This is standard offer from “Nordsee”, a fish store chain in Germany. I buy it sometimes when I am on the road. I discovered something there once. I like the taste of the fish and regularly ask the salesman to give me a version without tartar sauce. I just wanted a little piece of lemon with it.
You can’t imagine how the wonderful flavor of the baked fish comes out! A few drops of fresh lemon juice still refines it without affecting the taste of the fish.
Normally you can’t have the fish baguette without tartar sauce without ordering it separately!
Because it is part of the recipe and customers like to eat it that way. Although the tartar sauce which tastes very sweet, drowns out the taste of the fish and brings extra calories that I don’t have to have.
And this tartar sauce does not taste bad, but it is always served with the fish and if it is used everywhere, just tastes boring.
As if the “Nordsee” people were afraid that their products will not taste and they play it safe by simply smearing tartar sauce everywhere.
The question is, why do people like it? Because they get it permanently so served. They just want a quick snack and don’t want to waste too much time choosing seasonings. And because the remoulade always tastes somehow, they don’t care and their taste has been shaped that way. They simply cannot imagine eating it with other spices.
And now, when I invite someone to my house for a fish dinner, they always ask if the tartar sauce is there. I can also offer other sauces, but this must not be missing.
I think the whole discussion is somewhat crazy. Personally i think that dt should be most flexible and should start developing an image in a neutral way. Nothing fancy, no inherent filtering/denoising/coloring … whatever. So nothing to achieve an ooc-jpeg like look at all but something that tells me most about the image data itself. Also i think that is roughly what the dt policy and concept of long-time developers has been for many years. And there are very good arguments to do so.
If this is something users dislike or don’t agree on, very well - no problem. And if some of us think the learning curce of dt is too flat and people get frustrated within a short time this maybe might be saddening me but for sure this will not make me into coding for dt for a fool-proof target.
If some of you think dt is bad for people starting to do raw-editing, that may be so. But why wink at the developers? If there is a genuine interest in support for “beginners” there are much better and more productive ways: just one example. Develop styles and presets on your own. Document what they do, how they work and for what situations they would be fine. After making sure everything is of good quality offer them to others!
I’ve tried to avoid this becoming a dt thread. I also don’t intend to influence devs. Im a RT user and setting flat in camera style with small contrast and saturation boost allows the auto match curve in rt to give me a great starting point. Im not arguing in self interest.
I just see arguments being made about defaults, sooc jpgs etc. that seem not fully thought through or articulated. This thread was intended as a place to discuss those things.
You repeat the idea that good defaults are for beginners. I don’t think thats true but would be interested to hear more about why you think so.
If most users do 5 things upon opening the file perhaps its worth setting that default or otherwise enabling that out of the box look.
The rest of the post was a bit advanced for me but I think you are making assumtions about why people prefer tartar sauce. I think those assumptions are wrong.
That may be. And you see where a very abstract discussion can lead.
So my suggestion:
If you are having trouble finding a good default starting preset in darktable for yourself, please give me/us a couple of raw files as examples of your photo work and I will be happy to try to help you create/find that preset. Maybe this attempt will be enlightening for the other users and even developers here.
Anything else is just tedious talk that we’ve already had here enough times.
Besides being a deathtrap to me due to allergies, it isn’t the tartar sauce that is disturbing: it is putting breaded fish in bread! I know it is for texture, colour and to keep it together but no thanks. It is like how people ask Canadians why our milk comes in a bag with bags of milk and then (sometimes double) bag the bag of bags of milk at the grocer.
I am getting to something, honest! The problem I have commercial products like ACR/PS/LR is that the defaults are already OOC presets decided by the company and camera manufacturer. (Edit for clarification: You have no choice but to edit on top of / in addition to these presets.) Although there is a certain amount of play around this using sliders and other controls, you have no control over what they have decided for you, hampering your vision. We don’t have this problem with our favourite FLOSS apps, so presets can be friendlier in their context.
Using the fish baguette analogy again. My mom tends to customize and transform the fast food or takeout item so much so that she might as well make her own food to taste with no more time and effort. Of course, if she were to make her own baguette, it may be inferior because she now has to contend with determining how to make it, buying the ingredients and not messing up by over or under cooking the fish, etc. Takeout is usually more consistent and safe (if the kitchen staff adheres to food safety guidelines).
Thanks for the generous offer but this was intended as a meta discussion. You’ve put forward some ideas as to why you want a flat starting point which was interesting to hear. Ideally we would then be able to more precisely define what kind of flat but it seems we might struggle to get there. If you can it would be great to try though.
Do you want it to be as “correct” as possible or just low contrast/saturation. If you rework things from memory and intent does it matter?
Nice! But, while this effect most likely exists in perception, we do not know how visually strong it is in the cases we are looking at/which are discussed here. It’s a postulate that they are SO strong that users don’t learn anything if they apply them. I disagree with that postulate.
Then comes the question: well, if it is a strong effect, isn’t that an argument for several strong ‘looks’ and not one ‘neutral’ one which we know can never be fully neutral? (The funfunfun questions of how should out-of-gamut colors or luminance-levels be neutrally displayed on a limited gamut/peak-luminance display come to mind)
This effect (of unknown strength in the cases we discuss here) could also be helpful to demonstrate to newbies in order to form their skillset.
Just shouting at newbies: ‘NO! good looking is a bad starting point, GET SOME SKILL!!!11’ (I paraphrase, I hope this is not too toxic) certainly cannot be the answer. If people continue to bring this discussion up, there needs to be a way of adressing that and not a forum-wide gag-order or devs who are so stressed by just bringing it up that every discussions turns into…what we’ve seen.
There are so many edge cases were there is no ‘correct’ but only a bit less worse than others.
Could you define one …after all this discussion I can’t understand this statement…what would this default be and who would define it other than the individual. We have determined that each software has its default look so the user should migrate to that software not try to ask someone to make their software look like another product…this is like the dog chasing his tail…
Thats the topic for discussion! @s7habo has put forward some criteria but it needs fleshing out.
My ideal would probably be something you can export for view without further editing. It should also respect the exposure to a large extent. The perfect file would be like one of the more conservative basecurves but with emphasis on accurate colours and better handling of shadows and highlights, slight but effective sharpening and conservative noise reduction at high iso.
RT pretty much gives me that so personally I’m quite happy. I always have to tweak high dynamic range photos and perhaps more intelligence could give an even better starting point from that perspective. Like applying a light log on high dynamic range images. One thing to note is that my camera tweaks the jpeg slightly depending on the scene. The auto matched base curve won’t look identical for every file but this appears to be correct because the results match well.
The last comment of the day for me is that processing starts with vision and image capture. We have the wrong idea that what happens before raw processing doesn’t matter!
It absolutely does! It not only affects the outcome but the preparation, capture conditions and goals greatly affect the properties of the raw files you end up with, granted sometimes you have no control over that because they are old photos, someone else’s or there was some unforeseen or unfortunate issue. If I am taking a photo of a high dynamic range setting such as a sunset, what will I do with it in post? If I am going to use the filmic method, what do I need to consider to expose the subjects in such a way that filmic is able to accentuate what I want to pop? The same questions could be asked no matter what workflow or software you use. You see, there is much more to photography than putting images into a blender and hoping that they somehow turn out great.
That’s a good point. For the whole process to feel transparent the connection between the decisions on site and the pp process needs to be carefully considered. Of course we are stuck with what gear the manufacturers produce but the idea of a more continous workflow is a great one to consider.
There are a lot of discussions about these topics on pixls. Much of the reason for this thread is that I think these colour theory programming discussions could also inform discussions about what can be done with these methods to improve the out of the box experience. The last step with these methods is which one should be default and with what settings. It happens but the last step isn’t very explicit.
So I think that is what I was getting at RT matches both “your” goal and your visual style or preference so that is what you prefer and so that becomes your choice. I think as long as the software offers quality tools to allow users to craft their vision then its doing its job. However, how fast or easy or accurate those tools are in the hands of users are things that you can discuss and I think more important than subjective default starting points. The scenario that you offer is like JPEG+ basically look like the jpeg but have that with the raw data behind it to improve shadows and highlight editing and I think given the number of times this keeps coming up that is the desired result for a lot of users. In fact I believe many people that start shooting raw for the first time look at it with the mindset raw is just a higher performance setting in their camera that makes larger files and will look much better. When confronted with the reality that this is the process ( Developing a RAW photo file 'by hand' - Part 1 ) going on in the background and its not so straightforward they still want that ideal to come true. So they want software that will give this higher quality image as if it were as simple as making the setting choice on the camera…There is always room for debate about the tools in the software and usability, speed and quality of result but I still think trying to nail starting point that can be deemed “good” is pointless as taste perception and workflow is quite unique.
I think this is exactly the reason for the LR presets. Software comes with a bunch of looks many many and you sift through them and you find one appealing. Chances are you land back on that time and time again because that is the look you want. But there are 10’s or 100’s of looks in these catalogues and they are there for a reason. This is because there are that many and more ideas of what the user thinks the image should look like or the image will look good if it looks like that…
No, this is exactly what makes this discussion - and others about roughly the same topic - so never-ending fruitless. We both want a “good” starting point. The problem is how to define “good”. For me it’s a neutral image where i can see what i would have to do for a perfect endpoint. I can only guess but you seem to want something good-looking. This is definitely something else.
I don’t drive cars at all so i am also not personally affected. But -what is a good car? a) A Porsche? b) A truck? c) An old heckflosse? It’s all either a matter of taste or what sort of tool you want.
Raw-processing wise, we have different tools available, take the one suited best for you. And - again - if you want to improve one of those apps, go ahead! Write code, write docs, create content, create styles … just do it. I also started working on dt because i needed support for a camera i had and found a lot of fun in doing so. I found all people developing dt very friendly and productive, sometimes having different views than i have, but always constructive and i learned a huge lot. (About C programming, what makes such a large project working, about color science, demosaicing and mucht more).
So, keep going and enjoy your perfectly developed photos!
There are levels between nothing (which doesn’t exist all starting points are designed) and 2000 presets. What the presets do show is the enormous value in subjective decisions. Perhaps it depends on how you are trained and what you do but to me it’s pretty obvious that few things are more valuable than subjective choices made by well informed and talented people. I mean I get paid for my subjective decisions so I would say that…
Yes! All arts related professions and people spend their time with this question. It’s not mystical or impossible it’s just work. Like programming but different. Many here are obviously not used to it and react quite strongly to it but really its just work. Done by thousands of people all the time.
I have a copy of ON1 photo raw. THis is a commercial product with a catalogue of looks. You can dial in the opacity of them yada yada. It is a competitor with Luminar LR etc etc. They have a feature which they call AI automatch…so wow AI…it should be good. Its crap at least for my images. It almost always lightens the image no matter what the tone of the image and this is supposed to be an AI match of the jpg file. They supposedly have a decent budget and do a lot of research and the result is a far worse starting point than just neutral or linear. They have a second feature that you can use on jpgs as well called auto tone. It actually works better but still I end up adjusting most of the sliders if I try to edit an image after this so I might as will start myself as I could be fighting slider changes that I did not add to get to where I want to be.
Luminar is full of AI-ness and neither it or ON1 are expensive or require licences so if they were so great and AI was so great for automated starting points I think everyone would pony up 60-100 bucks and just have at it.
That doesn’t happen. There are lots of complaints about that software etc etc
So even software that markets itself really about how easy you can get to a great image really only works for a subset of users that like what it offers.
Darktable had something a bit like this with the old basic adjustments panel. There was an auto slider there that would move contrast brightness etc based on the whole image or a selection. You could by selecting different areas of the photo come up with a pretty pleasing solution quite often. Other times not so much but even then you still had to go to work so it might be better to build the image and add the layers so you know what is there and not have contrast coming from 4 spots that are pushing pulling one another…
