What does all this highlight recovery discussion have anything to do with seeing 4000 levels of data when editing raw files? Did you find my 4000 levels while you were editing the linked PEF? I still can't see the telltale behavior of editing high bit data that I had with 16bit neg scans.
Nevertheless, I think I found what may be part of the problem for the tinker toy/Fisher Price looking previews in ACR. It has something to do with my original complaint in the past I've had of ACR's clay like texture I get in similar images shot at certain distances to the lens. This one image of the cave just seemed to amplify this texture more so probably due in part to the prominent pattern exhibited in the rock that makes up the majority of the image.
So I opened the image in Raw Developer a DCraw based raw converter that renders previews similar to Guillermo's version. I got the same amount of detail out of the highlights of that rock as Guillermo while at the same time holding neutrality which is a new dropdown menu feature added in Raw Developer.
I could see quite a few more levels applying tightly spaced point curves in this area more than within the rest of the tonal scale of the image. So this is probably where a large portion of the 4000 levels landed. I just wasn't using a raw converter that allowed seeing this detail without making the entire image go dark having to apply an extreme negative adjust to the Exposure slider which is what I have to do in ACR. Also ACR's Recovery slider wouldn't reach up that high to affect the highlights in this particular image. It does work well in other images though.
Below are 100% zoomed view screenshots comparing how the odd texture in ACR appears compared to Raw Developer where I tried to get the same amount of detail tweeking tightly spaced curve points to bring out differentiating tonal levels close to the highlights in the rock. From this observation I suspect Raw Developer's tonal adjust shows a full rez pixel for pixel representation of the image's linear data compared to ACR's seemingly truncated almost jpeg appearance. Slowness waiting for updates to the preview in RD backs this up as well because ACR's adjustments are live and very quick.
So this is all about the previews and how they are rendered between two different raw converters is the cause of all of this misunderstanding of where most of the tonal levels reside.
[attachment=12129:01CaveNr...kLLforum.jpg] [attachment=12130:ACRcaveR...kLLforum.jpg]
Correction on the ACR image sample. I had the Clarity slider set to +30, but all it did is reduce local contrast when turning it off. It still shows the tiny posterized splotches making up rock detail.
