Zephyra
@zephyr.kampidh.com
Zephyr's journey of segmentation faults | Code dumps and rants, go here for art acc: @kampidh.com | Mostly works with C++ and webdev stuffs
January 7, 2026 at 7:52 PM
LRT: it was another direct upload to PDS to check which metadata got stripped by the CDN. Turns out the EXIF is intact, but sadly color profile got stripped (it was saved in R.2020 profile), well at least it got correctly converted to sRGB~
January 4, 2026 at 9:53 AM
LRT: it was another direct upload to PDS to check which metadata got stripped by the CDN. Turns out the EXIF is intact, but sadly color profile got stripped (it was saved in R.2020 profile), well at least it got correctly converted to sRGB~
Spicy undies straight to the balls, anyone?
November 17, 2025 at 5:16 PM
Spicy undies straight to the balls, anyone?
man that's a lot of garbage text data they scraped for the past month until today.. and it seems to occur in a weekly cycle
furthermore, this is with bandwidth limit enabled. can't even imagine if my hosting server had much beefier specs and bandwidth...
furthermore, this is with bandwidth limit enabled. can't even imagine if my hosting server had much beefier specs and bandwidth...
October 14, 2025 at 3:57 AM
man that's a lot of garbage text data they scraped for the past month until today.. and it seems to occur in a weekly cycle
furthermore, this is with bandwidth limit enabled. can't even imagine if my hosting server had much beefier specs and bandwidth...
furthermore, this is with bandwidth limit enabled. can't even imagine if my hosting server had much beefier specs and bandwidth...
*trying to figure out in what format the images are stored*
October 9, 2025 at 5:13 AM
*trying to figure out in what format the images are stored*
ughh, math bite me hard. apparently there was a slight miscalculation that gave me slight greenish tint even at 100%.
fixed pic, though non-linear colorspace scaling or chroma subsampling still give a greenish tint~
fixed pic, though non-linear colorspace scaling or chroma subsampling still give a greenish tint~
October 5, 2025 at 3:05 AM
ughh, math bite me hard. apparently there was a slight miscalculation that gave me slight greenish tint even at 100%.
fixed pic, though non-linear colorspace scaling or chroma subsampling still give a greenish tint~
fixed pic, though non-linear colorspace scaling or chroma subsampling still give a greenish tint~
close-up~ it's just a basic ordered / bayer dithering, but the bitdepth (R1G2B1) is kinda unusual, therefore it got a bit more shades of green available.
but I think that's not the main reason why it got a greenish tint when resizing or chroma downsampling~
but I think that's not the main reason why it got a greenish tint when resizing or chroma downsampling~
October 4, 2025 at 11:25 PM
close-up~ it's just a basic ordered / bayer dithering, but the bitdepth (R1G2B1) is kinda unusual, therefore it got a bit more shades of green available.
but I think that's not the main reason why it got a greenish tint when resizing or chroma downsampling~
but I think that's not the main reason why it got a greenish tint when resizing or chroma downsampling~
the original image is an indexed PNG, originally comes from reducing the image into 4 bit (R1G2B1).
it should be displayed correctly when viewed at 100%, or scaled when opened on a viewer that has a linear(?) interpolation, CMIIW.
here's a comparison when opening it in tev vs win photos at 50%
it should be displayed correctly when viewed at 100%, or scaled when opened on a viewer that has a linear(?) interpolation, CMIIW.
here's a comparison when opening it in tev vs win photos at 50%
October 4, 2025 at 9:26 PM
the original image is an indexed PNG, originally comes from reducing the image into 4 bit (R1G2B1).
it should be displayed correctly when viewed at 100%, or scaled when opened on a viewer that has a linear(?) interpolation, CMIIW.
here's a comparison when opening it in tev vs win photos at 50%
it should be displayed correctly when viewed at 100%, or scaled when opened on a viewer that has a linear(?) interpolation, CMIIW.
here's a comparison when opening it in tev vs win photos at 50%
lossless version, also dithering is magic.
might as well for checking whether the browser / viewer interpolates 'correctly' during resizing =p
might as well for checking whether the browser / viewer interpolates 'correctly' during resizing =p
October 4, 2025 at 9:16 PM
lossless version, also dithering is magic.
might as well for checking whether the browser / viewer interpolates 'correctly' during resizing =p
might as well for checking whether the browser / viewer interpolates 'correctly' during resizing =p
AND THE WAY HIS MOTHER CARESSES HIM AAAHHH-- SO PRECIOUS I CAN'T--
September 30, 2025 at 7:36 AM
AND THE WAY HIS MOTHER CARESSES HIM AAAHHH-- SO PRECIOUS I CAN'T--
another test with colorchecker patches from: babelcolor.com/colorchecker...
these pixels are dense. I throw some spatial dithering and it still looks good.
(and it seems pretty robust as well, I once reversed the supply polarity, got insanely hot and almost fried it~ but here it is, still alive)
these pixels are dense. I throw some spatial dithering and it still looks good.
(and it seems pretty robust as well, I once reversed the supply polarity, got insanely hot and almost fried it~ but here it is, still alive)
September 22, 2025 at 8:13 PM
another test with colorchecker patches from: babelcolor.com/colorchecker...
these pixels are dense. I throw some spatial dithering and it still looks good.
(and it seems pretty robust as well, I once reversed the supply polarity, got insanely hot and almost fried it~ but here it is, still alive)
these pixels are dense. I throw some spatial dithering and it still looks good.
(and it seems pretty robust as well, I once reversed the supply polarity, got insanely hot and almost fried it~ but here it is, still alive)
the results are surprising though. for a cheap IPS display (less than USD 3), having a low bitdepth, and subpar gamut (~60% sRGB), it behaves pretty well after calibration~
honestly, I was pretty disappointed at first with the washed-out look of this module~ :p
honestly, I was pretty disappointed at first with the washed-out look of this module~ :p
September 22, 2025 at 7:58 PM
the results are surprising though. for a cheap IPS display (less than USD 3), having a low bitdepth, and subpar gamut (~60% sRGB), it behaves pretty well after calibration~
honestly, I was pretty disappointed at first with the washed-out look of this module~ :p
honestly, I was pretty disappointed at first with the washed-out look of this module~ :p
behold, the smallest screen that I ever calibrate & profile
though this is absolutely far from ideal.. not only that I don't have the proper correction, a tiny bump to this setup can throw off the process lol
though this is absolutely far from ideal.. not only that I don't have the proper correction, a tiny bump to this setup can throw off the process lol
September 22, 2025 at 6:51 PM
behold, the smallest screen that I ever calibrate & profile
though this is absolutely far from ideal.. not only that I don't have the proper correction, a tiny bump to this setup can throw off the process lol
though this is absolutely far from ideal.. not only that I don't have the proper correction, a tiny bump to this setup can throw off the process lol
that should be pretty trivial with shaders
September 8, 2025 at 6:33 PM
that should be pretty trivial with shaders
Tried to make a calibration profile, basically just a 'simple' 1D LUT curve that's loaded into Luma3ds.
The result is pretty good, both the top IPS and bottom TN screen behaves well~ (but it does look funny when playing a game, as if playing through a desktop emulator)
before <-> after
The result is pretty good, both the top IPS and bottom TN screen behaves well~ (but it does look funny when playing a game, as if playing through a desktop emulator)
before <-> after
September 8, 2025 at 3:14 AM
Tried to make a calibration profile, basically just a 'simple' 1D LUT curve that's loaded into Luma3ds.
The result is pretty good, both the top IPS and bottom TN screen behaves well~ (but it does look funny when playing a game, as if playing through a desktop emulator)
before <-> after
The result is pretty good, both the top IPS and bottom TN screen behaves well~ (but it does look funny when playing a game, as if playing through a desktop emulator)
before <-> after
n3ds screen profiling anyone? =p
(inspired by devs on jpeg xl discord server)
(inspired by devs on jpeg xl discord server)
September 8, 2025 at 3:14 AM
n3ds screen profiling anyone? =p
(inspired by devs on jpeg xl discord server)
(inspired by devs on jpeg xl discord server)
so I dip my paws into modding..
September 7, 2025 at 2:49 PM
so I dip my paws into modding..
Why am I often hit a gold when it comes to a random string generator
August 22, 2025 at 10:44 AM
Why am I often hit a gold when it comes to a random string generator
Nintendo: "we'll brick your console if you do mods"
meanwhile Sega:
meanwhile Sega:
June 17, 2025 at 9:43 PM
Nintendo: "we'll brick your console if you do mods"
meanwhile Sega:
meanwhile Sega:
my web trap caught another (bot) victim.
didn't expect it to get into my main domain though, but heck it disguised as a normal browser.
as you can see here it manages to make almost 200k hits per hour (around 55+ hits/s or 3300+ hits/min), which is basically saturating my tiny server -_-
didn't expect it to get into my main domain though, but heck it disguised as a normal browser.
as you can see here it manages to make almost 200k hits per hour (around 55+ hits/s or 3300+ hits/min), which is basically saturating my tiny server -_-
May 27, 2025 at 6:51 AM
my web trap caught another (bot) victim.
didn't expect it to get into my main domain though, but heck it disguised as a normal browser.
as you can see here it manages to make almost 200k hits per hour (around 55+ hits/s or 3300+ hits/min), which is basically saturating my tiny server -_-
didn't expect it to get into my main domain though, but heck it disguised as a normal browser.
as you can see here it manages to make almost 200k hits per hour (around 55+ hits/s or 3300+ hits/min), which is basically saturating my tiny server -_-
"Instead of using a proprietary sized rechargeable battery, how about using an off-the-shelf 18650 cell for better repairability?"
May 11, 2025 at 11:06 AM
"Instead of using a proprietary sized rechargeable battery, how about using an off-the-shelf 18650 cell for better repairability?"
Mmmkay that was the second time that a "recertified" drive died pretty early on me...
March 31, 2025 at 1:52 PM
Mmmkay that was the second time that a "recertified" drive died pretty early on me...