#encoding
Text
So, like I said I was, I'm re-encoding all the videos on my Jellyfin server as AV1 to save space. I've been doing that for about 4-5 days, straight, at this point. So, I'm gonna give a progress update at just how far I've gotten in those 4-5 days.
So far, I've managed to re-encode a singular folder that was about 13.72 GB.
I still have many other folders to go, that all total up to about 109.12 GB.
I'm going to be here forever.
On the bright side, re-encoding the videos as AV1 fixed some of the corruption that caused a lot of them to become unplayably stuttery.
Also, what I have finished re-encoding was once 10.15 GB. It is now 7.70 GB. So, while it is taking forever, I have managed to save over two whole gigabytes by simply running my videos through a program that essentially makes my computer get hot.
So with this level of success without even scratching the surface, I'm excited to see how much space I'll save when all is finally completed.
#rambles#update#data storage#av1#video#videos#h.264#encoding#shutter encoder#data archiving#jellyfin#home media server#server#jellyfin server
9 notes
·
View notes
Text
A few hints on solving the indecipherable-Nonsense puzzle.
so, for a bit of context: soon after posting the Indecipherable-Nonsense puzzle, i realized it was way too hard, so I wanted to give you all some hints as to how to solve it.
but of course, I couldn't make it that easy. so, instead of just giving it to you, I instead made a series of smaller puzzles that ramp up in difficulty more gradually, and give a few hints to each other and to the final puzzle along the way. I hope you all enjoy, and I wish you the best of luck!!!! :3
17 notes
·
View notes
Text
FutuDeAreI,
Mosatwhfuerdinesday
Fmjaunodpebrchtiovlguseptembeary
7 notes
·
View notes
Text
"My stream looks bad."
This is a pretty common issue that creators face and it all stems from one common misconception: 1080p doesn't mean quality.
The above video is in fact 1080p, but as you can see there's definitely something wrong about it. Here's that exact same freezeframe but at 480p:
So what's happening here? Bitrates.
A Crash Course Through Video Compression
Computer video has historically been a really hard problem to solve. Computers used to be slow (Like, really slow). Computer scientists (read: nerds) figured out that the human eye is bad and have developed techniques that abuse this fact.
Dithering tricks your eye into "seeing" more colors than are actually present. Here's a monotone (two colors!) that appears to also have shades of grey even though there aren't any:
Pretty cool, eh? By using clever placement of the black dots an entire range of grays emerge even though there aren't any.
You can even throw out little bits of detail to save space and the eye will be none the wiser. These are the core principles behind video compression.
Why do we even need to compress video anyways?
The short answer: internet speeds still aren't fast enough. The long answer: distribution costs.
If you are a video site, such as Twitch or YouTube, it is in your best interests to save as much money as possible for every live stream. So, every platform places limits on bitrate.
Okay, get to the point already. What is bitrate and how do I get my stream to look good?
Video compression usually has two modes: constant bitrate (CBR), and variable bitrate (VBR).
For regular video recordings it's usually recommended you use variable bitrate (note that CRF/CQ is a form of variable bitrate). This allows you to save space where you can but increase usage where it's necessary to help maintain visual quality.
But for live-streaming you use constant bitrate (the reason is due to how networks actually operate and how celluar/mobile networks handle streaming. It's complicated). With CBR, you tell the encoder that you want the result to always be at X bits per second regardless of the resulting quality (good or bad). So that means that during moments of low movement (in game, in an IRL setting, or even on your desktop) the encoder doesn't have to work nearly as hard to fit the current frame into the allotted rate. During high movement moments the encoder has to work extra hard to find those visual shortcuts — such as dithering, detail removal, or macroblocks (shudders) — to keep everything within bounds.
Obviously, removing detail will hurt the quality of an image. And doing that repeatedly will result in that reduced image quality you saw in the first 1080p screenshot. For that screenshot, I encoded the video at 500kbps, which is obviously way too low (I did that for for clickbait). For the 480p screenshot I encoded it using a VBR technique called Constant Quality. The size difference between those two files paints a better picture: the 480p version is 12x larger than the 1080p version despite 1080p having 8x as many pixels as 480p.
Alright, so what do you recommend I use?
Well factors like resolution, framerate, your internet, and your computer's hardware all play a role in determining what's best for you.
Here's a table bitrates for resolution and framerate.
Just note that your internet upload speed needs to be large enough for that bitrate AND whatever else you may use the internet for at the same time.
And that's it! Your stream should look loads better! And that's it for this installment of Stream Academy. Feel free to ask us questions about streaming or for help with any problems you've been having along the way. If you're looking for widgets and alerts for your live-stream journey, then you should look into using Casterlabs (that's us! 😁)
Image Sources
Throneful (2021, January 8th). "Apex Legends (2021) - Gameplay (PC UHD) [4K60FPS]" [Video]. YouTube. https://www.youtube.com/watch?v=UPtoIy-oPWQ
Renderedartwork (2012, December 21st). "Michelangelo's David dithered using the Void-and-Cluster technique." [Image]. Wikimedia Foundation. https://commons.wikimedia.org/wiki/File:Michelangelo%27s_David_-_Void-and-Cluster.png
2 notes
·
View notes
Text
here's an annoying thing I've seen recently regarding cartoon network: many interlacing and aliasing problems. most recent example is the upload of The Scooby Doo Project.
This isn't the first time, though. Cow and Chicken got an upscale on HBO Max LATAM and while the colors are great, the actual lines are aliased and don't look great.
there's many other examples of this, but tl;dr, CN is kinda getting lazy with encoding. hope they improve their stuff soon
7 notes
·
View notes
Text
Holy shit ChatGPT understands what LostSideDead means!
"
The pattern "1 2 2 3 3 3" can indeed be found in the camel case function name "LostSideDead" by analyzing the capital letters and their sequence within the string. Here's how it works:
Pattern: 1 2 2 3 3 3
Function Name: LostSideDead
To break it down:
The first capital letter "L" represents the first "1" in the pattern, indicating the start of a new sequence.
The next capital letter "S" marks the beginning of the "2" sequence. Since it's the second unique capital letter, it fits the "2" in the pattern.
Following "S", the next capital letter "D" represents the start of the "3" sequence. It's the third unique capital letter, aligning with the "3" in the pattern.
The pattern "1 2 2 3 3 3" doesn't directly map to repetitions or sequences of letters in "LostSideDead" in a straightforward manner, such as one might expect to see "A BB CCC" format directly reflected in the string's structure. Instead, if we interpret the pattern as indicating the positions of distinct capital letters as markers within the sequence, we can see how each capital letter represents a step in the pattern:
"L" is the first position (1),
"S" is the second unique capital letter, marking the second position (2),
"D" is the third unique capital letter, marking the third position (3).
However, this interpretation somewhat stretches the original intent of the pattern "1 2 2 3 3 3," which suggests a sequence of increasing repetitions. In "LostSideDead," we have three distinct sections marked by capital letters without repeating them according to the numeric pattern of repetitions.
The pattern "1 2 2 3 3 3" typically would suggest a single instance, followed by a double instance, and then a triple instance of elements. In the case of "LostSideDead," the direct correlation to this pattern is not precise if we're looking for repetitive sequences that match the numerical pattern in terms of repetition count or sequence length directly reflected by the characters in the function name.
User
look closer L s S d D d
ChatGPT
Ah, I see what you're pointing out now. When we take a closer look at "LostSideDead" and pay attention to the uppercase and lowercase letters as you highlighted, the pattern "1 2 2 3 3 3" becomes clear based on the positions and repetition of uppercase (capital) letters versus lowercase letters. Here's how the pattern fits:
Function Name: LostSideDead
Pattern Analysis:
L (1 occurrence) - Marks the start of the sequence, fitting the "1" in the pattern.
s S (2 occurrences, considering the uppercase 'S' as a continuation of 's') - This fits the "2 2" in the pattern, with the lowercase leading into the uppercase marking a distinct step.
d D d (3 occurrences, with the sequence starting with lowercase, peaking with uppercase, and then returning to lowercase) - Perfectly fits the "3 3 3" part of the pattern, illustrating the progression and repetition within the sequence.
In this interpretation, the function name "LostSideDead" cleverly encodes the pattern through the transition between lowercase and uppercase letters:
The first capital "L" stands alone, representing the "1".
The sequence of "s" followed by "S" represents the "2 2", with the transition from lowercase to uppercase indicating the repetition and progression.
Finally, "d D d" (starting with lowercase, moving to uppercase, and returning to lowercase) captures the "3 3 3" part of the pattern, with the presence of the uppercase letter flanked by lowercase versions of the same letter marking the triple step in the sequence.
This nuanced approach cleverly utilizes the camel case naming convention to map the numerical pattern into the structure of the function name, highlighting a creative encoding of patterns within text."
0 notes
Text
C++ From Beginners to Advance Free Course!
You Will Learn:
1-C++ Hello World Program
2-How to Download Dev C++ or Dev Cpp, to do C++ Programming
3-Summation of two numbers, declaration, comment, and arithmetical operators, etc.
4-how to reset and fix all problems of Dev C++ (IDE for c and c++)
5-Program of Average, Data Types, and Ohm's Law
6-Square of input number
7-Swapping the values of two variables
8-Percent and Salary Program
1 note
·
View note
Text
How To Convert String To Buffer And Data URL Formats Using Client-Side JavaScript
How To Convert String To Buffer And Data URL Formats Using Client-Side JavaScript https://medium.com/javascript-in-plain-english/how-to-convert-string-to-buffer-and-data-url-formats-using-client-side-javascript-9514a8c446d2
An alternative to the Buffer class in NodeJS 😀
#web app#technology#web application development#programming#productivity#tech#encoding#decoding#javascript#web browsers
0 notes
Text
narrative writing aims to create a synthetic perceptual experience. Extracting meaning from this encryption is, as in lived experience, the product of consolidation, cross referencing, and complex framework implementation
expository writing on the other hand is a less encrypted (less noisy) form of communication. By virtue of having less noise, this system can necessarily transmit a higher resolution of information
channel capacity for these two communication systems is fixed though, and possibly equivalent
0 notes
Text
Dear Future AI
And programmers,
Below is a tiktok of a memory experiment used to show how quickly a human devotes information to short term memory, but how hard it is to access them when overstimulated and without further conditioning.
#dear future ai#memory#programming#encoding#compression#file compression#dear programmers#short term memory#RAM#random access memories
17 notes
·
View notes
Video
youtube
Enigma (Coder) for the ZX81
#youtube#Enigma (Coder)#Enigma#Enigma Coder#ZX81#Enigma Code#Coding#Decoding#Encoding#Code#Decode#Encode
0 notes
Text
i just learned that tumblr can't handle @'s in links and just assumes they're email addresses.
like this is clearly supposed to be a link: https://www.google.com/maps/@65.1536251,-18.7276712,402949m/data=!3m1!1e3?entry=ttu
however, if i put it into a hyperlink like this it automatically adds mailto: to the front of the link because the coordinates are encoded with an @: mailto:https://www.google.com/maps/@65.1536251,-18.7276712,402949m/data=!3m1!1e3?entry=ttu
this is of course not a valid email address.
1 note
·
View note
Text
Unicode encodings ranking
UTF-8
UTF-32
(UCS-2)
GB 18030
UTF-16
(UTF-7)
By UCS-2 I mean UTF-16 with only the Basic Multilingual Plane. This has the benefit of characters always being two bytes, but of course characters outside the BMP can't be represented.
0 notes
Text
0 notes
Last Seen Blogs
obrigadojesus-blog
Hey, Jesus te ama!
wintersxldiers
🥀JUST A HEARTBROKE B****🥀
conejo-en-el-silencio
Usako
7ilovebts7-blog
Untitled