Jothi Prasath

Search

Mon, 01 Jan 0001 00:00:00 +0000

How to enable Anime4k upscaling in MPV player

Mon, 29 Jan 2024 00:00:00 +0000

What is Anime4k?

Anime4K is a set of open-source, high-quality real-time anime upscaling/denoising algorithms that can be implemented in any programming language. This is also not replacement for SRGANs, as they perform much better on low-resolution images or images with lots of degradation.

Disclaimer

Anime4k is optimized for native 1080p anime. For lower resolution video file try using FSR upscaling.

Samples

Before

After

Summer Wars (2009)

Before

After

Summer Wars (2009)

Before

After

Ponyo (2008)

Step 1: Install MPV

Windows

Download the installer from the official MPV website and follow the installation instructions.

Linux

Ubuntu/Debian

sudo apt-get install mpv

Fedora

sudo dnf install mpv

Mac

brew install mpv

Step 2: Create a Shaders Directory

Windows

Navigate to appdata. This is typically located in %APPDATA%. You can quickly navigate there by entering this path into the File Explorer’s address bar.
Create a Folder “mpv” inside that folder.
Inside the “mpv” folder, create another folder named “shaders”.
Inside that “shaders” folder, create a folder named “anime4k”.

Linux/Mac

mkdir -P ~/.config/mpv/shaders/anime4k

Step 3: Download the Anime4k Shaders into MPV

Windows

Download this file https://github.com/bloc97/Anime4K/releases/download/v4.0.1/Anime4K_v4.0.zip
Extract the downloaded file, where you can find bunch of .glsl files.
Copy that those .glsl files into previously created “anime4k” folder.

Linux/Mac

Run this command to download the Anime4k shaders into mpv shaders folder

curl -Lo ~/.config/mpv/shaders/anime4k/anime4k.zip https://github.com/bloc97/Anime4K/releases/download/v4.0.1/Anime4K_v4.0.zip
unzip ~/.config/mpv/shaders/anime4k/anime4k.zip -d ~/.config/mpv/shaders/anime4k/
rm ~/.config/mpv/shaders/anime4k/anime4k.zip

Step 4: Modify MPV Configuration File

Windows

open the mpv folder and create a text file “input.conf”
Paste the following lines and save it.

CTRL+1 no-osd change-list glsl-shaders set "~~/shaders/anime4K/Anime4K_Clamp_Highlights.glsl:~~/shaders/anime4K/Anime4K_Restore_CNN_M.glsl:~~/shaders/anime4K/Anime4K_Upscale_CNN_x2_M.glsl:~~/shaders/anime4K/Anime4K_Restore_CNN_S.glsl:~~/shaders/anime4K/Anime4K_AutoDownscalePre_x2.glsl:~~/shaders/anime4K/Anime4K_AutoDownscalePre_x4.glsl:~~/shaders/anime4K/Anime4K_Upscale_CNN_x2_S.glsl"; show-text "Anime4K"
CTRL+0 no-osd change-list glsl-shaders clr ""; show-text "GLSL shaders cleared"

Linux/Mac

Run this command to create/open input.conf file

nano ~/.config/mpv/input.conf

Paste the following lines and save it.

CTRL+1 no-osd change-list glsl-shaders set "~~/shaders/anime4K/Anime4K_Clamp_Highlights.glsl:~~/shaders/anime4K/Anime4K_Restore_CNN_M.glsl:~~/shaders/anime4K/Anime4K_Upscale_CNN_x2_M.glsl:~~/shaders/anime4K/Anime4K_Restore_CNN_S.glsl:~~/shaders/anime4K/Anime4K_AutoDownscalePre_x2.glsl:~~/shaders/anime4K/Anime4K_AutoDownscalePre_x4.glsl:~~/shaders/anime4K/Anime4K_Upscale_CNN_x2_S.glsl"; show-text "Anime4K"
CTRL+0 no-osd change-list glsl-shaders clr ""; show-text "GLSL shaders cleared"

Usage Instructions

Play a anime video and press ctrl+1 to enable Anime4k and ctrl+0 for disable it.

How to enable FSR upscaling in MPV player

Fri, 26 Jan 2024 00:00:00 +0000

What is FSR (FidelityFX Super Resolution)?

FSR, or FidelityFX Super Resolution, is an upscaling technology developed by AMD (Advanced Micro Devices) to improve the visual quality of lower-resolution images or videos in real-time. It is designed to enhance the gaming and multimedia experience by providing a more detailed and sharper image without significantly impacting performance.

Here we are going to apply the same fsr techinque to mpv media player to upscale the lower resolution videos.

Step 1: Install MPV

Windows

Download the installer from the official MPV website and follow the installation instructions.

Linux

Ubuntu/Debian

sudo apt-get install mpv

Fedora

sudo dnf install mpv

Mac

brew install mpv

Step 2: Create a Shaders Directory

Windows

Navigate to appdata. This is typically located in %APPDATA%. You can quickly navigate there by entering this path into the File Explorer’s address bar.
Create a Folder “mpv” inside that folder.
Inside the “mpv” folder, create another folder named “shaders”.

Linux/Mac

mkdir -P ~/.config/mpv/shaders

Step 3: Download the FSR Shader into shaders folder

Windows

Open your web browser and go to this website https://gist.github.com/agyild/82219c545228d70c5604f865ce0b0ce5
Click on the Download ZIP button on the top right corner.
Extract the downloaded file, where you can find fsr.glsl.
Copy that fsr.glsl into newly created shaders folder.

Linux/Mac

Run this command to download the FSR shaders into shaders folder

curl -Lo ~/.config/mpv/shaders/FSR.glsl https://gist.githubusercontent.com/agyild/82219c545228d70c5604f865ce0b0ce5/raw/2623d743b9c23f500ba086f05b385dcb1557e15d/FSR.glsl

Step 4: Modify MPV Configuration File

Windows

open the mpv folder and create a text file input.conf
Paste the following lines and save it.

CTRL+1 no-osd change-list glsl-shaders set "~~/shaders/FSR.glsl"; show-text "FSR"
CTRL+0 no-osd change-list glsl-shaders clr ""; show-text "GLSL shaders cleared"

Linux/Mac

Run this command to create/open input.conf file

nano ~/.config/mpv/input.conf

Paste the following lines and save it.

CTRL+1 no-osd change-list glsl-shaders set "~~/shaders/FSR.glsl"; show-text "FSR"
CTRL+0 no-osd change-list glsl-shaders clr ""; show-text "GLSL shaders cleared"

Usage Instructions

Play a video which is lower than your display resolution and press ctrl+1 to enable FSR and ctrl+0 for disable FSR.

gzip vs zstd: A Performance Comparison

Sun, 21 Jan 2024 00:00:00 +0000

Introduction

I have a collection of files and folders containing JavaScript, Node.js modules, Python, and C/C++ files from various mini projects. The total size is 3 GB, and while I don’t need them all the time, I still want to archive them for occasional use. I usually use gzip for compression, but I’m currently experimenting with the newer and more efficient zstd compression. In this blog post, I’ll be comparing the compression performance of gzip and zstd.

System Specifications

Operating System: Arch Linux
Processor: Ryzen 5 5500U (6c/12t)
RAM: 8GB

Compression

gzip with pigz

Using pigz, a parallel implementation of gzip:

tar -c -I"pigz" -f temp.tar.gz temp

Time taken : 40 seconds
CPU usage : 50%
File Size : 848MB

zstd

The zstd compression algorithm offers a range of compression levels from 1-19, where 1 is least compression and 19 is highest compression. Using higher compression levels requires more cpu power and time. Therefore i have choosen a sweet spot and decided on level 3 and 19.

zstd (Level 3)

Using zstd with level 3(default)

tar c -I"zstd -T0" -f temp.tar.zst temp

Time taken : 27 seconds
CPU usage : 15%
File Size : 770MB

zstd (Level 19)

Using zstd with level 19

tar c -I"zstd -19 -T0" -f temp.tar.zst temp

Time taken : 4 minutes 2 seconds
CPU usage : 50%
File Size : 593MB

Extraction

gzip with pigz

tar -x -Ipigz -f temp.tar.gz -C temp

Time taken : 56 seconds
CPU Usage : 10%

zstd (Level 3)

tar -x --zstd -f temp.tar.zst -C temp

Time taken : 52 seconds
CPU Usage : 7%

zstd (Level 19)

tar -x --zstd -f temp.tar.zst -C temp

Time taken : 50 seconds
CPU Usage : 7%

Conclusion

Clearly, the zstd level 19 achieves a smaller size file but comes at the cost of increased CPU power and time consumption. However, using the default zstd (level 3) proves to be a more efficient choice compared to gzip, giving a balance between compression performance and resource utilization. Additionally, the decompression almost similar regardless of which level.

How I Optimized My Google Photos Library with the Power of AVIF

Fri, 14 Jul 2023 00:00:00 +0000

Introduction

As a user of Google’s free tier services, I have access to 15GB of storage. With a total of 2.7GB of Google Photos already occupying my storage, I started to worry about the future growth of my photo collection. Considering the rate at which my photos are accumulating, it became clear that my storage would likely fill up within approximately four or less years. To address this, I decided to convert all my images to a better optimized format. which is absolutely my favorite, AVIF format.

Why AVIF Format

When it comes to choosing a highly optimized image format, three major formats stand out: HEIC, AVIF, and JXL. However, considering browser and platform support, AVIF emerges as the clear winner. Let me explain why AVIF stands out among these formats.

JXL, or JPEG XL, offers impressive image compression efficiency. However, it doesn’t supported by popular browsers like Google Chrome and Firefox, which are widely used for daily browsing activities. Relying on JXL for image delivery may result in compatibility issues across various platforms.

On the other hand, HEIC, the format primarily used by Apple devices, has gained popularity due to its efficient compression capabilities. However, it is not supported by any browsers, except Safari. Accessing HEIC images on Android devices can be challenging and may lead to compatibility problems.

Therefore, considering efficiency, browser, and platform support, AVIF is the preferred choice for an optimized image format. Additionally, I have already written a blog post on the topic of “JPG vs AVIF: A Comparison”. In that post, I gone into more detail about the efficiency and visual difference between these two formats. Feel free to check it out for more information.

Google Photos Data Extraction

To convert all of my Google Photos, first I started by dumping the content from Google Photos using the Google Takeout service https://takeout.google.com/. Here’s a step-by-step description of the process

On the Google Takeout page, I deselected all services.

I selected Google Photos option as the service to export.

And initiate the data export and it informed to wait while Google prepared the download.
After approximately an hour, I received an email from Google containing a download link. This link led me to a zip file containing my Google Photos data.

Data Cleaning

After downloading and extracting the zip files containing my Google Photos data, I was left with a total of 32 folders containing photos and videos. The way they structured initially included individual folders for each album, labeled with the album name. Additionally, separate folders were created for each year, such as “photos from 2020,” which contained all the photos from that year. However, the album folders duplicated the respective year-specific folders, resulting in unnecessary redundancy.

So I performed a data cleaning process by removing all the album and archive folders. As a result, Now I have a more organized structure with folders dedicated to each year (e.g., “photos from 2020”) that contain the photos from that specific year.

Merging Metadata

Disclaimer

The "google-photos-exif" script used for merging metadata, only works well with photos that contain EXIF metadata. EXIF metadata means information like device model, date, and other camera settings

Each of the year-specific folders contains images as well as separate JSON files that hold corresponding metadata. These JSON files contain important information such as the date and time of each image, as well as geolocation data.

To properly organize the images by date and time, it was crucial to merge the images with their respective JSON files. Fortunately, I discovered a useful script called google-photos-exif on github that specifically performs this task. The script efficiently merges the images with their corresponding JSON metadata and saves them all in one big folder.

After completing the merging process, I cross-verified the results by checking some of the image properties. Specifically, I checked the “created on” information, where the original image date and time were explicitly printed.

This successful merging of image metadata ensures that each image is now associated with its correct date, time, and other relevant information. And it stores all images in one folder, it simplifies the organization and future access of the images.

Analying The Files Before Conversion

After successfully merging the images with their corresponding metadata, the next step was to analyze the types of files present in the folder. This identifies the number of various file formats such as JPG, PNG, AVIF, HEIC, and videos.

To accomplish this, I began writing a Python script that goes through the entire folder and analyze the file types and sizes. The script provided the number of files for each format and their total sizes.

According to the script’s output, the folder contained:

1,734 normal JPG and PNG images, occupying approximately 1.93 GB of storage space.
82 HEIC images, which are typically captured by Apple devices and are already compressed, so I decided not to convert them. They occupy approximately 97 MB of storage space.
Additionally, there were 77 video files, totaling approximately 1.21 GB. However, at this point, I am not yet familiar with video encoding formats and the longer encoding time required for videos on my laptop. Therefore, I excluded the videos from the conversion for now.

Converting Into AVIF

To convert the images into the AVIF format, I initially started writing a shell script. However, I soon realized the advantages of cross-platform compatibility and decided to scrap the shell script code. Instead, I began writing code in Python to perform the conversion.

For the conversion process, I used pillow_heif Python library, which provides support for handling AVIF images. To optimize the conversion speed and efficiency, I implemented parallel multi-processing techniques. The resulting code, along with all the necessary files, is available in a GitHub repository google-photos-compressor.

The conversion process takes approximately 45 minutes to complete. This can vary depending on the power and specifications of the system being used for the conversion. The processing speed and efficiency of the system play a significant role in determining the overall time required for the conversion. In my case, using a Ryzen 5 5500U processor with 6 cores and 12 threads running at 4GHz, with 8Gb ram

Analying The Files After Conversion

To calculate the impact of the conversion process, I used the same Python script to analyze the folder containing the converted images. Here are the findings:

The initial collection of 1.93 GB, consisting of 1,734 JPG and PNG images, has been compressed into 679 MB.
The total size was significantly reduced to 679 MB, resulting in a compression rate of approximately 65%.
As a result, the conversion process saved a total of 1.25 GB of storage space.

I have successfully optimized my image collection while ensuring that the visual quality of the photos is preserved.

Deleting All Images In Google Photos

When it came to deleting all the photos in Google Photos, I discovered that there was no built-in option to accomplish this task easily. Manually selecting and deleting each of the 1,734 photos would have been a time-consuming nightmare.

Fortunately, I came across a JavaScript solution that could be executed in the browser’s console to automate the selection and deletion of all the files in Google Photos. That script automatically selects and deletes the photos without the need for manual interaction. It saves a considerable amount of time and effort.

You can find the JavaScript code at the following link: delete_photos.js

Uploading Converted Photos To Google Photos

To upload the converted AVIF images, and other images and videos back to Google Photos, I followed a specific order. First, I uploaded the converted AVIF images, followed by the HEIC images, and finally the videos.

During the uploading process, I noticed that all the albums in the Albums section were empty. So, I proceeded to delete these empty albums.

After the upload was complete, I observed that a few images (5 in total) were screenshots and WhatsApp images, while some of the videos (15 in total) were received from WhatsApp. These images and videos were not properly sorted in terms of their date and time metadata.

To solve this issue, I manually adjusted the date and time for these particular images and videos, referring to the date and time information present in the original versions of the files. This helps me to arrange the images and videos in chronological order within Google Photos.

Conclusion

In this blog post, I shared my experience of downloading and converting my Google Photos collection to the AVIF format. Overall, by converting my Google Photos collection to the AVIF format, I was able to achieve significant storage savings while maintaining the visual quality of my images.

I hope my journey and the steps I followed to optimize my photo collection using AVIF format will inspire and guide others in their efforts to manage and optimize their digital media.

JPG vs AVIF: A Comparison

Mon, 10 Jul 2023 00:00:00 +0000

JPG and AVIF are the two main contenders in the battle, where JPG/JPEG are common image format we use daily in mobiles, laptops, cameras. It is mostly supported by all devices and platforms. And on the other hand AVIF the next generation image format. It is not supported by many social media platforms and browsers like Microsoft Edge and Opera Mini.

Methodology

I have selected four different test images to run through both the JPG and the AVIF encoders:

Flower - A normal image we take everyday using our smartphones.

Taken from Redmi 9 Power
Lemur - A normal image we take everyday using our smartphones.

Taken from free-images.com. Untouched 13Mb image
Text - A screenshot image of terminal.

A text image with small range of colors
Cartoon - A cartoon image.

A cartoon I designed by using Sketchbook

All images used in this post (source .jpg, encoded .avif)

To conduct the image encoding tests, the following system specifications were used:

Operating System: Linux (Arch Linux)
Kernel Version: 6.4.2
Processor: Ryzen 5 5500U (4 Ghz)
RAM: 8GB DDR4

Lossy compression

Lossy compression is a method of data compression that selectively discards or reduces some of the information from the original data in order to reduce file size. It is widely used in image compression algorithms to achieve smaller file sizes while maintaining acceptable image quality.

To encode the AVIF images I used avifenc provided by libavif

avifenc input.jpg output.avif --min 0 --max 63 -j 12

The --min and --max set the minimum and maximum quantizer for color. We use the full range here. The -j set the number of parallel threads. Here I am using 12 threads.

Flower

For this flower image I have cropped around the one of the flower. The crop is based on 200% zoomed image. Here are the results

File format	File size	Compression ratio	Time taken to encode
JPG	440Kb	-	-
AVIF	156Kb	- 64.54%	0.465s



Original	AVIF

Image (200% zoom crop)

We compressed the original 440Kb image to 156Kb image which is 64% lesser than original. And the AVIF version is more smooth in edges than JPG. Most of the smart phones take these images and the difference is not noticeable.

Lemur

For this lemur image I have cropped around the eye. The crop is based on 200% zoomed image. Here are the results

File format	File size	Compression ratio	Time taken to encode
JPG	13Mb	-	-
AVIF	1.1Mb	- 91.53%	3.547s



Original	AVIF

Image (200% zoom crop)

Comparing to other three images this lemur image is huge in size with resolution around 5K (4928x3264). We compressed the original 13Mb image into 1.1Mb image which is whopping 91.53% less than original image and There are some color artifacts in the AVIF format.

Text

For this image I have cropped around the text. The crop is based on 200% zoomed image. Here are the results

File format	File size	Compression ratio	Time taken to encode
JPG	392Kb	-	-
AVIF	40Kb	- 89.79%	3.54s



Original	AVIF

Image (200% zoom crop)

This image is very simple with limited colors. This will give you an overview of text based images. We converted the original image 392Kb to 40Kb image. Which is again a huge compression of 89.79%. The edges of the text are slightly blurry compared to original image.

Cartoon

For this cartoon image I have cropped center of the image. The crop is based on 200% zoomed image. Here are the results

File format	File size	Compression ratio	Time taken to encode
JPG	864Kb	-	-
AVIF	136Kb	- 84.25%	0.887s



(Original)	AVIF

Image (200% zoom crop)

The cartoon image is contain less colors and overfill colors in the edges. We converted the original 136Kb to 86Kb which is 84.25% lesser than original image. There are some blue color overfill in the edges of the original image which is not visible in AVIF format.

How to use AVIF in websites

You can use AVIF in websites using <picture> element. The <picture> element contains one <img> tag and one or more <source> tags. Put the normal JPG/PNG images in the <img> tag and add AVIF images to <source> tag. The browser first looks for the <source> tag, if the image format is supported by the browser it loads the image. If not supported by the browser then the <img> tag image loads.

You can verify which format of the file loaded by

Right clicking the page and select the inspect option.
Switch to the network tab and reload the page.
You can find the image name with extension .AVIF

Conclusion

According to the results presented above, my option is that AVIF is better compared to JPG in lossy compression. For haven’t tested for lossless compression because my original JPGs are compressed images. The main reason for this comparison is for website use.

How to Compile the Linux Kernel

Sun, 02 Jul 2023 00:00:00 +0000

What is Linux Kernel?

The Linux kernel can be defined as the core component of the Linux operating system. It is a piece of software that manages the system’s hardware resources and provides essential services to other software applications. In simple terms, the kernel acts as a mediator, allowing the operating system and the hardware to interact effectively.

Requiremented packages

git
fakeroot
build-essential
xz-utils
bc
ncurses-dev
libssl-dev
libelf-dev
bison
flex

Download Linux Source Code

To compile the Linux kernel, the first step is to obtain the source code. Which can be obtained from official website kernel.org. Download the required kernel version.

Disclaimer

It is recommended to select a stable version unless you have specific requirements for a particular version.

Extract the Source Code

tar xvf linux-6.0.7.tar.xz

Configuring Kernel

Navigate to extracted directory.

cd linux-version

Copy the current kernel config to current directory.

zcat /proc/config.gz > .config

Make changes using the following command.

make menuconfig

Building Kernel

Build the kernel by using the following command.

make -j$(nproc)

Install Kernel

The compiled kernel can be installed by using the following command.

sudo make -j$(nproc) modules_install
sudo make -j$(nproc) install

hello there!

Mon, 01 Jan 0001 00:00:00 +0000

Since you are already here, I think a short introduction is in order. I am Jothi Prasath, Computer Science graduate. I find immense joy in exploring various aspects of technology, particularly in the realm of open source and Linux. Most of the time trying to automate tasks.

I am currently seeking job opportunities in the field of Computer Science. With a strong background in computer science and a passion for technology, I am eager to contribute my skills and knowledge to a dynamic and innovative organization.

If you are looking for a highly motivated and skilled individual with a strong foundation in computer science, Please feel free to reach out to me through the provided contact information. I am excited to contribute my expertise and continue growing in the field of computer science.

You can find me on Linkedin

Here is a summary of the content in this site:

Posts list

Mon, 01 Jan 0001 00:00:00 +0000

Projects

Mon, 01 Jan 0001 00:00:00 +0000

Applications and software

Archlinux Ansible Playbook — A playbook for automating the installation and configuration of Arch Linux using Ansible. It provides a streamlined and reproducible method for setting up a fresh Arch Linux system with commonly used packages and configurations.

Auto-epp — Auto-epp manages the energy performance preferences (EPP) of amd-pstate-epp (CPU Scaling Driver). It adjusts the EPP settings based on whether system is running on AC power or battery power, helping optimize power consumption and performance.

Movie Radar — A Flask web application that displays latest now playing movies based on language using TMDB API.

qBittorrent Enhanced Edition Docker Image — A Docker image for qBittorrent Enhanced Edition, built automatically with GitHub Actions.

Movie Radar JS — A nodejs application that displays latest now playing movies based on language using TMDB API.

Debian Ansible Playbook — A playbook for automating the installation and configuration of Debian Linux using Ansible.

Password Manager — A GUI Password Manager using python with encryption.

Termux Development Setup — A script to set up a development environment in the Termux app for Android.

Gnomintosh — Macintosh inspired theme for GNOME Linux desktop environment.

Small Sur — Mac os Big Sur inspired theme for Xfce4 Linux desktop.

Bash Script Collection — A collection of useful Bash scripts that I have created for various purposes. These scripts are designed to automate tasks, simplify workflows, and improve productivity in a Linux environment.

Google Photos Compressor — Compress images stored in Google Photos to AVIF.

Wynk Music — Wynk music(unofficial) is an Electron-based lightweight client for streaming music.

Contributed Projects

code-server - VS Code in the browser
lazygit - Simple terminal UI for git commands
cue - A command-line music player
rtw88 - Linux Realtek Drivers
obsidian-telegram-sync - Transfer messages and files from Telegram bot to Obsidian

Resume

Mon, 01 Jan 0001 00:00:00 +0000

Jothi Prasath

Developer

Email: jothiprasath2@gmail.com
Born: 11, February, 2002

Education

The American College, Madurai

August 2020 – April 2023 B.Sc. in Computer Science.

Skills

Technical Skills

Python – Used extensively during my degree. And made some personal projects.
Shell scripting – POSIX sh, bash.
Databases – SQLite.
Devops tools – Git, Ansible, Docker.
VS Code – My main development tool, aided with handy plugins.
Web technologies – I am proficient in web technologies.
- CSS3 & HTML5 – The future of the web.
- Javascript – Node.js.
- Flask – Python web framework
- Others – Hugo (this site).

Social Skills

Teamworking.
Good communication and presentation skills.
Easily socialize, adept at connecting and engaging with others.

Projects

Please, see the projects sections of this webiste.

Gnomintosh – Github
Archlinux Playbook – Github
Auto-epp – Github
qBittorrent Enhanced Docker Image – Github
Password Manager – Github
Small Sur – Github
Termux Development Setup – Github
Bash Script Collection – Github
Google Photos Compressor – Github
Wynk Music – Github

Contributions

code-server - VS Code in the browser
lazygit - Simple terminal UI for git commands
cue - A command-line music player
rtw88 - Linux Realtek Drivers
obsidian-telegram-sync - Transfer messages and files from Telegram bot to Obsidian

Language skills

English
Tamil