pictures too dark in develop persona preview

[ALe5]

15 hours ago, Chris B said:

I wanted to make sure you weren't accidentally seeing the image in Unamanged mode. Can you please attach one of your raw files? 

The NEF file I attached shows the issue way more emphasized than others, posting a sample comparison, even if much less noticeable is still do it. Following the one with unmanaged mode.

P.S. the NEF file is from D90 the one in sample comparison is D850

_DSC2918.NEF

[lepr]

 

10 hours ago, ALe5 said:

_DSC2918.NEF 16.39 MB · 2 downloads

 

Hi ALe5,

As you know, this is not a view/display problem or an underexposed capture, and I'm posting to confirm that the same misbehaviour is happening with your D90 file in Develop Persona of AP 1.8.3 on macOS: the app is behaving as if it is working with gamma-encoded data instead of linear data.

My other raw developer apps do handle your file correctly.

With my Sony .ARW files and with a .NEF file from a D610 that I have at hand, Develop Persona correctly behaves as if it is working with linear data.

 

[ALe5]

2 hours ago, anon2 said:

 

 

Hi ALe5,

As you know, this is not a view/display problem or an underexposed capture, and I'm posting to confirm that the same misbehaviour is happening with your D90 file in Develop Persona of AP 1.8.3 on macOS: the app is behaving as if it is working with gamma-encoded data instead of linear data.

My other raw developer apps do handle your file correctly.

With my Sony .ARW files and with a .NEF file from a D610 that I have at hand, Develop Persona correctly behaves as if it is working with linear data.

 

Thanks for confirmation. I know it does encode same way to all my NEF files... that is why I believe is a common issue for many, some people it didn't see the difference yet cause is sometime minimal and get passed for good. As a fact at first I thought was normal, but then with some other NEF files get worse as the one you tested. Too bad does behave same on Mac, but they may found out what is happening.

[Chris B]

Thank you for the image.

Apple Core handles this correctly on macOS, however, Serif Labs does not on Windows. The D90 is also on the supported list so this is unexpected. I'll pass this back.

[lepr]

1 hour ago, Chris B said:

Serif Labs does not on Windows.

And Serif Labs does not on macOS

[Gary Cantwell]

I have the same problem, ie when I open a raw file in Develop persona it looks underexposed.

When image moved( with no adjustments) to Photo persona it appears properly exposed.

The underexposed problem does not happen when I open the same image in PS.

With this problem I can not rely on any adjustments I do in the Develop persona.

Therefore I cannot trust Photo persona as I have done no work on the image in Develop persona.

My intention was to cancel my PS subscription and use Affinity, but if this problem continues that will no longer be an option for me.

 

 

[Chris B]

26 minutes ago, anon2 said:

And Serif Labs does not on macOS

Agreed

[ALe5]

just 1 more thing, another file wrongly read by AP this time I develop from RAW to a working AP file the result is same when preview RAW you judge!
 

[ALe5]

4 hours ago, Gary Cantwell said:

I have the same problem, ... My intention was to cancel my PS subscription and use Affinity, but if this problem continues that will no longer be an option for me.

Is counterproductive to use AP, properly because the starting point already "launches" with an handicapped step.. and sorry if I begin to think do I have to ask for money back? Hope not!

If this can be of any help: Within the histogram there is a warning sign that says: "RAW histogram. Click to fine rendering". But after clicking there is no change but the yellow triangle disappear. The major difference between the 2 histograms is the way AP is showing the curves, if u look at the blue you may notice that is out the bound of diagram, that justifies the view and the cut of important data missing from the view. Details and clarity lost in "translation!" Hope you can get out of the quagmire.
 

[lepr]

39 minutes ago, ALe5 said:

just 1 more thing, another file wrongly read by AP this time I develop from RAW to a working AP file the result is same when preview RAW you judge!
 

Doesn't look like the problem that Serif Labs has with the D90 file. This time it looks like the raw data was correctly interpreted as linear data and the image is simply in need of a little exposure boost and a tone curve to give it some brightening and contrast.

[lepr]

4 hours ago, Gary Cantwell said:

I have the same problem, ie when I open a raw file in Develop persona it looks underexposed.

When image moved( with no adjustments) to Photo persona it appears properly exposed.

The underexposed problem does not happen when I open the same image in PS.

With this problem I can not rely on any adjustments I do in the Develop persona.

Therefore I cannot trust Photo persona as I have done no work on the image in Develop persona.

My intention was to cancel my PS subscription and use Affinity, but if this problem continues that will no longer be an option for me.

 

 

Your problem is not the problem that the Serif Labs raw engine has with D90 files.

Your problem is that the document view display is not being managed in 32 bpc mode despite the 32-bit Preview controls being set to manage the display.

That information doesn't help you, but I want to point out that there are two separate problems which are only superficially quite similar.

[ALe5]

43 minutes ago, anon2 said:

Doesn't look like the problem that Serif Labs has with the D90 file. This time it looks like the raw data was correctly interpreted as linear data and the image is simply in need of a little exposure boost and a tone curve to give it some brightening and contrast.

Are u in need of glasses? How could you say that? Look at the difference between the 2 pics... Then that is not D90 file. It seems the encoder is erratic!.. and wrong every time! But the results is unusable software and apparently not only for me. Why I have to start 2 or 3 steps "back?" Is like if I won't mind to start a race with flat tires, dirty windshield and whatever else someone won't like to start with!!

I don't even consider your post a suggestion as a solution.

[lepr]

2 minutes ago, ALe5 said:

Are u in need of glasses? How could you say that? Look at the difference between the 2 pics... Then that is not D90 file. It seems the encoder is erratic!.. and wrong every time! But the results is unusable software and apparently not only for me. Why I have to start 2 or 3 steps "back?" Is like if I won't mind to start a race with flat tires, dirty windshield and whatever else someone won't like to start with!!

I don't even consider your post a suggestion as a solution.

I'm well aware that your latest example was not from a D90. You might learn to talk a little more politely to someone trying to help you. Goodbye.

 

[ALe5]

12 hours ago, anon2 said:

I'm well aware that your latest example was not from a D90. You might learn to talk a little more politely to someone trying to help you. Goodbye.

 

Well thank you for trying to help...  you say politely? What did I say that is not polite? You say I might learn? You become my teacher at once? Well sorry for not being aware of that. Precisely what is useful to me of all of what you've said? Do I sound like I need a teacher? Or I may need a software that does its job? Just to understand ...

[Chris B]

On 4/20/2020 at 8:22 PM, ALe5 said:

Are u in need of glasses?

I would kindly ask you not to make comments like that when someone is just trying to help. In the past anon2 has been a font of knowledge from my point of view and I would greatly appreciate everyone being respectful to one another.

There are a few things in this thread that we need to look into. I am grateful for your help so far and the files you provided. Please give us some time to investigate. Thank you.

[Fritz_H]

On 4/16/2020 at 10:14 AM, Chris B said:

@Fritz_H - ROMM RGB is a wide gamut profile. Do you have a wide gamut capable monitor? If not, I'd recommend switching back to sRGB. Your Histogram will likely look correct then.

@Chris B

I do have a 10-Bit Screen:

for my understanding a histogram is just a "statistical" visualization of all the colors / the color-spectrum used in an image.
A wider color spectrum just adds more nuances to the histogram but dark remains dark and bright remains bright.
I do not see a reason, why a higher bit depth should shift the histogram.
Therefore I do not understand why my histogram is squeezed to the dark end  - just as if "10Bit" extended the histogram only at the bright end..

Fritz

[Fritz_H]

@all

@ALe5 posted some screenshots: comparing Photo with some Nikon-Software.
Did any of you take a look at the histograms?
for my understanding they should be identical (assumed there is no automatic optimization in place.) since its the same raw-data - but they are not.

I downloaded the demo-file he provided; this is my result:


FSViewer also displays the RAW properly (no screenshot)

Sorry, Serif - it seems that your RAW-Engine needs some improvement.

kind regards
Fritz

[lepr]

50 minutes ago, Fritz_H said:

@Chris B

I do have a 10-Bit Screen:

for my understanding a histogram is just a "statistical" visualization of all the colors / the color-spectrum used in an image.
A wider color spectrum just adds more nuances to the histogram but dark remains dark and bright remains bright.
I do not see a reason, why a higher bit depth should shift the histogram.
Therefore I do not understand why my histogram is squeezed to the dark end  - just as if "10Bit" extended the histogram only at the bright end..

Fritz

 

The histogram in Develop Persona is greatly affected by the choice to output a 16 bpc or 32 bpc image which is found in the Develop Assistant.

If 32 bpc, then the histogram is created from values in the linear image that you see in the document view.

If 16 bpc, then the histogram is created from the values in the gamma encoded image that will be created from the linear image that you see in the document view.

The histogram of linear values will appear to be squashed to the left when compared with the histogram of gamma encoded values.

 

[lepr]

22 minutes ago, Fritz_H said:

@all

@ALe5 posted some screenshots: comparing Photo with some Nikon-Software.
Did any of you take a look at the histograms?
for my understanding they should be identical (assumed there is no automatic optimization in place.) since its the same raw-data - but they are not.

I downloaded the demo-file he provided; this is my result:


FSViewer also displays the RAW properly (no screenshot)

Sorry, Serif - it seems that your RAW-Engine needs some improvement.

kind regards
Fritz

 

Serif representative Chris B has already acknowledged in this thread that the Serif Labs raw engine is failing to correctly process ALe5's rose raw file from a Nikon D90.

[ALe5]

2 hours ago, Chris B said:

I would kindly ask you not to make comments like that when someone is just trying to help. In the past anon2 has been a font of knowledge from my point of view and I would greatly appreciate everyone being respectful to one another.

There are a few things in this thread that we need to look into. I am grateful for your help so far and the files you provided. Please give us some time to investigate. Thank you.

Well, let me be as polite as I can be ...  You are another one that is giving me lessons. A moral lesson on how to behave... Do you really think I'm here to take moral lessons from you?

About I officially ask for a refund. As is this software is too much of a headache on a dual point of view.
 Now strange things, sometime happens, that does not go the way we have planned. There is a lot of work I really know of and some I'm not aware of, since I'm not a programmer. Believe me I know a bit and the part I don't I can imagine of what is behind such project of yours; I appreciated that so much that I put my money into it.  But I'm not here to take this kind of treat! I did not come here and have to explain my comments. To me you are missing a point and is under your eyes.
I don't accept anon2 comments cause are too "obvious" and are a copy of what I say and exposed... is not technically getting to the point; anon2 "try to help" comments are, basically,  saying to me that a rotten apple is not shining much but is good... when is, absolutely, not! That may explain my sarcastic comment; disrespect is another thing!

Thank you.

17 minutes ago, Fritz_H said:

@all

@ALe5 posted some screenshots: comparing Photo with some Nikon-Software.
Did any of you take a look at the histograms? for my understanding they should be identical (assumed there is no automatic optimization in place.) since its the same raw-data - but they are not...
Sorry, Serif - it seems that your RAW-Engine needs some improvement.
kind regards
Fritz

Thank you Fritz_H. Yes I did and I mentioned about the histogram. This, is just give one more "point of view" to them to investigate. Let me ask you how would you feel if I tell you that "your" file just needs exposure adjustment? (No need to answer)
To staff: See, this is a repetition of what I expose to you to evaluate. Is this Fritz_H post help to me? NO! Is his post said something different from mine few already posted? No! But, at least, confirm what I already exposed to you.

[ALe5]

57 minutes ago, anon2 said:

The histogram of linear values will appear to be squashed to the left when compared with the histogram of gamma encoded values.

If you notice some peaks are cut out the diagram area, means data is missing.

[ALe5]

On 4/20/2020 at 8:25 PM, anon2 said:

This time it looks like the raw data was correctly interpreted as linear data and the image is simply in need of a little exposure boost and a tone curve to give it some brightening and contrast.

This is the phrase get into my nerve. The 2 files comparison I posted above this comment of yours are so different.. The AP preview is totally out of focus that seems Norton Effect has been apply, Histogram curve's peaks are out of bound, to me data is missing and is slightly underexposed... Repeat my question to your comment: How you can say is correctly interpreted?

[lepr]

30 minutes ago, ALe5 said:

Thank you Fritz_H. Yes I did and I mentioned about the histogram. This, is just give one more "point of view" to them to investigate. Let me ask you how would you feel if I tell you that "your" file just needs exposure adjustment? (No need to answer)

I certainly did not say or imply that your rose image from the D90 was only in need of an exposure adjustment. 

I confirmed and agreed with your finding that the file was being very wrongly processed in Affinity and that it was being correctly processed in other apps.

 

[lepr]

19 minutes ago, ALe5 said:

This is the phrase get into my nerve. The 2 files comparison I posted above this comment of yours are so different.. The AP preview is totally out of focus that seems Norton Effect has been apply, Histogram curve's peaks are out of bound, to me data is missing and is slightly underexposed... Repeat my question to your comment: How you can say is correctly interpreted?

Regarding that comparison of a raw file from a D850 in Affinity and another app, I discussed only the tonality and not the sharpness because Affinity applies no sharpening by default, while I have no idea what your other app is doing with regard to sharpening. Please post a D850 raw file so I can compare the processing in various apps instead of viewing your screenshots which do not reveal many controls of the apps.

Clipped peaks in an Affinity histogram do not imply missing data. The histogram is simply not being vertically scaled adequately, and so the peaks are too high to fit in the available display space.

[lepr]

55 minutes ago, ALe5 said:

If you notice some peaks are cut out the diagram area, means data is missing.

The clipped peaks do not indicate that data is missing. If the histogram could be vertically reduced in scale, then the peaks would come into view at the top of the histogram display. Some apps will automatically scale the height of the histogram to completely prevent clipping of peaks in the histogram display.