Lightroom Classic: Would you please improve the auto enhancement functionality?

  • 2
  • Idea
  • Updated 4 weeks ago
  • (Edited)
Dear all,Years ago, Adobe made quite some effort to improve its auto enhance functionality. Nowadays I don‘t experience noticeable progress anymore even though paying for Lightroom Classic on a monthly basis.What I would like to see is a more advanced version of Sensei, meaning content sensitive image enhancements which is different for different types of photography as well as different times of the day (night shots don’t need to be as bright as shots taken in the sunlight!).Moreover, being a wedding photographer, I would highly appreciate a button to enhance all images from a wedding in one go and even align them with respect to Color and Tone! Today all this has to be done manually resulting in hours and days of work after a wedding.Realistically, there is no way a computer does the entire enhancements without any manual steps, I completely agree. Still, why is Adobe not investing in more advanced functionalities to provide us with a good basis to start with? And no, I do not want to upload all images to the cloud first. Having a computer with some reasonable hardware I won’t to start on my local machine as soon as I am back home. I do not want to waste my time by first uploading thousands of raw images to the cloud!Your opinions?
Photo of chrisb

chrisb

  • 11 Posts
  • 8 Reply Likes
  • frustrated that there are no noticeable improvements even though we are paying on a monthly basis

Posted 1 month ago

  • 2
Photo of Edmund Gall

Edmund Gall

  • 97 Posts
  • 30 Reply Likes
A lot there to unpack, but we'll give it a go.

"What I would like to see is a more advanced version of Sensei, meaning content sensitive image enhancements which is different for different types of photography as well as different times of the day (night shots don’t need to be as bright as shots taken in the sunlight!)."
  • You're asking Sensei to make an educated guess, and from the tone of the post, it sounds like you want it to be 99%+ accurate. How would Sensei do that? Let's use your example: Sensei has to make a guess about the time of day when the shot was created. Ok. Can't use tone: a dark photo could be shot at night, or shot in the day in a cupboard. It could use date/time metadata: but if you're like me, you'd leave your camera set to your home timezone. So, if I shoot a wedding in another country, night-time in my home country might be day time in the destination wedding. So, maybe it can work for those who don't travel outside their timezone.
"Moreover, being a wedding photographer, I would highly appreciate a button to enhance all images from a wedding in one go and even align them with respect to Color and Tone! Today all this has to be done manually resulting in hours and days of work after a wedding."
  • Do you know about LR's Match Total Exposures Develop feature? It does what you asked, and has been around for a while. Have you tried it? If so, does it work? What issues do you have with it?

"Realistically, there is no way a computer does the entire enhancements without any manual steps, I completely agree. Still, why is Adobe not investing in more advanced functionalities to provide us with a good basis to start with?"
  • Well, the definition of 'good basis' is debatable. If I shoot a wedding or any event with the odd test shot using a grey card (or, even better, a ColorChecker Passport), then I can generate at least the correct white balance (or a colour profile) that gives me a good basis (IMO) to start with for a series of photos (until lighting conditions change and then I just use the relevant test shot for that series).
  • AI can do a lot, but reading my mind is currently beyond it's capabilities when it comes to developing my photos.
"And no, I do not want to upload all images to the cloud first. Having a computer with some reasonable hardware I won’t (sic) to start on my local machine as soon as I am back home. I do not want to waste my time by first uploading thousands of raw images to the cloud!"
  • Sensei is based on machine learning. The early stages of developing an AI tool based on that needs a heckuva lot of data and processing resources. Maybe in time Adobe can release a smaller version of Sensei (perhaps called Genie? If so, Adobe, consider adding my name in the credits) that resides on our computers. But for now, your photos (or Smart Previews) have to be up on Adobe's CreativeCloud storage for Sensei to work (especially if you want it to make 99% accurate guesses). So, if you don't want to upload, well...

I'll give you the benefit of the doubt and say you aren't asking for a one button solution that allows you to import your wedding photos so Adobe LR can do everything magically and -hey presto! - you can send to your wedding clients before they wake up the next morning. Sensei has only been around for a relatively short time: so I'm sure Adobe's developers aren't sitting on their thumbs wondering what next to do with it. Your content-sensitive Auto Enhance is a nice idea in principle (but tough to execute, so might take some time to develop).

I also think any photographer would love their processing tool to take them some way to the final version, but even with AI-driven processing, I'd expect to have to put the hours in to get the quality I want. To produce exactly what I want, it's reasonable for me to expect to have to make the processing decisions: otherwise, what would differentiate me from you as wedding photogs?..
Photo of chrisb

chrisb

  • 11 Posts
  • 8 Reply Likes
Dear Edmund,

thanks for your thoughts!

Please don't get me wrong, I do not expect machine learning to work without learning. It's just that I don't see any further improvements in Adobe's auto enhance functionality. That's different with mobiles which take better images year after year - and taking / editing pictures is not even their only purpose.

Please allow me to continue the discussion based on your answer.

I guess your argument regarding the time zone is a reasonable restriction I haven't thought of while writing my post. This problem could be handled in multiple ways, e.g. a checkbox whether date/time is to be considered. Besides, its only a couple of clicks to adapt the timezone afterwards (something I usually do in order to align my photos with any photos being taken by mobiles). Cameras nowadays even have GPS modules included which could be used in addition.

Thanks for your tipp regarding Match Total Exposure! Yes, I tried it. However, for me it only works with photographs being taken in the exact same environment with the exact same people in the exact same clothes... I would love to use it when the same people are being photographed in different environments. Sure, the light in the scene has an impact on their skin tone, e.g. people look different in the sunlight or indoors next to a fluorescent lamp. What if Lightroom recognized Color-Checker Passworts automatically and adapted the white balance automatically? In that case, the color temperature could give some indication on whether a picture has been taken indoor or outdoor and how to correct color and tone. What if the GPS info was linked to some weather database?

Correct me if I am wrong, but I thought that the object recognition on Apple's iPhones is executed on the individual devices and not in the cloud. If that's correct, I would expect that Sensei does also work on our local computers (which are more powerful than little phones). If it helps the auto enhance functionality, I would even add some tags to my images (assisted learning).

Looking forward to hearing from you,
Chris
(Edited)
Photo of Edmund Gall

Edmund Gall

  • 97 Posts
  • 30 Reply Likes
Yep, agreed on most points.

The timezone example could be worked around by changing workflow to Edit Metadata date/time in LR in post before clicking auto enhance.

I don't work for Adobe, but from what I've read, my understanding is that Sensei operates only in CreativeCloud. No idea about plans to enable that to run on local device via LR Classic or LR CC Desktop. I also don't know about Apple's object recognition; my answer above was based on my knowledge of typical machine learning/AI-driven development.

Sensei-driven auto enhance is a reasonable request, and just because I can't figure out how to achieve it, it doesn't mean nobody else can. Looking forward to seeing what the developers at Adobe come up with (if it's added to their pipeline), even though the pragmatic side of me doesn't expect to be using auto enhance to get more than 25-33% (if so much) of my editing done...
Photo of Victoria Bampton - Lightroom Queen

Victoria Bampton - Lightroom Queen, Champion

  • 5326 Posts
  • 2105 Reply Likes
There's one other glitch you may have missed... if computers can make amazing professional photos by themselves, then people no longer need to pay wedding photographers. ;-) (Ok, I'm being cheeky, my Dad was a wedding photographer and I grew up in the business, but you get the idea!)

Long term, the request has legs, but in the shorter term, if you don't like spending days editing, there other professionals who can do it for you. It's a skill like any other. 
Photo of chrisb

chrisb

  • 11 Posts
  • 8 Reply Likes
If computers can make amazing professional photos by themselves...

Well that’s the point, they are not taking them but only a tool to make them look good and aligned. In my understanding professional does not mean to spend more time but to get to a good result. In any other job, computers also help to get the job done faster. So what’s different in photography? Starting with a solid basis does not mean a photographer is no longer allowed to add personal editing style?
(Edited)
Photo of Edmund Gall

Edmund Gall

  • 97 Posts
  • 30 Reply Likes
If you read Victoria's whole response, Christopher, I think you'd realise she's agreeing with you: her opening paragraph was just for fun...
Photo of Victoria Bampton - Lightroom Queen

Victoria Bampton - Lightroom Queen, Champion

  • 5326 Posts
  • 2105 Reply Likes
Tongue in cheek doesn't really work over text, does it! :)
Photo of chrisb

chrisb

  • 11 Posts
  • 8 Reply Likes
Sorry, didn‘t get that;)