Highlight Recovery and HR Range tools are busted.

[Tesselator]

Highlight Recovery and HR Range are almost completely broken.

Shoot or create a gradient from black to white and use those tools. Then use the same tools on just about any other (accepted as) professional editor. The one in AfterShot is busted and only ultimately reduces contrast across the entire image. I can demonstrate if need be. But it's pretty bad, anyone should be able to see the problem right off. And this is very very bad because every serious photographer needs this tool very very often. As it is no one serious can use AfterShot. The only people who would persevere are newbies who didn't know what to expect from a post processing application. This is almost as bad as if the Save tool were broken. :p

Likewise the Fill Light and Fill Range are dysfunctional. In this case however, instead of having too little of an effect and across the entire image, it's way over bearing and broad - to the point where the sliders aren't of a fine enough resolution to adjust the image properly. The Fill Range is weird as well.

Dev Team - please address this issue ASAP. These are hyper critical tools and need to be functioning properly.

[gareth]

That is why I now use CaptureOne where highlight recovery is needed - it recovers cloud structure where ASP creates a grey blur.
I really do like ASP in many ways...especially for adjustments that stop short of needing Photoshop ... just hoping they might sort out recovery and IQ.

[Tesselator]

Yeah. But just to clarify, it's an actual bug (err, Programmer mistake) where he's defined his ranges and clamps wrong. The code works, just on the wrong parts of the image. And actually his HR Range clamp is transposed. The clamp values are assigned to the opposite side of the range. LOL An honest mistake I supposed but it should be corrected immediately - as this is a very important tool.

[Tadjio]

Yeah. But just to clarify, it's an actual bug (err, Programmer mistake) where he's defined his ranges and clamps wrong. The code works, just on the wrong parts of the image. And actually his HR Range clamp is transposed. The clamp values are assigned to the opposite side of the range. LOL An honest mistake I supposed but it should be corrected immediately - as this is a very important tool.

Please can you show us an example so that I understand these concepts better? 'Clamp' is a new term to me.
Thanks for your input to the forum too.

[Tesselator]

Yeah. But just to clarify, it's an actual bug (err, Programmer mistake) where he's defined his ranges and clamps wrong. The code works, just on the wrong parts of the image. And actually his HR Range clamp is transposed. The clamp values are assigned to the opposite side of the range. LOL An honest mistake I supposed but it should be corrected immediately - as this is a very important tool.

Please can you show us an example so that I understand these concepts better? 'Clamp' is a new term to me.
Thanks for your input to the forum too.

I don't wanna sound too full of myself but I don't want to spend too much time creating elaborate examples and tutorials unless the Corel staff need it for some reason. Just for us users it's not needed. You can just use the tools in question and see how they malfunction by comparing the output with that from equivalent tools in other apps. For the definition of a clamp see: http://en.wikipedia.org/wiki/Clamping_(graphics) and in this case you can think of it as a definable end-point for the range which gets modified by the highlight recovery and the Fill Light tools.

[Tadjio]

Yeah. But just to clarify, it's an actual bug (err, Programmer mistake) where he's defined his ranges and clamps wrong. The code works, just on the wrong parts of the image. And actually his HR Range clamp is transposed. The clamp values are assigned to the opposite side of the range. LOL An honest mistake I supposed but it should be corrected immediately - as this is a very important tool.

Please can you show us an example so that I understand these concepts better? 'Clamp' is a new term to me. Thanks for your input to the forum too.

I don't wanna sound too full of myself but I don't want to spend too much time creating elaborate examples and tutorials unless the Corel staff need it for some reason. Just for us users it's not needed. You can just use the tools in question and see how they malfunction by comparing the output with that from equivalent tools in other apps. For the definition of a clamp see: http://en.wikipedia.org/wiki/Clamping_(graphics) and in this case you can think of it as a definable end-point for the range which gets modified by the highlight recovery and the Fill Light tools.

I've read Wikipedia and I've done an HR test varying the HR Range. I don't think HR Range has any effect at all, despite the Help. I can post an example tomorrow (or is it today in Japan?;)
In practice I find just applying a Highlights Recovery setting of 1 (one) is more than sufficient in most cases.
I didn't ask for elaborate examples and tutorials from you, just an example. Could you do so please?

[afx]

HR Range is very much subjective and image content and camera dependent. The default value was set by Dave from empirical testing.
While HR improved significantly with the first update last year, it still is nowhere near where it could be.
Posting about a specific problem without mentioning the camera used, makes it rather pointless to discuss specifics.

cheers
afx

[Tadjio]

HR Range is very much subjective and image content and camera dependent. The default value was set by Dave from empirical testing.
While HR improved significantly with the first update last year, it still is nowhere near where it could be.
Posting about a specific problem without mentioning the camera used, makes it rather pointless to discuss specifics.
cheers
afx

Thanks for your reply - I wanted to provoke a discussion, as well as to 'goad' Tesselator into providing evidence
I agree the HR has improved significantly although I find it produces a 'greying' effect too easily. Is Dave still with Corel?

I wrote late last night after running a series of tests using my Canon EOS 100D but could only come to the conclusion I wrote:

I've done an HR test varying the HR Range. I don't think HR Range has any effect at all... In practice I find just applying a Highlights Recovery setting of 1 (one) is more than sufficient in most cases.

With higher HR values ASP introduced the dreaded pink skies.
This was shooting a typical outdoor / architectural scene which had blown sky highlights and lots of shadows. I examined the Histogram (in ASP and PSP) and used the Clipping Warning ("W").

I've done a similar test this morning using my Olympus E-PM1. Again I usually only use HR=1 but I did find that as I raise the HR value, the HR Range does indeed have more of an effect. I am out on the road for the next 3 days but I will try to post examples.

I do find in general terms that the Olympus .ORF is better suited to ASP lighting / exposure adjustments than the Canon Rebel series .CR2.

[Tadjio]

Here is the example I was working on last night:
A photo taken with a Canon EOS Rebel 100D of Worcester Cathedral (RAW+JPG) with processing of the RAW .CR2 file in ASP.
I exported the original image and the adjusted one to PSP where I took these screenshots.
Image1 is the situation before any adjustments in ASP (none were done in PSP).

Before any adjustments

Image2 is after applying a HR setting of 5 followed by a Blacks setting of -5.

After HR and then Blacks adjustment

My findings were that any higher HR setting was unnecessary. In fact HR=1 would probably be acceptable.
Any change in the HR Range had no effect (from default = 35 down to 0 or up to 100).
I only applied the Black adjustment after seeing that HR had no effect on the Shadows, only the Highlights.
I used Clipping Warning to ascertain that HR=5 eliminated al blown Highlights.

[Tadjio]

I have carried out some further investigation of HR Settings and HR Ranges using my Olympus E-PM1 .ORF files, watching the Histogram and Clipping Warnings.
It's difficult to show specific examples but my main conclusions are:

HR Setting gives a 'crude' or 'rough' setting.
Often HR=1 is sufficient but I tried up to 12 and then 22 to eliminate Highlights Clipping.
HR Range refines this setting by adjusting the Histogram / Exposure.
Changing the Range from 'default' of 35 down to zero reduces the clipping but increases the overall highlights exposure.
Changing the Range up to the maximum of 100 reduces the Highlight Recovery the opposite way.
The HR Range in effect 'squeezes' the Histogram at the Highlights end.
In the end I settled on HR=10 and Range=10 in my example.

Oh and when I look at the final result in PSP the whole image is 'greyer' and the whole Histogram is 'subdued' especially the Reds

I hope this quick summary helps and contributes to the discussion before I return at the weekend!

[afx]

I agree the HR has improved significantly although I find it produces a 'greying' effect too easily.

Yup ;-(

Is Dave still with Corel?

Nope. Was sacked like the rest.

I do find in general terms that the Olympus .ORF is better suited to ASP lighting / exposure adjustments than the Canon Rebel series .CR2.

Probably. For example Canons and Nikons handle the ends of the Luminance spectrum quite differently which has a direct impact on how much can be recovered with any given algorithm.

cheers
afx

[Tesselator]

HR Range is very much subjective and image content and camera dependent. The default value was set by Dave from empirical testing.
While HR improved significantly with the first update last year, it still is nowhere near where it could be.
Posting about a specific problem without mentioning the camera used, makes it rather pointless to discuss specifics.

cheers
afx

No, the function is busted. It has nothing to do with camera specifics. I tested with 9 different cameras. (Yes, I'm a DSLR and EVIL freak and have that many and more - though downloading RAW samples off the net will work if you'd like to try for yourself.)

You either have not ever tried to use the ASP tools in question or, you have never used equivalent tools on ANY other app and have no idea what to expect. I guess neither is the case actually though right? Thus I'm not sure where you're coming from on this. I guess I'll have to eventually illustrate it huh? :p

[hpmickey]

I tested with 9 different cameras.

Hello, try the different camera profiles as well. It could be the night and day difference. I for one hate the default ASP profile for my Sony A-500 camera - wrong colors (especially the skin), wrong black point (too much noise in the dark areas), and frequently clipped highlights which are impossible to recover (pink blotches). Just out of curiosity I downloaded a dcp profile from some forum, converted to icc by dcp2icc utility, tweaked curves in ASP, and voila, suddenly everything is much better.

Btw, I stupidly deleted the original dcp and cannot find the download link again. I would really, really like to verify whether I can freely use it, so good the profile is (better said, so much I do like it ). If anybody knows about good Sony camera profile that will not cost me both hands, please, let me know.

Michal

[afx]

No, the function is busted. It has nothing to do with camera specifics.

Yes, it is busted, still the impact of the issue depends on the camera.

You either have not ever tried to use the ASP tools in question or, you have never used equivalent tools on ANY other app and have no idea what to expect. /quote]
I am using @/&!§/! Lightroom from §$@!ß! Adobe because of this issue for all images that need to be perfect...

cheers
afx