#lossless compression

[maybeasunflower]

PNG IS LOSSLESS COMPRESSION.

There is no reason to use another format, not in this day-and-age when 10MB files are considered "small".

JPG is lossy compression; don't do it unless you must.

We ask your questions so you don’t have to! Submit your questions to have them posted anonymously as polls.

[linkloka]

Optimize with Ease: ImageOptim

ImageOptim is a powerful online tool crafted to optimize and compress images, enhancing website performance and reducing load times. With its straightforward interface and effective optimization algorithms, it offers users a convenient solution to enhance their online content effortlessly. Features: Lossless and lossy image compression options Batch processing for optimizing multiple images…

[manjaro-official]

Bluetooth earbuds have convinced the public that 'lossless audio' is somehow expensive tech for true audiophiles when literally if you listen to a CD player with wired headphones you're getting lossless audio

[aldieb]

it’s so crazy when you find people who you can talk to in a way where the discussion feels like passing your ideas through a sieve that makes them sharper and more refined. for me 90% of the conversations i’m in can end up feeling like they’re doing the opposite, so it’s something really special

[booksandchainmail]

me watching a low-resolution stream: wow, lossy compression... just like in Exordia

[netscapenavigator-official]

My whole life has been a lie! This whole time I've been rocking WebP formatted photos on my Jellyfin server, they've been...

L O S S Y

Devastated. Kicking. Screaming. Crying.

I thought Shutter Encoder could detect when a PNG was input and encode the WebP file properly, but using Google's "webpinfo" command, I confirmed that all my formerly PNG photos are now lossy, so I guess it can't.

Fortunately, however, I found a semi-GUI macOS app which can batch convert photos into JPEG XL, and using some custom workflow rules, it CAN determine whether a photo needs lossy or lossless conversion. Looks like my data hoarding of the original formats is paying off, and I'm moving to JXL sooner than I expected.

[bettyweir]

convering all my behind the scene's JPG/JPEGs to PNGs so people can download whatever BTS they want without damaging the quality ^_^

[bearbleps]

"Cmon man, my friends are in there, I'm sure there must be a mistake."

"Oh yeah? I see you dont have an alpha channel. And what's that on your shirt, *lifts sunglasses and squints* some compression artifacts? Best move along, pal."

when you try to get into the png nightclub but you see the dotted rectangle surrounding the bouncer stretch him wider to block the entrance cus ur not on his clipboard

[tom2tec]

SAC ~ State-Of-The-Art Lossless Audio Compression

Sac is a state-of-the-art lossless audio compression model. Lossless audio compression is a complex problem, because PCM data is highly non-stationary and uses high sample resolution (typically >=16bit). That’s why classic context modelling suffers from context dilution problems. Sac employs a simple OLS-NLMS predictor per frame including bias correction. Prediction residuals are encoded using a…

[bluedrake121]

Found it

: 3

[stellophiliac]

how to build a digital music collection and stuff

spotify sucks aaaass. so start downloading shit!!

file format glossary

.wav is highest quality and biggest

.mp3 is very small, but uses lossy compression which means it's lower quality

.flac is smaller than .wav, but uses lossless compression so it's high quality

.m4a is an audio file format that apple uses. that's all i really know

downloading the music

doubledouble.top is a life saver. you can download from a variety of services including but not limited to apple music, spotify, soundcloud, tidal, deezer, etc.

i'd recommend ripping your music from tidal or apple music since they're the best quality (i think apple music gives you lossless audio anyway. .m4a can be both lossy and lossless, but from the text on doubledouble i assume they're ripping HQ files off apple music)

i also love love love cobalt.tools for ripping audio/video from youtube (they support a lot of other platforms too!)

of course, many artists have their music on bandcamp — purchase or download directly from them if you can. bandcamp offers a variety of file formats for download

file conversion

if you're downloading from apple music with doubledouble, it spits out an .m4a file.

.m4a is ok for some people but if you prefer .flac, you may wanna convert it. ffmpeg is a CLI (terminal) tool to help with media conversion

if you're on linux or macOS, you can use parameter expansion to batch convert all files in a folder. put the files in one place first, then with your terminal, cd into the directory and run:

for i in *.m4a; do ffmpeg -i "$i" "${i%.*}.flac"; done

this converts from .m4a to .flac — change the file extensions if needed.

soulseek

another way to get music is through soulseek. soulseek is a peer-to-peer file sharing network which is mainly used for music. nicotine+ is a pretty intuitive (and open-source) client if you don't like the official one.

you can probably find a better tutorial on soulseek somewhere else. just wanted to make this option known

it's bad etiquette to download from people without sharing files of your own, so make sure you've got something shared. also try to avoid queuing up more than 1-2 albums from one person in a row

tagging & organizing your music

tagging: adding metadata to a music file (eg. song name, artist name, album) that music players can recognize and display

if you've ripped music from a streaming platform, chances are it's already tagged. i've gotten files with slightly incorrect tags from doubledouble though, so if you care about that then you might wanna look into it

i use musicbrainz picard for my tagging. they've got pretty extensive documentation, which will probably be more useful than me

basically, you can look up album data from an online database into the program, and then match each track with its file. the program will tag each file correctly for you (there's also options for renaming the file according to a certain structure if you're into that!)

there's also beets, which is a CLI tool for... a lot of music collection management stuff. i haven't really used it myself, but if you feel up to it then they've got extensive documentation too. for most people, though, it's not really a necessity

how you wanna organize your music is completely up to you. my preferred filestructure is:

artist > album > track # track

using a music player

the options for this are pretty expansive. commonly used players i see include VLC, foobar2000, clementine (or a fork of it called strawberry), and cmus (for the terminal)

you can also totally use iTunes or something. i don't know what audio players other systems come with

i personally use dopamine. it's a little bit slow, but it's got a nice UI and is themeable plus has last.fm support (!!!)

don't let the github page fool you, you don't have to build from source. you can find the releases here

click the "assets" dropdown on the most recent release, and download whichever one is compatible with your OS

syncing

if you're fine with your files just being on one device (perhaps your computer, but perhaps also an USB drive or an mp3 player), you don't have to do this

you can sync with something like google drive, but i hate google more than i hate spotify

you can get a free nextcloud account from one of their providers with 2GB of free storage. you can use webDAV to access your files from an app on your phone or other device (documents by readdle has webDAV support, which is what i use)

disroot and blahaj.land are a couple providers i know that offer other services as well as nextcloud (so you get more with your account), but accounts are manually approved. do give them a look though!!

if you're tech-savvy and have an unused machine lying around, look into self-hosting your own nextcloud, or better yet, your own media server. i've heard that navidrome is a pretty good audio server. i unfortunately don't have experience with self-hosting at the moment so i have like zero advice to give here. yunohost seems to be a really easy way to manage a server

afterword

i don't know if any of this is helpful, but i just wanted to consolidate my personal advice in one place. fuck big tech. own your media, they could take it away from you at any moment

[charliejaneanders]

Imagine what it would look like if ChatGPT were a lossless algorithm. If that were the case, it would always answer questions by providing a verbatim quote from a relevant Web page. We would probably regard the software as only a slight improvement over a conventional search engine, and be less impressed by it. The fact that ChatGPT rephrases material from the Web instead of quoting it word for word makes it seem like a student expressing ideas in her own words, rather than simply regurgitating what she’s read; it creates the illusion that ChatGPT understands the material. In human students, rote memorization isn’t an indicator of genuine learning, so ChatGPT’s inability to produce exact quotes from Web pages is precisely what makes us think that it has learned something. When we’re dealing with sequences of words, lossy compression looks smarter than lossless compression.

Ted Chiang’s essay about ChatGPT is required reading

[utopicwork]

Trans girl invents new lossless image format

Over the course of a couple of days I put together something pretty significant: a new image format and compression technique. I'm calling it ShrinkRay (the extension being .sr) and it is made for a more efficient representation of pixel art. With this image: https://www.tumblr.com/potionslushie/759045761554055168/ as a test it compresses ~17% better then png and when converted back to png it still has some (albeit smaller) compression gains. Additionally since part of the compression is indexing colors I was able to trivially implement palette swapping with palette files (so far .hex only which means a lot of Lospec palettes).

So what does this have to do with p2chat? Well the compression gains will make sending and receiving chat images significantly faster but also the recommended limitations of ShrinkRay: maximum 256x256 px (with a scaling factor), maximum 10 colors including black and white as presets, will form the basis for p2chat's prototype limitations.

Whats next: cleaning up the code, documenting it, running further tests and putting up a repository. Later on though I have further compression ideas regarding simple shape implementation and a layer/frames system that's highly optimized. Though I'll be doing more important PierMesh work before that.

All that said I'll leave you with one last detail: all the code for this comes in at just under 128 lines.

Addition: it now supports animation/layers