[MarkJerling]

The whole detail in the red tree branches further down in the frame is certainly interesting. A fair bit of time was invested in trying to duplicate the detail in the jpeg, never really successfully. It's like the jpeg engine did some spot lightening and contrasting my software wasn't able to duplicate. although I could probably get the raw better if that was to be a point of focus. In my mind the willow was going to detract from the image no matter what I did, so I chose to concentrate on parts of the image that had more potential.The big surprise was how that spot contrast darkened the red cabin. To use the cabin as reference, the jpeg ended up a lot lighter than the raw. The original contrast was changed by the jpeg engine, and, if you wanted the same values as the raw, I could find no way to undo the change made by the jpeg engine. The jpeg engine guessed wrong on what my focus would be, and having made the changes and reduced the image from 14 to 8 bit, there just wasn't enough latitude left to change it back.

The willow was going to be a problem with this image no matter how it was shot. The three elements I chose to give preference to were in my mind more important. I'd love to have had a print worthy image to work on with no such weaknesses. Maybe then folks could have focused on the raw conversion vs jpeg, and not become distracted by the tree or other elements of the PP. However, I don't ordinarily keep jpegs from my files. This image was shot for the occasion with the jpeg recorded on the second card.

But my style being so unpopular with thread participants means the raw files are going to be at serious disadvantage no matter how good my starting image is. 90% said, they prefer the in camera processing. Making me a really poor candidate to run this test. If you can't show people something they want to do, you aren't the guy. And what they want to do, I find unappealing. Terms for a mutually agreed upon divorce.

Its really kind of humorous. I said "look what you can do with raw instead of jpeg" and everyone said "we don't want to do that."

Bottom line for me, anyone can post a OOC image. And if everyone on the craft show circuit does the same, then there's no incentive for them to buy your work, not the guy in the next booth's. You need to develop a style that differentiate yourself based on your preferences, and stick to it, so the people who buy from you will find something the next time they look at your work. But, I also get it, not many here come from that kind of experience. Many are looking for cheap and dirty, (and apparently colourless.)

Yes, what I take from this is that it's probably not bad idea to shoot raw + jpg. What would be interesting too, would be to see the out of camera raw image. When I was shooting with the K-5 I shot almost exclusively jpg, until I started using Lightroom. With the K-1ii, I've shot almost exclusively raw. I may need to do the two card thing and record raw and jpg!

[normhead]

Is the assumption here that the primary reason for photography with a "good" camera is to create a picture hanging from the wall, or is that merely the focus of this discussion?

I'm currently gearing all my photos towards display on my 55" 4K TV. But I still prepare like it's for a print. I've also bought my replace current ne when it dies. I go desktop laptop, and the new one is a laptop. I bought a 28" HD display to go with it. I don't find I need to do different processing for printong although when I send out print files I increase the exposure 10% to make up for the no screen lighting.

[GUB]

The whole detail in the red tree branches further down in the frame is certainly interesting. A fair bit of time was invested in trying to duplicate the detail in the jpeg, never really successfully

Wouldn't the difference in the trees be mainly in the colour temperature the jpg used to whiten the snow? As the colour temp goes up the red increases almost as fast as the blue goes down. That maybe gave enough detail boost in the leaves for the edge sharpening algorithm to pick up on. And the difference looks like there is a difference of sharpening there.

[reh321]

Yes, what I take from this is that it's probably not bad idea to shoot raw + jpg. What would be interesting too, would be to see the out of camera raw image. When I was shooting with the K-5 I shot almost exclusively jpg, until I started using Lightroom. With the K-1ii, I've shot almost exclusively raw. I may need to do the two card thing and record raw and jpg!

I routinely shoot "RAW+", and use the JPEG (saved three stars} unless I'm dissatisfied with it and can't 'tweak' it to get what I do want.

[stevebrot]

What would be interesting too, would be to see the out of camera raw image.

Look at the OOC JPEG. It is identical to the OOC RAW when processed in PDCU. I will post images later today (new thread) also provide a conversion using dcraw with demosaic, but no curves applied.


Steve

[reh321]

I'm currently gearing all my photos towards display on my 55" 4K TV. But I still prepare like it's for a print. I've also bought my replace current ne when it dies. I go desktop laptop, and the new one is a laptop. I bought a 28" HD display to go with it. I don't find I need to do different processing for printong although when I send out print files I increase the exposure 10% to make up for the no screen lighting.

The reason I asked is that I may be the "odd man out" here. I got into photography fifty years ago as an offshoot of my hobby as a 'railfan' {long story there}, so my primary goal has never been to display - my primary goal has always been to have an 'accurate' rendition in my files.

[PDL]

That sounds very good, but may prove frustrating.

Images are never saved as RAW in that RAW capture data is fundamentally non-viewable without being converted to pixel data in memory with possibility of being written to file (e.g. JPEG). With RAW+, the only image created is the one rendered to JPEG and that having being made according to the best efforts of the team that wrote the code used for both in-camera and with PDCU RAW processing.

If using PDCU for RAW processing, the first pass results will be virtually identical to the in-camera JPEG* when doing RAW+ minus information lost in moving from 14 to 8 bits. Try a different RAW converter and the first pass results will provide contrast between RAW converters and little more. (FWIW, I tried this just prior to posting.)


Steve

* First configure PDCU to process the RAW for the "Browse" view rather than using the embedded JPEG from the DNG/PEF file. This allows for easy comparison.

I disagree on the idea that RAW is non-viewable while JPEG is. In fact without in memory computing process a JPEG or RAW file can not be viewable on its own. When you look at a JPEG, you are looking a processed data interpreted by the device that you are using to view it on. When you look at a RAW file you are looking at processed information interpreted by the device your are using to view it on. Different OS's use different means of processing JPEG images, when you look at a JPEG in your favorite File Explorer, File Viewer, Finder process the OS is running the manufactures default JPEG/RAW process against it. This was the point of my argument, don't conflate RAW can only be viewed by being processed, with JPEG can be viewed by it self. JPEG requires processing also be it on a computer, tablet, phone or camera LCD screen.

My point is, that given the differences between basic interpretations used by different RAW converter manufactures it would be a better poll to look at how default profiles represent the same theme. Now given that using RAW+ to get the identical image, the taking the picture is all done at the same time. You have one exposure saved in two different formats. The you display the JPEG with all JPEG values set to 0 or whatever you choose as neutral, then you import the RAW file into whatever RAW processor you choose and Export it as a JPEG at 100% without making changes. There will be differences between the RAW "interpretations" because the algorithms from each manufacture are different. In fact, the manufactures state outright that their "colors/sharpness etc." are better than their rivals.

[swanlefitte]

So the Pentax software in camera and on computer is identical. And all raw converters are different. There is that thread 23 raw converters compared.

So my question is Is it impossible to get the same image with different raw processors all the time, in some cases, or it's never impossible with enough work?

[BigMackCam]

So the Pentax software in camera and on computer is identical. And all raw converters are different. There is that thread 23 raw converters compared.

So my question is Is it impossible to get the same image with different raw processors all the time, in some cases, or it's never impossible with enough work?

To obtain exactly the same results with every raw development tool you'd need to have:

- identical camera / colour profiles
- identical demosaicing algorithms and implementation thereof
- identical sharpening algorithms
- identical de-noising algorithms
- identical etc. etc. etc.

So the chances of being able to precisely reproduce a camera's own JPEG output given the raw file and any piece of software other than the manufacturers own (if indeed that software is truly capable of it) is awfully slim.

BUT: In-camera JPEG engines, whilst undoubtedly clever and improving all the time, don't - IMHO - provide the same level of user-driven intelligence in teasing the best from a raw image. And when they get it wrong, you're stuck with a massively-reduced set of image data to work with...

That doesn't mean JPEG is inferior, though. Assuming the basic exposure is more-or-less OK, then depending on the raw post-processing software available to us, our skills in using it, the time we have on our hands to process the image to our tastes etc., JPEG might be better for some folks, or in some situations. But everything in that JPEG image came from the raw data (obviously), yet resulted in an awful lot of data being discarded... so the raw file will always offer the possibility (if not always the probability) of obtaining the best possible image through post-processing. But you need to have software that is capable enough, and both the skills and time to apply it effectively.

[PDL]

So the Pentax software in camera and on computer is identical. And all raw converters are different. There is that thread 23 raw converters compared.

So my question is Is it impossible to get the same image with different raw processors all the time, in some cases, or it's never impossible with enough work?

I doubt that the Pentax software in camera and on the computer (if you mean DCU) is identical as DCU gets updates where your camera does not. I do not believe that the Prime JPEG engine is updated by firmware, it could be, but I really doubt it.

The whole idea of getting the same image with different RAW processors all the time comes down to, if you get the image you expect and are happy with, why continue to look? Along the lines of losing you keys, searching for them, finding them and putting them into you pocket and then you continue to search for them.

This whole discussion brings me to one of the best reasons to do software upgrades of RAW converters. Each version works differently, some improvements, some not so well worked out "features". Since I use Capture One for most of my RAW (which is pretty near exclusive these days) processing I can tell you outright that version 7 rendered my *ist Ds images quite differently than the current v11. Each time the RAW converter is updated/changed the rendering of RAW images is different with, believe it or not, greater capabilities of SUBTLE changes are much improved over previous versions. Phase One has done comparisons between its versions touting the "superiority" of the latest version over the previous one.

Yes, to answer your last question, each RAW converter does the "converting" differently than the other. The feature set of any given RAW converter is one of the criteria to be used in your workflow. When using OOC (out of camera) JPEG's you get the cooked interpretation, although you can modify parameters, that is programed into the Prime processor of your camera body. You are more or less stuck with that for the time you use that body.

As I said in a previous thread, I see LR (v6.14) as being more like the Fuji Chrome films, I see Capture One as more of a Kodak type rendering and DCU as Agfa. Each time a RAW converter is updated, it is like you are getting a whole new set of emulsions to work with and for the most part the differences are for the better. Add on to this, I can re-process my old (13 year old) RAW files and it is like shooting them again with all the potential provided by the new rendering capabilities of the converter. JPEG's can not do that.

[reh321]

My point is, that given the differences between basic interpretations used by different RAW converter manufactures it would be a better poll to look at how default profiles represent the same theme. Now given that using RAW+ to get the identical image, the taking the picture is all done at the same time. You have one exposure saved in two different formats. The you display the JPEG with all JPEG values set to 0 or whatever you choose as neutral, then you import the RAW file into whatever RAW processor you choose and Export it as a JPEG at 100% without making changes. There will be differences between the RAW "interpretations" because the algorithms from each manufacture are different. In fact, the manufactures state outright that their "colors/sharpness etc." are better than their rivals.

My point is that JPEG needs no "interpretation" - it is what it is.

[swanlefitte]

Thanks for the elaborations.
So a single pixel photo where I use the same colorspace and define it as 128,56,256 will look the same? It is in how the software generalizes it's applications that will create the differences. In other words if I define each pixel I can get exact copies but any process that can be defined by different interpretations have next to zero chance of being the same. Even a process saying draw a line 1 pixel wide in black could be different if not 45, 90, 180 degrees where exact precision isn't there.

Is this correct because it seems the rest necessarily has to follow.

[BigMackCam]

So a single pixel photo where I use the same colorspace and define it as 128,56,256 will look the same?

That's an interesting way of looking at it, but I see your point. If by colorspace you mean sRGB or AdobeRGB, you'd also need the colour and tone profile (which maps data to specific hue / saturation / luminance levels) to be equivalent between each raw developer / converter. Those camera-specific profiles can be manufacturer supplied, software developer supplied, user defined (through profiling tools) etc. The means with which the profiles are created has a fundamental impact on the resulting HSL of a pixel

[swanlefitte]

Wow so even 256, 256, 256 = white has a difference? I suppose it's kinda like fonts.

[BigMackCam]

Wow so even 256, 256, 256 = white has a difference? I suppose it's kinda like fonts.

HSL 256,256,256 probably isn't the best example... Although the translation of that could, I suppose, occur by default processing including profiles. I'd say it's the levels in between 0,0,0 and 256,256,256 that are more likely to be significantly different, due to the mapping provided by the profiles for each camera in each specific software used