Live data from Hacker News

DaVinci Resolve 21.1

blackmagicdesign.com

211–219 of 219 posts

Re: DaVinci Resolve 21.1

#211
post #23

I've been using it for many years, and it's grown in power and functionality with every release. BMD has offered free upgrades to Pro users (with no subscription) for many years, which is refreshing given the current state of software licensing. For those upset with the new agent integration, keep in mind that one of the big knocks on DaVinci Resolve is that it's hard for beginners. Powerful tools like color grading…

I view the agent integration (bearing in mind I haven't tried it yet) as an easier way to replace writing python scripts to do repetitive tasks, for people doing particularly repetitive work. A macro recorder might be more optimal to the anti-AI crowd, but this is probably easier to implement for the developers, while being more flexible for the end users.

I assume that this can be connected to local models, not just cloud models.

Re: DaVinci Resolve 21.1

#212
post #197
post #137

Earlier quoted context omitted.

Or buy the ~$400 speed editor and get the license for free. That's what many people do.

For me it was the other way around - bought the license and got the speed editor for free. Though it may be a different model, I think the stand-alone price for it was way less than $400.

The Speed Editor debuted at $295, so the same as a Resolve Studio license except you got the free keyboard.

Re: DaVinci Resolve 21.1

#213
post #119

Earlier quoted context omitted.

Someone was posting recently how it cost him tens of dollars of LLM Agent time to reverse engineer Linux Davinci and add codec he wanted back in.

This is one of the the cases where “the language model won't necessarily respect IP law” is really relevant. Individual users may be able to get away with this because they're too scattershot to sue, and users in some jurisdictions may actually be fine regardless, but broadly, the patent licensing around things like H.264 is not something BMD can afford to ignore. AFAIK the biggest reason Google was pushing WebM in t…

Linux Vivaldi builds use libffmpeg.so from chromium-ffmpeg, which includes decoders for most codecs used on the web (no h265/mpeg2 tho), and VA-API which will handle all hardware accelerated codecs on particular platform.

This looks similar to using Microsoft provided Media Foundation codecs. No idea why Black Magic doesnt do it like Linux Chromium browsers.

Re: DaVinci Resolve 21.1

#214
post #23

I've been using it for many years, and it's grown in power and functionality with every release. BMD has offered free upgrades to Pro users (with no subscription) for many years, which is refreshing given the current state of software licensing. For those upset with the new agent integration, keep in mind that one of the big knocks on DaVinci Resolve is that it's hard for beginners. Powerful tools like color grading…

Taking away some of the more monotonous and "uncreative" tasks in any job can be taking away some of the variety in the work. Variety is nice. Variety affords us downtime. As the job gets stripped down to its core function, while a machine keeps the pipeline full so that the work can continue uninterrupted, the downtime slips away. And great for beginners, now they get to pay some chatbot company to cover the gaps le…

It also takes away variety of the outputs as often it turns out those "uncreative" tasks actually do hide creativity.

Re: DaVinci Resolve 21.1

#215
post #17

Earlier quoted context omitted.

I have videographer friends, and this is exactly what they've been asking for. Their dream is to be able to say "sync the A and B camera footage, chunk the whole interview down into segments, and then pull together the highlights based on this outline." Not every application of AI is "do the creativity for me."

I would not be surprised if engineers who laid the groundwork for this back then would be happy with this workflow. I am reminded of an interview I saw about the Stranger Things intro (before mainstream audiences got weird about the show). The producers wanted a practical effects intro, and consulted with people who actually made 1980s intros the show was trying to replicate. The old guard said the show’s producers w…

> The old guard said the show’s producers were crazy to do it with practical effects and should just use CGI since this was what they would have done, if they had the technology.

As always, art benefits from limitations.

From the [dead] sibling

> Those old practical effects were incredibly labor intensive and took so long. The famous sequence where from the movie 2001 where he enters the monolith used slit-scan cameras and took up to 5 minutes per frame.

But those effects still hold up today, unlike most CGI from even decades later. So yes, like most modern "advancements" CGI lets you crank out imitations for cheaper but quality is lost. And the cost is itself a feature - it made directors really think about what effects are actually needed and which ones are just noise that don't benefit the film at all. And of course movie budgets on the whole haven't exactly been going down.

Re: DaVinci Resolve 21.1

#216
post #88
post #17

Earlier quoted context omitted.

I would not be surprised if engineers who laid the groundwork for this back then would be happy with this workflow. I am reminded of an interview I saw about the Stranger Things intro (before mainstream audiences got weird about the show). The producers wanted a practical effects intro, and consulted with people who actually made 1980s intros the show was trying to replicate. The old guard said the show’s producers w…

The irony being that a lot of older practical VFX artists aren't able to tell when CGI is terrible. They're often so enamored with what the tech enables, that they're the last to see how quickly it becomes shoddy. Not sure what the equivalent of the hedonic treadmill is for CG, but it ages very rapidly, versus the best practical effects which maintain a charming physicality - even when they're no longer convincing.

Yes and in general, taste can deteriorate with age. Just look at what Cameron did with the 4K remasters of his movies.

Re: DaVinci Resolve 21.1

#217
post #17

Earlier quoted context omitted.

I would not be surprised if engineers who laid the groundwork for this back then would be happy with this workflow. I am reminded of an interview I saw about the Stranger Things intro (before mainstream audiences got weird about the show). The producers wanted a practical effects intro, and consulted with people who actually made 1980s intros the show was trying to replicate. The old guard said the show’s producers w…

I just watched the entire cycle of Don Quixote films by/about Terry Gilliam. That is a master class in how to ensure that your magnum opus is the worst film you've ever made.

Self-referential art rarely ends well.

Re: DaVinci Resolve 21.1

#218
post #213

Earlier quoted context omitted.

This is one of the the cases where “the language model won't necessarily respect IP law” is really relevant. Individual users may be able to get away with this because they're too scattershot to sue, and users in some jurisdictions may actually be fine regardless, but broadly, the patent licensing around things like H.264 is not something BMD can afford to ignore. AFAIK the biggest reason Google was pushing WebM in t…

Linux Vivaldi builds use libffmpeg.so from chromium-ffmpeg, which includes decoders for most codecs used on the web (no h265/mpeg2 tho), and VA-API which will handle all hardware accelerated codecs on particular platform. This looks similar to using Microsoft provided Media Foundation codecs. No idea why Black Magic doesnt do it like Linux Chromium browsers.

Do you have any information on how the actual patent licensing is set up for Vivaldi? I wasn't able to find much on it, though I see the note in https://wiki.archlinux.org/title/Vivaldi that it “automatically downloads” the relevant files, which is at least suggestive that they can't just redistribute the binaries normally. The other closest thing I was able to find was https://github.com/chromiumembedded/cef/issues/3559 which describes a specific proprietary_codecs option in CEF, which is not quite the same thing but ought in some sense to be related… the way they describe having to set ffmpeg_brand=Chrome normally for that to be available doesn't sound promising?

I would be moderately surprised if whatever arrangement is going on there allows full use of the codecs for video editing software and not just for consumer decode of Web resources.

Re: DaVinci Resolve 21.1

#219
post #89

Earlier quoted context omitted.

Resolve is extremely RAM hungry. For 4K video I wouldn't want to be editing with less than 12GB of VRAM. I'm guessing that excludes most integrated graphics. Source - I work as a professional editor (amongst other things). Please commenters, do not come at me with your five minute low bit rate projects that required only a LUT and some contrast and 'worked fine' in DaVinci on a potato.

Integrated graphics use the main system ram so they basically all have access to more memory than the average GPU.

I should have also mentioned DaVinci also heavily uses CUDA on Windows, and metal on Mac. I'm assuming the functions implemented in these frameworks do have fallbacks on Vulkan or similar, but I doubt they're as optimised. I know there are some issues with Resolve on Linux for example. Any halfway decent Nvidia card with enough VRAM should work fine - i.e.: a 30 series card is grand. But I'd be skeptical integrated graphics would have the horsepower for denoising or a host of other build in effects. Rendering would likely be slow also. Although this is the least important element of the process on modern hardware.
Post reply on HN