#đŽâ ctt.cx website
3710 messages ¡ Page 4 of 4 (latest)
i think this is a bit better than what is there rn i am still working on it
*your videoâs
im not content with the term wiki
i gotta remove it from the readme and homepage description
what would it be instead
i dont have a proper term for ctt
alt text
Documentation: Explaining software
- Tutorials: How to achieve X
- Checklists: Good things to know / keep in mind
- Aggregators: List of resources
it's a bit of all of this
so a knowledge base
Frame interpolation, most commonly refered to as interpolation, is a video processing technique that creates new (imperfect) frames from existing ones by calculating how one would look like between point A and B. It's mostly known to make slow-mos of 24/60fps footage, here it's used in tandem with frame blending in order to have even more blur samples, to notice evzn less resample steps.
Keep in mind it creates frames from existing ones and has compromises:
- Smearing on static parts (e.g. overlays & video game HUDs) because of the gap between frame A and B being too distant, wether it's because the initial framerate is too low or the difference too much to predict
- Rendering times: The higher the fps (thus more seamless) you go, the more frames you need to interpolate and generate and blend the slower it'll take to render the same video length.
I tried to stay more concise and give in more details and terms to grasp it better
yeah but how to i present it to people who don't? i gotta use words
show don't tell maybe?
It's mostly known to make slow-mos of 24/60fps footage, here it's used in tandem with frame blending in order to have even more blur samples, to notice evzn less resample steps.
this is a bit weird
Smearing on static parts
this whole part could be summarized with just explaining artifacting then using smearing as an example
Frame interpolation, most commonly referred to as interpolation
i wouldn't say its commonly referred to as interpolation, maybe in this server it is, but not in general
"calculating how one would look like between point A and point B" is also weird
what point A and point B
the last paragraph is also unreadable for me
frame a frame b
"the slower it will take" is a misuse of that idiom
what about when i CANT
then say that?
*often/sometimes shortened to just interpolation
that'd fix it
even if u do it's pretty much a nothingburger imo
"generates a frame in between a and b by calculating what one would look like between a and b"
also we could just have a gif (by gif i mean webm) visually showing interpolation
like showing 2 frames and then an in-between frame
where else do you see people calling blur/smoothie "frame interpolation"?
yeah i think that'd explain it a whole lot better
just use what june sent + that
not here, but there is a lot of things that can be "interpolated"
and specfically we are speaking of frame interpolation
so to me it makes more sense to say that
yeah u have to mention that this wiki specifically means interpolation as in frame interp
it's not a thing people normally do
depending on ur background it can be very confusing to see if it's not explained clearly
we are not normal chat đ
this would be transcoding no
@crude elm confirm pls but i think so
encoding is taking the raw and putting it into a digital format
but encoding is often used in both ways
the raw is already a digital format
the "trans" part means that the format changes
in (re)encoding that doesn't have to happen
so it says video encoding but then says the definition of video transcoding (except that the output format of a transcode doesn't necessarily have to be compressed, just encoded but whatever)
focusing on finishing migrating and finishing my projects? ehhh idk
integrating this into ctt? hell yea!
Whatâs happening to the discord when the wikis done
We should just delete it on god
Like will this discord go from information base -> website based and render tests in discord
Because realistically you just need one channel after the whole wiki is done saying âall information has been migrated to ctt.cxâ
yeah, itâll pretty much just be a discussion server surrounding the topics of ctt.cx
Thatâs the plan afaik
Move all of the static/completed pages to an archive category
Things like #đŁď˝the-tweak-tips and #đâ links-and-news will obviously stay here
Same with #âď˝render-hall-of-fame
a
a lot of the stuff from the-tweak-tips will be re-included within categorized sections
Time posterization, most commonly refered as video interpolation or time interpolation is a video post processing effect that make virtual videos looks like motion blurry like real camera videos
frame blending makes it blurry
interpolation give more temporal detail to frame blending
Frame blending and intermediate frame generation are both motion interpolation
One is increasing the framerate the other is decreasing, it is a time interpolation not a frame
frame blending does not decrease frame rate
schizo thread
frame blending bledns frames creating motion blur effect
you are all wrong
here is 990 blended to 60FPS blended with 17 weights with the clip = havsfunc.ChangeFPS(clip, int(fbd['fps'])) commented out of the code
the video stayed at 990fps, and i slowed it down to 25% speed so you can play it on discord
Frame interpolation is a video effect, frame mean is an image effect
Yeah on gang shit
smearing on static elements of UI / smearing on static GUI elements / smearing on static elements on the screen
literally just shoved in an iframe
kinda like obsidian's graph view
kinda like what they got going on in https://publish.obsidian.md/hub
hear me out
add background color css on the map, make it the same as the docs background
try it out in a separate branch sure


what
on MAC
happens to me as well
howe much they pay you serfeant
when u scroll down in a page that doesnt have an emoji
the name of the website is replaced by the name of the page
though if the first heading actually has an emoji it's not scaled properly
i just realized this and am filing a but report asap
well shit i cant reproduce the damn bug
i just wrote one fancy ass bug report
having #đŹď˝suckless-cut, the GitHub README, and having it on https://ctt.cx
is making me realize documenting the same thing at three places is kinda redundant
3 places is fine if they all serve different purposes
- channel is for the discord community here that doesnât pay attention to other things
- readme is for the people on GitHub
- website is for the eventual transition to the website so we donât have to add it later
I think itâs fine
If you had it on two servers or two websites or two repos, yeah, thatâd be redundant
- that doesnât pay attention to other things*
wdym
Do you think that any more than 10% of this 18k member server even knows we have a website?
Thatâs who the channel here would be for
The website is just for when we eventually move everything else there
yet to move to ctt:
video
- x264 potato settings
- voukoder & nle settings
- 4K-Notifier
awesome
- lists & tweaking guides
- fmhy, privacyguides, tsd?
programs (or windows?)
- programs (100+ of them)
windows
- tweaklist
- link windows_for_noobs
"tweaking"
-
configuring optifine, lunar client, cheatbreaker
-
gpu resources, winutil
idk how imma organize software (programs), and operating systems (windows & linux)
@ocean ledge maybe useful to mention on wiki somewhere?
đ
happens in both chromium(brave, vivaldi) and firefox(firefox esr, floorp, mercury)
it's on squidfunk to fix that
volunteers for localizations:
german: tye2512
french: me lol
add ukrainian RAHHH
are you volunteering to translate?
i am a bit rusty with it but yes
honestly doubt anyone will ever use it, but would be cool
HELLO LOOK AT MY COOL AUTOMATIC SUPPORT BOT!!
case numba one: the user provided a full screenshot
OCR scanned the presence of both Traceback (most recent call last): AND Invalid data found when processing input.
case numba TWO: user is a dummy dummy and did NOT include a full error:
holy shit, cooking?
{
"^(?!.*Traceback \\(most recent call last\\):).*Invalid data found when processing input.*$": "Include a whole screenshot pwease"
}
litewawy this :3
why you a cat
@quiet olive @sharp pine look guys i made something super cool!!!!!!!!!!!!!!
no, internalmy it just appends what it OCR's to the original message
you could also select copy and paste from your terminal and process much faster
what if you make it automatically send the link to the error from https://rentry.co/smrs-errors (once i finish it)
why isnt this on the ctt wiki
you can also add the following
"remove the following from your recipe"
oh i just had to read below đ
Video track is unseekable https://discord.com/channels/774315187183288411/1121287665749340170
NotADirectoryError() https://discord.com/channels/774315187183288411/1106959205627347046
Cannot load nvcuda.dll
RIFE: failed to create GPU instance https://discord.com/channels/774315187183288411/1094996295082586292
me wen
gg
well played
useful resource thank you guys
should i move this to github so anyone can contribute
it has been couked
automatic support bot is crazy
Would be a good addition
@inner fossil is this worth of doing at all
@grave crater wrong person mb lol
i have yet to test the real world performance difference
it complies and optimizes the binary for the cpu the system is using
so in practice yes, its good but at the same time you cant redist the binary
since its cpu specific
i dont see how it would improve perf though
make benchmarks before saying its helpful
aside for anything computationally intensive like gfx apis
ill test this out on doom eternal (not as demanding) and portal rtx (demanding)
it's a really useless tutorial
you can do this with any program on your system, if you actually need better perf you do this on the system level through your makepkg config
you can just put the bloat behind a spoiler
want push perms for troubleshooting.md?

my dns dont like it
Not in DNS records
@ocean ledge could i make flowframes build guide for the wiki so ppl can get latest versions without paying !?
wouldn't a script be more useful
guide would be a good idea
yes probably
I would assume a build script being possible
what'd be the point of using flowframes
@ocean ledge
tried to explain the gaussian params better and removed all the custom stuff since thats byebye
also i will make gif showing difference of all the weights instead of just equal and ascending if u want
is it worth saying which weights differ from blurâs weights
idk how the weights differ exactly so someone else would write that
okie make sure to encode it well :3
@sileny erm
@âŹilrnt https://rentry.co/weighting-visualized
weighting
description
image
ascending
newer frames are more opaque than the previous ones
descending
older frames are more opaque than the newer ones
equal
all frames are equally opaque
gaussian
opacity peaks at the center frame and fades out towards the edges, mimicking a gaussian curve
gaussian...
why you frĂŠnch
its less understandable for new ppl
the descriotions are good tho
Should change ascending and descending since theyâre straight lines, not curves (unless theyâre actually not, please check src for me)
i was feeling ĹĂĹĹĂż
Gaussian and Gaussian symmetric are inaccurate
Also pyramid
Pyramid isnât symmetric by default
Itâs just one side
(Thatâs the issue Gaussian has too)
so itâs like ascending?
Yes
But showing it like that isnât very good since humans donât perceive the difference between gradients of light and dark linearly
Humans are much better at seeing the difference between lighter shades than darker ones
thank you
@silemt oh isee
idk i just called them curves because thats what the docstrings for the script do
but tbf they say linear ascending curve
linear curves arent curves
yes they are
i have curves
the gaussian one is wrong if this is supposed to explain default behavior
and the descriptions for those 2 are just paraphrased versions of each other
not very helpful
if you can tell me how f(x) = x where f: R -> R doesnt fit this definition from wikipedia
A curve is the image of an interval to a topological space by a continuous function.
then i will change it
it is in smoothie
gaussian is the lower half of the bell curve in smoothie (as opposed to upper half in blur)
pyramid in smoothie is equivalent to pyramid_sym in blur
those are the only changes going from blur to smoothie
oh no you caught me!
A linearly increasing curve isnât the same as a linear increase (which is a straight line)
Gah
Theyâre not known as curves in any context besides the definition given by atzur
Typically, a curve is a line which is curved
thats the topological definition
A like wh ich has no curve is not a curve
im not going by the colloquial one
Mathematically, you could call it a curve but it is very unusual
most (probably all) mathematical definitions of a curve dont disallow it being "straight"
yes this is what i learned
Please read both
Given this, I think itâs best to remove ambiguity
those are colloquial definitions
its supposed to be a mathematical library so i will use the common mathematical ones
i dont think its ambigous ive heard ppl say linear curve a lot
Theyâre not, unless you consider a definition used by the vast majority of a large country âcolloquialâ
yeah thats exactly what i would consider colloquial
Why use the confusing mathematical one if thereâs a possibility for a non-confusing and mathematical one?
Colloquial generally refers to smaller areas, like regions or cities
i dont think the word colloquial has any implications for how many people use it
Iâd replace linear with straight (or add it to that)
just that its used in informal contexts
from wikipedia curve page
a) curve is consistent with the descriptions for all the other functions
b) "line" sounds really weird
I donât think most users of smoothie looking for an explanation on weighting would see curve and think of a straight line
"straight curve" and "linear curve" dont really seem all that different to me
straight curve is borderline oxymoronic outside of mathematics
And Iâd consider them to be different, at least at a glance
Yes, theyâre the same thing, but the connotation feels different
it does, and i think "straight" sounds too related to shapes to make sense here
Also why donât we just include graphs of the functions in the wiki
Seems like an easy solution
eg. "straight increase" vs "linear increase"
The gradients are bad for this reason
straight is related to shape/form
not related to me
đŻ
but linear also has direct mathematical ties
So thatâs why it sounds better here
Iâd say âopacity of blurframes is increased linearlyâ
âď¸
or âthere is a linear increase in the opacity of the blurframesâ
sounds cool
I have the Gaussian functions graphed
In a pinned message in #đŹâ video-discussion I think
that woulkd be so funny for outsiders
"blur enthusiast choosing weighting"
with whole lot of graphs
works for the smoothie wiki sure but i cant write that in the docstrings
how about
i keep my vaguely elitist "curve"
and u guys write whatever is most intuitive on the wiki
đ
because saying curve is the most terse way i can put it and im not writing allat
If the typical smoothie user decides to look at docstrings, thatâs their fault
Pretty sure these variables are straight from smoothie so this should be accurate
Please do
lmk if anyone wants an embeddable geogebra applet for playing around with the gaussian/sym parameters to see how they affect the graph (i can try to figure it out)
u would think ascending is the most popular
Why on earth would you think this
i should probably remove gaussian_sym and just make gaussian default to being symmetric
altho that would break a lot of configs
good visualisation here
clicking brings you to desmos
for some of them
the images also don't have enough movement to really show the difference imo
you thinj he remembered?
wait wrtong imgs
someone make it a gif
i am lazy
@crude elm why does the ascending and descending look the opposite of what they are
could u do the same test with less blurframes
guhh ??
what were the settings exactly
[interpolation]
enabled: yes
masking: no
fps: 2x
speed: medium
tuning: weak
algorithm: 23
use gpu: yes
[frame blending]
enabled: yes
fps: 60
intensity: 1.0
weighting: descending
bright blend: no
did you do an oopsie on your mafs
ffmpeg can do that đ
ik it can but
i hate gif with all my soul
if you just do a basic command
it just looks awful
wait lemme gifski it
đ§
still looks bad
after some testing that took way too long because im slightly stupid: no
the new akarin expr frameblender just does it the wrong way around
works fine with vs-frameblender
who made the new one
idunno
okay done
-clips = [get_offset_clip(offset) for offset in range(-radius, radius + 1)]
+clips = [get_offset_clip(offset) for offset in range(radius, -radius - 1, -1)]
target/scripts/blending.py:34
2nd most smoothbrain fix i could come up with but it works
less braindead fix where i somewhat understand what im writing: ```diff
if offset > 0:
- return clip[0] * offset + clip[:-offset]
- return clip[offset:] + clip[-1] * offset
elif offset < 0:
- return clip[-offset:] + clip[-1] * (-offset)
- return clip[0] * -offset + clip[:offset]
`target/scripts/blending.py:19-22`
endofeternity, he makes vapoursynth plugins
inter esting
theres gotta be a better way to format this
id say this could be reworded a little bit because:
- hardware encoding is good for real time stuff like streaming or recording, but they are far inferior to software encoders for anything else
- H264 cqp 0 is not lossless
also yes av1 can be fast thats what the presets are for
why is h264 qp 0 not lossless
not for nvenc
Depends on what you mean by CQP
why
CQP (constant quantizer) 0 is lossless though
Iâm assuming youâre going off of something similar to this
Which is talking about CQP as in constant quality
Thatâs -cq, not -qp
And it should be mentioned that cq sucks for nvenc without qmin/max filters specified
you have to use the lossless option
the file size thing is such a "micro optimization"
if one already has a large drive
itâs not just about file size
hardware encoders just suck compared to software encoders for anything not realtime
worse quality, file size, features
yes they are fast which is good for a majority of use cases ppl have in this server, but if you truly want good quality you wonât be hardware encoding
please show comparisons in that case
i will ssim2 tmrw
you can get the same quality with both
hw enc is acceptable if u need to get a lot of content to be a bit smaller and don't feel like waiting an eternity
also saves electricity
@plain gale can vouch
ive still yet to get a source on why you can't just use qp 0
^ have used hevc nvenc to transcode about 2500 tv show episodes
it's very far from great, but it's also not bad
try it yourself
not true
can you provide some comparisons in that case
compare nvenc av1 features with svt av1
how does that matter?
honestly i was assuming
nvenc vs x264 h264 encoding
comparison in that sense
please provide actual comparisons
i am clearly a noob in this department but since you are saying the visual quality is poor with hw encoders please provide examples of it
more features doesnt mean better quality
you can get the same quality with both, its just that one will be able to do it more with less bits than the other
i have, looks lossless to me
just provide image comparison
grain synth = better quality?
no itâs just proof the hardware encoders lack
i am not buying it
yeah? im not saying they dont have less features
@crude elm is there any visual difference?
because i am very disagreeing on the quality part
or my eyeballs are broken and can't see the difference
or its a "micro" thingy
im saying that u can get the same level of visual quality with any two encoders that let you change how much quality the encoder will reduce
maybe not the exact same since they may use slightly different metrics but close enough
tldr, so no visual difference if everything is configured correctly
yeah and the hardware encs will be way less efficient at that
yeah
they are good for their speed nothing else
micro optimization again
itâs not a micro optimization
sw vs hw filesizes arent really a micro optimization
i guess it depends if someone wants it not
so preference
what did u mean by them inherently having "worse quality" then
u just agreed with me

they are less efficient at doing their job than hw encs
sure you could get the same quality but thereâs gonna be a tradeoff
visual quality too
comparison in that case again
hw encoders dont inherently produce worse visual quality than sw encoders
if you meant this then yeah thats correct, its just not a definition of "quality" ive ever seen anyone use
so itd be useful if you specified what you mean by the term
yes they do
on what basis
you can get the same quality with both, its just that one will be able to do it more with less bits than the other
exactly
color spaces?
and thats because one is more efficient than the other
" its just that one will be able to do it more with less bits than the other" wdym by less bits
not because one produces higher quality than the other
the tradeoff hardware encoders make is that they have nowhere near the features of software encoders
elaborate because i am not versed in this
if you mean efficiency when you say quality then you have to specify that
software encoders can make a video that is of visual quality X, and if you want to achieve X with a hardware encoder the video will have to be larger (ie. have more bits)
in general*
file size tradeoff
because software encoders typically have more features that they can use to get to X while using less bits (filesize)
đ
if we go off wikipedia definition
Coding efficiency is the ability to encode video at the lowest possible bit rate while maintaining a certain level of video quality
iâd say more efficient encoders will inherently produce higher quality than less efficient ones
that's at the bottom level
on the surface level its non-existent
?
"iâd say more efficient encoders will inherently produce higher quality than less efficient ones"
if the reference point is a given bitrate, yes
but usually its not, because we dont care about the bitrate that much
saying "encoder A inherently produces higher quality than encoder B" with no further context makes it sound like B cannot ever produce the same quality as A
and swapping A for B in your blur config for instance will immediately make the output look worse
which it wont, if youre using sane quality settings itll look more or less the same, only the filesize will change
if we instead used cbr or average bitrate everywhere then itd make sense to say "hw enc is worse quality than sw enc"
but we dont do that, for obvious reasons
honestly
i wouldnât be surprised if it did đ
youd have to convert the other settings accordingly too so assuming you were able to do that well enough then itd look the same
eg if we assume that x264 and h264 nvenc use the same qp scale then changing between the 2 shouldnt change how it looks
its not but thats besides the point
i donât think they use the same qp scale do they
also i swear nvenc is not lossless without the flag but i dont have ssim2 installed yet
im not sure its guaranteed that 2 identical videos will produce a perfect ssimu2 score
it might be but for eg. vmaf its not
why do you swear that tho
and it didnt look lossless?
so i could just be making myself look like an idiot
i wouldnât use how it looks to me as the deciding factor
is that a yes
looking lossless != lossless
??
did you look at the video and determine it doesnt look lossless
and based the claim on that
no
whats it based on then
ssim2
so u ran ssimu2 on an nvenc qp0 video and got an imperfect score
but got a perfect score with the lossless preset instead
well if thats the case and ssimu2 always produces a perfect score for identical videos then sure
but id wanna see a reproducible version of the test u did
ie input, methodology and results
quality and efficiency are 2 diff. things, also, av1 grainsynth is amazing but keep in mind it's definitely not 1 to 1
i'm not gonna say hardware encoders are good efficiency wise, but they're absolutely the best in the balance of speed/power efficiency/encoding efficiency
Hardware wins when it comes to performance:quality, software when it comes to quality:bitrate
I did not know this was a controversial thing though
https://media.xiph.org/video/derf/y4m/big_buck_bunny_1080p24.y4m.xz
-c:v h264_nvenc -tune lossless -c:a copy
-c:v h264_nvenc -qp 0 -c:a copy
nvenc can only do it with -tune lossless
encoding rule #1 never trust intuition when it comes to nvenc
@ocean ledge
import subprocess
from PIL import Image, ImageDraw, ImageFont
weights = [
"equal",
"vegas",
"gaussian",
"gaussian_sym",
"ascending",
"descending",
"pyramid",
]
for weighting in weights:
output = f"{weighting}.png"
subprocess.run([
"smoothie-rs",
"-i", "./swing.mkv",
"--override", f"frame blending;weighting;{weighting}",
"--peek", "31",
"-o", output,
])
image = Image.open(output)
font = ImageFont.truetype(r"C:\Windows\Fonts\arial.ttf", size=96)
draw = ImageDraw.Draw(image)
draw.text(text=weighting, font=font, xy=(30, 20), fill=(255, 255, 255))
image.save(output)
outputs like this and then u can concat those into a gif or wtv for the website
do u got the original clip
here
wait its 2mb i can just send it
smh
its like 5 sec
can u send em all imma make em into a gif
are there plans to accept my pr or not
yeah
cause these are pretty misleading as they are if so
ascending means the newest frame has highest opacity percentage
is that the case with your fix
yeah
i thought i mentioned that in the pr
i much prefer automating that in powershell
idk powershell well enough to add text to imgs
could you make it just take args[0] and do this
from PIL import Image, ImageDraw, ImageFont
from os import path
basename = path.basename(args[0]).split('.')[0]
image = Image.open(args[0])
font = ImageFont.truetype(r"C:\Windows\Fonts\arial.ttf", size=96)
draw = ImageDraw.Draw(image)
draw.text(text=weighting, font=font, xy=(30, 20), fill=(255, 255, 255))
image.save([args])
i think its not relevant to spend much time on a script like this
oh ive done this sorta stuff before
so why would i have to...... go back and change mine and re send it.. if you just... did that for me
and yes it is, to make comparisons
wdym
people can test how weights look with their own media with this
well i guess yeah but thats pretty easy to do anyway
i have 0 pwsh knowledge besides one liners so i cant remake it in powershell
ill do it in pwsh
just
the arg parse bit in python
id appreciate
ion wanna touch python for anything else
yes but they have to know what frame
which im assuming most avg users dont know how to get
trial and error?
still nice to have tho either way
yea
you can get the frame from a timestamp
this what i used to get frame mpvnet --osd-msg1='${estimated-frame-number} / ${estimated-frame-count}' '.\swing ~ Lemon.MP4'
no its literally just gonna apply text
cuz idk how to do that in PowerShell
i could also man up and do it all in python
frost has experience
or you could just do it in eff eff em peg
i just didnt know
OHY EA
i mean if u gonna be testing smoothie
u have ffmpeg
the reason i didnt is because mine has trouble getting fonts
and im lazy to figure out why
@sent very nice must try
why did i just write sent
opinions on restarting this
like clearing the repo and starting over
current docs r messy af and full of errors
idk about clearing the entire repo but maybe we could go page by page making improvements?
i don't think it's bad but i do think that formatting/clarity could be improved
we also need to fix instant loading cuz the current docs is slow when switching pages
except from that, imo the main issues rn are:
- no official smoothie installer (ik Aetopia has one but it isn't merged into the main repo or documented in CTT)
- the discord channels haven't been replaced with the docs, like originally intended
- inactivity, but obviously Couleur has his own priorities which is completely fine
big problem is color doesnât proof read his commits đĽ´
A lot of the pages are also a bit unintuitive
Itâd be a lot easier to just hand users the barebones information on a silver platter
Hereâs the link to the H264 NVENC OBS settings, copy as shown in the screenshot
screenshots are kinda annoying to maintain tho
if the text is formatted well enough, it should have the same effect
ugh the sm post script is also broken
to anyone: feel free to send a few screenshots with what's selected you deem confusing
which
the scoop one?
yes a vscode plugin that places images in the right place and prompt you for a filename when you paste an image would really really help
same as what i replied to frost a few msgs above
true but i was more thinking about UI updates and inconsistencies between updated screenshots
as in, an app's UI might update, so any screenshots would have to be remade, unlike if it was just text. when people remake the screenshots, people also take screenshots in different resolutions, Windows settings, etc
last gen? what is last gen
executable means an executable? what is an executable
âmeans beâ?
who is i? why have a page about it if you donât use it and are just gonna link to the official docs?
the only ide here is visual studio
and itâs not the only ide
most of jetbrainsâ software, arduino ide, and xcode are ides
to be fair i think the "i" would make a lot more sense if these were all discord messages sent by couleur
which is what the original point was i believe, to have the docs mirrored here more so than it being a website in its own right
i tried mirroring and it'd be a nightmare to create
thanks for the feedback june
then maybe change it since it's (or looks like it's supposed to be) documentation and not someone's personal blog
it's couleurs tweak tips
it is but you're not the only writer
oh, but thereâs multiple writers correct? (atleast thatâs what i thought, thereâs a writer role) so a first person pov would be confusing
you could, but anyone can contribute to this right? so if someone modifies something you wrote, then it wouldnât make much sense
how do other people do it
well i think it just is a bit confusing for docs that are written by multiple people to have some pages with first person pov writing
if you really do want to keep the first person pov, id recommend adding an author widget so people know who wrote what
@sinful estuary can i get le criticism
yes but then you would have a huge ass vertical list to scroll through
i think its better to spread horizontally
if u have the ability to collapse them i think it would be fine
i dont, curious to hear others take on this
btw are u not redoing the site with astro
no he3als tried it out with a fork named rewrite
which sounds like its been concluded
we havent agreed on making ctt starlight-based yet
hes just trying it out in a branch
if i were u id rly recommend starting from scratch even if u use the same framework
current wiki is a big mess
yup thats ur take on it
i mean as far as iâve heard everyone seems to agree on restarting it with proper guidelines this time
most the pages on the wiki that have been completed are now outdated or are inaccurate in places
@rugged scarab @past dew lmk ur opinions
i've gone through the wiki fixing what i found to be outdated/inaccurate #đąâ sauce-feed message
the only thing i'm not satisfied yet is the codec guide
let me know
sigh
where should cobalt.tools be linked
mozilla extension to format the last-downloaded file to markdown
the meat and potatoes in background.js:
'use strict';
let timer;
const videoExtensions = [".mp4", ".mov", ".avi", ".mkv", ".flv", ".wmv", ".webm"];
const imageExtensions = [".jpg", ".jpeg", ".png", ".gif", ".bmp", ".webp", ".tiff", ".svg", ".ico", ".heic"];
browser.browserAction.onClicked.addListener(async () => {
const downloads = await browser.downloads.search({limit: 1, orderBy: ['-startTime'], state: 'complete'});
if (downloads.length > 0) {
const path = (downloads[0].filename).replace(/^.*?\\docs\\/, "").replace(/\\/g, "/");
if (imageExtensions.some(ext => path.toLowerCase().endsWith(ext))) {
navigator.clipboard.writeText("");
browser.browserAction.setBadgeText({text: 'img'});
}
else if (videoExtensions.some(ext => path.toLowerCase().endsWith(ext))) {
navigator.clipboard.writeText("<video width=\"720\" height=\"405\" controls><source src=\"" + path + "\" type=\"video/mp4\"></video>");
browser.browserAction.setBadgeText({text: 'vid'});
} else {
navigator.clipboard.writeText(path);
browser.browserAction.setBadgeText({text: 'path'});
}
} else {
browser.browserAction.setBadgeText({text: 'fail'});
}
if (timer) {
clearTimeout(timer);
}
timer = setTimeout(() => browser.browserAction.setBadgeText({text: ''}), 1500);
});
@rugged scarab @river star fellow markders :)
what icon should I put for a page dedicated to ISO files
đżđ˝đ one of these
should I even make an installation guide for windows? theres countless of them already
atlas', revios'
dead
ayo
they rewrote from scratch
I hope this and jj can help reduce the friction that slows down ctt.cx writing and maintenance
- set up IDE and local mkdocs server
- write changes whilst constantly looking back at how the end result looks
- review them in github desktop and push,
it's quite a bit more work vs. simply editing a discord message
oh yeah I havent touched the website in 8 months đ¨
wowza!
I gotta figure out some more efficient workflow
- they claim backwards compatibility https://zensical.org/compatibility/
I am once again plugging this https://github.com/couleur-tweak-tips/CTT/pull/32
Couleur if you are writing review comments they are not visible
Afaik reviews are only visible to people with permissions to merge unless you publish them
yerss
https://squidfunk.github.io/mkdocs-material/blog/2025/11/05/zensical/
probably should've linked that first
-
yup admonitions is a p cool way to tuck info into some kind of drawer
-
yup gonna get around to fix these two lines today
