.

[kh1234567890]

Just putting out there, that RAW Therapee has no issues with either the DNG or PEF files.

In that it can read and demosaic one of the constituent RAW files. It just ignores the other three.

[Class A]

The data is encoded as from a hypothetical 96mpix (4 x 24mpix) bayer sensor. That makes sense.

How does that make sense?

The data is not meant to be demosaiced, as data from a 96MP Bayer sensor would be.

The proper representation would have been a linear DNG with RG(G)B data for each pixel.

There IS a bright side, if you use the PS option and the PS failed due to movement then you have access to your normal workflow with the big PS DNG and can create a normal image without the artifacts.

That's a small consolation for the fact that every software other than Pentax's own Silkypix version will fail to properly interpret PS-RAW data.

You could achieve the same effect by stripping away some of the RGB information from a linear DNG. The standard Bayer-Array pattern is known, so by removing any RGB data beyond it (or one of its shifted variants), you could retrieve single exposures as well.

I'm with Barry Pearson: Forgoing a perfectly valid representation by using an inappropriate representation that either means you are not going to be able to benefit from PS when using RAW files or be forced to use SilkyPix is a big fail. I'm not holding my breath waiting for Adobe to rewrite ACR/LR to interpret Ricoh's files, when all Ricoh had to do was to use a proper encoding.

Just putting out there, that RAW Therapee has no issues with either the DNG or PEF files.

But most likely, you won't see the benefits of PS (Pixel Shift) because it will make use of only a fourth of the data.

---

Last edited by Class A; 05-27-2015 at 05:37 PM.

[kh1234567890]

Looking at one of the PS PEF files with a hex editor, it seems to be not much more than four separate PEF files joined together, headers and all. Should be trivial to split them.

[eraser]

As you are one of the first with a k3ii, have you already looked at the Pentax supplied raw software?

Does it have an export option for DNG or TIFF (16 bit)?

So, opening a DNG in DCU I can save the file as JPG, TIFF (8bit), or TIFF (16bit).

[JohnBee]

But most likely, you won't see the benefits of PS (Pixel Shift) because it will make use of only a fourth of the data.

Indeed that is the case.

[interested_observer]

As I skim through here, with the resulting file apparently consisting of the combined shifted image, along with the 4 individual contributing images, once you have separated the individual images, what is the plan to process them back together in post?

LightRoom does not deal with layers, even though its newest version 6, does process HDR now. So, you are left with PhotoShop or Elements, that can handle layers. There are several other post processing utilities that support layers. Pixel Shifting is not quite HDR where there is a cottage industry of a number of independent software developers providing supporting capabilities.

I have to wonder what the plan was that Ricoh was considering? They must have considered that folks were going to go off he beaten path and start experimenting. Not having a support path in place is a recipe for disaster. This is starting to look like the tethering fiasco all over again. Ricoh/Pentax could have designed an API to tie into any one of the third party tethering packages - especially with the 645 and would have received everyone's complements. Instead its just - Pentax dropped the ball yet again.

Going back to the HDR approach, Pentax rolled out a special version of Silkyplex that processed the combined file into its component images, for other post processing utilities to handle. This is a bit more complex, pull out the component images and find a way to post process them back together with additional controls that folks are going to want to try to apply next. Pentax is going to need to come out with a whitepaper indicting what post processing works (or NOT) on the individual layers. Some guidance and suggestions. The also need to find a post processor (Silkyplex) to support the entire image processing chain with the shifting capability.

So, you really need to consider all of this going into the roll out of the full frame. This could just be a small reason to overshadow the ff body, with all the reviewers saying - excellent, but with the PS capability unresolved, what could have been absolutely GREAT, is now just a excellent body with some unresolved issues.

Pentax/Ricoh does NOT need this. All of this needs to be addressed publicly in the very near future. The body has not even been released in the US yet. What is the recovery plan here. They need to get out in front of this situation before it gets worse.

They need to head off a potential Lemon and turn the situation in to Lemonade.

[stevebrot]

At this early stage it's not at all clear to me whether they are (a) subverting the DNG format in a way that fails to meet or goes beyond the spec, or (b) just generating a perfectly compliant DNG that current software packages (including the validator) just don't know how to handle yet. They may have even had a dialog with Adobe on this. Who knows?

Those were my thoughts as well. Barry Peterson Pearson is sort of the forum expert on DNG, especially as it pertains to Pentax, but Adobe has a long history of funky support of of their own specification in regards to tools. Has anyone attempted a read of the DNG internal file structure using something other than Adobe's validator tool to confirm the ill-advised IFD chaining?*

I, for one, am a little very disappointed that Ricoh/Pentax has opted to merge the pixel shift as a post-process function rather than do this at capture time with the result being a normal Pentax DNG/PEF. The in-camera HDR was and is a kludge and it appears that Ricoh has taken the same short-cut with the Pixel Shift technology. The one redeeming factor might be if in-camera conversion to TIFF is available. The K-3II manual seems to indicate that it may be.


Steve

* I am have only a passing understanding of the DNG spec, but it is my understanding that any "stacked" image data should be represented by a tree structure as opposed to "chaining". It is possible that Adobe's tool has failed on a corner case.

[stevebrot]

In which case, the 'idiots' may be the Adobe LR and ACR developers.

It has happened before.


Steve

[lm4187]

Just putting out there, that RAW Therapee has no issues with either the DNG or PEF files.

Nor does PaintShop Pro X7 or DxO OpticsPro 10.

Cheers,
Lou

[eraser]

The one redeeming factor might be if in-camera conversion to TIFF is available. The K-3II manual seems to indicate that it may be.

So this does work in camera, you can convert a single file, select multiple files, or a whole folder to TIFF then save wherever you choose. Yields a ~70MB TIFF.