[m8o]

infosyn, welcome, and lol ... yes, many HDR threads on the board; thanx for that too tho.

[button]

Incidentally, there's a much better way to get High Dynamic Range if you're willing to do a bit more work: take a set of up to 5 RAW shots using Exposure Bracketing at +-2EV. This will give you an additional 8EV (4 at each end.) Then run the set through a $99 post-processing utility called PhotoMatix Pro for both Windows and Mac:
HDR photo software & plugin - Tone Mapping, Exposure Blending & HDR Imaging for photography

Photoshop CS2 and CS3 also have HDR-related commands and tools that try to do the same thing.

The whole HDR concept and workflow is explained very well at:
NatureScapes.Net - The High Dynamic Range (HDR) Landscape Photography Tutorial

Those methods work quite well for static subjects. For moving subjects, however, multi-shot methods either greatly (exponentially?) increase the work required to get a usable photo, or fail altogether. The holy grail for higher DR lies in some form of single capture. If the k20d can provide a meaningful boost in DR (>2 EV over current DLSRs), then that feature alone would warrant my buying one.

Given the cam's live view, I wonder if the Pentax engineers can (or already have) come up with some kind of real time digital masking filter, which could continuously map high EV value areas on the sensor, and concurrently reduce the ISO at those sites. I'll bet, though, that it would require quite a bit of computational horsepower.

John

[infosyn]

Those methods work quite well for static subjects. For moving subjects, however, multi-shot methods either greatly (exponentially?) increase the work required to get a usable photo, or fail altogether. The holy grail for higher DR lies in some form of single capture.

I'll bet! But think about it: in any still camera, whether or not you end up with multiple discrete files on the card, the shutter is going to need to be open for longer than normal to extract enough detail from shadows. That's why the bracketing needs to go in both directions. This is a physical limitation of any sensor. Thus moving objects (ghosts or motion blur, or both) are always going to be a problem for HDR, even if the extended range is extracted as part of multiple reads or multi-channels of the sensor's analog-digital output.

If the k20d can provide a meaningful boost in DR (>2 EV over current DLSRs), then that feature alone would warrant my buying one. John

Me, 2. For me, 1EV seems like just a joke/gimmick:ugh:, or maybe it's more a gleam in some engineer's eye who's simply testing the market reaction before building the real thing.

Anyway, thanks, folks, for your suggestions. I guess I'll just have to get a K20D next month, to see how it's really done.

[codiac2600]

I'll bet! But think about it: in any still camera, whether or not you end up with multiple discrete files on the card, the shutter is going to need to be open for longer than normal to extract enough detail from shadows. That's why the bracketing needs to go in both directions. This is a physical limitation of any sensor. Thus moving objects (ghosts or motion blur, or both) are always going to be a problem for HDR, even if the extended range is extracted as part of multiple reads or multi-channels of the sensor's analog-digital output.


Me, 2. For me, 1EV seems like just a joke/gimmick:ugh:, or maybe it's more a gleam in some engineer's eye who's simply testing the market reaction before building the real thing.

Anyway, thanks, folks, for your suggestions. I guess I'll just have to get a K20D next month, to see how it's really done.

Increasing the Dynamic range is more than just doing multiple exposures and combining them and actually the K20D can take multiple images and combine them in camera for HDR, but only in Jpeg.

The enhanced Dynamic Range mode in the K20D increasing the sensitivity of the actual sensor to allow better control of shadows and contrast, but the downfall is it can add noise to the final product. This is nothing like HDR, where you have multiple images at different exposures and combine them for a final result. This may be a gimmick to some people, but the real world applications are there you just need to believe a little. Oh and it tried to maintain the highlight as much as possible in this mode, which is the other end of the dynamic range spectrum.

[falconeye]

Now, with K20D in our hands, did anybody find out what this feature really means (200% more dynamic range with loss of ISO 100)? Sample shots in Raw?

[thePiRaTE!!]

Quick test. Hand held, manual focused. 400ISO, f2.0 1/180 w/Voigtlander 58mm. No post processing (except make JPG from RAW)

No Dynamic Range (DR) No High ISO Noise Reduction (NR)


No DR, strong NR


DR, no NR


DR, strong NR


pixel crop samples of same images -

[thePiRaTE!!]

FWIW, my completely subjective opinion is that DR adds too much noise for not enough gain, and that NR doesn't do enough to counteract it. This could be circumstantial and I may discover more use for it, but on my first night, thats my take. I've taken a bunch of pictures here and the noise is very distracting from DR.

I would really like to ask if anyone else has noticed that the noise seen on the screen of the K20 seems more than you were used to seeing on the screen of the K10, and what reasons for that might exist? Frankly, I am feeling a little underwhelmed by my upgrade at the moment. I feel like this is perhaps a lateral move to a new technology rather than a move up to an improvement.

EDIT: Doom and gloom on my part aside, I will say that in PS, the images are very similar between K10 and K20, that the perceived increase in noise on their relative back-screens seems exagerrated on the K20. Could it be that the K20 screen being of high res can simply 'see' more noise that the K10 screen glossed over? It'll take some getting used to for me.

---

Last edited by thePiRaTE!!; 03-07-2008 at 11:07 AM.

[Groucho]

I seem to notice more "noise" when looking on the LCD than on my K100D or DL. However, the photos don't show the noise as noticeably on the PC.

The K20D should be the same resolution (230k pixels) as the previous 2.5" LCD Pentaxes, I believe, so that shouldn't matter. The pixels are larger, though.

[Peter Fang]

The K20D should be the same resolution (230k pixels) as the previous 2.5" LCD Pentaxes, I believe, so that shouldn't matter. The pixels are larger, though.

Actually K10D's 2.5" LCD has 210K pixels so the K20D's LCD is an improvement in the resolution area but it's nowhere close to a 3" VGA LCD featured in either the D300 or the A700.

Peter

[mcjm78]

I just went to my local camera store to play with K20D for 30mins. During the test, I tried the 200% DR Fn, and toogle it on and off a few times. From my limited time observation, the pictures from DR enabled cast a yellow tone to the flourescent lights in the store. I find it quite odd, seem as if the WB is wrong.

I've tried DR off, and the flourescent light is white again. All on Auto WB. I wonder anyone have similar experience?

[WendyB]

I just went to my local camera store to play with K20D for 30mins. During the test, I tried the 200% DR Fn, and toogle it on and off a few times. From my limited time observation, the pictures from DR enabled cast a yellow tone to the flourescent lights in the store. I find it quite odd, seem as if the WB is wrong.

I've tried DR off, and the flourescent light is white again. All on Auto WB. I wonder anyone have similar experience?

And that is why it is HDR which some speculated I had no clue about. This setting is selected by the end user and is quite specific in the range of the RAW jpeg conversion format. I am quite surprised that you all have not figured that out yet.

I expect that a technical explanation will be asked for and am writing one up right now.

[WendyB]

[QUOTE=thePiRaTE!!;191143]FWIW, my completely subjective opinion is that DR adds too much noise for not enough gain, and that NR doesn't do enough to counteract it. This could be circumstantial and I may discover more use for it, but on my first night, thats my take. I've taken a bunch of pictures here and the noise is very distracting from DR.

I would really like to ask if anyone else has noticed that the noise seen on the screen of the K20 seems more than you were used to seeing on the screen of the K10, and what reasons for that might exist? Frankly, I am feeling a little underwhelmed by my upgrade at the moment. I feel like this is perhaps a lateral move to a new technology rather than a move up to an improvement.

EDIT: Doom and gloom on my part aside, I will say that in PS, the images are very similar between K10 and K20, that the perceived increase in noise on their relative back-screens seems exagerrated on the K20. Could it be that the K20 screen being of high res can simply 'see' more noise that the K10 screen glossed over? It'll take some getting used to for me.

Is this really considered a Point and Shoot Camera, I think not. One has so many variables and user control of taking the art of photography within ones hands. I could waste numerous hours finding the problems and issues with this piece of equipment but I refuse to take that route. I focus on how to use the tool to "photograph" and show our talent to the best of my ability and most often find simple workarounds to accomplish the task in hand.

I do not speculate about the technical I let the engineers take that end to worry about then explain the issues that I can not work out after the fact. Then they most often fix it.

I take images of life and the world around me , what do all of you do?

[Richard Day]

This question is from a decidedly non-professional 6-month K10D user: What's the "Dynamic Range Enlargement by 1EV" in the K20D really mean, technically?

Is this just a somewhat-misleading marketing buzzword for post-processing you could do yourself as well in Camera Raw/Photoshop Saturation, Contrast, Exposure, and/or Shadows? Or is it honestly a fundamentally improved way to extract both better highlights and better shadows without sacrificing any mid-range nuances, as the picture is snapped? In other words, does it affect the RAW file in any way, or just the JPG version?

Also, I'm confused by the fact that this feature can be turned on or off.(?) Seems like a very significant benefit you'd always want. What might be the drawbacks? Would your images look unnatural on media not specifically set up for HDR? Would the RAW files be significantly larger, as implied by a larger digital range? Can the resulting RAW files still be read by the current version of Adobe Camera Raw? Perhaps this explains the apparent uncertainty here as to whether the new ADC converter is 12 bit or 14 bit?

According the Ben's recent post on the K20D review in CI (France) the DR enlargement is where the Raw files are 14 bit instead of 12 bit:
https://www.pentaxforums.com/forums/197361-post12.html

I haven't seen this information before, so I have no idea as to the accuracy of this statement. Hopefully Ben or someone can validate this via their contacts with Pentax, I will try to do the same.

Edited addition: If the Raw files are indeed 14 bit as opposed to 12 bit, then the Raw file size for a given shot would be larger with DR enlargement 'on'. It would be worth testing this in both PEF and DNG format just in case it only occurs in one of the formats.

[janneman]

a few shots with and without.....
edr_or_not_edr Photo Gallery by jl2 at pbase.com

On some monitors the evidence is clear, on others not....

This thread also has some LCD posts. Regarding to that, the LCD on camera sometimes shows banding (yes!!) I cannot reproduce on my monitor, unless I doo extreme PP....

[flyer]

I seem to notice more "noise" when looking on the LCD than on my K100D or DL. However, the photos don't show the noise as noticeably on the PC.

The K20D should be the same resolution (230k pixels) as the previous 2.5" LCD Pentaxes, I believe, so that shouldn't matter. The pixels are larger, though.

I noticed that too. But I think it might be moire pattern on the screen due to the compression used to show 14.6M pixels on a screen that has only 230,000 pixels.