For all the potshots about AI, this update is huge even if you take away the AI features. They basically added lightroom to this release. There's some polish before you'd want to change your subscription, but its really tempting. It may be the best photo management/editor on linux. Yes, I know about darktable and rawtherapee and I stand by what I said. They also added a ton of motion graphics stuff which from the beta seem to be enough to undercut a lot of basic uses of after effects out. The later two features are in the free release as well!
I made the switch from premier to resolve a few years ago and it feels like such a breath of fresh air. Being able to do the same with Lightroom would be amazing so can't wait to check this out. I've been using the free version and honestly never needed the pro features but I think I'll make the one time purchase today just to support a non-subscription based product of this caliber
> ...before you'd want to change your subscription...
For anyone not in the know, Resolve has an exceptionally capable and feature rich free version. A lot of the AI features (and >4k editing) are locked to the Studio licence which is a one-time payment, but works simultaneously on two computers (including different OS's) and allows upgrades across major versions.
I spent less than $300 on it a decade ago and my licence works fine on new v21 released this week. My least-regretted software purchase in 3 decades.
I run a photography biz on the side, and this is huge tbh.
I already used Davinci to make a custom LUT, blended it with the Adobe camera standard profile to use in lightroom as a base for all my edits as Davinci's color tools are much better, and doing it this way just lets you get tones that you couldn't other wise get with Lighroom/ACR alone. This basically removes the need to have lightroom in there as a finishing step.
The only downside is by now I've got a really solid ImagenAI profile based on the 30k photos I've fed it over time, and that obviously relies on a lightroom catalog, and is also capable of applying my LUT since I've made it a slider in LR. I hate adobe, but I'm not sure I could go back to not using Imagen as now I can turn around a full wedding gallery in about 2 days.
I haven't had a chance to look at the non-beta version but in the 21 Beta, the Photo page didn't support Lumix or Olympus raw formats, and I own two cameras: a Lumix and an Olympus. :-(
I assumed they would add them later, I hope I'm right!
edit: nope, still only supports Canon, Fuji, Nikon, Sony and iPhone ProRAW.
Is the Lightroom stuff the same as was discussed on HN a couple months ago? If so, then, see also 296 comments about the new photo mode: https://news.ycombinator.com/item?id=47760529
Can you run batch processing with the same settings over a large sequence of images like in Lightroom as well? And/or make slight changes to some parameters at certain points of time like correcting light intensity during sunsets the transitions smooths when your hardware made a step change when capturing it?
I haven't tried these new image features at all, nor do I know how well exposed it is in the scripting API in the new version, but Resolve does have a Scripting API I've used for batch processing (of videos) in the past, I'm guessing this gist will soon be updated for version 21 then you can take a peek there: https://gist.github.com/X-Raym/2f2bf453fc481b9cca624d7ca0e19...
Output size is limited to 4K in the free version, I think. Which is not nothing (8MP or something, good for a reasonably large print) but it might make you question how much editing belongs inside Resolve.
It might be best editor on linux but running it on linux is not easy. You basically need pick correct hardware pick right distro. It might be pretty easy on the oficial rocky linux but on other distros good luck. Also no AMD support.
I am pretty experienced linux user and i would have to buy new NVIDIA card and pray things work out.
Sadly i found out it will be much cheaper to buy refurb mac mini. So now i have dedicated machine for editing video.
That wasn’t an AI tool at all, though. Neither is pre-2023 content-aware fill, AFAIK.
They are both PatchMatch (well the healing brush certainly is), which is a heroic bit of code. Entirely deterministic statistical algorithm. Not AI by really any definition (including back then)
Probably not you, but on this very forum there are plenty who will argue that LLMs / AI are entirely deterministic and that given enough time, a chisel, and sufficient clay tablets, any AI output can be calculated by hand.
I don’t know enough about anything to determine which opinion is correct.
Is AI output not just matrix multiplication and a random seed?
I don't disagree necessarily on the fundamentals; I am slowly catching up on what LLMs do and don't do but that sounds right to me.
But what I would observe is that the healing brush does not have a random seed. It will always do the same thing if applied at the same pixel.
(I am actually less sure if content-aware fill randomises; I always got the impression it did not)
This makes it both incredibly powerful and occasionally frustrating.
Because on the one hand, you can learn to apply your judgement to precisely control what it will do, and change the radius or position if you learn it is likely to fail, which becomes instinctive. I absolutely love using it to fix scratches in film scans; it's a quick, precise, controllable tool that can be used in a way that is amazingly convincing, and it ends up quite a "zen gardening" thing as a result. It'll sell you on the cheapest wacom pen once you know how efficient it can be.
On the other hand there are situations where it simply cannot work the way you want because it will always find a pattern you don't want it to.
(You can sometimes use the clone brush tool first, to manually break up the pattern that patchmatch will find)
> But what I would observe is that the healing brush does not have a random seed. It will always do the same thing if applied at the same pixel.
Given a model architecture that supports it, greedy decoding + the same inputs + prompts, that's true for most LLMs today too, I don't think people consider them less/more AI because of that.
Is there any RAW processing software for Linux that works for DJI drone photos?
I have a bunch of photos from a Mini 3 Pro and Mini 4 Pro -- 4 years' worth -- that I haven't published because I don't have a way to process them.
No tools on Linux (RawTherapee, Darktable, RapidRAW) render their colors correctly no matter how you mess with the sliders, and all of the Github issue pages are dismissive of the problem. There is something fundamentally wrong about how Linux RAW libraries are reading DJI photos.
Lightroom on a Mac tested on another computer renders them all correctly, but I don't own a Mac nor Lightroom.
>They also added a ton of motion graphics stuff which from the beta seem to be enough to undercut a lot of basic uses of after effects out.
I moved my team to Resolve about 2.5 years ago and I can say with absolute confidence Fusion > AE. Resolve Studio is all you need, period. Lightroom is clearly better right now than their Photo editor but like all their other offerings (ehhh except fairlight lol still mediocre), it's only a matter of time
This has not been my experience. I have frequent crashes with Fusion that seem to be related to specific nodes. Checking my computers usage it doesn't seem like I'm running out of juice either.
When it works, it's really nice and I'm sold on the node based workflows, but it's not all entirely roses.
Resolve itself is excellent, Fairlight works just Ok and is still heaps better than Audition ever was. But Fusion I feel like hasn't seen enough love in the past few years.
The whole Reactor setup for installing Fuses feels really broken and for the free users they disabled those entirely in version 19.1.
That's like saying all fine art would be photography, all film would be CGI, or all music would be synthesized electronica.
That's not how aesthetics seem to work. Artists will make more or less good use of generative AI in their work, and it will probably seep into most media in some way or another, effecting them, but arts mostly don't get replaced and AI doesn't really offer an exception to that history.
> That's like saying all fine art would be photography
Or is it like saying most portraits will be photographs rather than paintings? There are still a lot of portraits painted (maybe even as many as the pre-camera days), but by raw numbers most portraits are created by photographers.
There will also be a longer or shorter period of time in which such technology will be abused by artists (because it's new) and at some point it will stabilize.
Aren’t the cameras they are making aimed at professional productions? Those are probably going to replaced last, the first thing will be (or are) TikTok clips shot on smartphones.
I don’t think they’ll see a decline in cinema camera sales due to AI soon.
Yeah they don't make a single consumer grade product. Prosumer at its lowest, and even then they're way too technical for the uninitiated. No one is picking up a BMPCC and just shooting/posting online.
I doubt it because I wouldn't waste my time watching it and I can't imagine all that many people today that don't already watch AI videos are going to suddenly change their mind and decide they like AI produced Hallmark movies.
Even the frontier models running on insanely powerful hardware could only generate 15 second clips in low resolutions.
And yeah, I saw some demos from Seedance 2.0, and they were awful. It's ridiculous how much people on Xitter were like "You can't even tell it's AI!" and I was like "It's trivial to tell it's AI" and could easily pick out all the markers. An individual screenshot could look good, but every time the camera angle changed, there would be a glaring inconsistency.
You people are either blind, delusional, or outright insane. AI might be used for a quick clip, or used to enhance something recorded by a camera, but "most video" is definitely wrong.
You are only looking at your own consumption at the moment. There are a lot of problems that still need to be fixed especially with rope artifacting. The 4k most models taunt isn't equivalent to a real 4k image or video as well at the moment you need a quality factor of two to get the equivalent result of a shot image or video. Resolution does not indicate quality.
So much respect for Black Magic. They are absolutely World Class and their business model is extremely generous.
Having said that, for all the AI features, the big one would be setting key frames etc. with an agent, driving the general editing workflow with text,etc. I realize this is non trivial but it's certainly viable for a team of this calibre.
I think if BM added a paid for agent which helped execute their traditional video editing tools (even if it "only" supported a subset) then that's a subscription a lot of people would be willing to pay for, especially as their core tool is so generous.
You might be interested in a tool we're working on which does some of what you're wanting https://sparkfxstudio.com/ It's presently in beta but is an AI tool for helping speed up video workflows using agentic AI.
There's a Linux amd64 version and a Windows arm64 version. However, I use a Linux arm64 machine, and arm64 machines are going to be a lot more common going forward. I wonder if there are plans to release a Linux arm64 build?
people complaining about AI features have clearly never wasted hours editing video or lost time and money discovering a technical flaw in a rush shot three days ago. For actual workflows, these tools are lifesavers
Because those people see the phrase "AI features" and the first thought is those sloppy generative AI stuff where things shift.
Where as marketing at all these corporation is trying to genericize "AI features" into anything using an algorithm.
"Content aware fill", something we've had for over a decade is now "AI object removal"
I’m not sure I agree? For each of your examples there are algorithmic approaches and neural network approaches. Companies have certainly been loose and wild with how they market these, but there remain distinct approaches and implementations for each. Very generally speaking, the neural network based approaches (aka “generative AI”) perform better but with much worse degenerative cases and a higher baseline rate of unwanted side effects (that are normally not immediately visible but tend to cause issues down the line).
My bigger concern is that these neural network based solutions have taken the place of the former rather than supplemented them. Many tools no longer provide the algorithmic/kernel-based approach at all, and have marketed the “AI” (née ML) alternative as a strict superset/upgrade, despite its potential drawbacks.
(Interestingly while the inference-based implementations generally have higher latency (or infinitely worse, cloud and pay-as-you-go requirements), for some computationally difficult kernels the inference-based approach is actually faster!
I really don't understand why people are complaining about the AI features. These all mostly seem like solid quality of life enhancements and CGI-like tweaks.
some of these genuinely excite me, like the slate recognition (chore-reduction), clip search (although I'd want to see how reliable it is), and deblur (a typical PITA post fix). anything that makes masking, tracking, and level-matching easier saves hours on hours, but only if they're either reliable or easy to fix what the automation gets wrong (and it'll always get something wrong by the director, no matter how good it is, because telling editors what they did wrong is how directors make money).
the more SFX-end ones like facial aging and face reshaper, or the talent-replacing ones like speech cloning/ADR, feel both too prescriptive for a director to dial in what they aesthetically want, and also not good enough for the final cut. so I struggle to find where they'd be actually useful in a workflow as opposed to being a trap, looking just fine enough at a glance to sneak into a final cut while looking poor when viewed by the audience.
likewise the focal adjustment and upscaling just feel gross. the kinds of things a good cinematographer can already do, and it'll be so so tempting to use tools instead of taking time to do it right in camera because it'll look good enough in the editing bay, but I feel like it'd really stand out as fake in the final cut unless you're targeting like, heavily compressed social ads. the less you use them the better they'll work, which isn't ideal for a marquee feature.
if the blemish remover really does respect continuity it'd be a nice-to-have, but it also feels like another trap to be careless/cheap on things like makeup or lighting at the shoot, at the expense of looking fake in post
I think that's the broader angle that bugs me the most. all of these tools are convenience tools for editors, but in the end they'll really be justifications for directors/producers/studios/agencies to cheap out and do shittier work faster on the shoot. a cheap, shitty shoot covered in AI bandaids is still going to hit an audience like a cheap, shitty shoot.
Artists appreciate and use time saving tools. The AI features in Lightroom, for example, are very well received by photographers (myself included). Automatic subject, background, sky, body part masking, content-aware/generative fill and generative remove are genuinely helpful and time saving.
Photographers have already delegated their art to pushing buttons on a machine. They are the most receptive to AI tools, but not representative of artists in general.
What are you talking about? I don't know anyone grinding away deep in DaVinci or Flame that wouldn't love AI assistance. If anything, the studio heads will come in and insist it's all done by hand with the same energy they keep people up to 4am changing the color and shape of the eyebrows of the camel on a box of cigarettes just to deal with their neuroticism.
I felt the same when I got Vegas and Sound Forge, but they never got released on any platform other than Windows, so eventually outgrew them. I totally understand what you mean; I use it, but also happy with Blender!
One really neat thing is that you get free major updates. I bought a Blackmagic camera back when resolve was version 14 I think, and today I still use the very same license with Resolve 21.
Not that I spent any extra money on the license compared to what the camera itself costed, but I also feel like the 0 money I spent was well spent :) Harder about the time commitment to move from something you know really well to something new, but the time I spent on that was very well worth it too.
Anyone using this headlessly got a read on how much of this an agent could do without human intervention? Would love to have a gut check on "sure, spend the $295 and you'll get some benefits for free if you have an agent run your videos through this before shipping them"
To be clear, my use case is making weekly online videos suck a little less - not grading feature films :)
You can use the free version (no need to pay), and it should be possible to drive Resolve using the Python API -- see this MCP server built on top of that API:
You can also try pointing Claude to the API surface directly.
The result might be chaotic, but you should be able to automate editing... maybe. :-) Ha. (You'd need a good loop to make sure Claude can see what it's doing..? Beyond screenshots I mean - ie how would Claude know its doing a good job editing?)
They could remove the word "AI" from each one of those feature titles, and the titles would be just as descriptive without them. At this point, it's just marketing noise, more distracting than informative. Maybe like "cyber" in the 1990s. Would you like some AI tea with your cybercrumpets?
You're right to be tired of it, and of course I haven't tired these specific features yet, but Davinci was already lowering the barrier to entry for filmmakers, and if 21 works as they say it will, then you're looking at a major lowering of said barrier.
I think they've made lots of great practical choices! 100% in agreement with running local models for these tasks.
My opinion is that, for end users, if you name your feature "AI" to market it, you kind of already failed to read the room. You're writing to VCs while hoping it convinces customers.
Name what the feature does, what it gains them. Call it "smart" if you must imply some black box treatment.
Naming AI as the selling point for everything feels a lot like that Android tablet ad circa 2010:
"Your wife will love the new dual core Tegra™ chipset!"
> My opinion is that, for end users, if you name your feature "AI" to market it, you kind of already failed to read the room. You're writing to VCs while hoping it convinces customers.
100% this.
Maybe I live in a bubble, but consumer sentiment regarding AI seems extremely negative. Boasting "AI" features is more likely to lose sales than to create them.
And each is very specific to a use case, not a "general chat prompt for triggering API calls" but things like "ML model to categorize video clips and assigning tags + names, so you can find it faster" and similar.
I'd also get tired if it was "AI ala Microsoft/Google" where the goal is to get you to write forever with a chat bot somewhere else, but these features are very different from that.
Eventually, in moviemaking, generative AI is going to be seen the way CGI is. That is, how people complain about CGI when it's obvious/distracting/noticeable, but the best usages of it won't be noticeable.
Sure, and like CGI, it will change the nature of the media entirely.
Different stories shown with different treatment. With CGI, scenes zoomed out to wider shots and effects swelled even louder over lighting, intimacy, acting, etc.
Old styles didn't disappear or stop evolving entirely, of course, but the center of attention profoundly shifted and the "big" production money went with jt.
Generative AI will likely drive some kind of analogous shift in dominant film aesthetics. I don't know where, but I'm not particularly excited by it myself yet.
I don't know much about the movie industry, but I think the biggest commercial successes for early CGI was possibly Jurassic Park? I don't think it would have been nearly as good if it had all just been done with puppets.
Likewise, there's been a ton of movies since that could in theory have been done purely with SFX instead of VFX, but which is probably must better from having used VFX/CGI, titles like The Matrix, Lord of the Rings, Inception, Avatar, etc.
Obviously there was also charm in titles like E.T. and Gremlins, and I think there might still be a market for movies like that, such as the 2019 Netflix series The Dark Crystal: Age of Resistance (I loved the 1982 version as a child, though this sequel wasn't for me).
I guess the question is: will generative AI allow new movies to be made, the same way that CGI has? Or will it just be an economic shift: the same quality CGI but at a fraction of the price?
There may still be people out there who believe that AI will never be able to create CGI as good as humans can, these might be the same people who used to say that CGI can never look realistic. And if you work in VFX, I bet that you can spot a fake mountain with the slightly incorrect shadows in the distance easily, but as a simple movie watcher, I really don't see it, especially when it's only on screen for like 3 seconds.
If you look at the ten top-grossing films in the past 15 years, it's almost exclusively computer animations or SFX showcases with some token, quip-based acting thrown in every now and then (all the Marvel universe stuff).
Basically, I'm not even sure if replacing actors with AI would be stylistically perceptible for the average moviegoer. There are interesting films still being made, but almost no one is watching them, because you have the option to watch some SFX explosions and flying superheroes in Avengers VII.
I just pulled in 10 iPhone RAW files. It did a really nice job of processing them, I did the usual pulling-up shadows and highlights, and played with some of the sliders. Noise reduction and sharpening tools are primitive. Yet the photos look great. I finally managed to export ("render") them.
It's a baffling process flow if you're coming from Lightroom or a manual ACR workflow. But I'm excited to see where this goes. Quite simply, the output results are great. And free!
One of the rare pieces of software that actually gets you excited with each new release. Moved to Resolve from Final Cut a few years back and I've never been happier. Looks like this release just continues the already great experience.
For people using Resolve, would you recommend someone already quite well-versed in KDenLive to switch, for some non-profit work on cutting together educational content with some animations, some talks etc?
Will it allow me to drastically improve my workflow (save time for some tedious tasks), increase quality of the outputs etc?
If you're organizing/having hundreds of clips you want to put together, or overall want a more opinionated workflow, then I'd say give it a try at least, the free version doubles as a trial :)
I'm a Premiere migrant to Resolve (Studio) some years ago, biggest hurdle is the opinionated workflow, it basically wants you to use the tabs in the bottom to go from "Media > Cut > Edit > Color > Fusion > Audio > Deliver" (simplified) so different tools available in different areas, made for different use cases, but in general once you've learnt the overall and high-level concepts, it makes editing really easy and smooth.
Besides, it's probably the most stable video editor that runs natively on Linux since ever, I think I've had it crash once, and the Fusion 3D text doesn't work properly for me, but besides that, runs like a dream and UX is miles ahead anything else available.
Stability sounds interesting. While KdenLive has worked OKish for me, certain versions have broken my projects, and there are some long-standing bugs, like subtitles disappearing if I do a specific operation in the wrong moment. While the latter is fixed by a simple Ctrl+Z it is giving me second thoughts about using this for really large projects.
Resolve was a much better experience for me than kdenlive. But you can easily try it out for yourself because most of it is completely free (in fact you probably won't ever need the paid features for what you do)
I'm no expert (relatively new to the field myself), but I was trying to put together some simple videos with animations in Final Cut Pro and decided to try DaVinci Resolve, and I'm glad I did. The Fusion stuff bundled into it is incredibly powerful for animations.
It does take some getting used to, but the amount of tutorial content on YouTube is another reason I'm happy I made the switch. A lot of really good stuff on there. (Search on 'DaVinci Resolve Fusion' to see some examples of it in action if you want to get a feel).
Does someone know if Blackmagic Pocket Cinema Camera is a good camera since it comes with the resolve studio ? Instead of buying a separate camera in the same price rangen (sony) for studio or indoor recording ?
Resolve is an incredible tool, and I wish they improved the Linux support especially on AMD. It's the last reason why I have a windows machine, and Win11 made it unbearable to use.
Pretty sure that does not cover a face database indexing your own photos/videos, running locally on your own computer. If it did, that would be extremely silly.
> BIPA establishes standards for how companies must handle Illinois consumers’ biometric information. In addition to its notice and consent requirement, the law prohibits any company from selling or otherwise profiting from consumers’ biometric information.
Edit: Still garbage, required 100-lines script and LD_* shim to make it even run at Linux Mint 22.3 on AMD CPU + GPU. UX is even worse than Darktable, don't bother, not even close to Lightroom.
At this point we need a Kickstarter campaign to make Lightroom run in Wine/Proton (no, no matter how much you try, it will not work so far). Edit: or GSOC to support Darktable to improve their UX.
I don't hate it. I'm just not compatible with it I guess. It's like GIMP in its early days - has most of the features of competition (Lightroom vs Darktable, Photoshop vs GIMP), even more features, and tweaks, and more knobs. But it misses USER EXPERIENCE part, it's basically unusable, unless you use it since early days.
If Darktable had a grant/GSOC just to improve UX, it could be a valid competitor to Lightroom. Currently, it's not. It's bunch of Python/Lua scientific code with some UI, that processes pictures.
For all the potshots about AI, this update is huge even if you take away the AI features. They basically added lightroom to this release. There's some polish before you'd want to change your subscription, but its really tempting. It may be the best photo management/editor on linux. Yes, I know about darktable and rawtherapee and I stand by what I said. They also added a ton of motion graphics stuff which from the beta seem to be enough to undercut a lot of basic uses of after effects out. The later two features are in the free release as well!
I have both the latest Lightroom and Davinci Resolve.
Recently I edited a few images requiring removing extra people from the frame and I was able to do all editing in Lightroom, in seconds.
As much as I dislike Adobe, Lightroom’s shortcuts and flow are now habits.
I will likely continue using both.
I made the switch from premier to resolve a few years ago and it feels like such a breath of fresh air. Being able to do the same with Lightroom would be amazing so can't wait to check this out. I've been using the free version and honestly never needed the pro features but I think I'll make the one time purchase today just to support a non-subscription based product of this caliber
> ...before you'd want to change your subscription...
For anyone not in the know, Resolve has an exceptionally capable and feature rich free version. A lot of the AI features (and >4k editing) are locked to the Studio licence which is a one-time payment, but works simultaneously on two computers (including different OS's) and allows upgrades across major versions.
I spent less than $300 on it a decade ago and my licence works fine on new v21 released this week. My least-regretted software purchase in 3 decades.
I run a photography biz on the side, and this is huge tbh.
I already used Davinci to make a custom LUT, blended it with the Adobe camera standard profile to use in lightroom as a base for all my edits as Davinci's color tools are much better, and doing it this way just lets you get tones that you couldn't other wise get with Lighroom/ACR alone. This basically removes the need to have lightroom in there as a finishing step.
The only downside is by now I've got a really solid ImagenAI profile based on the 30k photos I've fed it over time, and that obviously relies on a lightroom catalog, and is also capable of applying my LUT since I've made it a slider in LR. I hate adobe, but I'm not sure I could go back to not using Imagen as now I can turn around a full wedding gallery in about 2 days.
I haven't had a chance to look at the non-beta version but in the 21 Beta, the Photo page didn't support Lumix or Olympus raw formats, and I own two cameras: a Lumix and an Olympus. :-(
I assumed they would add them later, I hope I'm right!
edit: nope, still only supports Canon, Fuji, Nikon, Sony and iPhone ProRAW.
They mentioned in notes the support will come. They support lumix videos.
Is the Lightroom stuff the same as was discussed on HN a couple months ago? If so, then, see also 296 comments about the new photo mode: https://news.ycombinator.com/item?id=47760529
Can you run batch processing with the same settings over a large sequence of images like in Lightroom as well? And/or make slight changes to some parameters at certain points of time like correcting light intensity during sunsets the transitions smooths when your hardware made a step change when capturing it?
I haven't tried these new image features at all, nor do I know how well exposed it is in the scripting API in the new version, but Resolve does have a Scripting API I've used for batch processing (of videos) in the past, I'm guessing this gist will soon be updated for version 21 then you can take a peek there: https://gist.github.com/X-Raym/2f2bf453fc481b9cca624d7ca0e19...
I was not aware that we could edit photos in DaVinci Resolve in Linux. Thank you for this info, I'll certainly give it a try!
Output size is limited to 4K in the free version, I think. Which is not nothing (8MP or something, good for a reasonably large print) but it might make you question how much editing belongs inside Resolve.
You couldn't until this release. :-P
Laat time I checked, deleting an image on DaVinci didn’t delete on disk. That’s something I’m really missing
It might be best editor on linux but running it on linux is not easy. You basically need pick correct hardware pick right distro. It might be pretty easy on the oficial rocky linux but on other distros good luck. Also no AMD support.
I am pretty experienced linux user and i would have to buy new NVIDIA card and pray things work out.
Sadly i found out it will be much cheaper to buy refurb mac mini. So now i have dedicated machine for editing video.
I suspect main reason resolve supports linux is render farms with purposefully bought hw.
Also, nobody would have complained about "content aware face fill", "AI" image edition tools have been standard for ten years in Photoshop
I remember CS2 making big news with the "healing brush", that's like 20 years ago or something like that.
That wasn’t an AI tool at all, though. Neither is pre-2023 content-aware fill, AFAIK.
They are both PatchMatch (well the healing brush certainly is), which is a heroic bit of code. Entirely deterministic statistical algorithm. Not AI by really any definition (including back then)
In my mind, it's as much "AI" as "AI Slate ID" introduced in this release, which I guess was kind of the point of what GP said.
Probably not you, but on this very forum there are plenty who will argue that LLMs / AI are entirely deterministic and that given enough time, a chisel, and sufficient clay tablets, any AI output can be calculated by hand.
I don’t know enough about anything to determine which opinion is correct.
Is AI output not just matrix multiplication and a random seed?
I don't disagree necessarily on the fundamentals; I am slowly catching up on what LLMs do and don't do but that sounds right to me.
But what I would observe is that the healing brush does not have a random seed. It will always do the same thing if applied at the same pixel.
(I am actually less sure if content-aware fill randomises; I always got the impression it did not)
This makes it both incredibly powerful and occasionally frustrating.
Because on the one hand, you can learn to apply your judgement to precisely control what it will do, and change the radius or position if you learn it is likely to fail, which becomes instinctive. I absolutely love using it to fix scratches in film scans; it's a quick, precise, controllable tool that can be used in a way that is amazingly convincing, and it ends up quite a "zen gardening" thing as a result. It'll sell you on the cheapest wacom pen once you know how efficient it can be.
On the other hand there are situations where it simply cannot work the way you want because it will always find a pattern you don't want it to.
(You can sometimes use the clone brush tool first, to manually break up the pattern that patchmatch will find)
> But what I would observe is that the healing brush does not have a random seed. It will always do the same thing if applied at the same pixel.
Given a model architecture that supports it, greedy decoding + the same inputs + prompts, that's true for most LLMs today too, I don't think people consider them less/more AI because of that.
Is there any RAW processing software for Linux that works for DJI drone photos?
I have a bunch of photos from a Mini 3 Pro and Mini 4 Pro -- 4 years' worth -- that I haven't published because I don't have a way to process them.
No tools on Linux (RawTherapee, Darktable, RapidRAW) render their colors correctly no matter how you mess with the sliders, and all of the Github issue pages are dismissive of the problem. There is something fundamentally wrong about how Linux RAW libraries are reading DJI photos.
Lightroom on a Mac tested on another computer renders them all correctly, but I don't own a Mac nor Lightroom.
Resolve runs on Linux. What am I missing?
>They also added a ton of motion graphics stuff which from the beta seem to be enough to undercut a lot of basic uses of after effects out.
I moved my team to Resolve about 2.5 years ago and I can say with absolute confidence Fusion > AE. Resolve Studio is all you need, period. Lightroom is clearly better right now than their Photo editor but like all their other offerings (ehhh except fairlight lol still mediocre), it's only a matter of time
This has not been my experience. I have frequent crashes with Fusion that seem to be related to specific nodes. Checking my computers usage it doesn't seem like I'm running out of juice either.
When it works, it's really nice and I'm sold on the node based workflows, but it's not all entirely roses.
Resolve itself is excellent, Fairlight works just Ok and is still heaps better than Audition ever was. But Fusion I feel like hasn't seen enough love in the past few years.
The whole Reactor setup for installing Fuses feels really broken and for the free users they disabled those entirely in version 19.1.
> For all the potshots about AI,
Most video is going to be AI in the near future. They see the writing on the wall. Their camera business line is going to sharply decline.
> Most video is going to be AI in the near future
That's like saying all fine art would be photography, all film would be CGI, or all music would be synthesized electronica.
That's not how aesthetics seem to work. Artists will make more or less good use of generative AI in their work, and it will probably seep into most media in some way or another, effecting them, but arts mostly don't get replaced and AI doesn't really offer an exception to that history.
> That's like saying all fine art would be photography
Or is it like saying most portraits will be photographs rather than paintings? There are still a lot of portraits painted (maybe even as many as the pre-camera days), but by raw numbers most portraits are created by photographers.
There will also be a longer or shorter period of time in which such technology will be abused by artists (because it's new) and at some point it will stabilize.
Aren’t the cameras they are making aimed at professional productions? Those are probably going to replaced last, the first thing will be (or are) TikTok clips shot on smartphones.
I don’t think they’ll see a decline in cinema camera sales due to AI soon.
Yeah they don't make a single consumer grade product. Prosumer at its lowest, and even then they're way too technical for the uninitiated. No one is picking up a BMPCC and just shooting/posting online.
I doubt it because I wouldn't waste my time watching it and I can't imagine all that many people today that don't already watch AI videos are going to suddenly change their mind and decide they like AI produced Hallmark movies.
"AI makes real world obsolete." I think that's enough hacker news for today.
Why do you believe this? Are you expecting all media to become fully AI - sports, TV, movies, youtube?
They've been spamming slop project submissions the last few months. But then again, who isn't?
No way.
Even the frontier models running on insanely powerful hardware could only generate 15 second clips in low resolutions.
And yeah, I saw some demos from Seedance 2.0, and they were awful. It's ridiculous how much people on Xitter were like "You can't even tell it's AI!" and I was like "It's trivial to tell it's AI" and could easily pick out all the markers. An individual screenshot could look good, but every time the camera angle changed, there would be a glaring inconsistency.
You people are either blind, delusional, or outright insane. AI might be used for a quick clip, or used to enhance something recorded by a camera, but "most video" is definitely wrong.
God I hope not. That’s depressing.
Even more reason to check out.
So live A/V is just dead now? Movies are dead? I really don't get where this take is coming from.
You are only looking at your own consumption at the moment. There are a lot of problems that still need to be fixed especially with rope artifacting. The 4k most models taunt isn't equivalent to a real 4k image or video as well at the moment you need a quality factor of two to get the equivalent result of a shot image or video. Resolution does not indicate quality.
So much respect for Black Magic. They are absolutely World Class and their business model is extremely generous.
Having said that, for all the AI features, the big one would be setting key frames etc. with an agent, driving the general editing workflow with text,etc. I realize this is non trivial but it's certainly viable for a team of this calibre.
I think if BM added a paid for agent which helped execute their traditional video editing tools (even if it "only" supported a subset) then that's a subscription a lot of people would be willing to pay for, especially as their core tool is so generous.
You might be interested in a tool we're working on which does some of what you're wanting https://sparkfxstudio.com/ It's presently in beta but is an AI tool for helping speed up video workflows using agentic AI.
There's a Linux amd64 version and a Windows arm64 version. However, I use a Linux arm64 machine, and arm64 machines are going to be a lot more common going forward. I wonder if there are plans to release a Linux arm64 build?
people complaining about AI features have clearly never wasted hours editing video or lost time and money discovering a technical flaw in a rush shot three days ago. For actual workflows, these tools are lifesavers
Because those people see the phrase "AI features" and the first thought is those sloppy generative AI stuff where things shift.
Where as marketing at all these corporation is trying to genericize "AI features" into anything using an algorithm. "Content aware fill", something we've had for over a decade is now "AI object removal"
"Noise suppression" is "AI voice extraction"
Motion unblur is now "AI motion unblur".
I’m not sure I agree? For each of your examples there are algorithmic approaches and neural network approaches. Companies have certainly been loose and wild with how they market these, but there remain distinct approaches and implementations for each. Very generally speaking, the neural network based approaches (aka “generative AI”) perform better but with much worse degenerative cases and a higher baseline rate of unwanted side effects (that are normally not immediately visible but tend to cause issues down the line).
My bigger concern is that these neural network based solutions have taken the place of the former rather than supplemented them. Many tools no longer provide the algorithmic/kernel-based approach at all, and have marketed the “AI” (née ML) alternative as a strict superset/upgrade, despite its potential drawbacks.
(Interestingly while the inference-based implementations generally have higher latency (or infinitely worse, cloud and pay-as-you-go requirements), for some computationally difficult kernels the inference-based approach is actually faster!
I really don't understand why people are complaining about the AI features. These all mostly seem like solid quality of life enhancements and CGI-like tweaks.
some of these genuinely excite me, like the slate recognition (chore-reduction), clip search (although I'd want to see how reliable it is), and deblur (a typical PITA post fix). anything that makes masking, tracking, and level-matching easier saves hours on hours, but only if they're either reliable or easy to fix what the automation gets wrong (and it'll always get something wrong by the director, no matter how good it is, because telling editors what they did wrong is how directors make money).
the more SFX-end ones like facial aging and face reshaper, or the talent-replacing ones like speech cloning/ADR, feel both too prescriptive for a director to dial in what they aesthetically want, and also not good enough for the final cut. so I struggle to find where they'd be actually useful in a workflow as opposed to being a trap, looking just fine enough at a glance to sneak into a final cut while looking poor when viewed by the audience.
likewise the focal adjustment and upscaling just feel gross. the kinds of things a good cinematographer can already do, and it'll be so so tempting to use tools instead of taking time to do it right in camera because it'll look good enough in the editing bay, but I feel like it'd really stand out as fake in the final cut unless you're targeting like, heavily compressed social ads. the less you use them the better they'll work, which isn't ideal for a marquee feature.
if the blemish remover really does respect continuity it'd be a nice-to-have, but it also feels like another trap to be careless/cheap on things like makeup or lighting at the shoot, at the expense of looking fake in post
I think that's the broader angle that bugs me the most. all of these tools are convenience tools for editors, but in the end they'll really be justifications for directors/producers/studios/agencies to cheap out and do shittier work faster on the shoot. a cheap, shitty shoot covered in AI bandaids is still going to hit an audience like a cheap, shitty shoot.
Artists HATE AI. I fully expect some sort of DaVinci Resolve backlash, artists refusing to cooperate with those using this software.
I doubt it. Artists hate fully AI generated slop.
Artists appreciate and use time saving tools. The AI features in Lightroom, for example, are very well received by photographers (myself included). Automatic subject, background, sky, body part masking, content-aware/generative fill and generative remove are genuinely helpful and time saving.
Photographers have already delegated their art to pushing buttons on a machine. They are the most receptive to AI tools, but not representative of artists in general.
These tools are very specific and keep the artist in control. Slop generators aren't reallly controlable.
What are you talking about? I don't know anyone grinding away deep in DaVinci or Flame that wouldn't love AI assistance. If anything, the studio heads will come in and insist it's all done by hand with the same energy they keep people up to 4am changing the color and shape of the eyebrows of the camel on a box of cigarettes just to deal with their neuroticism.
“WE WANT TO DO THINGS THE HARD WAY AND WE WON’T BUDGE”
Got a copy of the Studio version a few months later I opened my YouTube channel: among the best money spent in software of my life.
I felt the same when I got Vegas and Sound Forge, but they never got released on any platform other than Windows, so eventually outgrew them. I totally understand what you mean; I use it, but also happy with Blender!
What do you like most about it?
One really neat thing is that you get free major updates. I bought a Blackmagic camera back when resolve was version 14 I think, and today I still use the very same license with Resolve 21.
Not that I spent any extra money on the license compared to what the camera itself costed, but I also feel like the 0 money I spent was well spent :) Harder about the time commitment to move from something you know really well to something new, but the time I spent on that was very well worth it too.
Anyone using this headlessly got a read on how much of this an agent could do without human intervention? Would love to have a gut check on "sure, spend the $295 and you'll get some benefits for free if you have an agent run your videos through this before shipping them"
To be clear, my use case is making weekly online videos suck a little less - not grading feature films :)
You can use the free version (no need to pay), and it should be possible to drive Resolve using the Python API -- see this MCP server built on top of that API:
https://github.com/mhadifilms/dvr/
You can also try pointing Claude to the API surface directly.
The result might be chaotic, but you should be able to automate editing... maybe. :-) Ha. (You'd need a good loop to make sure Claude can see what it's doing..? Beyond screenshots I mean - ie how would Claude know its doing a good job editing?)
The whole first section: 9 features, 9 titles with "AI" in them.
I don't think their use of it is bad at all, I'm just tired.
They could remove the word "AI" from each one of those feature titles, and the titles would be just as descriptive without them. At this point, it's just marketing noise, more distracting than informative. Maybe like "cyber" in the 1990s. Would you like some AI tea with your cybercrumpets?
"Background removal" or "Face aging" without AI were done before, and they were shit.
Putting "AI" into the feature title means "this time it actually works"
You're right to be tired of it, and of course I haven't tired these specific features yet, but Davinci was already lowering the barrier to entry for filmmakers, and if 21 works as they say it will, then you're looking at a major lowering of said barrier.
Its all local if that helps?
I think they've made lots of great practical choices! 100% in agreement with running local models for these tasks.
My opinion is that, for end users, if you name your feature "AI" to market it, you kind of already failed to read the room. You're writing to VCs while hoping it convinces customers.
Name what the feature does, what it gains them. Call it "smart" if you must imply some black box treatment.
Naming AI as the selling point for everything feels a lot like that Android tablet ad circa 2010:
"Your wife will love the new dual core Tegra™ chipset!"
> My opinion is that, for end users, if you name your feature "AI" to market it, you kind of already failed to read the room. You're writing to VCs while hoping it convinces customers.
100% this.
Maybe I live in a bubble, but consumer sentiment regarding AI seems extremely negative. Boasting "AI" features is more likely to lose sales than to create them.
But their target audience are creators, not consumers. Creators love AI.
And each is very specific to a use case, not a "general chat prompt for triggering API calls" but things like "ML model to categorize video clips and assigning tags + names, so you can find it faster" and similar.
I'd also get tired if it was "AI ala Microsoft/Google" where the goal is to get you to write forever with a chat bot somewhere else, but these features are very different from that.
Eventually, in moviemaking, generative AI is going to be seen the way CGI is. That is, how people complain about CGI when it's obvious/distracting/noticeable, but the best usages of it won't be noticeable.
Sure, and like CGI, it will change the nature of the media entirely.
Different stories shown with different treatment. With CGI, scenes zoomed out to wider shots and effects swelled even louder over lighting, intimacy, acting, etc.
Old styles didn't disappear or stop evolving entirely, of course, but the center of attention profoundly shifted and the "big" production money went with jt.
Generative AI will likely drive some kind of analogous shift in dominant film aesthetics. I don't know where, but I'm not particularly excited by it myself yet.
I don't know much about the movie industry, but I think the biggest commercial successes for early CGI was possibly Jurassic Park? I don't think it would have been nearly as good if it had all just been done with puppets.
Likewise, there's been a ton of movies since that could in theory have been done purely with SFX instead of VFX, but which is probably must better from having used VFX/CGI, titles like The Matrix, Lord of the Rings, Inception, Avatar, etc.
Obviously there was also charm in titles like E.T. and Gremlins, and I think there might still be a market for movies like that, such as the 2019 Netflix series The Dark Crystal: Age of Resistance (I loved the 1982 version as a child, though this sequel wasn't for me).
I guess the question is: will generative AI allow new movies to be made, the same way that CGI has? Or will it just be an economic shift: the same quality CGI but at a fraction of the price?
There may still be people out there who believe that AI will never be able to create CGI as good as humans can, these might be the same people who used to say that CGI can never look realistic. And if you work in VFX, I bet that you can spot a fake mountain with the slightly incorrect shadows in the distance easily, but as a simple movie watcher, I really don't see it, especially when it's only on screen for like 3 seconds.
Maybe, if there's anything left to change?
If you look at the ten top-grossing films in the past 15 years, it's almost exclusively computer animations or SFX showcases with some token, quip-based acting thrown in every now and then (all the Marvel universe stuff).
Basically, I'm not even sure if replacing actors with AI would be stylistically perceptible for the average moviegoer. There are interesting films still being made, but almost no one is watching them, because you have the option to watch some SFX explosions and flying superheroes in Avengers VII.
It really challenges my stance of human supremacy over AI when I only see these rote, shallow, unimaginative defenses of AI.
Likewise when I see rote, shallow, unimaginative critiques of defenses of AI. ٩(^‿^)۶
I just pulled in 10 iPhone RAW files. It did a really nice job of processing them, I did the usual pulling-up shadows and highlights, and played with some of the sliders. Noise reduction and sharpening tools are primitive. Yet the photos look great. I finally managed to export ("render") them.
It's a baffling process flow if you're coming from Lightroom or a manual ACR workflow. But I'm excited to see where this goes. Quite simply, the output results are great. And free!
One of the rare pieces of software that actually gets you excited with each new release. Moved to Resolve from Final Cut a few years back and I've never been happier. Looks like this release just continues the already great experience.
For all the issues with AI, these features aren't so bad. The "AI" search is possibly one of the more useful ones. That'll save me a fair bit of time.
For people using Resolve, would you recommend someone already quite well-versed in KDenLive to switch, for some non-profit work on cutting together educational content with some animations, some talks etc?
Will it allow me to drastically improve my workflow (save time for some tedious tasks), increase quality of the outputs etc?
If you're organizing/having hundreds of clips you want to put together, or overall want a more opinionated workflow, then I'd say give it a try at least, the free version doubles as a trial :)
I'm a Premiere migrant to Resolve (Studio) some years ago, biggest hurdle is the opinionated workflow, it basically wants you to use the tabs in the bottom to go from "Media > Cut > Edit > Color > Fusion > Audio > Deliver" (simplified) so different tools available in different areas, made for different use cases, but in general once you've learnt the overall and high-level concepts, it makes editing really easy and smooth.
Besides, it's probably the most stable video editor that runs natively on Linux since ever, I think I've had it crash once, and the Fusion 3D text doesn't work properly for me, but besides that, runs like a dream and UX is miles ahead anything else available.
Stability sounds interesting. While KdenLive has worked OKish for me, certain versions have broken my projects, and there are some long-standing bugs, like subtitles disappearing if I do a specific operation in the wrong moment. While the latter is fixed by a simple Ctrl+Z it is giving me second thoughts about using this for really large projects.
Resolve was a much better experience for me than kdenlive. But you can easily try it out for yourself because most of it is completely free (in fact you probably won't ever need the paid features for what you do)
Great to hear.
I'm no expert (relatively new to the field myself), but I was trying to put together some simple videos with animations in Final Cut Pro and decided to try DaVinci Resolve, and I'm glad I did. The Fusion stuff bundled into it is incredibly powerful for animations.
It does take some getting used to, but the amount of tutorial content on YouTube is another reason I'm happy I made the switch. A lot of really good stuff on there. (Search on 'DaVinci Resolve Fusion' to see some examples of it in action if you want to get a feel).
Wanted to expand on the animation part, so will check this up for sure, thank you.
Does someone know if Blackmagic Pocket Cinema Camera is a good camera since it comes with the resolve studio ? Instead of buying a separate camera in the same price rangen (sony) for studio or indoor recording ?
Discussion from April when this was announced: https://news.ycombinator.com/item?id=47760529
Resolve is an incredible tool, and I wish they improved the Linux support especially on AMD. It's the last reason why I have a windows machine, and Win11 made it unbearable to use.
Hey Blackmagic, just be sure you're not in violation of Illinois BIPA with the face search thing. They can and will come after you.
Face search? What do you mean?
Pretty sure that does not cover a face database indexing your own photos/videos, running locally on your own computer. If it did, that would be extremely silly.
Yeah, that’s been a thing in photo software for at least like 20 years. I remember using it as a teenager for my parents
I think you're right in this case, but Illinois lawmakers are generally extremely silly, so I wouldn't put it past them.
Why would the law make exemptions for running the search on on a computer that has the database.
Doesn't seem to apply here.
> BIPA establishes standards for how companies must handle Illinois consumers’ biometric information. In addition to its notice and consent requirement, the law prohibits any company from selling or otherwise profiting from consumers’ biometric information.
https://www.aclu-il.org/campaigns-initiatives/biometric-info...
Still a public beta?! Not sure why this is news ... the AI features?
I think it actually released today. I keep an eye on the forums to see the updates - https://forum.blackmagicdesign.com/viewtopic.php?f=21&t=2369...
(Can confirm - I just opened it on my laptop (I had the latest beta installed) and it prompted me to download the release version)
The Lightroom competitor would be a big (and welcome) change to the market if it delivers.
The Lightroom competitor working on Linux is a huge news.
(Darktable doesn't count, it's a scientific software with some wobbly UI).
Edit: Still garbage, required 100-lines script and LD_* shim to make it even run at Linux Mint 22.3 on AMD CPU + GPU. UX is even worse than Darktable, don't bother, not even close to Lightroom.
At this point we need a Kickstarter campaign to make Lightroom run in Wine/Proton (no, no matter how much you try, it will not work so far). Edit: or GSOC to support Darktable to improve their UX.
The Darktable hate is real. I guess scene-referred workflows (vs display-referred popularized by Lightroom) are too difficult for most people to grok.
I don't hate it. I'm just not compatible with it I guess. It's like GIMP in its early days - has most of the features of competition (Lightroom vs Darktable, Photoshop vs GIMP), even more features, and tweaks, and more knobs. But it misses USER EXPERIENCE part, it's basically unusable, unless you use it since early days.
If Darktable had a grant/GSOC just to improve UX, it could be a valid competitor to Lightroom. Currently, it's not. It's bunch of Python/Lua scientific code with some UI, that processes pictures.
Wrong kind of “resolve” haha. This one is more “please don’t let my AI-generated code leak keys before deploy” than video editing.
so, it looks like all the AI features run locally (via DaVinci AI Neural Engine)?
could use a little more AI. have they considered replacing users altogether?
dAvInci Resolve
who is going to use all this stuff to make what movies... where is the audience for any of it?