 Okay, I'm excited to present the first talk of today, which is by Jean-Baptiste Eugen and Massimo. Will tell us about how KDN Live became the great and awesome video editor it is today. Okay, thanks for your welcome. I am Jean-Baptiste Mardel, I have been the KDN Live maintainer for the past 18 years and I'm happy to be here with you at Academy. Some of you might already know me as I have been on stage a few times already, but it's the first time I'm not alone on stage, so I have some team members with me, so it makes me really happy because KDN Live is really a community and a team project, so it's great to not be alone on stage. So I'm just going to quickly introduce you with the team. So we have been mostly having online meetings in the last years, we hope to be able to make a stream soon. In the team who couldn't make it today, there is Farid, Julius, Remy and Kami, so I just want to say hi to them today. So this is the whole team with Massimo and Eugen. So I will now let the other members of the team introduce themselves and keep up. Good morning everyone. My name is Eugen Moore, I'm joining the KDN Live core team in 2018. I'm mainly behind the scene, the coder is Jean-Baptiste Mardel. I'm more about documentation and bug tracking and the homepage and all what is behind the scene needed that we can keep KDN Live alive. Hello everybody, I'm Massimo, I'm a video editor and filmmaker and a consultant of the project, it's ten years that I'm on the project and I don't know what to say, I'm very happy to be here because I'm very fond of this project and of the guy that works with us to create what we are doing and as we are a video application to me the best way to introduce ourselves is by showing you a video. KDN Live is a community project which aims to deliver a free and open video editing software application to allow everybody to produce quality content in order to increase the democratization of the media. The application is a graphic interface written in C++ with Qt and KDE libraries for the MLT framework written by Dandenity which relies on FF, MPEG to decode and encode almost all the video and audio formats that are out today and which host effects libraries like Freyor and Mubit for video and Ladspot and SOX for audio. By following the official documents that we can find on the internet the project was launched by Jason Wood who released version 0.2.3 in October 2003. Soon KDN Live 0.2.4 followed but the community did not yet exist and the group was very small. The project stopped for two years. Before the end of 2005 Jean-Baptiste Mardel who heard that the project would be reactivated offered his help and in 2006 he signed the post of the new release 0.3 of the program. From this moment on he became the main reference point for the project. Version 0.4 and 0.5 are soon distributed but there is an issue. A refactoring of the code is needed to go forward. The program has to be moved from KDE3 which is not compatible with MLT to KDE4. The rewriting ends in 2008 followed by several new releases but in 2011 a further migration from KDE4 to KDE5 was needed to allow the program to grow. In 2012 a crowdfunding campaign is launched to fund the operation and before the end of 2014 the goal is reached. Then finally in 2015 KDE4 becomes an official KDE application. Jean-Baptiste was invited to Academy to present 10 years of activity. The new perspective is to make the project even bigger. Before the end of the year the first KDE4 is announced with the aim of getting more people involved. During this virtual meeting it immediately became clear that in order to grow KDE4 needed to be cross-platform. The Windows version was announced and 2016 began with a sprint meeting followed by the new logo and the new site. But as soon as the development of new features started it was evident that the code for the timeline had to be rewritten because it was too old and no longer fit. Everybody knew that the refactoring could take several months or years but the community did not lose enthusiasm. In 2018 a new roadmap was written and in 2019 the refactored code was distributed. Although some fine-tuning was still needed. Then finally since 2020 and continuing up to now the long-awaited new features along with a plethora of new effects and dozens of important improvements have been continuously added to create the most powerful, free and open video editing application ever. We have about 100,000 users using Canelife daily or weekly or frequently. If you have saw a lot of users you get a lot of feedback as well so if the software is running well you get no feedback. If we have a release like 23 or 4 the bell is ringing after 30 seconds. You have issues, a lot of bugs not usable so have a lot of user feedback very soon if the software is not running well. So how to communicate with all these users? We try to keep the communication channel as lean or as simple as possible so we have the communication channel for notification of the software is our home page. It's Twitter, sorry it's Twitter and Musto and Mustodon. The communication itself with our user, we have these channels it's Matrix, it's our forum discuss it's the feedback when something happens or for wishes it's Baccala and we have quite a big community on Reddit that give us feedbacks. And it's sometimes really a challenge to feed and check all these channels to have some other quad feedback to the supplier as well. So this one keeps about 2 up to 3 people really busy behind the scene just to communicate with the users. Downloads, you see here a curve over time these are the downloads of Windows and Mac and Appimages we have a huge peak on the beginning of COVID it looks like users like to have some video editing possibilities and as you can see we can keep the amount of downloads or users over time as well. What you do not see on this curve is there's a similar curve for the unique visits on our home page and we just found out the people going on the home page looking for the download link, clicking on that and download the application that leads to us that we refurbish the home page to make the user easier to find the download page. What is missing on this curve is flatback downloads and the PPA downloads and we do not really know as the distribution is via mirror server we do not know the download numbers from the mirror servers. Who is using CadenLive? It just made an actual poll last week and found out about 40% is social media, it's YouTube, it's PeerTube, Tiktok. I know schools using our software as well so I get asked years ago that we can prepare a download version which is possible without additional software so I download the packed Windows version and repack it in a self extraction find so that the schools do not need additional software installed and do not need administration rights to install the application. This gives about 10% of our users and interestingly we have about 7% professional users, I think. It's wedding, editing, most probably, some TV we know, some movies we know as well and then for the majority only up to 45% is personal use doing their vacation holidays and editing like this one. What are the platforms used with CadenLive? Mostly half-half Windows, Linux and a growing community on macOS we have about 5000 macOS downloads on the macOS it works only on the Linux CPU at the moment. With such a community you give a lot of feedback and you get interesting feedbacks as well from the user when you're listening to it. Remember the YouTube part we have. So the community asks about subtitles, the community asks about markers the video to get the chapters in YouTube. So we investigate and found out there is a framework FOSC which can generate text from speech and can generate automatically subtitles and even you have the possibility then going in the text field click on the text and the playhead jumps on the exactly same position. Yeah, subtitle generation and even we improved that one founding another thing lately is whisper. It's really crazy. This founds the notation of a sentence, the punctuation and it's really more accurate than FOSC I would say translation and even you can make a translation to English and it's about 90-95% accurate in the meantime. So people coming and asking how to donate to KDNK Life. There was a lot of asks from the community. They do not like to donate in a big pot KDE. They like to spend their money directly to KDNK Life. So we go to the KDE headers, I don't know. And KDE starts a pilot project with KDNK Life that users can donate directly to KDN Life. We started this donation on the 20th of September 2022 and reached our goal within a month. We collect 15,000 euros within a month. It's absolutely crazy. It just went up. I made a forecast by myself and my forecast said within a month we should reach the goal. In the meantime it's about 27,000 I heard yesterday because it runs in the background further on without some additional needs let's say. This gives us the possibility to implement really bigger features like the nested timeline which we'll talk about. We can now work on stability and we will work on performance as well to make the software even more performant. Documentation, first we thought. Documentation, secondary thing. It's not so important. It steals just our time because it's a huge amount to spend just within documentation. We checked the statistics and found out we have 3,500 views per week on the documentation itself. And then we can further down analyze the statistics. We found out a third. It's only checking out the effect part and another third checks only out getting started. The interesting thing is the people stay one up to one and a half minutes on the page itself which is for internet let's say really, really long. We calculated that roughly 10 up to 15% of our users use our new documentation manual and at the moment we are ongoing with updating the whole effects part that the user has a better overview of effects, what they are doing, how to implement these effects and with this new possibility here in Sphinx it's really nice you can put in some keywords and have really a list where these keywords are found in the documentation. This is much more easier than with the old system. Yeah, copyright. This is a bit of weak part of CadenLive let's say because the more users you have, you have 100,000 maybe even more, they start stealing CadenLive meaning Google advertisements, links to other software. We have such cases. Or just take the name behind is just a money collector something like this one. So we start protecting CadenLive as good as possible. The logo and the name is copyright on the KDE at the moment and we going on and having the source code each file with a copyright, an SPDX copyright I think really sure we protect our work, our 20 year long work now with a proper copyright which is really tricky. The goal is to keep CadenLive free without cost for the future. Thank you. What makes CadenLive so successful? One thing that I think is very important for CadenLive success is that we have a vision for the project. This is where Massimo I will talk a bit later about it has a very important role because we have a focused vision of what we want to achieve, what we want users to be able to do for CadenLive so I think it's an important part to have a vision for the project. We also try to listen to our users, so the two things are linked I think it's important to have an idea of where we want to go but also to listen to user input and to do the best of the two things. And also at one point we also realized that doing too much things was not good. At one point we had a video capture, we had a DVD wizard, we had a stop motion widget so we had lots of features but everything was working not so good. So we tried at some point a few years ago we really decided to refocus to have less features but to try to make them work better. I think that's one important thing also in the project evolution. Now the things that we don't do so well, it was really quite, the last release 23 or 4 was a good demonstration of what we don't do so well. So we'll try to improve automated testing. We have some automated testing since a few years but we need to improve to increase the testing. We also, since a few years we've talked about organizing test days before the release because we have a large user community to organize test days to gather feedback before the release. It's something that we never really managed to do. Another issue that we have is that KDNLive is a big project with many, many dependencies so we depend on MLT as a video engine which itself depends on FFMpeg, OpenCV, Friar and tens of other projects. So if one thing changes in our dependency it can affect the whole application. So if, for example, a default parameter for an effect changes then when we open KDNLive this can lead to a change. This is something that we cannot detect in our testing. So we have plans to improve. One thing that we want to do is continue improving our current test base. Another thing to prevent this issue with dependencies affecting our application is to have a rendering test shoot that's something that we started to work on. It means that we have automated scripts that open a project, render the video file and compare it to a reference rendering. And this way we can detect if anything changes in our dependencies that affects the rendering this will be detected by this test shoot. So this is a project that we want to definitely implement before the end of the year that will also bring better regression testing. Because that's also one thing with increase in popularity. We have lots of users and also the expectations are a bit higher so we just can't keep breaking everything on each release so we're really trying to go in a more professional way and have better testing, regression testing. Another thing that we want to explore is something that will be talked at Academy is UI testing with Selenium. This is something that we don't do at all but I think it's also an interesting approach because this is something that we absolutely don't test and some parts of the UI would really need to be tested too. So now what changed over the years? I already talked about the popularity of the software. Another thing is that before ten years ago we only had Linux versions so packaging was very easy. We just prepared an archive of the file and we let distributions do the work. Now we provide packages from many platforms. We have Windows, Mac, App Image, Flatpak, PPA, Snaps, a few more. So it's a huge job to do this packaging. This is something that really changed over the years. I think it has some really nice advantages. It brings our code faster to the users so we can also react faster to the feedback but the problem is that it's a lot of work. We use the KDE tools to build these packages. We have craft and we have support from KDE people and KDE infrastructure but still for a small team like we are it sees a huge work that we can't do as well as we would like. So this is still an issue. Another thing that changed over the years is us, the team. We decided also a few years ago that we wanted to go outside of our Linux users' bubbles to grow, to have Windows versions, macOS versions. I think it's important for software if we want to grow. But the problem is that we are a small team. Life also makes that, for example, I have a job, now I have a kid, so I have a bit less time. So we had to somehow find ways to make sure that KDE Live can keep growing. So we launched Fundraiser. The idea behind the Fundraiser was also that I had the opportunity to reduce my time at my day job. Currently I have one day per week that I can spend on KDE Live so it's not so much to maintain such a huge project and I had the opportunity to lower and the goal is to have two days per week for KDE Live. So we launched this Fundraiser, a fantastic Fundraiser. It was very successful as Eugene said. We collected money and, yes, as it was also mentioned yesterday, we have not yet spent any money. So why is this? As I said, I had the opportunity to lower my day job but it takes some time. So I hoped I could just switch like this but it takes some more time. So the idea is really that this money will be used when I have some extra time. We also want to hire a few people, maybe for some dedicated tasks, but that's something also that we are not very sure how to manage the hiring. All these things take some time and since managing KDE Live is already a big, huge thing so things take more time than we planned or than we would like. And now it's demo time. I think we talk a lot but I think it's also nice to demonstrate a few of the things that we can achieve with KDE Live. It can come, it's going to demonstrate a few features that were recently implemented, like nested timeline that caused a lot of regressions but we are working on it to fix it and it has some nice advantages and also speech-to-text. Thank you for the question. We should know why you can use this one because we come to the list too. Okay, so we load here our video that we just watched before so we can scroll it and we can have the first timeline already load here in this space so for example, we can decide that we want to edit in our space, for example, this animation so we start for the first frame, we play it maybe faster, so... We go back and we select the in-out point who knows already how to work the program, knows what I'm doing, so I set where I want to start where I want to end my clip, I can drag it on the timeline or I can insert straight away with our automatic system so I can select it on the point, for example, it could be the conclusion, for example, of our video, starting from here and go to the end. I just do it, I set in point here and for example, we stop here so now we have a first timeline with our clip now I create, for example, another timeline in our sequence folder this is to create a new sequence so we can select with the title of the sequence the number of video tracks, the number of audio tracks following our needs we create a new timeline and we can see we have two tabs here we can switch from one to the other for example, we load again our... sorry, we load again see, we have one and two we load again our clip and for example, we can decide now we need the beginning of the video and so we set in point on the beginning of our animation here we can pull up to the... the title and then we can, for example, select when we speak about Academy we are Academy, so well, we don't speak about Academy so we select the part... first I have to insert the part I selected then I go here I select all the situation about Academy then... now what we have? we have two different timelands but not just two different timelands also two different clips because if I can load the clip inside our monitor and you can see we can scroll and actually trim again in an out point on our new clip and now with the second clip that is as I show you contain exactly the part we decide to have you know we can, for example, add the sequence one inside the sequence two for example, so we now have our more the other part or maybe we can even create another third sequence where we can of course we decide the title where we can put for example our sequence all together have another another sequence I don't know why I didn't take the audio to all the sequence one we do it ok and now we have both the clip of course we can trim it, we can slip for example I can click and trim that one I can reverse it, they work exactly they work exactly like with this keyboard ok I can do I don't know what I did here oh ok I lost myself anyway I can trim and refine the trimming of the clip and so on so what is need this system this system for example if you have several footage for a project for example that is the same footage that you can share for different you make to a series of episodes of the same footage you can load in the same project all the footage you need and then you edit different version of episode 1, episode 2, episode 3 and then you can export each timeline differently separately in other videos and you can also recycle edit part from a video in another video for example if you created a timeline with a main title of something you can reuse the main title in all the series without redoing the same piece of timeline in the other timeline another stuff I want to show you is the subtitle we speak about before so we create another timeline at this point just to have a free space and we take a piece of our video for example when we speak about I have no audio here now ok I go blind so I don't know we are selecting because we speak about subtitle we should listen a little bit of audio but it's ok so just to make a very short test we have this clip here so in our point ooh what I'm doing ok so we open the subtitle track that is here we click on magic wand and we can timeline zone as you can see there is the VOSC model or the model we can select here now we have VOSC activated so the selected clip the timeline zone and we can proceed on the subtitle generation so in this moment ok we have just this table because it's what I selected we are very short so if we have more more time a bigger clip we have all the synchronized part of this subtitle can be moved can be resized can be copied and can be also edited in this part also the text can be changed in the area with custom text change the font, change the position change the color it depends what you want to do this automatic subtitle we have the clip here we have the text editor and then by using the for example we have an interview we can extract the text and edit the interview by just using the text and so we can export the edited the final interview based on the section of the text we have lighted and by dumping the rest of the text we don't need what we want to do what comes from the idea that for one side we listen to users because they use the community to use it maybe have some ideas that maybe very the idea that we never had so we collect the input that come from the community that make the software stronger we know that even property software is fine to us because we introduce first what they are watching at us and from the other side we have experienced people like me that are working in the field so they are in contact all the time with the software of big corporation software so what editor needs on the field in the company that uses software so we try to take best practice with the fund around to make it the best possible so we this idea to implement the best practice that video editing needs is our vision and for the future we have to finish between the end of the year what the donators aim to have so we are in the road to reach the before the end of the year the goal is what we will create and for the next year we think that we will go forward with insert some AI tools and mostly collaborating in space online to I finish to make a collaborative job between editors around the world so make it so if you want to proceed with question where do you see in the special video editing in Linux in the future when most default major companies that make cameras have proprietary formats that are not supported on Linux there is just it doesn't work it won't work and there is absolutely no documentation to start building the infrastructure so you can use formats such as red, cinemas, digital cameras that take raw footage Sony has another format Apple is famous for their products there is no support for that and those companies are doing nothing to open source the standard so we can have support so where do you see that professionally coming how can we fix that you go there is a problem about that companies try to close more possible but they have also arrangement with other companies that provide software so if you buy sometimes some tools you have to finish to work on some specific platform this is actually it's not something that professional like too much some like because of old mentality like they don't think about opening most close so I buy something that works like I buy a fridge a freezer so yes to work this way but a lot of people mostly younger don't think this way anymore so they would like to be open to use what they prefer because it means I change the camera I have to migrate to another software no I want to use my software and it's always actually a form of scam because actually at the end of the day they are always the same most of the containers are close property stuff but what is inside the container because video files are made by containers so the caps that contain the stream and the stream that is encoded by MPEG or iFrame etc so mostly they use the same way to compress the stream so that the codec is encoded but maybe they try to have different caps that are contained to close but I see that at the end of the day for example FFNPEG is able at the end of the day to work around the problem there is some problem of patents of course of copyright around this stuff but overall so far I was able to manage to use almost every stream I found in the worst of the option but this happened also in Avid before the software that sometimes you have to make conversion to match the specific software so nothing is changed for real so it is a challenge of course and we have to be open to new format to change all the time also because new compression new way to broadcast etc. I ask the man to be so far we are doing pretty well because I am able to use it on the field and it is working so far maybe one thing also that we should improve and do more we have already a small presence in schools but we really should do more also we had a standard all-american film festival I think this at the beginning of this year so making also students aware of these issues I think is something that will help in this field thanks I would also like to point out to the person who asked before there is also the academy software foundation which depends on the linux foundation that tries to open source it tries to get industry companies to open source as much as possible and they are quite successful they have got many formats that are open for video effects and video effects tool chains and for timelines etc I think that you are actually implementing some of these technologies already, is that correct? we for example incorporate open time it is true that there is an attention of open software mostly also in Hollywood for example the fact that the company want to customize their needs so they look for Pixar or Disney so they BBC so they deliver what they do so because they develop in the open source way they deliver so they create standards this way that have a return from them because if everybody adopt their standards they do not control their standards but anyway they create an integration so I hope personally that the direction will be that at the time and also in companies like Sony that I quote Sony they will not stand the impact of this mentality that is becoming more spread around the industry in my opinion there is a question from online will you use any collected funds to support some of the upstream projects and infrastructure that you depend on might you be able to get support from video and computer equipment manufacturers? so yes that is something that we we planned we had contacts with MLT maintenance video framework that we use to propose them to use some of the funds unfortunately they were not able they didn't have the time or wish most of the time of the need for this but yes that is something some of the core issues in KDNL I related to our dependencies so one of the tasks in the first fundraising was a task that has to be implemented in MLT the video framework yes that is something that we will do I think you said that you are working 18 years on KDNL now is it still fun can you keep going yes it's still fun that gives me going because also the world of video editing really really changed during these 18 years so there is always new things to try to test so I really enjoy it maybe I can add something most important for him is that he has a team behind who helps him I think years ago we had almost without doing everything by themselves and now he has a team behind who helps him keep him away the daily tasks like home page updates stuff like this one so he can really focus on coding we feed him what is needed or what is a bug could be as good as possible that he just can really hack on the code and always new challenges and I think this is really important that you have a group behind and not a solo guy because this doesn't work over such a long time and just thinking about team of course we are open so everybody wants to help us it's welcome in every aspect with packaging coding and supporting testing using whatever so this is an interesting point without advertising we have about two or three peoples coming from outside and ask can I help on this and that and we open them and say yes you can help because a good example is the documentation we have a guy now who is heavily reading writing documentation we have other guys like Davy on the background on the windows side who checks the forum part so it's really interesting that you get on certain stage some help from outside people ask ask you how can I help and this is something I think keeping a chomp up this the life yes yes thanks no more question I guess ok thank you very much for the great talk