A D300 DR Challenge

Started Oct 14, 2010 | Discussions
Tony Beach Forum Pro • Posts: 11,967
A D300 DR Challenge

This is a continuation of a discussion begun in another thread. I thought it would be interesting to explore the question of how to extract the most DR out of the D300, so I am offering an NEF file here: http://photos.imageevent.com/tonybeach/mypicturesfolder/sharing/_3A40964.NEF Since I do not want to force people to open a 1.5 MB file when they open this thread, I'm including some crops I made from the above file in the reply immediately below this one.

In order to put some context to this thread as it relates to the previous discussion that I left off here: http://forums.dpreview.com/forums/read.asp?forum=1039&message=36618725 I would like to start this thread out by replying to what was written in that thread (or you can just skip that and offer up your best effort at extracting the shadow detail from the file above).

sem wrote:

Did you notice the example?

[See the link above to this post.]

First what is missing here is an NX Capture conversion of the file. Second, we do not have much to go on in determining just how useful the recovered detail is, or what was even captured (or not) in the original RAW file.

Tony Beach wrote:

sem wrote:

At base ISO, with careful exposure, one surely can capture more practically useful DR than the Nikon conversion is willing to chew at once.

From my experience, the DR of the D300 starts to be less than useful around 7 stops (some are more critical, and some are less critical). It's a little more than what I get from my D200 and noticeably less than I get from my A850. Regardless of what RAW converter I use, what I see past 7 stops is degraded resolution and noisy and near-monochromatic tonality.

Well, depends on your quality standards. Of course multiple AEB exposures will produce a better result than squeezing the most out of one NEF using pseudo-HDR.

Flash-fill or reflectors can also help, particularly when the subject you are shooting is nearby. As for my standards, my first reply in this thread will demonstrate what I'm referring to.

First, I'm talking about Nikon's conversion of the RAW data, and there is a difference between what can be accomplished in NX and what can be accomplished in the camera. On that I'm 100% in agreement with you.

Also NX conversion is limited in DR. NX is the best at improvisation with partially blown highlights, but only as long as you set EC low enough (so the shadows are blown).

Regardless of what RAW converter you use, blowing highlights leads to unrecoverable colors. One solution I sometimes use is to do my own improvising by taking what is essentially a white area and burning the channels individually to recreate a believable color there.

But if you have a 10+ stop non-blown NEF, the NX conversion will be blown at least at one end no matter how low you set contrast etc. It seems that best results in harsh light do not rank highly on Nikon's list priorites - they care more for providing good results in good light, that's all.

First, you are not going to pull 10 stops of DR out a D300 NEF no matter how hard you try, at least not what I would consider 10 usable stops. Second, even with HDR, when you start to overexpose large areas of the shot, you get so much flare in the shadow areas that the results there are also unusable.

Second, one has to wonder about those ludicrous DR numbers in DPR's reviews. My guess is that the settings used to attain those numbers would make for a dreadful photograph.

They operate with the term "useful DR" but they take any stop, noisy or not. Their DR score of the D90 is nearly flat line 8EV over all ISO values; while those seven-and-a-half stops are almost clean at ISO 100, they are miserably noisy at ISO 6400.

Again, I do not think the seventh (or even the sixth) stop of DR is " almost clean." Maybe you can clean it up with some NR and pump it full of saturation to make it look acceptable, but you are dealing with extremely compromised data there.

I haven't tested the latest version of ACR, but the version in LR2 and CS3 exhibit poor resolution compared to NX.

Well, everybody seems to appreciate the improvement. I've only test-driven LR3b. It seems to work fine, aside for some hue twists to gray with recovery (I hear solvable with special profiles). The best thing is that working with NEFs isn't as much hassle as with NX.

I have offered a D300 NEF file that has just over 6 stops of DR in it, and I look forward to seeing what anyone here can do with it with any RAW converter to produce a 100% crop that looks remotely as good as the one I show in the reply below which was taken at a more optimal EV.

OP Tony Beach Forum Pro • Posts: 11,967
Here's mine (caution: 1.5 MB file embedded)

Here are two crops from the file in the previous post, along with a crop from another exposure that was optimized to capture the patch:

The middle and right crops are my conversions using Capture NX 1.4 and Capture One 5.2. If anyone can do better with either of those programs, that would be great. Whatever RAW converter you use, show the crops and a small representation of the full image, and it would be very helpful to us is you could describe what you did to optimize the file. I consider my results so pedestrian that I didn't bother to describe the adjustments I made in Capture NX or Capture One.

_sem_ Veteran Member • Posts: 5,033
Re: A D300 DR Challenge

Tony Beach wrote:

First what is missing here is an NX Capture conversion of the file. Second, we do not have much to go on in determining just how useful the recovered detail is, or what was even captured (or not) in the original RAW file.

Here's the thread related to that sample: http://forums.dpreview.com/forums/readflat.asp?forum=1034&message=36110058

It is certainly not an objective comparison and my limited skills with various converters may have an impact. But I've tried to be fair and produce about the maximum reasonable degree of shadows lifting with all converters.

Haven't used CNX2 but I hear from others that the harsh-light DR situation is similar as with VNX. I mean, the U-point thing might have been useful here if it offered localised exposure compensation. Don't know if it provides better noise reduction.

Flash-fill or reflectors can also help, particularly when the subject you are shooting is nearby. As for my standards, my first reply in this thread will demonstrate what I'm referring to.

Yes, fill-flash may be interpreted as hardware HDR (including that it may make a scene look unnatural if used inappropriately).

Regardless of what RAW converter you use, blowing highlights leads to unrecoverable colors. One solution I sometimes use is to do my own improvising by taking what is essentially a white area and burning the channels individually to recreate a believable color there.

Certainly. I mean, with an exceeding DR and the limitation to a single exposure (movement) it may be necessary to blow the sun and some other highlights to get the shadows acceptable. With sharp-edged specular highlights this is not such a problem, but there are often halos in the smooth transitions. Yes you can get rid of them by blowing the remaining channels. But if you want to keep the white patch as small as possible, NX seems to do the best improvisation job - but then you must blend exposures to take care of the shadows.

First, you are not going to pull 10 stops of DR out a D300 NEF no matter how hard you try, at least not what I would consider 10 usable stops. Second, even with HDR, when you start to overexpose large areas of the shot, you get so much flare in the shadow areas that the results there are also unusable.

To be honest, I don't know the exact DR value I'm working with. Once I was playing with a D90 NEF sample that showed nothing blown in Rawnalyze and was full recoverable by merging two VNX conversions about 2EV apart (Photomatix). If the default latitude of VNX is about 8 stops, that would be about 10 stops all together. There was some noise in the shadows, but I didn't find it highly objectionable (well, comparable to my recent sample ... ADL does much worse sometimes). DxO processed that same image very similarly in a single conversion, and since then I was too lazy to use Photomatix, although it seemed to offer a bit more control.

I have offered a D300 NEF file that has just over 6 stops of DR in it, and I look forward to seeing what anyone here can do with it with any RAW converter to produce a 100% crop that looks remotely as good as the one I show in the reply below which was taken at a more optimal EV.

You have a good sample that heavy lifting does not work for vivid colours, obviously multiple exposures is much better. I guess it could be a bit better if you exposed 0.6EV higher into the highlights headroom, but that's difficult in practice and wouldn't help miraculously either. But I wander how do you estimate 6 stops of DR, I'd guess 9+. Don't know in which histogram I can trust the tick-marks. I guess Nikon histograms show about 8 stops alltogether, and your NEF spills over the shadows edge. Photobola shows your NEF is almost all over their 10-tick histogram, only about one tick of the highlights is empty.

So here's my take with DxO 5.3

Modified vivid preset, EC -0.3, Lightning: 70 (default), BP 9, WP 243, gamma 2.65, preserve shadows off, contrast global 48 local 14, WB: 7539K, -21, Vibrancy 0, Color: Camera body; Neutral colour, realistic tonality 2.2 (this matters a lot - totally different other settings with different profiles, and also different resulting colours), noise: default, CA: default
(no lens corrections - DxO lens module for 45mm/2.8 not available)

OP Tony Beach Forum Pro • Posts: 11,967
Re: A D300 DR Challenge

sem wrote:

You have a good sample that heavy lifting does not work for vivid colours,

It also doesn't work well for fine detail. Note the lack of texture in the thread of the patch where the detail is being buried by noise.

I guess it could be a bit better if you exposed 0.6EV higher into the highlights headroom, but that's difficult in practice and wouldn't help miraculously either.

Yeah, I thought I was doing ETTR as I used uniWB and customized settings, but I was too conservative on the histogram (got to trigger blinkies, that's a 1/3 of a stop right there). Still, maybe we're talking more like 7 stops here rather than 6 stops, it's still well below the kinds of numbers we see in DPR's reviews.

But I wonder how do you estimate 6 stops of DR, I'd guess 9+.

I spot metered the scene with plus 1 EC. The highest EV was 1/1250s whereas the patch was 1/20s, that's 6 stops. This is where I should have been alerted to how much I was "underexposing" the shot, there should be about 2 stops from a middle gray metering with my method to the minimum headroom, and I only allowed for one.

Photobola shows your NEF is almost all over their 10-tick histogram, only about one tick of the highlights is empty.

What makes you think a tick equals a stop? If the scale is linear, then half the scale should represent one stop, if the scale is non-linear then establishing the values of the ticks is unclear to me. As I read Photobola, the bold ticks are a stop apart and the other ticks (only available for the highest three stops) are a third of a stop apart, so that's 7 stops in the histogram.

So here's my take with DxO 5.3

Modified vivid preset, EC -0.3, Lightning: 70 (default), BP 9, WP 243, gamma 2.65, preserve shadows off, contrast global 48 local 14, WB: 7539K, -21, Vibrancy 0, Color: Camera body; Neutral colour, realistic tonality 2.2 (this matters a lot - totally different other settings with different profiles, and also different resulting colours), noise: default, CA: default
(no lens corrections - DxO lens module for 45mm/2.8 not available)

Looks similar to the NX conversion to me. I sincerely appreciate your post and the time you spent on it, so thanks for that.

TITCHY Senior Member • Posts: 1,313
Re: A D300 DR Challenge

not a tekkie .nor do I pretend to be ,I am curiuos as to what we are trying to achieve ,is the objective to recover clean vivid colours ,or is the emphasis on what appears to be detail lost to noise .

this is from your original NEF ,its probably the best I could do and is probably a bit over sharpened too .
[IMG]

I apologize if I have the wrong impression and dont want to waste anyones time .
so please be gentle.

edit ,almost forgot , NX2

 TITCHY's gear list:TITCHY's gear list
Nikon Df Nikon Z6 Nikon AF-S Nikkor 24-70mm f/2.8G ED Nikon AF-S Nikkor 70-200mm f/2.8G ED VR II Nikon AF Nikkor 50mm f/1.4D +59 more
OP Tony Beach Forum Pro • Posts: 11,967
Re: A D300 DR Challenge

TITCHY wrote:

not a tekkie .nor do I pretend to be ,I am curiuos as to what we are trying to achieve ,is the objective to recover clean vivid colours ,or is the emphasis on what appears to be detail lost to noise.

Both, and to retain highlight colors too.

this is from your original NEF ,its probably the best I could do and is probably a bit over sharpened too .

It looks pretty good actually, I could see doing two conversions then blending them because your highlights look pretty much blown in that version. I would be interested in what steps you took and if they apply to NX 1.4. Thanks for your contribution.

TITCHY Senior Member • Posts: 1,313
Re: A D300 DR Challenge

sorry ,but I was not taking much notice of settings ,I did not think it was that good ,I will have a look at the original in NX it should show the ticked boxes of what I used ,to be honest though I tried just about every setting in there trying to squeeze more quality in the darkened detail ,I did not notice I had blown the highlights so much .

but it shows at least that this is possible in NX2 ,there may have been a touch of Nik tonal contrast filter used too ,I tend to use that a bit too freely these day .
[IMG]

blown highlights not so bad but its lost a bit of zip in the vibrance department .
thats the trouble with me ,cant just stop when its as good as its going to be .

 TITCHY's gear list:TITCHY's gear list
Nikon Df Nikon Z6 Nikon AF-S Nikkor 24-70mm f/2.8G ED Nikon AF-S Nikkor 70-200mm f/2.8G ED VR II Nikon AF Nikkor 50mm f/1.4D +59 more
OP Tony Beach Forum Pro • Posts: 11,967
Re: A D300 DR Challenge

TITCHY wrote:

sorry ,but I was not taking much notice of settings ,I did not think it was that good ,I will have a look at the original in NX it should show the ticked boxes of what I used ,to be honest though I tried just about every setting in there trying to squeeze more quality in the darkened detail,

[Snip]

thats the trouble with me ,cant just stop when its as good as its going to be .

The real point of this is for all of us to practice, and learn from one another; so it's win-win as far as I'm concerned.

_sem_ Veteran Member • Posts: 5,033
Re: A D300 DR Challenge

Tony Beach wrote:

sem wrote:

It also doesn't work well for fine detail. Note the lack of texture in the thread of the patch where the detail is being buried by noise.

Agreed. Though the lack of texture in my version is partially because I killed it with the black point setting. I cut that bit of deepest shadows DR to hide the noise. But with other two colour profiles I tried this wasn't needed.

Yeah, I thought I was doing ETTR as I used uniWB and customized settings, but I was too conservative on the histogram (got to trigger blinkies, that's a 1/3 of a stop right there). Still, maybe we're talking more like 7 stops here rather than 6 stops, it's still well below the kinds of numbers we see in DPR's reviews.

But I wonder how do you estimate 6 stops of DR, I'd guess 9+.

I spot metered the scene with plus 1 EC. The highest EV was 1/1250s whereas the patch was 1/20s, that's 6 stops. This is where I should have been alerted to how much I was "underexposing" the shot, there should be about 2 stops from a middle gray metering with my method to the minimum headroom, and I only allowed for one.

I'm still inclined to believe that there must be some ground behind the high DR numbers. It could be that the spot meter is not a pinpoint spot so that even when measuring the black part of the patch it is averaging the visible brighter and darker threads of the textile, so that you don't get the reading for the deepest black?

What makes you think a tick equals a stop? If the scale is linear, then half the scale should represent one stop, if the scale is non-linear then establishing the values of the ticks is unclear to me. As I read Photobola, the bold ticks are a stop apart and the other ticks (only available for the highest three stops) are a third of a stop apart, so that's 7 stops in the histogram.

I don't really know about Rawnalyze ticks, I just know that there's ten of them (may be shown on a lin or log scale). I just know this histogram is a bit wider than Nikon's and Adobe's. And another guess is that their histograms are about 8 stops wide because this is believed to be the DR that may be viewed on computer displays.

Looks similar to the NX conversion to me. I sincerely appreciate your post and the time you spent on it, so thanks for that.

I still think it is a bit better than NX. But only a scratch compared to the blended exposures. I appreciate your time too.

TITCHY wrote:

not a tekkie .nor do I pretend to be ,I am curiuos as to what we are trying to achieve ,is the objective to recover clean vivid colours ,or is the emphasis on what appears to be detail lost to noise .

Well it should be admitted that the objective is not universally clear in such DR contests. If you can see the scene while processing, one possible aim is to get the final image (with restricted DR) look as close as possible to the original scene. If you have never seen the original scene, the aim is to make the image look best, to make as much DR as possible discernable (open the shadows as much as noise and colour degradation permit), not blow the highlights preferably (but this may not always hold). In my take I took care not to blow any highlights - not only in the enlarged section but the whole image. I guess if this restriction is relaxed too much, this is no longer a DR challenge

However your sample shows that shadow detail is better recoverable when the highlights may be blow. A HDR blend of two normal-DR conversions is likely to give a bit better result - well, just another scratch closer to the result of blended bracketed exposures. A considerable part of the degradation is due to DR compression from the covered input range to the viewable target DR and will reappear when you do the exposure blending - you are losing contrast somewhere, you can only choose where it hurts least.

OP Tony Beach Forum Pro • Posts: 11,967
Re: A D300 DR Challenge

sem wrote:

I'm still inclined to believe that there must be some ground behind the high DR numbers.

No more or less so than when DPR and other sites measure high ISO image quality, given that it's pretty much an oxymoron to refer to quality and high ISO in the same sentence. Indeed, what you see when you boost the ISO 6 stops from base ISO is pretty similar to what you see in those boosted shadows that are 6 and 7 stops down from what would be an optimal exposure.

It could be that the spot meter is not a pinpoint spot so that even when measuring the black part of the patch it is averaging the visible brighter and darker threads of the textile, so that you don't get the reading for the deepest black?

The deepest black may be 10 stops, but the the brighter colored parts are not, and they are also exceeding the DR of the sensor.

What makes you think a tick equals a stop? If the scale is linear, then half the scale should represent one stop, if the scale is non-linear then establishing the values of the ticks is unclear to me. As I read Photobola, the bold ticks are a stop apart and the other ticks (only available for the highest three stops) are a third of a stop apart, so that's 7 stops in the histogram.

I don't really know about Rawnalyze ticks, I just know that there's ten of them (may be shown on a lin or log scale).

Look closer, there are 6 bold ticks and the first bold tick is halfway from the right edge, that is consistent with a single stop measured linearly. The smaller ticks are consistent with third of a stop increments, and they stop showing those as the stops become too crowded tfor them to be useful, and likewise the bold ticks stop being shown when they too would simply all be bunched together at the left side of the histogram.

I just know this histogram is a bit wider than Nikon's and Adobe's. And another guess is that their histograms are about 8 stops wide because this is believed to be the DR that may be viewed on computer displays.

It's not that the histogram cannot reflect more stops of DR, it's that it isn't large enough on the screen to meaningfully do so. As for how much DR your monitor can show, a print will show even less.

TITCHY wrote:

not a tekkie .nor do I pretend to be ,I am curiuos as to what we are trying to achieve ,is the objective to recover clean vivid colours ,or is the emphasis on what appears to be detail lost to noise .

Well it should be admitted that the objective is not universally clear in such DR contests.

Just to be clear, it is a challenge and not a contest.

If you can see the scene while processing, one possible aim is to get the final image (with restricted DR) look as close as possible to the original scene.

The scene looked very much like the blended version I showed. My eyes had no problem seeing the patch and the background like that, although the DOF I saw was immeasurably more.

If you have never seen the original scene, the aim is to make the image look best, to make as much DR as possible discernable (open the shadows as much as noise and colour degradation permit), not blow the highlights preferably (but this may not always hold). In my take I took care not to blow any highlights - not only in the enlarged section but the whole image. I guess if this restriction is relaxed too much, this is no longer a DR challenge

Yes, ultimately the goal is to optimize how we use our tools to get a better photograph, and reality is only one consideration as far as that goes.

However your sample shows that shadow detail is better recoverable when the highlights may be blow. A HDR blend of two normal-DR conversions is likely to give a bit better result - well, just another scratch closer to the result of blended bracketed exposures. A considerable part of the degradation is due to DR compression from the covered input range to the viewable target DR and will reappear when you do the exposure blending - you are losing contrast somewhere, you can only choose where it hurts least.

This does confirm your point about conversion being a limiting factor. That said, the sensor still is the limiting factor since we can do two conversions from one exposure and blend them, but we still can clearly see better results when blending two exposures. When I do HDR, I set the EV about 2.5 stops apart on my A850 and find that I get good shadows and good highlights that match about what my eyes see, on my D300 I would figure that number would be more like 3.5 stops apart.

rufusm
rufusm Veteran Member • Posts: 3,046
Re: A D300 DR Challenge

Here is my go at it, Capture NX2 and finished in CS3
CNX2: WB 5200
Quick fix: Bring in the blacks and highlights, shadow +100 and contrast -25
LCH: chroma +20
Levels and curves: toe and shoulder adjusted to taste
Color booster +75 +75 opacity
USM: 25,5,4
Output to CS3, -100 blue, USM 15,30,1
Full image:

Closer crop:

 rufusm's gear list:rufusm's gear list
Nikon D750 Fujifilm X-Pro2 Fujifilm X-T4 Nikon AF-S Micro-Nikkor 105mm F2.8G IF-ED VR Tokina AT-X Pro 100mm f/2.8 Macro +10 more
OP Tony Beach Forum Pro • Posts: 11,967
Re: A D300 DR Challenge

Thanks, the patch looks good. The background is blown out, and I think we're still looking at some sort of dual conversion followed by blending the two in post processing in order to get optimal results.

rufusm
rufusm Veteran Member • Posts: 3,046
Re: A D300 DR Challenge

Dual conversion and a merge in CS3 would certainly go a long way in saving this image. Another way I could have gone, although I like the background to my taste only using global adjustments and 10min work, is local background adjustments or a burning layer in CS3. Several ways to get the results as shown and discussed in this thread. Now a real challenge would be to save an image of a person and try to recover that, very difficult to maintain proper skin tone and detail, thank goodness for Nikon flash to rein in the wide exposure range in daylight.

Tony Beach wrote:

Thanks, the patch looks good. The background is blown out, and I think we're still looking at some sort of dual conversion followed by blending the two in post processing in order to get optimal results.

 rufusm's gear list:rufusm's gear list
Nikon D750 Fujifilm X-Pro2 Fujifilm X-T4 Nikon AF-S Micro-Nikkor 105mm F2.8G IF-ED VR Tokina AT-X Pro 100mm f/2.8 Macro +10 more
_sem_ Veteran Member • Posts: 5,033
Re: A D300 DR Challenge

Tony Beach wrote:

Look closer, there are 6 bold ticks and the first bold tick is halfway from the right edge, that is consistent with a single stop measured linearly. The smaller ticks are consistent with third of a stop increments, and they stop showing those as the stops become too crowded tfor them to be useful, and likewise the bold ticks stop being shown when they too would simply all be bunched together at the left side of the histogram.

I was counting the ticks on the "mapped" version of the histogram. Any official interpretation of them (I see Photobola is gone)? I do realize that the crowdedness at the left end of the raw histogram is one reason for the miserability of the bottom stops (noise being a separate one).
Btw isn't the recorded WB odd?

Hoddo Contributing Member • Posts: 559
Re: A D300 DR Challenge

I might have the wrong end of the stick here and frustratingly, I could upload these images directly into DPR...argh!

Anyway, utilised NX2 for this and concentrated only on the badge. Taking little care, check out the halo!, I performed the following

WB to Daylight 5500
added 40 points of colour to the whole image in quick fix section

Very loose selection made round badge and applied LCH step...white slider moved right to 160, mid to 0.92 and changed opacity of this layer to screen.

Holding Shift key added another LCH step also set to screen and took the mid point to input 88 output 157

and finally added another step (hold shift key) only this time Photo Effects.

Brightness to 33

Ignoring the sloppy selection the background remains as the original and could of course be altered. here I only concentrated on the badge.

OP Tony Beach Forum Pro • Posts: 11,967
Re: A D300 DR Challenge

sem wrote:

Btw isn't the recorded WB odd?

It is uniWB.

OP Tony Beach Forum Pro • Posts: 11,967
Re: A D300 DR Challenge

Thanks Martin, except for the halo around the badge, you have shown an excellent conversion.

Chris Walden Senior Member • Posts: 2,394
My version...

Good challenge... something I also explored a while back and posted my findings here (link below). I tested a pretty dynamic scene with lots of variance in proper exposure of highlights and shadows (probably 10+ stops) and found that even though 1 .nef showed no clipping in the shadows it was very hard to regain the details in the shadows. However with a 7 sequence bracketed shot I was able to see the details perfectly. Pseudo HDR didn't work as well as I thought it would. Pay attention to the third or fourth image down with the circles and notice the difference in the shadows.

http://forums.dpreview.com/forums/read.asp?forum=1039&message=33005462 .

On to your challenge... I use LR3 and played around for about an hour with various settings and even did a few exposure blended variants. Honestly I wasn't very happy with most of them so I just started over fresh and first thing I did was slide the Fill Light over to 100. Huge improvement. From there I turned down the sharpening to kill some of the noise, changed the tone curve to medium contrast and that was it. Sometimes simple is better... This is the result of the last go which looks pretty good. Seems like the best balance of pulling out the detail in the shadows and not blowing any highlights. Oh and changed the WB to daylight. You can see most of the settings on the right side of the screen shot.

-- hide signature --

Chris
http://www.33g.com/photos

The difference between a successful person and others is not a lack of strength, not a lack of knowledge, but rather a lack of will.

OP Tony Beach Forum Pro • Posts: 11,967
Re: My version...

Chris Walden wrote:

I use LR3 and played around for about an hour with various settings and even did a few exposure blended variants.

I have LR2, so my expectation is that your results are similar to what I might have done if I had put as much effort into it as you did. Thanks a bunch for taking so much time on this, I like to think that this is a collective effort here at getting better conversions from high DR NEF files.

Honestly I wasn't very happy with most of them so I just started over fresh and first thing I did was slide the Fill Light over to 100. Huge improvement. From there I turned down the sharpening to kill some of the noise, changed the tone curve to medium contrast and that was it. Sometimes simple is better... This is the result of the last go which looks pretty good. Seems like the best balance of pulling out the detail in the shadows and not blowing any highlights. Oh and changed the WB to daylight. You can see most of the settings on the right side of the screen shot.

I haven't come to any conclusions yet here, but based on your results LR does seem like a good choice for this sort of conversion. I'm going to have to go back and use my older versions of the software being used in these conversions and the settings you guys have used, see what I get, and show some side by side comparisons.

Keyboard shortcuts:
FForum MMy threads