The kind of ai he’s referring to in the thumbnails has actually been around for a while its ai upscaling essentially just having an ai go through all the pixels in an image and split them into 4 then color them in based on the context of the picture
@@jadesoda5305 I don't mind that kind of AI, just generative AI. It's the fact that everyone (including me) is so annoyed by the entire concept of AI (because of generative AI) and people are talking about it so much that Joe has to specify whether or not the tools are AI.
The science (and implications) behind this is fascinating - we’ve known this in video decompression since H.264 development (20 years ago), that “correct” isn’t the end of the story when it comes to decoding. It means more if you know “which correct” the decoder chooses, and then tailor the encoder to match. There are many ways to skin a cat. It gets even more fundamental than that when considering simple upscaling - even a purely uncompressed image. There are many ways to upscale, none of which involve AI. They all produce different quality of results, even though ALL of them use ALL the data and contain all the source information. There’s “correct”, and then there’s “optimal”.
I'm glad I _didn't_ know about these tools previously. My dumb ass would have spent even more time on pointlessly refining the irrelevant graphical quality of the numerous powerpoints we did in school. But now that that's over, these tools are neat!
LOL! I was in school before PowerPoint existed (there are such people), but I've spent hours trying to clean up images pointlessly. Thanks for the laugh.
Although I would say you can get even better looking images with AI up samplers Though if you care about the image retaining to the original image, these are still better
The most important piece of information in this video is not about these tools, but rather that you can open a folder in Windows, type "cmd" in the address bar, and it'll open the command line in that folder!!!
IDK about Explorer, but Directory Opus, which Joe has mentioned a few times (and I believe said he uses, and I do too) pops up a "Command prompt here" option in the right-click context menu in any folder., which does this same thing. So useful!
@@philuribe7863The default Windows 11 explorer (at least in the latest OS updates) also has the option to open the current folder in the new command prompt window. Now I try to teach myself to use that instead of typing cmd in the address bar.
Yeah, that was a brand new piece of info for me too. It's the 2nd piece of info I've learned from him: the first was what F5 gives you in notepad. It's gives you the current time and date. As I am updating a few txt files these days, it has saved me from having to type the info into the file
I knew about this but still prefer to open cmd from the search bar, copy the folder path and paste it after "cd", which is quick enough. The problem with the toolbar method is after that if you want to move back to a previous folder or copy the full path, it kinda stops working as expected (at least in W10).
6:18 "chroma upsampling" that is mentioned here isn't an unnecessary upsampling of the color information, it is necessary if your image uses 4:2:0 pixel format ! The chroma upsampling is done to get the chroma part of the image up to the luma (i.e. actual displayed image) definition, so except if your JPEG is 4:4:4 you always have to do chroma upsampling to display the final image. I guess this quality 6 mode just uses a better algo for upsampling chroma in that case.
Another tip related to the CMD one: If you are in a command prompt window and want to open file explorer to your current directory, type "start ." (with the period and without the quotes) into the command prompt.
I wonder. Due to the nature of JPG, namely that its divided into discrete chunks that dont influence each other, surely the processes employed by these tools could be parallelised on the GPU when available, meaning automatically doing this when opening a jpeg could be viable.
I knew totally opposite - PNGGauntlet and PNGquant. The former compress PNG with or without quality loss according to settings, the letter compress PNG with quality loss. Sometimes, programs(like iOS screenshots or uncompressed 3D render) make PNG without any compression and these programs helped me when there're file size restrictions but I need to avoid JPEG artifacts. However, they will take several minutes to process.
I use a program called *Upscayl* for that. It uses AI that is run locally (so no need to upload the images anywhere, but you need a decent GPU for high-resolution images), and it tries to "invent" new information into the image It is originally meant for real-life photographs, but I don't recommend it because the AI makes everything look smooth, but like too smooth. It looks like someone just used the _Fill bucket_ in MS Paint. The upscaled real-life objects don't have any texture to them... However, it's excellent at "simple" graphics, basically vector images, where the objects are made from simple lines. The AI can sharpen these types of images extremely well. You can get from 256x256 to 2048x2048, and the image will look completely sharp. It's not that good for text, however, because if the letters are very low quality, they will look deformed, so I will definitely try out programs from this video 👍 PS: If anyone wants to use Upscayl, *_remember to change the output format in the settings from JPG to PNG!_* Otherwise the program will not only compress the result image, but you will also lose all transparency values
One thing I'd definitely add to this chain of programs is good png recompression (lossless) with tools like oxipng, a rust program for recompressing png files more efficiently. The output of jpeg2png can be shrunk by like 7-8% for free, and the defaults of many other programs that output png files can be shrunk by 15% or more, especially if they leave in the alpha layer but have no alpha data (photoshop does this very often so I see a lot of that).
@@CorrosiveCitrus If you do the fancy export, yeah, but a lot of people just seem to click "quick export as PNG" which has some pretty stupid default behaviors that bloat the file size of the average png online a lot. Though even when you tell photoshop to do maximum compression from the full screen oxipng can usually find a few % filesize, and use far less CPU time than photoshop will.
I didn't know these tools existed. I was still using the old Vector magic for designs that can be vectorized and just some waifu2x for other things. This one helps removing the artifacts a lot
@@ThioJoe i have to disagree. adobe enhance sounds very "gross" to me, and very obvious. i've had to stop watching certain youtube channels because they switched to using adobe enhance instead of traditional noise removal.
Cider, a third party program for Apple Music, uses an algorithm (Adrenaline Processor) that is supposed to remove artifacts commonly associated with AAC and makes it sound better. Specific use case but it does seem to exist.
@@agiverreviga4592 I don't think this would be possible in a similar way. The reason that this works well for text on jpg is twofold: 1. The boundary around text is very low frequency (i.e. there are few transitions from one colour to another - it is easy to identify the outlines) 2. The parameters of jpg compression can be easily deducted, if they're not even explicitly encoded in the file. This provides a way to test any "guesses" about which noise is due to artifacts or organic to the image. With low bitrate audio, I think there is more loss of the original signal and fewer hints to tell signals and noise apart. So you pretty much have to rely on solutions like AI that "guess" more instead of using a specific algorithm.
As far as text is concerned, I usually just determine what font was used and redo the text and realign it as best I can. This works most of the time. It's very rare that I don't come right with the font.
Back in the early days of the Internet, I dabbled with software that made use of the dynamic compression of jpeg. I would use a strong compression of the image overall, but then mark certain zones to have no compression, usually where text existed in the image. … all I need to do is remember what software I was using for that task 25 years ago 🤔
Another tool that do this is the discontinuated Microsoft Office Picture Manager that comes with office 2003-2007, was incredible making zoom without loss
Ironically, you could phrase the problem shown in the video as initializing the parameters of a very particularly defined, mostly frozen neural network with naive guesses for the unrounded DCT values and then training it to minimize a loss metric concerned with "noisiness"
JPG compression is something really insane technically. How the heck they came out with it! There is a video on TH-cam worth checking out about JPG compression. Insanity.
@@hopperelec That's not how it works. And that shouldn't make someone's deliberate offline image browsing deliberately suck, so it's mediocre quality just for convenience. It's the bluetooth "buds" vs. wired headset argument all over again.
@@dominic.h.3363 That is mostly how it works. Of course, some it could be done asynchronously, but not all of it. And you can't just render the images one way in browsers and one way offline- there needs to be a standardized way to render them so that the same file looks the same anywhere.
Depending on what the user is doing, a few seconds could add up quickly. Like if you are opening a folder of images to go though them, a few extra seconds when looking at each image could be significant overall. Also because this is only needed when the image has artifact issues, there wouldn't be a need to do this with every image so it still wouldn't make since to do it all the time anyway.
@@hopperelec You are moving goalposts. I never said this should be the DEFAULT method of rendering images, I said the OPTION should exist to MAKE it the default, offline. Viewing images I have on my computer. I couldn't give two shits about websites.
Kind of off topic here but I keep thinking "AI" is "AL". Doesnt help if you're slightly dyslecic like I am. I dont know why but an upper case "l" and a lower case "I" are the exact same, you can't even tell witch is witch. Why they don't put the extra bits in so you can differenciate between the two letters is beyond me.
Many people don't know in what cases a PNG is smaller. And I hate it when people make screenshots and paste them directly on Discord. Everything becomes a PNG, where the size is often a few MB, instead of like 400 KB.
People are lazy and jpg with 80ish % quality still is the default file format in a lot of programs Without Google and their bs we'd be using JXL on all our devices by now
I wonder how much difference these tools make when a JPEG is encoded by JPEGLI. JPEGLI is a new encoder that makes the same JPG where the artifacts are still there but less visible.
Not super related to the video but recently I've discovered that a fully black picture is smaller in size on png compared to jpg which is unusual for the 2 formats. It probably makes sense if you look at how those 2 compressions work but I still found it neat.
I really appreciate your efforts! Just a quick off-topic question: I have a SafePal wallet with USDT, and I have the seed phrase. (alarm fetch churn bridge exercise tape speak race clerk couch crater letter). Could you explain how to move them to Binance?
word of warning , do not watch this after a long day of work dealing with excels or reading materials on your laptop or computer..... i almost forgot i was watching a video at full screen
When I started watching this video, I figured that I wouldn't be able to use either program, since I have a VERY old computer, running a VERY old version of Windows. I was pleasantly surprised to find that both JpegQuantSmooth and Jpeg2Png (original) ran just fine. However when I tried Jpeg2Png on a screenshot, it gave me an error. An error, I see your patched version has fixed. So I downloaded the patched version and tried it. Naturally, your patched version wouldn't run at all, and Windows told me that it wasn't a valid program. Why did this happen? Because you compiled it with a program, or options that just automatically turn out executables that are are broken on older versions of Windows. People will say "You're using an obsolete version of Windows, you can't expect people to still support it!" yet, the fact that the original version works on older version of Windows proves that it doesn't actually need a newer/modern version. The compiler just doesn't bother to produce a backward compatible binary. In other words, it's not planned obsolescence, it's forced obsolescence. While I don't really expect authors to go out of their way to support older versions of Windows, I find it extremely annoying to see programs that are incompatible for no other reason than a compiler intentionally breaking backwards compatibility.
jpeg is much better for photos at creating a smaller file than png without losing as much noticeable difference... but for screenshots, your statement is true.
I'm not familiar with Github. I see there a list of folders and files (if that's what they are) but I don't see the exe to download. So how do I proceed in practical terms to use this software?
![SAFEPLANET - ดาวเคราะห์ [ THE PLANET ]](http://i.ytimg.com/vi/4X1XL3I_k3c/mqdefault.jpg)
I can't believe we live in the age where you have to specify that you're not talking about AI.
Also the first ad I got on this video was for the Rayban AI glasses where someone just cheats at pool
I can't believe we have to say that as a good thing.
Welcome to brave new world, pal.
The kind of ai he’s referring to in the thumbnails has actually been around for a while its ai upscaling essentially just having an ai go through all the pixels in an image and split them into 4 then color them in based on the context of the picture
@@jadesoda5305 I don't mind that kind of AI, just generative AI. It's the fact that everyone (including me) is so annoyed by the entire concept of AI (because of generative AI) and people are talking about it so much that Joe has to specify whether or not the tools are AI.
The science (and implications) behind this is fascinating - we’ve known this in video decompression since H.264 development (20 years ago), that “correct” isn’t the end of the story when it comes to decoding. It means more if you know “which correct” the decoder chooses, and then tailor the encoder to match. There are many ways to skin a cat.
It gets even more fundamental than that when considering simple upscaling - even a purely uncompressed image. There are many ways to upscale, none of which involve AI. They all produce different quality of results, even though ALL of them use ALL the data and contain all the source information. There’s “correct”, and then there’s “optimal”.
I'm glad I _didn't_ know about these tools previously. My dumb ass would have spent even more time on pointlessly refining the irrelevant graphical quality of the numerous powerpoints we did in school.
But now that that's over, these tools are neat!
omg frrrrrrr
LOL! I was in school before PowerPoint existed (there are such people), but I've spent hours trying to clean up images pointlessly. Thanks for the laugh.
wait, I can open to the current directory by typing cmd in explorer?!!!! 🎉
YO HOLD UP THATS FIRE
Not only that, you can open any executable in the address bar of a folder and it will start with that folder as current directory
It's basically address bar like in web browsers that handles protocols
Can't wait for the "It has been a thing in Linux for [ridiculous amount of years]" comment :}
‘Not Ai’ what has the world come to Thio?😂
Although I would say you can get even better looking images with AI up samplers
Though if you care about the image retaining to the original image, these are still better
So says the guy who uses the phrase "game-changer" in his video. Theo might be an AI!
The most important piece of information in this video is not about these tools, but rather that you can open a folder in Windows, type "cmd" in the address bar, and it'll open the command line in that folder!!!
IDK about Explorer, but Directory Opus, which Joe has mentioned a few times (and I believe said he uses, and I do too) pops up a "Command prompt here" option in the right-click context menu in any folder., which does this same thing. So useful!
I knew this one, but after 30 years of using and abusing Windows, I sometimes get surprised by gems like this. Sometimes, it's pretty annoying. 😂
@@philuribe7863The default Windows 11 explorer (at least in the latest OS updates) also has the option to open the current folder in the new command prompt window. Now I try to teach myself to use that instead of typing cmd in the address bar.
Yeah, that was a brand new piece of info for me too. It's the 2nd piece of info I've learned from him: the first was what F5 gives you in notepad. It's gives you the current time and date. As I am updating a few txt files these days, it has saved me from having to type the info into the file
I knew about this but still prefer to open cmd from the search bar, copy the folder path and paste it after "cd", which is quick enough. The problem with the toolbar method is after that if you want to move back to a previous folder or copy the full path, it kinda stops working as expected (at least in W10).
4:45 whoah are you kidding me? Doing it the hard way for 30 years :D
@@the-niker me too! Always copied from there then windows r cmd “ CTRL v “ - great time saver !
@UXXV Too bad you can't seem to elevate the terminal this way.
Also if you are in the command prompt you can use either "start ." or "explorer ." to open an explorer window with the current directory
@@the-niker You can do sudo lately, you have to enable it first though in the settings.
@@the-niker You can elevate an existing CMD but it's kinda tedious without a script or Alias.
6:18 "chroma upsampling" that is mentioned here isn't an unnecessary upsampling of the color information, it is necessary if your image uses 4:2:0 pixel format !
The chroma upsampling is done to get the chroma part of the image up to the luma (i.e. actual displayed image) definition, so except if your JPEG is 4:4:4 you always have to do chroma upsampling to display the final image.
I guess this quality 6 mode just uses a better algo for upsampling chroma in that case.
Another tip related to the CMD one:
If you are in a command prompt window and want to open file explorer to your current directory, type "start ." (with the period and without the quotes) into the command prompt.
WAIT I COULD HAVE DONE THAT INSTEAD? My dumb ass be typing in `Explorer .`This whole time
@@breadles5 You should be able to type explorer and have it open the working directory but no, Windows is stupid.
Space before dot?
@johanponken yes, as dots specify cwd
@@breadles5 i always just use "explorer .", it's more clear what it does and doesn't really have any issues
Glad to know this. I always use PNG, GIF, and GIF
why does everyone call it JIF
use jpeg-xl and webp also
@@tauon_ use JXL don't use WebP
@@BurnerAccount101-ui4et The eternal GIF pronunciation mystery 😂
@@BurnerAccount101-ui4et Cause it's like giraffe, and what furries do.
I wonder. Due to the nature of JPG, namely that its divided into discrete chunks that dont influence each other, surely the processes employed by these tools could be parallelised on the GPU when available, meaning automatically doing this when opening a jpeg could be viable.
GIMP with the G'MIC addon installed also hasa function to remove JPEG artifacts called compression smoothing out something like that
There should be a way to implement these tools as GIMP plugins too
doesnt work this good tho
These algorithms depend on the actual JPEG data to work, a GIMP plugin probably won't have access to that.
@@-ion Ah, that makes sense
Why you write like a Mario
As a graphic designer, this will be quite useful. Thank you for sharing this.
Thio manages to make the most useful tools known to Windows~
It's insane!Thanks for helping out!
I knew totally opposite - PNGGauntlet and PNGquant. The former compress PNG with or without quality loss according to settings, the letter compress PNG with quality loss. Sometimes, programs(like iOS screenshots or uncompressed 3D render) make PNG without any compression and these programs helped me when there're file size restrictions but I need to avoid JPEG artifacts. However, they will take several minutes to process.
I use a program called *Upscayl* for that. It uses AI that is run locally (so no need to upload the images anywhere, but you need a decent GPU for high-resolution images), and it tries to "invent" new information into the image
It is originally meant for real-life photographs, but I don't recommend it because the AI makes everything look smooth, but like too smooth. It looks like someone just used the _Fill bucket_ in MS Paint. The upscaled real-life objects don't have any texture to them...
However, it's excellent at "simple" graphics, basically vector images, where the objects are made from simple lines. The AI can sharpen these types of images extremely well. You can get from 256x256 to 2048x2048, and the image will look completely sharp. It's not that good for text, however, because if the letters are very low quality, they will look deformed, so I will definitely try out programs from this video 👍
PS: If anyone wants to use Upscayl, *_remember to change the output format in the settings from JPG to PNG!_* Otherwise the program will not only compress the result image, but you will also lose all transparency values
I'm sure you've gotten a ton of thanks already but I have to say these are incredible tools and thank you for introducing them to us!
One thing I'd definitely add to this chain of programs is good png recompression (lossless) with tools like oxipng, a rust program for recompressing png files more efficiently. The output of jpeg2png can be shrunk by like 7-8% for free, and the defaults of many other programs that output png files can be shrunk by 15% or more, especially if they leave in the alpha layer but have no alpha data (photoshop does this very often so I see a lot of that).
Doesn't photoshop give you the option?
@@CorrosiveCitrus If you do the fancy export, yeah, but a lot of people just seem to click "quick export as PNG" which has some pretty stupid default behaviors that bloat the file size of the average png online a lot. Though even when you tell photoshop to do maximum compression from the full screen oxipng can usually find a few % filesize, and use far less CPU time than photoshop will.
I do photo editing every day in my work. This was such a helpful video, Joe. Thanks!
I didn't know these tools existed. I was still using the old Vector magic for designs that can be vectorized and just some waifu2x for other things. This one helps removing the artifacts a lot
The price to performance would be crazy
Does such thing exist for low bitrate audio files?
Actually yes adobe has their “enhance speech” tool which I’ve talked about a while ago. It does use AI but it is extremely good
@@ThioJoe i have to disagree. adobe enhance sounds very "gross" to me, and very obvious. i've had to stop watching certain youtube channels because they switched to using adobe enhance instead of traditional noise removal.
Some Cowon players have an option like this called "MP Enhance"
Cider, a third party program for Apple Music, uses an algorithm (Adrenaline Processor) that is supposed to remove artifacts commonly associated with AAC and makes it sound better. Specific use case but it does seem to exist.
@@agiverreviga4592 I don't think this would be possible in a similar way.
The reason that this works well for text on jpg is twofold:
1. The boundary around text is very low frequency (i.e. there are few transitions from one colour to another - it is easy to identify the outlines)
2. The parameters of jpg compression can be easily deducted, if they're not even explicitly encoded in the file. This provides a way to test any "guesses" about which noise is due to artifacts or organic to the image.
With low bitrate audio, I think there is more loss of the original signal and fewer hints to tell signals and noise apart. So you pretty much have to rely on solutions like AI that "guess" more instead of using a specific algorithm.
Your videos are so often such a hidden gem!
Those programmes look amazing. I will definitely give them a try.
Great video, I might actually use this on a regular basis.
Waifu2x for cleaning up jpegs is the one valid use of AI, that being said, is neat that there are non-ML tools that work really well
Great find! Never heard of this, thanks! Will be fun to play with.
Imagine if browsers automatically did that when you download JPEGs, and also converted WEBPs to PNG, sounds like heaven
I needed this. Been doing it manually in Photoshop for a while now. Extremely time consuming.
it's absolutely insane!
Thanks, will come in handy!
As far as text is concerned, I usually just determine what font was used and redo the text and realign it as best I can. This works most of the time. It's very rare that I don't come right with the font.
Back in the early days of the Internet, I dabbled with software that made use of the dynamic compression of jpeg.
I would use a strong compression of the image overall, but then mark certain zones to have no compression, usually where text existed in the image.
… all I need to do is remember what software I was using for that task 25 years ago 🤔
Another tool that do this is the discontinuated Microsoft Office Picture Manager that comes with office 2003-2007, was incredible making zoom without loss
I thought he was gonna explain JPEGMAFIA's tools
i hope this is implemented on all pcs and all internet so that we dont even know the image was actually bad quality
Thank you for this video!
Ironically, you could phrase the problem shown in the video as initializing the parameters of a very particularly defined, mostly frozen neural network with naive guesses for the unrounded DCT values and then training it to minimize a loss metric concerned with "noisiness"
JPEG 50% quality looks like VHS
It feels like your channel was made for me
JPG compression is something really insane technically. How the heck they came out with it! There is a video on TH-cam worth checking out about JPG compression. Insanity.
I've used these. Results from my tests seemed quite similar to Selective Gaussian Blur with careful control of the threshold.
very helpful, thanks
It would take SECONDS to open every single jpeg image? The horror...
Wait until you find out that many websites have HUNDREDS of jpeg images on them, so you'd be waiting HUNDREDS of seconds
@@hopperelec That's not how it works. And that shouldn't make someone's deliberate offline image browsing deliberately suck, so it's mediocre quality just for convenience. It's the bluetooth "buds" vs. wired headset argument all over again.
@@dominic.h.3363 That is mostly how it works. Of course, some it could be done asynchronously, but not all of it. And you can't just render the images one way in browsers and one way offline- there needs to be a standardized way to render them so that the same file looks the same anywhere.
Depending on what the user is doing, a few seconds could add up quickly. Like if you are opening a folder of images to go though them, a few extra seconds when looking at each image could be significant overall.
Also because this is only needed when the image has artifact issues, there wouldn't be a need to do this with every image so it still wouldn't make since to do it all the time anyway.
@@hopperelec You are moving goalposts. I never said this should be the DEFAULT method of rendering images, I said the OPTION should exist to MAKE it the default, offline. Viewing images I have on my computer. I couldn't give two shits about websites.
Kind of off topic here but I keep thinking "AI" is "AL". Doesnt help if you're slightly dyslecic like I am. I dont know why but an upper case "l" and a lower case "I" are the exact same, you can't even tell witch is witch. Why they don't put the extra bits in so you can differenciate between the two letters is beyond me.
*Very useful thank you* 👍👍👍👍
Incredible wow. I use either png or webp for my screenshots on my site but this is mind blowing
Thanks for the cmd tip.
Many people don't know in what cases a PNG is smaller.
And I hate it when people make screenshots and paste them directly on Discord. Everything becomes a PNG, where the size is often a few MB, instead of like 400 KB.
256 colors does the trick (actually in IrfanView doing 65535 first can get you down to 50-90 colors, but then do 256 to get palette).
I never know you could type cmd in the file explorer and open that path!
I will never type cd again🤣
People are lazy and jpg with 80ish % quality still is the default file format in a lot of programs
Without Google and their bs we'd be using JXL on all our devices by now
Now we need a program that transform lossy audio into lossless :)
I wonder how much difference these tools make when a JPEG is encoded by JPEGLI.
JPEGLI is a new encoder that makes the same JPG where the artifacts are still there but less visible.
That's amazing. I wonder if you can make a custom button in Directory Opus that applies this tool to selected images.
Thank you, I need this to improve some poor quality Screanshots(SSs) of Book Pages mostly
Upscayl is what I would use
Not super related to the video but recently I've discovered that a fully black picture is smaller in size on png compared to jpg which is unusual for the 2 formats.
It probably makes sense if you look at how those 2 compressions work but I still found it neat.
Who needs artificial intelligence when I have so much natural stupidity? ¯\_(ツ)_/¯
That's why I subscribe this channel
4:49 I did not know this pro tip relating to cmd. Not that I use it often, but good to know, I suppose.
Helpful man so damn helpful.
I'm stuck on TH-cam watching a video about cutting edge dithering. Seems appropriate.
THANK YOU SO MUCH
I CAME FOR THE PIXEL PEEPING
In about 8 hours your Github repo went from 90 ⭐ to a 115 :) [ _edit_ CET 10-19 if being very specific matters ]
4:10. *Info starts here. EVERYTHING before is just blather and a guy who just owes his own voice.*
The only reason I clicked on this videos is because it has "not ai" in the title.
Forget iPhones, what about charging Android phnes?
"Do I look like I know what a JPEG is?"
Cool! I've had PNG as my default screen shot format for ages, jpeg is really bad for that use case.
I really appreciate your efforts! Just a quick off-topic question: I have a SafePal wallet with USDT, and I have the seed phrase. (alarm fetch churn bridge exercise tape speak race clerk couch crater letter). Could you explain how to move them to Binance?
this is the funniest thing ever when watching this while drunk
word of warning , do not watch this after a long day of work dealing with excels or reading materials on your laptop or computer..... i almost forgot i was watching a video at full screen
Can’t wait for the JXL revolution
Do I look like I know what a JPEG is? I just want a picture of a god dang hot dog!
Finnally a bnot AI video.
When I started watching this video, I figured that I wouldn't be able to use either program, since I have a VERY old computer, running a VERY old version of Windows. I was pleasantly surprised to find that both JpegQuantSmooth and Jpeg2Png (original) ran just fine. However when I tried Jpeg2Png on a screenshot, it gave me an error. An error, I see your patched version has fixed. So I downloaded the patched version and tried it.
Naturally, your patched version wouldn't run at all, and Windows told me that it wasn't a valid program. Why did this happen? Because you compiled it with a program, or options that just automatically turn out executables that are are broken on older versions of Windows.
People will say "You're using an obsolete version of Windows, you can't expect people to still support it!" yet, the fact that the original version works on older version of Windows proves that it doesn't actually need a newer/modern version. The compiler just doesn't bother to produce a backward compatible binary. In other words, it's not planned obsolescence, it's forced obsolescence.
While I don't really expect authors to go out of their way to support older versions of Windows, I find it extremely annoying to see programs that are incompatible for no other reason than a compiler intentionally breaking backwards compatibility.
Oh my god, I'm too late to learn about this app, Instead I converted the most old JPG graphic design with some of compression into SVG xD
Would be nice if someone would make a gui version
jpegqs64 has a couple of exes with GUI in the names, and also a plugin to IrfanView
Can you provide documentation or a link on how to compile.
I've just gone and added some detailed instructions on the Wiki of my jpeg2png repo fork that should help.
Video starts at 4:11
So, you dumped the audio translation project for good?
Why did ugreen sponsor thiojoe, michael mjd and bringus in a span of 2 days?
very cool!
Do they use BI or CI?
DI ... maybe?
Can you please tell me the name of the personal cloud you showcase once in one of your video? I have been looking for it and I can't find it.
Crazy how "AI" means "evil" in some people's mind.
Actually both of them use the foundation of ML aka AI.
Is that true?
@@CorrosiveCitrus no, it's not. They don't use ML or AI. The closest you'll get is that they all use maths.
@@mattsephton That's what I thought -- don't know why there's so many people saying it
please explain how to run it after downloading it .
Thank you!!! This is great.
Then please make an alternate Windows Photo Viewer with a checkbox to view jpgs de-artifacted.
Not gonna lie, even at 1080p, I couldn't really tell the difference. I had to really look. So I guess this is just kind of nitpicking.
Why doesn't every image viewer use that normally?..
He said why in the video. It's because it would take longer to open the image
Why does jpeg get used ever when png exists?
jpeg is much better for photos at creating a smaller file than png without losing as much noticeable difference... but for screenshots, your statement is true.
All I learn from this video you shouldn't use Jpg for screenshots and Ugreen makes good power banks
I just saw the noification
I'm not familiar with Github. I see there a list of folders and files (if that's what they are) but I don't see the exe to download. So how do I proceed in practical terms to use this software?
i am watching you twice
Ever since you've stopped posting fake/joke videos you've been posting banger after banger, you're joke videos were still funny though