This is Your Brain On Surveillance: New Study Reveals How Awareness of Being Watched Alters Our Brains - The Debrief
Tags: tech, surveillance, psychology, cognition
It looks like it’s not only impacting negatively our privacy. The linked paper (good to read as well) hints at negative impacts on mental health as well. Still needs to be fully validated but it doesn’t look good already.
You love artists and their music? You probably should get off Spotify then… because they’re clearly at war to reduce even further how much they pay artists. Clearly it’s not about discovering artists anymore, it’s about pumping cheap stock music to increase their margin. Also its clear the remaining musicians trapped in that system will be automated away soon… you don’t need humans to create soulless music.
Like everyyear I take a couple of days off at the end of the year to wind down and spent time with the family. The year has brought many major changes, both to KDE and to me personally: We did the KDE MegaRelease 6, the next major update to KDE’s software suite. Plasma 6 further made Wayland the default graphical session. I also spent a lot more time in Qt itself, particularly Qt Wayland, rather than KDE code. Anyhow, between family visits and feasts there’s always some time for quality KDE hacking.
That’s right: Monitoring task progress in Konsole while busy doing something else
I’ve always been a huge fan of Windows 7’s task bar with its progress reporting and Jump Lists. Nine years ago (wow, really?!) I added support for the Unity Launcher API to Plasma’s task bar in order to display download and copy progress. The other day I was browsing systemd changelog when I stumbled upon:
The various components that display progress bars […], will now also issue the ANSI sequences for progress reports that Windows Terminal understands. Most Linux terminals currently do not support this sequence (and ignore it), but hopefully this will change one day.
I hope so, too! Guess whose Konsole understands ConEmu-specific OSC (Operating System Command), the stuff systemd uses, for progress reporting now? There’s still a few quirks to be worked out since Konsole allows you to have multiple split views within the same tab. Nevertheless, we’ve got plenty of time until the next KDE Gear release in April 2025 to finalize it. Moreover, I asked kde-builder (KDE’s meta build system and spiritual successor to kdesrc-build) to support it, so you could monitor KDE compile progress at a glance.
I’m a scratch-your-own-itch type of guy. When I finally got fed up with Element (a Matrix chat client) in a browser window eating my CPU, I gave our own NeoChat application a try. The first thing I added was a “Copy Link Address” context menu when hovering a link in addition to fixing the missing “Edit” entry. Next, I had the window title include the chat room name since that’s what I am usually looking for in my task bar. Finally, Kirigami’s Avatar component can now load its image asynchronously which should speed up scrolling through the timeline and lists of rooms and users.
Ink marker colors shown again in redesigned Printer Manager
Speaking of Kirigami, Qt 6.8 added an animateClick function to buttons. It briefly flashes and then triggers it. This is now used throughout Kirigami in keyboard shortcut handling, bringing it in line with the Qt Widget world. Qt 6, too, has a concept of “accent color” for a few releases. Plasma’s accent color system predates it, though, so there’s some friction between the two. While we don’t have a proper Kirigami Theme API for it yet, at least setting the highlight now also sets the accent color. With that, ink cartridge levels have the appropriate marker colors in printer settings again. Speaking of accent color, I just backported some changes we made for Frameworks 6 to Frameworks 5 to ensure that KF5 apps can interpret Breeze Icons from KF6 properly, notably fixing the black folder icons.
I hope you also got the chance to spend some time with your loved ones. If you enjoyed what the KDE Community brought you this year, please consider donating to our Year End Fundraiser or to me personally, so we can continue rocking in 2025!
AudioTube now shows synchronized lyrics provided by LRCLIB. This automatically falls back to normal lyrics if synced lyrics are not available. (Kavinu Nethsara, 25.04.0. Link)
Quickly renaming multiple files by switching between them with the keyboard arrow keys now correctly starts a renaming of the next file even if a sorting change moved it. (Ilia Kats, 25.04.0. Link)
Fixed a couple of regressions in the 24.12.0 release. (Akseli Lahtinen, 24.12.1. Link 1, link 2, link 3)
We improved how we are displaying the signature and certificate details in the mobile version of Okular. (Carl Schwan, 25.04.0. Link)
When selecting a certificate to use when digitally signing a PDF with the GPG backend, the fingerprints are rendered more nicely. (Sune Vuorela, 25.04.0. Link)
It's now possible to choose a custom default zoom level in Okular. (Wladimir Leuschner, 25.04.0. Link)
Merkuro Mail now lets you search across your emails with a full text search. (Carl Schwan, 25.04.0. Link)
Additionally, the Merkuro Mail sidebar will now remember which folders were collapsed or expanded as well as the last selected folder across application restarts. (Carl Schwan, 25.04.0. Link)
This week, Joshua spent some time improving Tokodon for mobile and
in particular for Android. This includes performance optimization, adding
missing icons and some mobile specific user experience improvements. (Joshua Goins, 25.04.0. Link 1, link 2 and link 3). A few more improvements for Android, like proper push notifications via unified push, are in the work.
Joshua also improved the draft and scheduled post features, allowing now to discard scheduled posts and drafts and showing when a draft was created. (Joshua Goins, 25.04.0. Link)
We also added a keyboard shortcut configuration page in Tokodon settings. (Joshua Goins and Carl Schwan, 25.04.0. Link 1 and link 2)
Finally, we created a new server information page with the server rules and made the existing announcements page a subpage of it. Speaking of announcements, we added support for the announcement's emoji reactions. (Joshua Goins, 25.04.0. Link)
For a complete overview of what's going on, visit KDE's Planet, where you can find all KDE news unfiltered directly from our contributors.
Get Involved
The KDE organization has become important in the world, and your time and
contributions have helped us get there. As we grow, we're going to need
your support for KDE to become sustainable.
You can help KDE by becoming an active community member and getting involved.
Each contributor makes a huge difference in KDE — you are not a number or a cog
in a machine! You don’t have to be a programmer either. There are many things
you can do: you can help hunt and confirm bugs, even maybe solve them;
contribute designs for wallpapers, web pages, icons and app interfaces;
translate messages and menu items into your own language; promote KDE in your
local community; and a ton more things.
You can also help us by donating. Any monetary
contribution, however small, will help us cover operational costs, salaries,
travel expenses for contributors and in general just keep KDE bringing Free
Software to the world.
To get your application mentioned here, please ping us in invent or in Matrix.
Lots of KDE folks are winding down for well-deserved end-of-year breaks, but that didn't stop a bunch of people from landing some awesome changes anyway! This will be a short one, and I may skip next week as many of us are going to be focusing on family time. But in the meantime, check out what we have here:
Notable UI Improvements
When applying screen settings fails due to a graphics driver issue, the relevant page in System Settings now tells you about it, instead of failing silently. (Xaver Hugl, 6.3.0. Link)
Added a new Breeze open-link icon with the typical "arrow pointing out of the corner of a square" appearance, which should start showing up in places where web URLs are opened from things that don't clearly look like blue clickable links. (Carl Schwan, Frameworks 6.10. Link)
Notable Bug Fixes
Fixed one of the most common recent Powerdevil crashes. (Jakob Petsovits, 6.2.5. Link)
Recording a specific window in Spectacle and OBS now produces a recording with the correct scale when using any screen scaling (Xaver Hugl, 6.2.5. Link)
When using a WireGuard VPN, the "Persistent keepalive" setting now works. (Adrian Thiele, 6.2.5. Link)
Implemented multiple fixes and improvements for screen brightness and dimming. (Jakob Petsovits, 6.3.0. Link 1, link 2, link 3, and link 4)
Auto-updates in Discover now work again! (Harald Sitter, 6.3.0. Link)
Vastly improved game controller joystick support in Plasma, fixing many weird and random-seeming bugs. (Arthur Kasimov, 6.3.0. Link)
For printers that report per-color ink levels, System Settings' Printers page now displays the ink level visualization in the actual ink colors again. (Kai Uwe Broulik, 6.3.0. Link)
Pager widgets on very thin floating panels are now clickable in all the places they're supposed to be clickable. (Niccolò Venerandi, 6.3.0. Link)
Wallpapers with very very special EXIF metadata can no longer generate text labels that escape from their intended boundaries on Plasma's various wallpaper chooser views. (Jonathan Riddell and Nate Graham, Frameworks 6.10. Link)
Fixed one of the most common Qt crashes affecting Plasma and KDE apps. (Fabian Kosmale, Qt 6.8.2. Link)
125 KDE bugs of all kinds fixed over the last week. Full list of bugs
Notable in Performance & Technical
Significantly reduced the CPU usage of System Monitor during the time after you open the app but before you visit to the History page. More CPU usage fixes are in the pipeline, too! (Arjen Hiemstra, 6.2.5. Link)
Plasma Browser Integration now works for the Flatpak-packaged version of Firefox. (Harald Sitter, 6.3.0. Link)
How You Can Help
KDE has become important in the world, and your time and contributions have helped us get there. As we grow, we need your support to keep KDE sustainable.
You can help KDE by becoming an active community member and getting involved somehow. Each contributor makes a huge difference in KDE — you are not a number or a cog in a machine!
You don’t have to be a programmer, either. Many other opportunities exist:
You can also help us by donating to our yearly fundraiser! Any monetary contribution — however small — will help us cover operational costs, salaries, travel expenses for contributors, and in general just keep KDE bringing Free Software to the world.
I started this blog back in 2010. Back then I used Wordpress and it worked reasonably well. In 2018 I decided to switch to a static generated site, mostly because the Wordpress blog felt slow to load and it was hassle to maintain. Back then the go-to static site generator was Jekyll, so I went with that. Lately I’ve been struggling with it though, because in order to keep all the plugins working, I needed to use older versions or Ruby, which meant I had to use Docker to build the blog locally. Overall, it felt like too much work and for the past few years I’ve been eyeing Hugo - more so since Carl and others migrated most of KDE websites to it. I mean, if it’s good enough for KDE, it’s good enough for me, right?
So this year I finally got around to do the switch. I migrated all the content from Jekyll. This time I actually went through every single post, converted it to proper Markdown, fixed formatting, images etc. It was a nice trip down the memory lane, reading all the old posts, remembering all the sprints and Akademies… I also took the opportunity to clean up the tags and categories, so that they are more consistent and useful.
Finally, I have rewritten the theme - I originally ported the template from Wordpress to Jekyll, but it was a bit of a mess, responsivity was “hacked” in via JavaScript. Web development (and my skills) has come a long way since then, so I was able to leverage more modern CSS and HTML features to make the site look the same, but be more responsive and accessible.
Comments
When I switched from Wordpress to Jekyll, I was looking for a way to preserve comments. I found Isso, which is basically a small CGI server backed with SQLite that you can run on the server and embed it into your static website through JavaScript. It could also natively import comments from Wordpress, so that’s the main reason why I went with it, I think. Isso was not perfect (although the development has picked up again in the past few years) and it kept breaking for me. I think it haven’t worked for the past few years on my blog and I just couldn’t be bothered to fix it. So, I decided to ditch it in favor of another solution…
I wanted to keep the comments for old posts by generating them as static HTML from the Isso’s SQLite database, alas the database file was empty. Looks like I lost all comments at some point in 2022. It sucks, but I guess it’s not the end of the world. Due to the nature of how Isso worked, not even the Wayback Machine was able to archive the comments, so I guess they are lost forever…
For this new blog, I decided to use Carl’s approach with embedding replies to a Mastodon. I think it’s a neat idea and it’s probably the most reliable solution for comments on a static blog (that I don’t have to pay for, host myself or deal with privacy concerns or advertising).
I have some more ideas regarding the comments system, but that’s for another post ;-) Hopefully I’ll get to blog more often now that I have a shiny new blog!
Yet another long piece in this interesting and in depth conversation about Bluesky. The fact that it stays civil is called out explicitly and this is appreciated.
Bluesky is already hitting growth pains regarding moderation and its guidelines. By being centralized it is also more at risk within the current US political climate.
Kind of unsurprising right? I mean LinkedIn is clearly a deformed version of reality where people write like corporate drones most of the time. It was only a matter of time until robot generated content would be prevalent there, it’s just harder to spot since even humans aren’t behaving genuinely there.
A good balanced post on the topic. Maybe we’ll finally see a resurgence of real research innovation and not just stupid scaling at all costs. Reliability will stay the important factor of course and this one is still hard to crack.
It looks like analog chips for neural network workloads are on the verge of finally becoming reality. This would reduce consumption by an order of magnitude and hopefully more later on. Very early days for this new attempt, let’s see if it holds its promises.
I wouldn’t use it as much as advocated in this article, still this is a good reminder that Java became way more approachable for smaller programs in recent years.
How do you do, fellow web developers? A growing disconnect.
Tags: tech, career, complexity, learning
It tries hard at not being a “get off my lawn” post. It clearly points some kind of disconnects though. They’re real. I guess it’s to be expected with the breadth of our industry. There are so many abstractions piled onto each other that it’s difficult to explore them all.
Visitor Pattern Considered Pointless - Use Pattern Switches Instead
Tags: tech, design, pattern, java, type-systems
One of my favorite of the traditional design patterns in object oriented languages. Now obviously when you get pattern matching in your language… you don’t need the visitor pattern anymore.
Estimating projects sells them short (and that’s okay)
Tags: tech, project-management, estimates
I don’t exactly use this approach to factor in the uncertainty… but I guess there’s something to be made out of this proposal. I’ll keep it in mind for my next project.
Interesting ideas about leadership lacking in impact. Indeed it should be seen as a communal function, it’s not about individuals leading each in their own directions. Think about it in a systemic way.
In recent weeks we have been working on transferring LabPlot’s documentation to a new format.
We decided to move the documentation from the DocBook and MediaWiki format to the Sphinx/reStrcutredText framework. In our perception Sphinx offers a user-friendly and flexible way to create and manage documentation. Easy math typing and code formatting also come along. Additionally, Sphinx supports basic syntax checks, and modern documentation practices, such as versioning and integration with various output formats like HTML, PDF and ePub.
The new user’s manual is available on a dedicated page: https://docs.labplot.org. Please check it out and let us know what you think.
The manual still needs to be supplemented with new content, so we encourage you to contribute to the documentation, e.g. by fixing and adding new sections, updating images, as collaborative efforts can lead to a more comprehensive resource for everyone. Please check the Git repository dedicated to the documentation to find more details on how to help make it better.
We started this project with the intent of providing users a tool helpful in inking sketches. It is based on a research article by Simo & Sierra published in 2016, and it uses neural networks (now commonly called simply AI) to work. The tool has been developed in partnership with Intel and it’s still considered experimental, but you can already use it and see the results.
In the section below there are some real life examples of use cases and the results from the plugin. The results vary, but it can be used for extracting faint pencil sketches from photos, cleaning up lines, and comic book inking.
Regarding the model used in the tool, we trained it ourselves. All the data in the dataset is donated from people who sent their pictures to us themselves and agreed on this specific use case. We haven’t used any other data. Moreover, when you use the plugin, it processes locally on your machine, it doesn’t require any internet connection, doesn’t connect to any server, and no account is required either. Currently it works only on Windows and Linux, but we’ll work on making it available on MacOS as well.
Use cases
It averages the lines into one line and creates strong black lines, but the end result can be blurry or uneven. In many cases however it still works better than just using a Levels filter (for example in extracting the pencil sketch). it might be a good idea to use Levels filter after using the plugin to reduce the blurriness. Since the plugin works best with white canvas and grey-black lines, in case of photographed pencil sketches or very light sketch lines, it might be a good idea to use Levels also before using the plugin.
Extracting photographed pencil sketch
This is the result of the standard procedure of using Levels filter on a sketch to extract the lines (which results in a part of the image getting the shadow):
Another possible result is to just stop at the plugin without forcing black lines using Levels, which results in a nicer, more pencil-y look while keeping the lower part of the page still blank:
Here in the pictures above you can see the comic book style inking. The result, which is a bit blurry compared to the original, can be further enhanced by using a Sharpen filter. The dragon was sketched by David Revoy (CC-BY 4.0).
Cleaning up lines
Examples of sketches I made and the result of the plugin, showing the strong and weak points of the plugin. All of the pictures below were made using the SketchyModel.
On the pictures below, on the scales of the fish, you can see how the model discriminates lighter lines and enhances the stronger lines, making the scales more pronounced. In theory you could do that using the Levels filter, but in practice the results would be worse, because the model takes into account local strength of the line.
(Optional) Install NPU drivers if you have NPU on your device (practically only necessary on Linux, if you have a very new Intel CPU): Configurations for Intel® NPU with OpenVINO™ — OpenVINO™ documentation (note: you can still run the plugin on CPU or GPU, it doesn’t require NPU)
Run the plugin:
Open or create a white canvas with grey-white strokes (note that the plugin will take the current projection of the canvas, not the current layer).
Go to Tools → Fast Sketch Cleanup
Select the model. Advanced Options will be automatically selected for you.
Wait until it finishes processing (the dialog will close automatically then).
See that it created a new layer with the result.
Advice for processing
Currently it’s better to just use the SketchyModel.xml, in most cases it works significantly better than the SmoothModel.xml.
You need to make sure the background is pretty bright, and the lines you want to keep in the result are relatively dark (either somewhat dark grey or black; light grey might result in many missed lines). It might be a good idea to use a filter like Levels beforehand.
After processing, you might want to enhance the results with either Levels filter or Sharpen filter, depending on your results.
Technology & Science behind it
Unique requirements
First unique requirement was that it had to work on canvases of all sizes. That meant that the network couldn’t have any dense/fully or densely connected linear layers that are very common in most of the image processing neural networks (which require input of a specific size and will produce different results for the same pixel depending on its location), only convolutions or pooling or similar layers that were producing the same results for every pixel of the canvas, no matter the location. Fortunately, the Simo & Sierra paper published in 2016 described a network just like that.
Another challenge was that we couldn’t really use the model they created, since it wasn’t compatible with Krita’s license, and we couldn’t even really use the exact model type they described, because one of those model files would be nearly as big as Krita, and the training would take a really long time. We needed something that would work just as well if not better, but small enough that it can be added to Krita without making it twice as big. (In theory, we could do like some other companies and make the processing happen on some kind of a server, but that wasn’t what we wanted. And even if it resolved some of our issues, it would provide plenty of its own major challenges. Also, we wanted for our users to be able to use it locally without a reliance on our servers and the internet). Moreover, the model had to be reasonably fast and also modest in regards to RAM/VRAM consumption.
Moreover, we didn’t have any dataset we could use. Simo & Sierra used a dataset, where the expected images were all drawn using a constant line width and transparency, which meant that the results of the training had those qualities too. We wanted something that looked a bit more hand-drawn, with varying line-width or semi-transparent ends of the lines, so our dataset had to contain those kinds of images. Since we haven’t been aware of any datasets that would match our requirements regarding the license and the data gathering process, we asked our own community for help, here you can read the Krita Artists thread about it: https://krita-artists.org/t/call-for-donation-of-artworks-for-the-fast-line-art-project/96401 .
The link to our full dataset can be found below in the Dataset section.
Model architecture
All main layers are either convolutional or deconvolutional (at the end of the model). After every (de)convolutional layer except for the last one there is a ReLu activation layer, and after the last convolution there is a sigmoid activation layer.
Python packages used: Pillow, Numpy, PyTorch and Openvino
Numpy is a standard library for all kinds of arrays and advanced array operations and we used Pillow for reading images and converting them into numpy arrays and back. For training, we used PyTorch, while in the Krita plugin we used Openvino for inference (processing through the network).
Using NPU for inference
This table shows the result of benchmark_app, which is a tool that’s provided with Intel’s python package openvino. It tests the model in isolation on random data. As you can see, the NPU was several times faster than the CPU on the same machine.
On the other hand, introducing NPU added a challenge: the only models that can run on NPU are static models, meaning the input size is known at the time of saving the model to file. To solve this, the plugin first cuts the canvas into smaller parts of a specified size (which depends on the model file), and then proceeds to process all of them and finally stitch the results together. To avoid artifacts on the areas next to the stitching, all of the parts are cut with a little bit of a margin and the margin is later cut off.
How to train your own model
To train your own model, you’ll need some technical skills, pairs of pictures (input and the expected output) and a powerful computer. You might also need quite a lot of space on your hard drive, though you can just remove unnecessary older models if you start having issues with lack of space.
Drivers & preparation
You’ll need to install Python3 and the following packages: Pillow, openvino, numpy, torch. For quantization of the model you will also need nncf and sklearn. If I missed anything, it will complain, so just install those packages it mentions too.
Moreover if you want to use iGPU for training (which might still be significantly faster than on CPU), you’ll probably need to use something like IPEX which allows PyTorch to use an “XPU” device, which is just your iGPU. It’s not tested or recommended since I personally haven’t been able to use it because my Python version was higher than the instruction expects, but the instruction is here: https://pytorch-extension.intel.com/installation?platform=gpu&version=v2.5.10%2Bxpu . The sanity check for the installation is as follows: python3 -c "import torch; import intel_extension_for_pytorch as ipex; print(f'Packages versions:'); print(f'Torch version: {torch.__version__}'); print(f'IPEX version: {ipex.__version__}'); print(f'Devices:'); print(f'Torch XPU device count: {torch.xpu.device_count()}'); [print(f'[Device {i}]: {torch.xpu.get_device_properties(i)}') for i in range(torch.xpu.device_count())];" It should show more than 0 devices with some basic properties.
If you manage to get XPU device working on your machine, you’ll still need to edit the training scripts so they’ll able to use it: https://intel.github.io/intel-extension-for-pytorch/xpu/latest/tutorials/getting_started.html (most probably you’ll just need to add this line: import intel_extension_for_pytorch as ipex to the script on the very top, just underneath “import torch”, and use “xpu” as the device name when invoking the script, and it should work. But as I said, the scripts hasn’t been tested for that.
Dataset
You’ll need some pictures to be able to train your model. The pictures must be in pairs, every pair must contain a sketch (input) and a lineart picture (expected output). The better quality of the dataset, the better the results.
Before training, it’s best if you augment the data: that means the pictures are rotated, scaled up or down, and mirrored. Currently the data augmentation script also performs an inversion with the assumption that training on inverted pictures would bring the results faster (considering that black means zero means no signal, and we’d like that to be the background, so the models learn the lines, not the background around lines).
How to use the data augmentation script is explained below in the detailed instruction for the training part.
For quick results, use tooSmallConv; if you have more time and resources, typicalDeep might be a better idea. If you have access to a powerful GPU machine, you might try original or originalSmaller, which represent the original description of the model from the SIGGRAPH article by Simo-Sierra 2016, and a smaller version of it.
Use adadelta as the optimizer.
You can use either blackWhite or mse as the loss function; mse is classic, but blackWhite might lead to faster results since it lowers the relative error on the fully white or fully black areas (based on the expected output picture).
In the folder, run: python3 [repository folder]/spawnExperiment.py --path [path to new folder, either relative or absolute] --note "[your personal note about the experiment]"
Prepare data:
If you have existing augmented dataset, put it all in data/training/ and data/verify/, keeping in mind that paired pictures in ink/ and sketch/ subfolders must have the exact same names (for example if you have sketch.png and ink.png as data, you need to put one in sketch/ as picture.png and another in ink/ as picture.png to be paired).
If you don't have existing augmented dataset:
Put all your raw data in data/raw/, keeping in mind that paired pictures should have the exact same names with added prefix either ink_ or sketch_ (for example if you have picture_1.png being the sketch picture and picture_2.png being the ink picture, you need to name them sketch_picture.png and ink_picture.png respectively.)
Run the data preparer script: python3 [repository folder]/dataPreparer.py -t taskfile.yml That will augment the data in the raw directory in order for the training to be more successful.
Edit the taskfile.yml file to your liking. The most important parts you want to change are:
model type - code name for the model type, use tinyTinier, tooSmallConv, typicalDeep or tinyNarrowerShallow
optimizer - type of optimizer, use adadelta or sgd
learning rate - learning rate for sgd if in use
loss function - code name for loss function, use mse for mean squared error or blackWhite for a custom loss function based on mse, but a bit smaller for pixels where the target image pixel value is close to 0.5
Run the training code: python3 [repository folder]/train.py -t taskfile.yml -d "cpu"
On Linux, if you want it to run in a background, add “&” at the end. If it runs in a foreground, you can pause the training just by pressing ctrl+C, and if it runs in a background, find a process id (using either “jobs -l” command or “ps aux | grep train.py” command, the first number would be the process id) and kill it using “kill [process id]” command. Your results will still be in the folder, and you’ll be able to resume the training using the same command.
Convert the model to an openvino model: python3 [repository folder]/modelConverter.py -s [size of the input, recommended 256] -t [input model name, from pytorch] -o [openvino model name, must end with .xml]
Place both the .xml and .bin model files in your Krita resource folder (inside pykrita/fast_sketch_cleanup subfolder) alongside other models to use them in the plugin.
We are happy to announce Kdenlive 24.12. This release focuses on bug fixes, improved stability, and usability enhancements across the board. Numerous crashes and glitches have been addressed, including issues with audio capture, effect zones, high DPI display rendering, and subtitle editing. Proxies, rotoscoping, and project management workflows have been significantly refined, resolving lags, incorrect EXIF orientation handling, and archiving problems. We’ve managed to sneak in some little nifty features as well like the ability to resize multiple timeline items, Shift + Del shortcut to extract clips from the timeline, added actions to quickly add Marker/Guides in a specific category and mixes (same track transitions) can be 1 frame long.
Under the hood, we’ve dropped support for Qt5 and now require Qt6, alongside updated dependencies (MLT 7.28 and KF 6.3). This release comes with a lot of code cleanups and refactored Whisper settings. Optimized threading and memory management. Additionally, fail-safe measures have been taken to prevent invalid project profiles and script names.
Subtitles
We’ve added support for Advanced SubStation Alpha (ASS) subtitles, a widely used text-based format renowned for its flexibility in creating highly styled and customizable subtitles. ASS subtitles support advanced features such as font family, size, and color; text outlines and shadows; alignment and positioning; scaling and rotation; margins and spacing; and effects, including masking and other enhancements. This feature was developed by Chengkun Chen as part of Google Summer of Code (GSOC).
Subtitles
Subtitle Manager
The new subtitle manager is now integrated with style management and has been divided into four sections: Files, Layers and Content, Style, and Info, which correspond to the four main components of ASS subtitles.
Files – create, import and export subtitles
Layers and Content – create/remove subtitle tracks and apply styling
Styles – create and manage styles
Info – displays information about subtitles
Subtitle Style Editor
The new and powerful Subtitle Style Editor allows you to control all the styling capabilities of the ASS format.
Animated Subtitles
The ASS format supports three types of effects: Banner, where the text scrolls sideways across the screen; Scroll, where the text moves vertically; and Karaoke, where each word is highlighted in sync with the audio.
Currently, only the Banner and Scroll effects are accessible through the user interface, but additional styling, including Karaoke effects, can be applied using ASS tags.
Speech-to-Text
We’ve polished the Speech to Text features ensuring a smoother and more reliable experience. Seamless installation, GPU translation and threading issues have fixed. We’ve also resolved issues with the display of Vosk, Whisper and Seamless model folder sizes on Windows. Added the ability to update all virtual environment packages have updated to the latest version of Whisper. Lastly, the Whisper settings interface has been refactored.
Effects
With this version, we complete the final task of our fundraiser: builtin effects and a redesigned effects interface. Rendering of keyframe types like Bounce, Circular, and Exponential has been improved, alongside fixes for zone-based effects, rotoscoping lag, shape filter rendering, improved precision for time remapping, motion tracker models and prev/next seeking in monitor. It is also now possible to have single-frame mixes (same track transitions).
Interface redesign
The new Effect Stack redesign enhances usability with clearer organization of keyframeable and non-keyframeable parameters, improved layout consistency, more compact and clean. We’ve also added info buttons in effect headers for quick access to documentation.
Built-in Effects
To make your workflow much more fluid, the new effects panel gives direct access to effect parameters, allowing to quickly and easily adjust them. Currently built-in effects are Transform and Flip for video clips and Volume for audio clips. Built-in effects can be enabled/disabled in the settings.
New Effects
As usual there is always room for some eye candy, so we’ve added two color correction effects, HSL Primaries and HSL Range as well as GPS Effects (Images below displaying Distance, Altitude and Speed among many other values).
Other Highlights
Fix audio capture issues
Added Shift + Del shortcut to extract clip from timeline
Fix clip monitor history menu not showing up on audio clips
Fix spacer tool leaving a few frames after last clip
Implement resizing multiple timeline items
Fix Pexels Videos provider
Fix Alt+click to loop between clips using an effect in project monitor
Titler: ensure only plain text can be pasted
Titler: added support for tabulations
Add Actions to quickly add Marker/Guides in a specific category
Together with Intel, we have been working a new plugin for Krita: the fast sketch plugin, or maybe, better, a fast inking plugin. This is an experimental plugin that makes it (sometimes) possible to automatically ink a sketch, using neural networks.
This plugin uses models to figure out how to ink a sketch: the included models were trained on openly available data: there was no scraping or stealing involved! The plugin comes with a manual that explains how to get the scripts you can use to create a model trained on your own data: what you need are before and after images of your sketch and your uncolored inked drawing, and the training software can run on your own hardware (it will take a lot of time, though).
Throughout the development process we've been discussing this plugin with artists on the Krita Artists forum.
The plugin can be downloaded and extracted in a Windows Krita 5.2.6 folder and should then be enabled in the plugin manager in Krita's settings dialog.
There is also a download of Krita 5.3.0 pre-alpha available that includes the plugin for Windows and Linux. Currently, we don't have a working MacOS version ready, and since the plugin is implemented in Python, there will be no Android packages.
ervin
@kbroulik:kde.org
