I opened your JPEG screenshot, then in-turn cropped the darktable side, then the LR side, so I could compare the histograms. The difference I can discern would be attributable to a “tone curve” on top of the initial processing. Indeed, I tried a curve on the darktable crop, and I was able to make an equivalent image just by dragging the top-right point over to the left, just to scale the image data.
Ah! You’re starting to ask the same sort of questions I asked about three years ago when I started to re-explore photography. What I found is that raw processors, including what your camera does internally, do various levels of processing to present you with what you might call a “photograph”. LR and its ilk tend to do more, RT and dt a little less, depending on how configured. Even dcraw, the basic command line raw processor, scales your image for output by default, and you have to explicitly tell it not to.
What I found works for me was to come to understand Every Single Thing that has to be done to a raw image, and I mean ‘really raw’, in order to make it presentable. It’s not a lot:
white balance,
demosaic,
color conversion,
whatever other processing to make the image presentable.
#4 is where so-called ‘secret sauce’ of cameras and softwares resides, but I’ve found a simple linear scaling is all it takes to make my proof images, and a lot of times those are good enough for the family.
Why don’t you just store the dng’s?
Whenever you need, you can always pick one of those, do a quick edit and voilà, there you have a high quality photo.
Do you have in mind just storing high quality jpg’s?
If so, basically, your image is about 1.5 stops more exposed when you open it in LR and Windows image viewers than when in dt. This doesn’t mean that your DNG is any less good; i.e., the data is still there. What is different is one of two things or both:
The DNG has tags that indicate the 1.5 stop setting.
LR or the image viewer adds 1.5 of exposure by default.
How to solve this problem is simple. Open it in LR / viewer, lower the exposure and generate a new thumbnail or preview. AFAIK, you can do that in LR, PS, Adobe DNG Converter and likely your phone’s camera and image editor apps.
Alternatively, just edit the DNGs, export TIFs or JPGs and decide whether you want to keep the DNGs.
As I said I want to avoid editing any photos. Just quickly browse through DNGs the same way I would browse JPGs. I am looking for a Windows and Android apps to view DNGs quickly without them adding any exposure shifts. I saw IrfanView mentioned. Anything for Android?
Adobe Bridge is fast but it adds that nasty overexposure.
It doesn’t take that much more work to make the adjustment or generate new previews for LR or Bridge. You choose one item from one dialog, apply it to every file and voila. Same goes for dt and all other apps.
Other than that, underexpose your images by 1.5 EV but that would hurt your dynamic range.
PS If you really don’t want to touch anything, then JPG is still an option. There is no shame in using JPGs. We share and print them all of the time, not the raw files. If you have an app like OpenCamera, you probably have enough settings to satisfy your photography needs without needing to go into raw.
I tested IrfanView and there is a major color difference. Blue shows an obvious purple hue (see pic). I tried playing with the settings in the Plugins section (camera white balance,…etc), but it doesn’t fix it. If I uncheck the camera white balance, I get a red hue over whole image.
As I understand it, the dng contains a camera profile which has a colour matrix and a tone curve. dt seems to be using the colour matrix but not the tone curve. The other apps you mention are probably using both.
If you really don’t want to edit your photo’s then you’re best getting them right in the camera and saving as jpegs. Take the photo and look at it. If the exposure isn’t how you wanted it then change it and take the photo again until you get something you like.
Does opencamera have the option to save as raw+jpeg? That would probably be your best bet.
Yes, use darktable.
As pointed out several posts ago, you never see the raw information in an imageviewer. At least some basic processing is done - in most cases with an ‚one size fit‘s all‘ approach.
Endless discussion about different colourscience of different manufacturers and which ist better is an indicator, that there are a lot of decisions to make between capturing light on a sensor, store it and process it to achieve the intended look. Most manufactures uses different picture styles to cope with different users expectations.
So to enable IrfanView to display the raw files like your cameras manufacturers jpg they have to implement all theses decisions your camera manufacturers jpg engine made …
The jpgs look unnatural to me (esp the ones out of my phone) with all the denoising. The dngs have a higher bit depth too, isn’t it? They look more life-like, esp at 100%.
This is a common experience when people first encounter Darktable (DT) and free software RAW development. Here’s my own post about it, from almost exactly a year ago:
In my case, the RAW development was very far from the JPG one. It turned out there were all sorts of issues with the shots, partly due to LED lights giving out of gamut colors which confused even the in-body JPEG rendering…
One of the problems we’re facing, if I understand this correctly, is that camera manufacturers (Nikon, Canon, etc) collaborate with Adobe to have their JPG development process reproduced in Lightroom (LR) more faithfully. This means the RAW development LR users first see is often close to the out of the camera (OOC) JPEG, because proprietary development algorithms from the camera are reproduced in LR. I found this particularly striking when working on pictures from my Fujifilm X-T2 camera.
An entire play RAW thread discussed the finer details of how that bridge (and its environment) could be rendered:
In the end, I think the experience of free software RAW processing is that you are not trying to reproduce the JPG or LR development. I understand you are not aiming to do that, but the reality is that development is a personal, maybe even artistic process that will yield vastly different results depending on your objectives and (in my case, lack of) skill.
The way I browse RAWs is through Darktable, and I made it explicitely use the JPG thumbnails. This makes sense to me because it’s how I saw the shots when I was taking them, it’s often the shot I want. It does mean that I sometimes have to go uphill to go back to that JPG, but it’s a small price to pay to not have Adobe and/or Microsoft crap on my computers.
I came to about the same conclusion, but from a different journey. Knowing from my computer science background that editing JPEGs rapidly degrades the image (well, all editing degrades the image, it just becomes noticeable faster with 8-bit JPEGs, a discussion for a different thread…), I set out to control my in-camera processing to produce JPEGs that didn’t require any editing other than maybe a crop. Long-story short, I was not successful to my satisfaction, so I worked to make a raw workflow that met my needs. Another long-story short, at that I was successful, but my particular path involved writing my own software.
Now, I don’t think you need to go that far. The mainstream FOSS tools provide sufficient control and flexibility to manage any raw workflow I can envision, including my own. But, more importantly, they allow you to manage your images on your own terms, not some arbitrary default that a camera manufacturer or Adobe figured out. Even then, RawTherapee now has a histogram-matching tool that will make your raw development look substantially close to the JPEG that the camera embedded in your raw file, so to an extent you can compare.
Your camera and Adobe essentially ‘held your hand’ in making an image that, really, just keeps you from asking them a bunch of questions like “Why is my image dark?” “Why doesn’t my image ‘pop’?” and such. The FOSS tools are much less about that, and more about providing you control in how your images are rendered.
I believe someone did that already for the Fujifilm camera: they are “styles” that reflect the different film emulation… Unfortunately, the results, while close in terms of colors, do not really match the other parameters (lighting, contrast, etc)…
Your DNG doesnt have a colour space tagged in the EXIF (at the time of capture), so Irfanview doesnt know what colour space to show it in. Darktable is assigning it one when it shows the DNG for editing (as does ACR, Lightroom, and Rawtherapee). I get purple skies in the DNG you have shared with Irfanview and FastoneImage Viewer and I guess I would get the same with the purple hut roofs in your second picture too. It’s probably more technical than this, but its a colour space issue imo.
The metadata and transforms are there. It is just that different apps read and process tags differently. Even the humble JPG looks different on different apps one way or another, unless you are a real pro at this stuff. Although I use Irfanview, it isn’t the best app for colour management, just as Windows isn’t the best OS for it.
