Why "0" should be Zero

[TK2142]

Hi,

for those who didn't see the point of starting from neutral develop settings (in the respective LR4 discussions), watch this demonstration by Michael Frye (<- this is a link).

He nicely shows how starting from non-neutral develop settings does not suit some images at all and that these images require undoing all default develop settings before proper editing can start. I hope the demonstration shows the sceptics as to why neutral RAW develop settings can be useful.

I'll try to edit my camera profiles with the DNG profile editor once LR4 has developed to a point when it becomes usable. Hopefully, this will work as desired. However, I still think that no one should be required to jump through such hoops only to get a neutral develop setting. Setting sliders to "0", where "0" means zero, should be all that is required. Everyone else, who wants to start with some default contrast and other enhancements, should see those enhancements reflected in the slider settings.

The curve in the camera profile should just compensate any deviations of the camera from a desired norm. It should not implement someone's idea of a useful starting point for RAW editing.

[Jeff Schewe]

TK2142 wrote:

I'll try to edit my camera profiles with the DNG profile editor once LR4 has developed to a point when it becomes usable. Hopefully, this will work as desired. However, I still think that no one should be required to jump through such hoops only to get a neutral develop setting. Setting sliders to "0", where "0" means zero, should be all that is required.

We've already covered this at length in the now gone LR4 beta forum...the old "neutral" wasn't really neutral and the old "linear" wasn't really linear...there are always baseline adjustments to a raw file. A truely linear (meaning zero tone mapping) isn't really useful since it's so dark.

If you want a replication of PV 2010's linear curve, you can roll your own using the DNG Profile Editor, yes, that works.

So, it'll be interesting to hear what Michael has to say about LR4 and PV 2012. You be sure to ping us when he redoes his video, huh?

:~)

[Guest]

Adobe should just provide a flat preset and/or profile so this stops being an issue. Sheesh.

[Jeff Schewe]

Mark Alan Thomas wrote:

Adobe should just provide a flat preset and/or profile so this stops being an issue. Sheesh.

Why? The OP has drunk the Cool-Aid of a little known "instructor" who advocated a rather minority approach; the "Curves First" approach which with PV 2012 is pretty much out the window.

As it stands, you can pretty much do whatever you want to do with PV 2012...you just have to learn how to use the controls. It's all about tone mapping...what you start with really doesn't have an impact on what you end up with. With the new image adaptive controls in PV 2012, doing the tone mapping with a point curve edit is, well, ignoring the considerable effort the engineers went to to improve PV 2012.

If you want to go down that rabbit hole, there's a way…with a bit of effort. If you want to take advantage of the PV 2012 image adaptive adjustments, no, you'll need to figure out a way to regress your approach to some sort of "linear" or "neutral" starting point. To be honest, I could care less about how the image starts out, what I care about is how the images ends up.

Also note that this OP hasn't, it seems actually updated to LR4...he's waiting till LR4 is usable for him. Whenever that might happen...

And yes, there was a posted issue regarding upgrading a LR3 catalog that extensively used the LR3 point curve editor...

(does anybody else see the delicious irony here :~)

With regards to actually changing the way LR4's curves editor behaves, sorry, that ship has sailed...it ain't gonna change. There was a better potential for changing with the beta, but now that it's shipping, it's a foregone conclusion. So the OP is only trying to cause agitation because the functionality ani't gonna change and the OP knows it.

These are not the droids you where looking for...

[Cornelia-I]

Jeff Schewe wrote:

Mark Alan Thomas wrote:

Adobe should just provide a flat preset and/or profile so this stops being an issue. Sheesh.

Why? The OP has drunk the Cool-Aid of a little known "instructor" who advocated a rather minority approach; the "Curves First" approach which with PV 2012 is pretty much out the window.

Is it necessary to put a demeaning label to Michael Frye? Minority approach it may be, but it receives some attention among Craft&Vision.

I think neither you nor Adobe complains if Lightroom is advertised as a potential means of achieving the photographer's vision in post processing by David du Chemin and his authors at Craft & Vision.

Jeff Schewe wrote:

...

And yes, there was a posted issue regarding upgrading a LR3 catalog that extensively used the LR3 point curve editor...

(does anybody else see the delicious irony here :~)

"Delicious irony" - do you mean that people deserve to have lost their point curve edits ?!

For following such a minority approach i.o. learning LR3-basic sliders?

Jeff Schewe wrote:

...

With regards to actually changing the way LR4's curves editor behaves, sorry, that ship has sailed...it ain't gonna change. There was a better potential for changing with the beta, but now that it's shipping, it's a foregone conclusion.

I disagree: there was never any potential for changing anything in LR4-beta. This was no beta, it was a preview, as someone from Adobe nicely admitted.

The real drama is only that they managed to introduce bugs into the final version which had not been present in beta. That is tantamount to Fukushima nuclear desaster for a software company.

I am not advocating any zero-matters.

But I am repulsed by your attitude towards people who fail to appreciate *your consulting baby* about LR4 defaults and basic sliders with numericals that are meaningsless due to image-adaptive.

Cornelia

[Keith Reeder]

Cornelia-I wrote:

Is it necessary to put a demeaning label to Michael Frye? Minority approach it may be, but it receives some attention among Craft&Vision.

You're taking the point completely out of context, which is that the OP quoting this guy (who I freely admit I'd never heard of until this thread) as "proof" of the validity of his complaint, isn't going to wash - in this respect he is indeed a minority voice, and as such is no more an arbiter of what's right or wrong here than anyone else.

Speaking personally, I'm going to take Jeff's opinions of how Lr hangs together over those of "some guy on the internet" every time.

[Jeff Schewe]

Cornelia-I wrote:

Is it necessary to put a demeaning label to Michael Frye? Minority approach it may be, but it receives some attention among Craft&Vision.

Hey, Michael may be a real nice guy...he's a pretty good photographer, but I'm not sure how skilled he is at raw processing...can't really tell since he didn't actually finish the images.

Did you actually watch the linked video? I did...his first image where he talked about reducing the Contrast to 0 and putting the curve to linear just ended with no additional image adjustments...I didn't see the point of using that image if he wasn't going to finish the editing on the image to make his case. The second image he reduced Brightness to zero and and then set the curves to essentially redo what he undid with Brightness and Contrast. Again, I didn't see that image as arriving at an optimal result that was based on starting at zero and ending up with an improved result.

I didn't see that video as a compelling case for starting at "zeroed" settings...and zeroing the settings before readjusting everything is a minority position. Now with the image adaptive adjustments in PV 2012 Basic, I think trying to do all tone mapping using the point curve is wrong...it defeats the purpose of the image adaptive adjustments. But hey, if somebody want to go down that hole, it's still possible...just not optmal.

[Cornelia-I]

My opinion on Michael Frye is founded on his eBook http://craftandvision.com/books/light-and-land/.

And no, his approach is not the one&only to achieve to a good final result.

My post was not about how compelling his case might be or not.

It was about attitude towards people doing things differently and schadenfreude about their fate at the hand of an insidious bug, which is entirely Adobe's inexcusable fault.

[Jeff Schewe]

Cornelia-I wrote:

....schadenfreude about their fate at the hand of an insidious bug, which is entirely Adobe's inexcusable fault.

Had to look that one up...to be honest, I take no joy because of the upgrade bug with LR3 > LR4 catalog upgrades with regards to the point curve edits being "lost". Perhaps I should not have used the word "delicious" to describe the irony–but it is indeed ironic. And I agree that it's an "insidious bug"...and Adobe has worked hard to determine how it happened and how to fix it and are distressed that it happened in the first place.

But regarding the philosophy of zeroing out the image and making an image look worse before you end up making it better is, I think wrong headed. It complicates the process and in the case of a "linear curve" puts too much emphasis on the point curve editor and not enough emphasis on properly using the Basic tone mapping controls. And now with PV 2012, this approach acts to diminish the importance of the new image adaptive adjustments.

I think starting with an image that has been normalized so it's appearance is not intentionally bad is a better approach. It allows a user to better evaluate what adjustments the image needs rather than starting from some arbitrary artificially poorer starting point. PV 2012 is the result of a lot of research and engineering…seems a shame to ignore that.

[Keith Reeder]

Mark Alan Thomas wrote:

Adobe should just provide a flat preset and/or profile so this stops being an issue. Sheesh.

It's not an "issue" now.

[TK2142]

Keith_Reeder wrote:

It's not an "issue" now.

For you, it isn't.

[Keith Reeder]

Neither you, nor Michael Frye, nor anyone else, has made anything like a convincing case for this being any more than a matter of personal preference.

As Jeff has pointed out numerous times, it's not as if the default starting point is cast in stone anyway - and as he also says, it's how the image looks at the end that matters, not how it looks at the start.

[hsbn]

Michael Frye updated his workflow for LR4, you can watch the video here: http://www.michaelfrye.com/landscape-photography-blog/2012/04/13/lightroom-4-the-new-tone-controls/

He also showed how to get "zero" value in LR4 as you did in LR3.

[TK2142]

hsbn wrote:

Michael Frye updated his workflow for LR4, you can watch the video here:

Thanks a lot for that pointer.

Note that at 9:14 Michael Frye says "Does that mean that Adobe finally came to its senses and took out the boost in midtones that the brightness default used to give and the boost in contrast from the contrast slider and the medium contrast tone curve? Ah, no. "

He states "The defaults are the same, just the numbers have changed." (which is another way of saying "Now the numbers are lying").

He furthermore says "I'm not thrilled about losing a true linear point curve. But I can live with it."

He thinks that overall the PV 2012 improvements are worth a trade off but there is no doubt that he regards the new lying numbers as a negative.

In particular, he notes that the new equivalent of a linear curve is now an inverted S curve with many points and refers to that as a "Very impractical starting point".

[dorin_nicolaescu]

Who is Michael Frye?

[Keith Reeder]

TK2142 wrote

(which is another way of saying "Now the numbers are lying").

No it's not.

It's another way of saying "the numbers are entirely arbitrary and can mean whatever Adobe wants them to mean".

In other words, this proves nothing, and Frye's opinion continues to be just that, and worth no more  and no less, than anyone else's...

[CSS_Simon]

Michael Frye takes nice photos, but I think it is a mistake to assume that the old (PV2010) settings at zero were somehow neutral in an absolute sense.  He talks of "setting contrast to zero and a flat tone curve to see what the picture really looks like" (my emphasis added).  The old zero settings were no more an absolute zero/flat than the new ones.  It's rather like saying that a temperature of 0 Fahrenheit is a real zero but 0 Celsius isn't.  No, they're both arbitrary, and only 0 Kelvin is an absolute, and that isn't very useful for most every-day purposes.  Similarly, a flat raw image that has been converted from Bayer array pixel data to RGB data (but has had no tone curve added) isn't very useful as a starting point for photographers. 

It's a matter of subjective judgement and personal preference as to where you like to start.  Some starting points are better than others, but the only true "absolute zero" of demoisaic-ed sensor data with no tone mapping is probably not one of the better ones. 

If I'm wrong, could someone please define mathematically what tone curve to apply to sensor data to create this absolute-zero tone, true neutral image? 

[Keith Reeder]

Exactly, Simon. As has been explained before, all Raw files, when first "opened" - in any converter - must necessarily have some (again arbitrary) tonal/contrast/saturation etc. values applied to them just to give the photographer a place to start from.

I'm not knocking Michael Frye, incidentally - I'm sure his work is a fine example of the genre - I just don't see any compelling reason for considering his opinion on this point to be definitive, and I imagine that the only reason his views are being given particular credence by the OP in this thread is because they happen to agree with the point he's trying to make.

[TK2142]

CSS Simon wrote:

Michael Frye takes nice photos, but I think it is a mistake to assume that the old (PV2010) settings at zero were somehow neutral in an absolute sense.

There is no denying that there are at least more neutral than PV 2012 at zero.

CSS Simon wrote:
It's rather like saying that a temperature of 0 Fahrenheit is a real zero but 0 Celsius isn't.  No, they're both arbitrary, and only 0 Kelvin is an absolute, and that isn't very useful for most every-day purposes. 

As a matter of fact neither of them is arbitrary as they are both well-defined. But there is something fundamentally "absolute" about "0 Kelvin". It doesn't get colder than that. It is not arbitrary at all.

CSS Simon wrote:
If I'm wrong, could someone please define mathematically what tone curve to apply to sensor data to create this absolute-zero tone, true neutral image?  

The only tone mapping that is required is a gamma encoding because the RAW data is linear and at the display end a gamma decoding takes place. Without the respective gamma encoding, the image will look too dark.

No additional brightness / contrast boost is neutral. Some may consider such boosts necessary all the time so they could have them as their default develop settings. I wouldn't mind if LR came with such boosts out of the box, but then the boost should be reflected in non-neutral slider settings.

Finally, I'm not saying that a completely neutral starting point is always the best. I have my own non-neutral starting point in LR3. But it is own I worked out for myself and that Iintimately know. If an image is not suited for starting with this default, I know exactly what to do to go back to neutral and start over. This is not easily replicated with LR4 because no neutral settings correspond to some arbitrary negative slider settings and most likely they will not exactly undo the hidden tone curve in the camera profile either. Surely, I can tinker with my camera profiles so that they have linear tone curves, it just shouldn't be that complicated.

[CSS_Simon]

"There is no denying that they [PV 2010 zeros] are at least more neutral than PV 2012 at zero."

Quite a few of us do deny it!  As Jeff and others have explained, the zero settings in PV 2010 and in PV 2012 are both arbitrary. 

The old PV2010 "zero" settings were pretty flat, but "flat" and "neutral" are themselves subjective terms, and your idea of neutral is not necessarily mine.  Again, my point is that choosing an arbitrary rendering and describing that as "neutral", the "absolute zero" is a highly subjective choice.  If it works for you (or Michael Frye) then great, but it is an arbitrary choice. 

"As a matter of fact neither of them [Fahrenheit and Celsius] is arbitrary as they are both well-defined."

They are both well-defined, but they are also arbitrary!  The choice of the freezing point of water as zero (Celsius) or the coldest temperature that anyone could at the time create (Fahrenheit) are both arbitrary. 

"The only tone mapping that is required is a gamma encoding because the RAW data is linear and at the display end a gamma decoding takes place. Without the respective gamma encoding, the image will look too dark."

Without gamma encoding, a raw image will certainly look too dark, I quite agree.  But gamma encoding is an arbitrary choice!  What value of gamma?  If you (and Michael Frye) are saying that a good starting point for adjustment is a very low-contrast image, then that's fine - and in many cases I agree with that.  But it is an arbitrary choice. 

Have you seen the "Spinal Tap" movie, where the lead guitarist insists that their amps are louder as the volume control goes to 11 rather than 10?

[TK2142]

CSS Simon wrote:
The old PV2010 "zero" settings were pretty flat, but "flat" and "neutral" are themselves subjective terms, and your idea of neutral is not necessarily mine.  

It is not "my" idea of neutral. It is a technical definition. People can repeatedly claim that there is no "neutral" rendering for RAW files till their blue in the face, it doesn't make it true.

CSS Simon wrote:

But gamma encoding is an arbitrary choice!  What value of gamma? 

The gamma encoding is not arbitrary. In the past there was a difference between Macs (1.8) and PCs (2.2) but nowadays Apple also uses 2.2. ACR could even access the current display gamma from the display profile and use it accordingly.

[CSS_Simon]

"It is not "my" idea of neutral. It is a technical definition."

Can you point me to that technical definition?

"The gamma encoding is not arbitrary. In the past there was a difference between Macs (1.8) and PCs (2.2) but nowadays Apple also uses 2.2."

So, you mean it used to be arbitrary, but it isn't any more? 

The main purpose of gamma correction is to provide perceptually uniform encoding.  It's not about providing any particular "look" to an image.  And, in fact, the gamma curve is removed before display by an inverse curve (allowing for the response of the display device).  Macs before Snow Leopard used a look-up table in Quick Draw with a transfer function of approximately 1.8 gamma.  Now it's the same as PCs with 2.2 gamma. 

"ACR could even access the current display gamma from the display profile and use it accordingly."

Now that really would be a bad idea!  The gamma information from the monitor profile is correction information to overcome non-linearity in the monitor.  In addition, a gamma of (typically) 2.2 is conventionally imposed - but the colour management process of mapping an image from the program's working space to monitor space is designed to deliver an overal linear gamma. 

The purpose of photography (before we apply "artistic interpretation") is to produce an image that matches the original scene.  Therefore the overall system gamma from camera through to to display must be linear.  Otherwise the image won't look like the original scene. 

[TK2142]

CSS Simon wrote:
Can you point me to that technical definition?  

I have already explained what neutral RAW development means in this thread.

CSS Simon wrote:

So, you mean it used to be arbitrary, but it isn't any more? 

No. There used to be two incompatible systems and images could only be correctly encoded for either of them.

CSS Simon wrote:
The purpose of photography (before we apply "artistic interpretation") is to produce an image that matches the original scene.  Therefore the overall system gamma from camera through to to display must be linear.  Otherwise the image won't look like the original scene.  

What you are trying to say is that the gamma encoding must match the gamma decoding value. Otherwise the non-linear encoding inbetween will result in tonal distortation.

That is why the gamma encoder (here, the RAW converter) must always use the same value as the gamma decoder (here the display). That's why I said the RAW converter could consult the display about which gamma value to use. In most cases 2.2 would be just fine to use, but some people prefer 1.8 or other values because they want to optimise for printing output or compensate for viewing conditions.

[CSS_Simon]

TK, we're probably talking slightly at cross purposes.  The gamma curve applied to an image is a matter of coding efficiency to minimse quantisation error and noise.  This doesn't realy matter in 16-bit systems, but in 8-bit systems (such as jpeg) then if you don't apply a perceptually uniform coding system (a gamma of 1.8, 2.2 or the quaint sRGB curve) then you end up with more noise.  The exact gamma number you use doesn't really matter, so long as the display system removes it.  In a colour-managed system, the software will remove any curve in the source image and apply the curve that the destination device is expecting, to make the overall system linear.  The only time a user needs to interfere with this and specify a gamma is if the system isn't end-to-end colour managed. 

I think this is the same as you are saying, forgive me if it isn't.

The tone curve used internally in a program's working space is somewhat arbitrary.  For example: Lightroom uses linear mapping for most processing, but with an sRGB tone curve for displaying histograms and Adobe RGB with 2.2 gamma for the library previews.  For exporting it uses whatever you say (e.g. sRGB tone curve) and for display it uses whatever the monitor profile says. 

A separate issue is tone mapping for subjective or artistic reasons, and where this thread started (I think), was "is there a neutral starting point for subjective/artistic tone mapping?". 

In this respect, in theory you're right and I was wrong (sorry about that), but I think not in practice!

That is, in theory, a neutral starting point is a raw iamge with any sensor non-linearity removed and no other tone curve applied (other than as part of the colour management process).  This is where my knowledge of colour science as applied to image (sensor) data rapidly runs out, and a wiser man than I would shut up at this point.  However, my (limited) understanding is that in practice a flat scene-rendered tone mapping (as close as possible to the tone mapping of the original scene), won't necessarily look neutral.  This is where the subjective element comes in to decide "what is a neutral tone rendering?"  Adobe go further and make that default mapping adaptive to optimise the two ends of the scale. 

I'm not aware of any raw convertor that attempts a truly neutral scene-related tone mapping.  Neither PV2010 or PV2012 do that, and I'm not sure it would be very useful as a starting point. 

[TK2142]

Keith_Reeder wrote:

It's another way of saying "the numbers are entirely arbitrary and can mean whatever Adobe wants them to mean".

Sure, Adobe can choose to let "0" mean "25" or "-100" or whatever, that's correct.

Whether certain assignments are more misleading than others, is another question though.

People tend to associate "neutral" with "0", in particular when it is in the middle of a scale ranging from -n to n.

To associate that middle "0" point with anything else than "hands off" is misleading in my book. Your mileage obviously varies.

There is really only one question which is why Adobe changed from being open about certain default develop settings being applied (LR brightness & contrast values which were not "0") to switching to "0" but keeping the non-zero effects. I understand it has to do with the fact that they wanted to align LR with Revel and the respective customers who do not care about what is right and what is wrong but only want to have things be "intuitively" right.

Keith_Reeder wrote:
In other words, this proves nothing, and Frye's opinion continues to be just that, 
and worth no more  and no less, than anyone else's...

I have never said that Frye is the ultimate authority on image processing. That's your own strawmen you created there. Knock it over as many times you like.

I only pointed to Frye's video because it demonstrates how certain default develop settings are unsuitable for certain images and that it is better for these images to start developing without any default boosts of brightness and contrast. For some images it will be better to start without auto-highlights and auto-blacks so these non-optional auto-adjustments should be made optional.