Skin retouching, AI, Client-Server, etc.
I have a lot of ideas that might improve Capture One and make it "ready for the future". What I mean by this is, from my point of view there will be a time in the future where "just being a RAW developer" might not be good enough anymore. Don't get me wrong I definitely like Capture One and think it mostly does things right, but it could really benefit from additional features. I for example would like to use only one software for the complete image processing workflow and getting rid of Photoshop completely even for people photography and skin retouching. I also think there will be a heavy emphasis on AI in the future seeing what for example Apple is doing with their new M1 chip having 16 special neutral learning cores dedicated to AI.
Creating Masks: First request and basis would be a computer supported creating of masks for skin, lips, eyes, hair, etc. It's rather time consuming doing this manually and here AI definitely has the potential of assisting and maybe even doing it completely by the software using AI.
Skin Retouching: Second request would be the necessary tools for skin retouching like getting rid of pimples, wrinkles, blemishes, etc. in a very easy way and still keeping the results natural.
Skin Smoothing: Adding to that I would like to see skin smoothing tools without making the result just soft like reducing clarity. And still the goal would be to keep results natural.
AI & Multiple Images: The next request would be the most time saving of all and tying above mentioned requests together. With the help of AI I would like to be able to do skin retouching in "my way" and getting the results I prefer, and than applying these adjustments and retouching to the rest of the images from that shoot, and let AI figure out the adjustments in the correct place. Let's say I remove a pimple in a specific location of the face I would like Capture One to figure out the location of this pimple and the retouching steps I did to it, then figure out the location of this pimple in the other shots even when the model has moved the head and apply the same adjustments to the other shots.
The same is true for all other adjustments I did, so basically I would like to apply all kinds of adjustments to one photo and let the AI figure out the necessary adjustments and the locations to apply these. And I know this is definitely a huge request and will take time and probably multiple versions to get into a reasonable and working state.
HDR Sliders: Another completely different request is, I would like to have the HDR sliders in a different way. Instead of having sliders for highlights and whites on one side and slides for shadows and blacks on the other side, I would like to have a single slider and using the luminosity range to chose the range of adjustments. And let users create as many of these adjustments/adjustment layers as necessary. This way the user has full control over the adjustments. This could even be done with the colour to adjust the brightness for a particular colour or colour range.
Number of layers: I would also like to see the maximum number of layers be opened up significantly especially keeping the above mentioned feature request in mind, since these will definitely increase the need for "more layers" to not hit a rock.
Client-Server Architecture: One more request would be the possibility of having a real client-server architecture. By that I mean having the RAW files on a NAS for example and the catalog on one or more client computers like desktop and laptop for example, and still having the catalog consistent on all of these computers, and at best this being done in real-time to use these machines at the same time and switch between them.
-
The alternative view is that it is still better for Capture One to concentrate on being primarily a raw developer, and leave the other kind of work to tools designed for the job. There are good reasons why Adobe have Lightroom and Photoshop and not some hybrid Lightshop or Photoroom.
(I accept of course that there is a discussion to be had about how far being a raw developer should go, and what features are suitable for that aim. And I suppose that it depends on the kind of work you do. For myself, I only occasionally find it necessary to go on from Capture One to something like Photoshop or Affinity Photo, but I am aware that there are some genres of photography where progressing from Capture One to Photoshop is the norm for most images, not the exception. But for those of us who only need PS or AP a small proportion of the time, it might be preferable to not have Capture One bloated with extra features that we would not use much.)
Ian
0 -
I respect everyone has different needs and therefor not everyone is in line with "my" feature request and ideas, that's completely fine with me. Still I would like to see these features in Capture One and maybe even other ones. And this might or might not be where programs are heading to in the future. Some programs even start doing this and even if they are not there yet especially in terms of quality, but they will probably improve over time as well.
And honestly I don't see it negative in a way "when too many features are added they will most likely be implemented very poor". I don't think it has to be that way, but it will take time and only one by one since there are only so many resources available.
Not sure, maybe something like turning on/off additional features would be an idea. So everyone wanting them can turn them on and they get downloaded and installed, users not wanting them just don't activate them. Just thinking, this way at the end everyone would mostly get what she/he wants.
0 -
Isn't the "optional use" where plug-ins came from a long time ago in Adobe and other products?
Also have you taken advantage of Portrait Pro? (or whatever it may be called these days).
I'm looking forward to a time when an Elon Musk instigated brain controller understands what I am thinking and simply creates the image without the need for a camera at all, complete with an assessment of its general acceptability (so as not to create something that might upset someone) and a complete analysis of its possible global offensiveness and revenue generating potential. Also whether someone else has already "thought" of it. I would not want to break any Imagination rights regulations.
0 -
I like the idea of AI for skin retouching. If I watch all the C1 tutorials, I have the feeling this is one of the main features of C1. So why not use the technology of 2021. Even some little iPhone apps can do it, why not C1? I don’t know why LR not already have included it, because they have face detection…
-1 -
I support the request to increase the upper bound for number of layers.
Regarding the use of AI, there are generally two types of applications:
1. Multi-step employment of operations with predictable outcomes. The effect of operations is known and the user employs them accordingly. To achieve a compound affect it is sometimes necessary to combine several predictable operations.
2. Single-step invocation of complex, "smart" operations whose outcomes vary per image content. "One click"- or "single slider"-operations are designed to achieve a complex result which, however, will have variable results depending on image content.
I regard the first category as "tools" and the second category as "shops". The shop will typically be quicker than yourself but the product will often not quite be what you would have produced while having full control over operations with predictable outcomes.
The moment Lighteoom started to convert from a tool into a shop (with "smart sliders" that claimed to know better what the user wants than the user itself) I left it and went looking for an alternative.
There is of course a legitimation for both tools and shops, but I for one hope that C1 will remain a tool.
P.S.: I'm not at all against automation as long as the outcome is entirely predictable. Anything requiring judgement, however, should be left to the user.
P.P.S: Obviously "smart" functionality on top of reliable tools does not hurt those who don't like to use it, but I'd rather see C1 development resources targeted at removing the many remaining productivity hurdles, than spend on turning C1 into a shop.
0 -
@Class A: I’m all in for the users having full control of the adjustments and not letting it up to some fancy yet sometimes weird AI doing all kind of stuff.
I’m more in for AI supporting the user. For example when creating masks like skin using some “clever algorithms” taking colour, tonality, and things like this into account. Even more so when applying adjustments to one image by the photographer and afterwards copying these adjustments to the rest of the shoot. For example removing a pimple in someone’s face in the first image, this would be the job of the user. Than copying the same adjustments to the other 149 images would be the job of the software, that would be a tremendous time saver.
Don’t get me wrong I still think it’s a lot of effort and by no means it’ll be perfect from the start, but I still think it’s worth it, and I also think that’s where a lot of software will head to in the next years.
1 -
@All, I'm almost scared about the new (fake) AI technology how to make unsharp photos sharp again. I rather like to have automatic masks and maybe auto levels, but just as an option. I don’t like these new influencer "make me beautiful – skin clean – make eyes bigger" etc. software. I also prefer to have to contol over everything;-)
Like back in days using filters in Photoshop. The best way to use them, is you don’t see, you have used them a lot…
0 -
As I said I don’t want AI to adjust images on its own. I want to make adjustments myself, but only on one image and than copy these adjustments to the rest of the shoot letting AI figure out exactly what and where and how to adjust to match the adjustments I have made at the first image.
1 -
okay i got it like my panoramic software ptgui is able to recognise the same spots in different images, so, it would be great if c1 is able to match same spots too.
0
Post ist für Kommentare geschlossen.
Kommentare
9 Kommentare