For those looking to actually edit video professionally on Linux you can’t really beat the free version of Davinci Resolve. Over the past 5 years it has become an incredibly competent non-linear editor alongside being the industry standard for color grading. I’ve honestly tried everything under the sun on Linux for video editing and nothing beats it. If you’re looking to stay open source I think blender is starting to become more competent but still a bit of a task getting it setup for doing serious NLE video work.
I have pretty modest video editing needs and have been using shotcut on linux. I'm aware of Davinci Resolve but had always assumed (for no reason other than few commercial developers do) they wouldn't have a linux version let a lone a free one. Thanks for the tip off!
I really liked Lightworks several years ago when I was evaluating all the video editing packages, except for the pricing structure at the time. It was really easy to pick up, and worked great under Linux.
My memory was at the time the only pricing option was a subscription model, and I don't do videos very frequently. I knew I'd be annoyed at paying $20/mo for the many months when I wasn't using it. More recently I've seen that they have a "buy outright" option for around $300, which I would have lept at back when I was first looking at it. But I've since moved on to Resolve.
For clarity, this is because Resolve seems to require OpenCL support. I haven't tried it myself, but the beignet package might serve to provide OpenCL support. I had a similar issue when using the OS AMD drivers and using Resolve, which don't come with OpenCL support.
Several years ago I spent several days trying to get Resolve working on Linux and had absolutely no luck. More recently I found a script that would repackage the Resolve installer as a Deb package and that seemed to work pretty slick.
That might be so, but I have been bitten with every single "Free" closed source tool I've ever used when it went away in some way or the other. Unless the free tool is totally unusable or the non-free one 5x more "good", I'd rather use the time to get good at a os tool that will just get better year after year.
I have heard about Resolve for color grading, but how good it is when it comes to editing? Is it simple and intuitive enough to make few edits, add titles for YouTube videos?
I'll echo another comment that a few Youtube videos and off you go. It's pretty easy, I'm a video editing neophyte and thought it was easy to pick up.
The thing about Resolve is: You aren't going to outgrow it. It's easy enough for doing simple edits, but if you want to do fancy stuff with it, it can do it. You won't outgrow it.
That said, I spent a stupid amount of time with performance issues in Resolve. My cameras output H.264 or something, and I'd have to convert it to ProRes to edit, a process that took huge amounts of space and time. This was under Windows. Even with that, performance was still pretty bad. One kids concert video took me around 40 hours of editing to get an hour of final footage.
Prior to that video, my desktop machine did a better job with ProRes, but it got milk spilled on it during the home schooling of 2020.
However, and this isn't a solution that everyone can or should do: Resolve on the new Macbook Pro boxes is amazing and will handle anything I throw at it, without having to go through the hoops of converting to ProRes and the like.
I've not used Resolve but I've tried pretty much all Linux NLEs. Here are some questions arising from my workflow:
How about resource file support for non-standard (as in TV/cinema) parameters, such as video framerate, video resolution, video formats? How about integrating these into the timeline, including per-clip pan/zoom?
Have you used Vegas Pro (for Windows)? I feel like no Linux NLE is able to replicate the flexibility Vegas Pro allows you in treating each clip as an independent, resizable, transformable asset. Instead, many NLEs enforce an "all clips must keep the same parameters as the project file" rule, where e.g. you can't incorporate a 60FPS file into a 50FPS project.
I'm not really sure about your second paragraph, that's not something I've tried doing.
Resolve works like this: You set a project parameters (resolution, framerate being the big ones), either manually or it is set by the first clip you bring in. Then you can bring in clips that don't match those values, and they are just, at render time, converted to the target resolution/frame rate.
You can take clips in your timeline and make a composite clip out of it, basically rendering it within the project to a new artifact, I usually use that primarily for combining the 5m clips my camera breaks things down into into a single large clip, but I think you can also put panning on that. I haven't used Vegas Pro, that sounds pretty slick.
Not OP but its editing is also pretty advanced and with some YouTube tutorialling, it can muster some effects you might find in Adobe After Effects. Resolve has replaced Adobe Premiere Pro for me. I'm happy to have saved the subscription fee!
I was a old Final Cut Pro guy (never really got into the new X version). I was able to pickup resolve just fine and they make great improvements each year. If all you are doing is YT videos with titles this may be overkill, maybe Kdenlive.
I generally stick with the GTK/GNOME ecosystem but for my (admittedly quite limited) video editing needs I've found Kdenlive to be easier to figure out.
For me Shotcut won in usability. Openshot often had no tools for what I wanted, and Kdenlive was just so clunky with the interface I wanted to throw my laptop away and pay someone else finish the project. Shotcut is by no means perfect and for example simple masking ends requiring sliders in the plug-in to adjust the rectangle size on the display... but it still won.
I have found Shotcut to be entirely stable over the last year on both Windows and Linux, whereas the last version of Kdenlive I just tested this morning experiences freezes on Windows. I'm quite happy with Shotcut.
Weirdly, in my use, Openshot was the stable one among those. Not discounting your experience, but wondering if it could be, e.g., distribution-specific.
During the first lockdown wave I settled with kdenlive for did some light video podcast productions. For a free program I think the UX and features are awesome, plus there is a large enough community and video tutorials to answer most questions a beginner might have.
If developers can't be arsed to mention on front page what platforms their product supports then I don't think they really know what "intuitive" means. They might be cool guys and nice people and all but interacting with users is not their strength.
I agree. However, it was refreshing to find that it meant it was for Linux, which is the defacto OS to ignore in the "PC and Mac" world.
It's also not a bad exercise for Windows and Mac users to learn that the best way to determine this is to go to the downloads section and/or look for system requirements.
I kind of agree. It's very annoying to read about a program and find out on the download page that it's only available for Windows or for macOS and the open source community is starting to do the same with the assumption that just "open source" is enough to signal that it's Linux only.
Of course with modern versions of Windows you can probably run this program through WSL relatively easily as well, but that is far from the default for running open source programs.
As this is GTK based, I suppose there's no technical reason why it can't be made to work cross platform if someone would put in the effort. It'd still be nice to see it mentioned, though.
At first I did this as a form of revenge: "If they just announce their software, I'll just announce the software I use regardless of the OS they use!"
Then I started thinking: "Well, their are probably not malicious or ignorant. This is just probably an acceptable behavior." And continued to behave that way.
I don't think we should complain when FLOSS non-multiplatform software (that support a FLOSS OS) should be criticized when announced without disclosing the lack of portability. People who choose to not use a FLOSS OS should be able to try it more easily and without the limits imposed by usual proprietary software licenses restrictions.
Now, if you announce something, even if it is FLOSS, that only works on windows or macos, the thing is entirely different. A FLOSS user will have difficulty trying it or will have to agree with the restrictions of proprietary software licenses. These are the software that, when announced, should state very clear which platforms it supports.
That's also strange. Clearly the comment was coming from someone not familiar with the project. Responding "I just submitted the change, here's a link" would have been more reasonable.
But I also understand that the true beauty of open source is the entitlement of free work from devs on a project receiving less than 8 euros a month from donations.
Someone posted the project to Hacker News. A commenter pointed out that they didn't put the supported platforms on their website. That has nothing to do with "entitlement of free work". If the devs don't want anyone else discussing their work, they have every right to keep it private. If it gets posted to HN, it will be discussed.
> Responding "I just submitted the change, here's a link" would have been more reasonable.
This is entitlement to free work. Raising an issue is much more reasonable than an HN user submitting a change on behalf of another random HN user. I do understand that the OP wasn't asking for someone to do this for them, though, and not displaying supported OSes is a totally valid critique.
Sure, it was once common practice to have a list of supported platforms on your homepage, but you have to excuse people for not knowing that. Most people just get their apps (including Linux apps) from an "app store" sort of thing.
That said, the download page is where I'd definitely expect something like this, and the download page is useless. It tells you that various Linux distros have Pitivi in their package managers (which ones? links?), and has a Flatpak download (what's flatpak?)
I used Pitivi, it's okay! Not 100% stable, but that was many years ago, so it's probably better now.
The news on the frontpage make it look like the last release was almost two years ago; but actually there was a release in 2021 and development is ongoing.
I prefer blender. It is brutaly packed with features, but still free and can get the simple job done. But you will not outgrow it...
It is 3D modeler, renderer, 2D painter. But it can even do nonlinear video editing/cutting. You can combine 2D video with motion tracked 3D scenes... Certainly worth learning the basics.
I've learned basics from youtube.
First you have to switch blender to video editing "workspace". This opens the internal non-linear video editting GUI, which makes things little bit more obvious.
I came to say the same thing. The video editing features were very usable in Blender in 2015. Given the open short movies they do are used to drive development, I would be surprised if the video editing hasn't been improved since.
Yeah... i beleive the blender 3 got rid of that weird thing of using right mouse button for everything and went with some more sane default settings which prefer left button for basic operation.
Blender seems to be the best long term solution in that it's got good bones and a great license. It got me back into video editing after I moved all of my machines to Linux ages ago. The ability to do basically anything inside of the program with python is superb.
Wow the new version will have node-based video editing like Blender or Unity. After using Shotcut a while back I started wondering if node-based video editing was feasible, glad to see some projects trying it out
That is great, I edit videos from time to time and I have been using Openshot for that. However it is so buggy that I ended up fixing one myself.
When I entered into the code I admired one of the most spaghetti software I have ever encountered. No tests, all is verification is made on opening it and "checking" if it was working. Several major bugs that nobody have an idea why, I ended up certain times commenting parts of the code to make it work.
You just cannot beat Cinelerra. Occasionally I try to use some heavily advertised one (like Kdenlive or Pitivi) but immediately realize you cannot "Generate Keyframes While Tweaking". This is essential.
Cinelerra has been splitted to two^H^Hthree but only Cinelerra CG seem to work.
Those who remember old-time Cinelerra can be amazed with the render-list of 20, which all seem to work. In the olden golden times you just made big MJPEG and used FFMPEG for coding.
Maybe you mastered it better than me but I always found Cinelerra-CV a bit clunky, especially if I was dealing with oddball formats (HDV) that I didn't understand. Maybe I would get the hang of it now but I've kinda moved on.
Clunkiest part was [ and ] - keys, which are not readily available on Nordic keyboard. It has been finally fixed (after 20 years) and you can used < and > - keys instead.
I get where you're coming from. There are many Linux video editors already, and many of them could use slight improvements. If people contributes to those instead of going off and making their own project from scratch, it would benefit everyone. That's true.
But I think that it's impractical in reality. For various reasons, such as politics, differences in vision, BDFLs, death by committee, etc., many open-source projects find themselves reaching a plateau and stagnating. At that point, it becomes easier to just make your own competing project with the features you want, than try to fight to get your features added upstream.
I think that this, counter-intuitively, leads to the stagnant project improving, as the "competition" with new features helps un-deadlock whatever bureaucracy was preventing improvements.
Not sure how much this matters, but it took me quite awhile to find a link to the source code. I missed the gitlab logo at the bottom, clicked around on the Contribute page for awhile, until the "Development Environment" page brought me one step further, where I eventually found a link to the code.
Obviously, finding it is a one time cost, but given the landing page praises the cleanliness of the code, I'd expect a link to it. Make the "clean codebase" link to https://gitlab.gnome.org/GNOME/pitivi and that would be perfect!
A little OT but why do flatpaks take so much size? I have a small partition for linux and last time I ran out of disk space it was because these flatpaks were taking nearly 1GB for simple programs that I reckon should not be more than 10MB executables.
Is this like iTunes on windows where it downloads the whole apple operating system to run the software? I know disk space is cheap but it's not always the case especially with these new laptops with fixed and small nvme drives.
edit: just found out this too requires 2.4GB disk space :/
You chose the worst possible way to express a concept many of us could agree about; let me attempt to expand it a bit.
GTK choice of putting everything, including buttons, on the windows title bar kills intuitiveness and slows down the workflow. When a window/dialog/whatever opens, I look at the title, then if necessary I move it to another place by grabbing that same title bar, which is immediate because that is where I was already looking to, then start reading its contents downwards. When finished I'd expect to see choice buttons just below the contents of that frame, I'm looking just there so that it would come very easy, but no, now I have to go back upwards and find the title bar to push a choice button. To me, this is against any interface writing logic; it trades usability for a dubious more modern appearance, and I would very much welcome a patch to correct it.
There's always a hypervocal minority in any a GNOME related thread claiming that GNOME 3 was a regression, it has too much padding, it's too slow, it takes too much memory, it doesn't allow you to configure something, etc etc etc.
Meanwhile, every big distro chooses it and millions get on with it just fine.
The button layout in GTK dialogs makes essentially no sense from a UX perspective, because, yes, people read things left-to-right (in LTR languages), top-to-bottom, and the program flow should reflect that. Apart from that problem, buttons on top makes little sense for other dialogs due to the awkward mouse movements it forces. Mouse moves down, selects a file, maybe picks the file type (do I want a jay-peg or a jif?) and then... has to move all the way up again?
I'm mostly not in a direct line-of-fire for user complains due to Gtk/Gnome (but actually received some, mainly around, you guessed it, file dialogs), but the impact zone is basically next door. And it's not exactly happy noises and good smells coming out that door. This is from users who couldn't care less what the thingamajing is called and don't know or care if their Linux explorer is called Dolphin or Thunar. What they do notice is that they've got an OS upgrade from IT and now they're angry because two thirds of the UI are gone and the remainder makes less sense to them.
But if very basic components like file choose and file save dialog are incredibly painful to be used with buttons scattered through all window corners (fitts law anyone?) then "nothing" starts resounding in your mind.
No, there is an extra settings button before select which looks to me totally secondary not to be present by the primary button. Primary actions should be together:
- Choose letter
- Choose size
- Click select
...In a row or L movement
We will get this extra UI darkening: now is not evident you can type font size numbers cause all controls are flat.
In gtk3 at least inputs look like inputs and buttons look like buttons.
The non-standard and objectively poor positioning of the buttons is fascinating. What was the logic behind this?
It seems to me that designers are much less likely to give their time for free to a software project than programmers. Is that why OSS suffers from problems like this? A junior UI designer would never make these kinds of decisions.
I can understand why designers wouldn’t work for free. What I don’t understand is why an average coder couldn’t do some basic googling and learn the fundamentals. You can get a long way by imitating commercial work and following standard conventions. So is it lack of skill, or some sort of mistaken belief that the conventions established by the experts are wrong?
I’m wondering here if the idea is to imitate a forward/back step from a wizard (albeit with the buttons in the “wrong” vertical position) Not a paradigm appropriate for repeated everyday interactions, mainly because wizards are not designed for efficiency.
Without strong leadership and buy-in from everyone involved, it's very hard for design or UX to be involved in open source projects in a meaningful way.
Implementing a design requires discipline and a shared vision, which is often against the bazaar-like nature of open source projects. Everyone's working on their own particular piece and can't agree about anything at all – just look at the internecine battles that are still going on about which init system to use.
In my experience, any attempt at making an interface better, even if it's proven with data and testing, results in extreme pushback about dumbing down. The attitude seems to be that making software accessible is not 'real computing' and therefore should be fought tooth and nail.
Designers often don't work on open source projects not because they don't work for free, but because it's a miserable experience.
Having said that, good counter-examples to this are the recent Blender redesign, where significant effort has been put into making the software more consistent and easier to use, and the Elementary distro where one of the goals is to create a good user experience.
Reason? i can only think like:
what? space is not optimized! there are useless corners but all buttons at the bottom. also aesthetically it looks too heavy. let's visually balance everything moving every button to every unused corner space.
yeah, looks great! Check done!
Now, let's get rid of all visual clues for actions/controls...
The positioning is not objectively wrong. Where do you get this from? Did you make tests with people who never used a PC before? Who are not biased? Or is this your opinion, which would be subjectively by definition.
UI usage is mostly learned.
Just to be clear, I don't think GTK is the perfect UI system. In my opinion, there is none.
UI usage is mostly learned, you're right. All UIs are fairly arbitrary, helping users understand the current state of their computer and perform actions, by using metaphors to help construct a mental model of what's happening.
But we need to be aware of those learned habits, metaphors and conventions and build those affordances into our own designs if we want people to use our software efficiently.
A convention often used in modals and other popup windows is to put the close icon in the top right. So putting the confirm button there could cause a user to click it wanting to close the window and confirm a destructive action instead.
Another convention is to put the confirm/cancel buttons at the bottom of a dialogue window. Buttons like those don't normally go in the title/toolbar, and putting them there goes against the top-down, left-to-right way we read things.
Putting things in non-conventional places is not 'wrong' per se, but adds extra friction to every interaction.
No, it's not true. On the top of my head there is Inkscape which truly is intuitive (despite GTK) and, arguably with their latest betas, beautiful.
I wish we could have something like it with Qt/KDE. Calligra's Karbon once was a nice promise but it is now only a shadow of what it was meant to be and even living a slow death.
The OP is saying that the comment that says GTK is unintuitive is useless is true, and you are saying that GTK is beautiful and intuitive, which is just confirming the OP's comment. I think you miss the point here.
Their CDN certificate seems to have expired 3-4 months ago, which is a quite scary warning for a user to see when downloading an exe from the internet. No Linux download either it seems (Pitivi is a Linux program)
Thanks for that. I've been a mac user, so haven't downloaded it in ages now. I'll let them know about this. And you're right - it is indeed a scary warning to see when downloading an executable. They've been a Windows only shop since 2001 when it was started. The product to look for is "muvee Reveal Encore".
I pinged this 'cos the approach is very different from most editors especially from the perspective of what a user needs to know about video editing (much less with muvee). There are editing "styles" that you get to choose which will adapt whatever video/photos content you throw in and adapt it to the music track you supply. The styles have many parameters you can tweak too. Otherwise, it's a button to produce a video. The first time you use some media item, it'll analyze the audio & video .. which will take some compute time.
Only the (scheme based) scripting language on top of which the automatic editing system is built is open source - https://github.com/srikumarks/muse . However, it's editing and effects engine is programmatically accessible within the application though, so you can make your own styles and even write programs to make specific compositions in case you don't want to use the automatic features. Documentation here - https://srikumarks.github.io/muvee-style-authoring/index.htm... .
It's also not free. The shop is also convoluted, its not clear what version you should even be buying. This might be a great case for the reasons behind the original OPs share.
