[rechmbrs]

I believe they have licensed FAT32 which includes the use of long file names. The problem is that the JPEI designed the suggested format back in DOS days and haven't updated it.

All I want is that the folder and file names be unique.

RON C

[dlacouture]

Another one: if a RAW HDR file is not possible, then what about producing 3 pre-aligned RAW files when using the HDR feature in RAW mode?
The camera could fill the crop data of the DNG files (just as ACR does when you crop), so as to align them perfectly when loaded into an editor.
Full RAW data would still be available if you need it, and it would ease post-processing greatly (no need to align them!).

(Damn, seems like I keep finding possible firmware fixes, must mean this camera is really crap!)

[falconeye]

Another one: if a RAW HDR file is not possible, then what about producing 3 pre-aligned RAW files when using the HDR feature in RAW mode?
The camera could fill the crop data of the DNG files (just as ACR does when you crop), so as to align them perfectly when loaded into an editor.
Full RAW data would still be available if you need it, and it would ease post-processing greatly (no need to align them!).

(Damn, seems like I keep finding possible firmware fixes, must mean this camera is really crap!)

The difficult part is alignment, not hdr. Raw hdr in tripod mode would be feasible.
Once you align, you have to demosaice first. After demosaicing, you cannot write mosaic data anymore.

Only solution is to write so-called linear DNG which is the only raw format supporting non-mosaic data. However, you must add the encoder to the firmware and linear raws are three times as big. Bottom line: much trouble for little benefit.

The better option would be to mark an HDR bracket as such in DNG and LR autoaligns them to a single virtual image on import automatically. Not difficult and somebody should tell Adobe to innovate.

[dlacouture]

Once you align, you have to demosaice first. After demosaicing, you cannot write mosaic data anymore.

In my proposal, the DNGs would be left intact, just as with standalone DNGs.
Only change made to them would be that they are already properly "virtually" cropped so as to be aligned. The RAW data would not be cropped or changed in any way, of course, it's just a matter of DNG tags being properly filled.

[falconeye]

In my proposal, the DNGs would be left intact, just as with standalone DNGs.
Only change made to them would be that they are already properly "virtually" cropped so as to be aligned. The RAW data would not be cropped or changed in any way, of course, it's just a matter of DNG tags being properly filled.

I see what you mean. But you may make an assumption here.
What you propose implies that alignment is done +/- 2px and with no rotation correction. Having written an aligner myself, I can say that this isn't good enough. A good alignment has subpixel accuracy and must include rotation too.

If you could live with a 2px accuracy (I couldn't), then you can actually combine all raws into a single full 16bit raw.

You may think that alignment is the major job. Well, somewhere on lumolabs you find a report with own software. 2px alignment w/o rotation could be made so fast that I couldn't care less if it's done in camera or not. Even full alignment is pretty fast if done properly.

[dlacouture]

I see what you mean. But you may make an assumption here.
What you propose implies that alignment is done +/- 2px and with no rotation correction. Having written an aligner myself, I can say that this isn't good enough. A good alignment has subpixel accuracy and must include rotation too.

?

You can crop with a rotation in ACR, no? So the information must be in the DNG tags.

Regarding the +/- 2px, I'll take your word on it But a quick test with ACR shown me that you can have what looks like sub-pixel accuracy with the crop tool (zoomed in at 400%, you can see that the exact boudary of the crop inside the pixel is kept between sessions, even if the crop result is pixel-related). Don't know how it can be used, though.

[jogiba]

Match the 60D in manual video /audio mode settings , Eye-Fi support and price.
[YT]http://www.youtube.com/watch?v=jbym30NNWqw[/YT]

[falconeye]

?

You can crop with a rotation in ACR, no? So the information must be in the DNG tags.

Regarding the +/- 2px, I'll take your word on it But a quick test with ACR shown me that you can have what looks like sub-pixel accuracy with the crop tool (zoomed in at 400%, you can see that the exact boudary of the crop inside the pixel is kept between sessions, even if the crop result is pixel-related). Don't know how it can be used, though.

Ok, now we're on to something. Yes, the firmware could write DNG with cropping meta data of sufficient precision. That's different from a cropped raw though.

Nevertheless, I still don't see the exact benefit. LR has much better resources to compute the proper alignment upon import. It's rather a wish for Adobe than Pentax. BTW, I've seen the Pentax hdr autoalignment fail which means that their algorithm probably is a compromise.

[dlacouture]

Oh, another one : give us the Fn menu back!!!!!

I'm tired of not knowing if I am in Focus selection or Fn mode.
I know, there is a little icon in the viewfinder, or you can see it in the status screen, but when I want to enter one of the Fn menus, I'm not looking into the viewfinder anymore, and as I disable the status screen because it's really disturbing, I never see it...

A simple solution would be to reinstate the Fn menu when in AF point SEL mode (and only in this mode, as the problem doesn't arise in Auto or Center), if the status screen is not displayed.
One long press on OK (just as now) would display the Fn menu (see K10/K20).
Of course, it should be slightly modified to follow the current menus available on the Pad, by replacing ISO by custom Images.

Ok would go back to focus point selection (just like now).

In the end, the only change needed for this to work is to display an image of the pad with the functions available, if the current screen is not the status screen.

We could keep the proposal made above, that the camera would revert to focus point selection on its own if shutter or AF buttons are pressed or at the end of a timer.

[1connu]

Match the 60D in manual video /audio mode settings , Eye-Fi support and price.

C'est vraiment dommage que pentax ne s'en inspire pas plus. Si pentax pouvait sortir un firmware qui s'approcherait de canon...

[Beau]

I would like exposure bracketing with bigger EV steps. 4 or even 5 EV.
With a DR of 14EV, a 2EV difference is to small. At 14.1EV, a single K-5 file should almost match a K-7 [-2,0,2] bracketed exposure so you could just as well skip the [-2,2] and take a [-4,0,4] shot. Oh, and an option to skip the 0 shot altogether. If you take a -2 0 2 shot, the 0 adds nothing because it's well within the overlapping range of the other shots. [-2,2] would be great for handheld HDR.

[Trig]

I too would love to have the ability to fire off more than 5 shots in AEB. The ability of 7, or even better, 9 would be perfect for HDR work.

[andersju]

Surely (or hopefully) this has already been mentioned, but if so I'd still like to add my +1. Coming from Canon's 5Dmk2 I really like the Pentax K-5, but some video shortcomings irk me:

I would like to have the possibility to fix ISO in video mode. ISO jumping around in a scene because the meter is fooled is annoying and doesn't look good. The ability to set everything manually would be ideal. Pretty please with sugar on top?

I would like to have the possibility to turn off auto-gain control, rather than having to record sound separately and sync afterwards.

These (especially the first) are my main gripes at the moment. They seem like relatively trivial changes. There are tons of indie filmmakers buying 5Ds and 7Ds at the moment... people that Pentax would be able to easily attract with its rugged, compact K-5 and excellent selection of vintage manual focus lenses.

[Beau]

I'm asking for bigger steps in-between each shot, not more shots. Less shots for the same result actually. This would shorten the time between your first and last shot, and thus make hand-held HDR easier.

An example:
Suppose you have a scene which ranges from 2LV to 24LV. You'd need a DR of 22EV.
Each shot can take 14EV.

You would expose your 0 shot to capture 6-20LV
-2 captures 4-18LV
-4 captures 2-16LV
+2 captures 8-22LV
+4 captures 10-24LV

Since the -4 shot captures LV2-16 and the +4 captures LV10-24, you could get the same DR using just these two shot. (disclaimer: I haven't actually tried this)If the full 14EV were to be captured perfectly noise-free you could use a difference of 14EV between each shot and a 2 shot AEB-HDR would give you a DR of 28EV. I'm not sure how much the ranges of each shot have to overlap in practice, but I don't see why the difference can't be a lot more than 2EV.

[secateurs]

Oh, and an option to skip the 0 shot altogether. If you take a -2 0 2 shot, the 0 adds nothing because it's well within the overlapping range of the other shots. [-2,2] would be great for handheld HDR.

If you just want to take 2 shots you can use 2 shot bracketing and set the 2 points wherever you want them.