Hacker News new | past | comments | ask | show | jobs | submit login
Shotcut – A free, open-source, cross-platform video editor (github.com)
182 points by oridecon on Aug 5, 2015 | hide | past | web | favorite | 47 comments




Looks fairly interesting for some of my more casual use, I might give it a download.

Any big gotchas that you have come across or any tricks using it (special plugins, etc)?


Like most (all?) of the Linux video editors you have the occasional crash. It happens way less than Openshot or Pitivi but it's still present. I don't know if it also happens on Windows or OSX.

You can version control your edits and do text operations since it's a MLT file so that helps.

Not really my thing but you can insert HTML5 overlays and animations. There's a demo on the youtube channel but it's quite "verbose" for most users.

GPU accelerated filters are really fast and I have no problems working with 1080p 15GB+ files. But that's nothing compared to what professionals use so I can't really comment on that.


> the occasional crash

I don't get this. Why is this true for the genre? I've been trying various video editors for 10 years, and they all crash. Every last one of them. I recently gave up on Windows Moviemaker because it crashes rendering anything over 30 minutes long. (Not only does it crash, it freezes Windows so bad you have to powercycle.)


Exact same experience here. I too would like to understand why this happens. Am guessing the excuse on Windows would at one time have been the huge variety of supported video hardware, but that no longer seems to be the issue. And that's only a guess. Not sure if it's DirectX or whatever they call it, device drivers, application-level code, or what. These experiences have made me loyal to exactly zero video editor programs.


Is that first screenshot showing QT and the last one GTK, or is it just QtCurve or some other theme-trickery?


Per github page it's a Qt5 application. Qt can use Gtk2 to render its widgets, so that's probably what's happening in the ubuntu screenshot.


thank you, you've provided exactly what every software website, open source or not needs and often doesn't have. Good large screenshots, and a true walk through, not some wonky thing using animated paper cutouts or some other showing of benefits displacing actual features. If a realtor wants to sell a house, they can talk up benefits in the car all they want, but ultimately they need to show the nooks and crannies of the house for any traction.


> Leap Motion for jog and shuttle control.

I love this idea. I added the ability to scrub through the history of my paint app [1]. The circle gesture provides the direction, size, and speed which can be used to create a very expressive scrubbing tool. Fast tight circle scrub really fast and large slow circles scrub really slow with high precision. Can't wait to check out their implementation.

[1] http://youtu.be/DM1JtYyr88k


Not to mention Kdenlive, which is being actively developed again and has been the most capable and stable of the free video editors I've used this past year (including Openshot, Pitivi, and Flowblade)


And Blender! I've found it more intuitive, feature-rich and stable than anything else. And it installs straight from Debian's default repos.


Blender could almost replace aftereffects for a lot of us, but it does so much it would probably confuse a new user as to how you could replace AE.


is there a way to get Kdenlive in a single folder with all it's dependencies, isolated? or even the new version that's using KF5

I'm not using KDE and had a few problems in the past after pulling too many things.

> 0 upgraded, 109 newly installed, 0 to remove and 30 not upgraded. After this operation, 289 MB of additional disk space will be used.

edit: nevermind, daily-builds https://kdenlive.org/download-development but it still needs some things installed


Folks may also be interested in the Openshot cross platform video editor (no affiliation other than as a user). They use (and helped develop) some of the libraries used by Shotcut. http://openshot.org/

However openshot did commit the cardinal sin of second system effect. Version 1 was doing well, with a whole bunch of changes queued for version 2 including going cross platform, changing the UI toolkit, updating/changing the media toolkit, and a thousand other features. Version 2 has been imminent for well over a year now, and still doesn't seem too close to release.


The relationship extends further than that. Openshot 1.x was developed on top of the MLT (Media Lovin' Toolkit) Framework. MLT and Shotcut are developed by the same person.


Or if you want something that works now, you can use Blender, which has a very full featured video editor built in to it, and has had for a long time.


Blender is a good option too. You can cut and do things very fast using full timeline with overdrop or the regular flow.

IMO for anything more complex than that the UI gets on your way. I even learned most of the keybinds but it just doesn't feel intuitive.


+1 for Blender. Easy to install, stable, intuitive and feature rich. I've tried a lot of free video editors and Blenders been the only one that I've actually achieved anything in.


Partially OT, but does anyone know any open source video editor or better library for merging multiple videos into a 360 degree video? (python would be awesome)


MoviePy might do what you need, particularly the CompositeVideoClip class. See https://zulko.github.io/moviepy/ref/VideoClip/VideoClip.html...


Thanks, I'll look into it!


Is there a simple way to do a picture in picture effect (with unlimited levels of PIP?)

Trying to do this for a recent project with anything else just did my head in with complexity...

(I guess the best i can think of as a comparison is the TED intro? where lost of different clips fly across the screen)


aftereffects could probably be scripted for this.


Name looks too annoyingly like a typo for "shortcut".

How about "CTTC".

"With CTTC, you can trim your video of boring dialog, and just keep the exciting chase at the end!"

"CTTC is pronounced 'see-tick', or by uttering its expansion: Cut to the Chase."


I thought the same. At first I thought typo. Then I thought they took a shotcut on spelling shortcut. Then I realized it was for video apps.


also very interesting: https://github.com/jliljebl/flowblade


I've been using shotcut for over a year and I highly reccomend it. It is leaps and bounds ahead of most Linux-supporting video editors.


From UI perspective as well as the architectural design behind I would rather recommend Pitivi, as it's based on GStreamer.


Comparisons to iMovie?


Wish there was a Web app to edit video. Upload to S3 and away I would go creating (jump cut) edits which would be fed to ffmpeg to dice up.


youtube can do this. kaltura has something, too. nice for kids and building joke videos or maybe for building your next presentation.

however, pro video files are BIG, and upload will take longer than actual editing process with usual upload rates.

More important: nobody wants to have a bloated webapp that runs in an already bloated web browser (all of them are!) and eats all your cpu power just by executing some cool javascript gui framework - you need every bit of performance for rendering.


That sounds like an incredibly wasteful way of editing. Why would you want to use a web app?


Although video editing requires a level of performance that javascript is not equipped for, a web video editor is not a terrible idea.

I would imagine all the heavy lifting would need to be done by a plugin or a local media encoding server. Portability and a rich editing interface would be the key features. Collaboration and version control would be a nice feature too. I am surprised how terrible the top commercial video editing programs are at simply sharing a project file with another person. Video projects really need the cloud backing the source files and version control.

All of this can be done in native applications, but imagine being able to work on a video project from any modern browser. Maybe you don't even need a local media encoder, the server could render your changes and stream them as needed.

Edit: I would love to be able to edit video on my Macbook Air and have a GPU cluster in the cloud doing all my rendering.


I'm having trouble finding a source at the moment, but I believe that the BBC attempted this a few years ago. They eventually abandoned the project because as the editors needed the full resolution file of everything they were using to ensure the shot was in focus etc., storing them locally was a better option. I can't really see that changing in the future either.

An NLE that checked out files from a remote server and then made edits to a text based file from which anything could be shared would be very useful though.


Nice! I'm excited about this.

Why GPLv3? Why not make the license permissive as possible?


Although permissive licenses are great for libraries, they aren't that good for applications. Why would you let someone make a closed source version of your open source video editor? No, you never ever want that happens, that's when copyleft comes into play, it's ideal for applications like this.


> Although permissive licenses are great for libraries, they aren't that good for applications. Why would you let someone make a closed source version of your open source video editor?

Because you believe that open source actually works and is efficient in practice, such that even if someone makes and sells a proprietary derivative, it is more likely to grow the pie than to harm the open source base (and likely, if the downstream vendor is smart, to result in substantial contributions back to the core, because the downstream vendor can't get community maintenance to control costs without contributing stuff back to the community.)

Works for a variety of existing permissively licensed software for which their proprietary downstream versions, the sellers of which are also sponsors of and code contributors to the open source project.


Personally speaking: I think open source/permissive licensing works for libraries. I think it's been amply demonstrated that that doesn't work well at all for applications, and honestly that's the one place where I'll throw a bone to the free software folks.


> I think open source/permissive licensing works for libraries. I think it's been amply demonstrated that that doesn't work well at all for applications

PostgreSQL is an application, not a library. It is permissively licensed. It has at least one major proprietary downstream derivative that contributes back significantly to the core. Its usually recognized as a successful open source RDBMS.

Please explain to me how it is "amply demonstrated" that permissive licensing "doesn't work at all for applications"?


I am surprised that you would attempt to stretch the definition of "application", in the consumer-software context in which this discussion very obviously is, to server software.

Well, that's a lie, I'm not surprised.


> I am surprised that you would attempt to stretch the definition of "application", in the consumer-software context in which this discussion very obviously is, to server software.

I am still seeing a complete lack of support for the claim that it is "amply demonstrated" that permissive licensing "doesn't work at all for applications", even restricting "applications" to consumer applications.


OSX, Safari, apple never contributed back to the core Windows used (Net?) BSD network stack Google did not contribute much back to the linux kernel even though it's GPL though. So many products/compagnies are using FFmpeg and not contributing back. With that many risks already existing, it is safe to assume that most compagnies would not contribute back. Besides, postgresql is not an application, no Would Oracle have bought Mysql if it was BSD ? Probably not


How about Eclipse? I know non-programmers who use Bioclipse and Knime, just for example.

Maybe those aren't the greatest applications ever, but imagine what it might be like given a different upstream product with similarly permissive licensing?


> Why would you let someone make a closed source version of your open source video editor?

GPLv3 is a lot more restrictive than GPLv2 which is already restrictive.

Honestly, if had a GPLvX software and some guy came to me asking for a different license with the agreement that he would support it commercially I would be 100% ok with that (especially if he would agree to push bug fixes back to me - it would be a win-win for everyone).

How would that be any different than hiring a consultant to configure/support it in house? If you are using GPL software privately then you don't have to release the changes or the source. The only difference being is that if a whole company supported the one product they could have multiple people to work on it.


> Why would you let someone make a closed source version of your open source video editor?

How is that a bad thing? I think that's a zero sum game way of looking at it.

No damage is done if someone makes a closed source version. They can add value and stimulate competition. More often, they fail.

Occasionally, core developers may enrich and provide additional functionality upon a permissive core. Why not look at Postgres as an example?

And what harm is done? Lost code contributions? Some have no interest in collaborating on software they can't incorporate / copy from later without restriction. It takes time to wrap ones brain around a codebase.

Why would someone take time to understand the internals of a GPLv3 app, when they could never incorporate pieces of it in their code later on?

> it's ideal for applications like this.

Ideal is subjective. Judging by the above, it seems as if enforcing envy trumps programmers doing as they wish with the code.


> Why would someone take time to understand the internals of a GPLv3 app, when they could never incorporate pieces of it in their code later on?

...to make it better for their uses, rather than their products? "Enforcing envy," though--please, be more of a jerk.

I'm curious as to what large-scale applications have you developed and released under a permissive license that make you such an authority on what others should do with tens of thousands of their man-hours.


If there is no damage done when someone makes a closed source version, then there is no damage done when some pirates that closed source version. Some have no interest in paying for the closed sourced version anyway, and why should the closed source developer enforce his envy against users doing as they wish with the code?




Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact

Search: