#off-topic-tech
1 messages ยท Page 108 of 1
I have a little issue
There are two tiny small objects on top of my monitor
(They look like dead pixels but they're physical objects)
Idk what they are, could be mosquito eggs or some shit
looks like burned pixels
LCD burn
Nah, particles between the LCD layers.
Or between the LCD and the outer glass/plastic.
I once literally had burned spots on my macbook screen
And they are not between anything, I can see them protruding out of the monitor
and it was common issue for that type of screen back then.
Ah, impact marks on the outer material.
Pixels got too hot or something and literally burned through everything.
And since then I had a little crater in my screen.
Yeah my original question was how to clean them
Well, probably you don't.
I am scared a wet paper towel will damage the monitor screen
Alcohol solution?
Paper towel, microcloth or like.
Dry paper towel just slides over them
With just bit of water over most alcohols.
Hmm blue alcohol would work?
Never with dry paper.
I don't think I have isopropyl At home
Except the final gentle drying.
Is blue alcohol fine
Try with just water first.
I use alcohol to clean nails
Try
Depending on alcohol, it can damage the plastic (if not glass surfaced), or the coatings.
Water doesn't.
Never had any problems with my monitors since I use it.
Um @twin dew what the actual fuck is this
A lot of those monitor special towels are just alcohol solution ๐คทโโ๏ธ
Said "depending on alcohol".
Why is it brown
I told you it's mosquito eggs
So that is just denaturated ethanol.
But again, the denaturation fluid might cause issues for some other brand.
Well you asked so we answered.
Yeah thanks, my original question was if using a wet paper towel is safe
And since baldur said it's safe I went with it
Only real possible damage with that is if you drip something into control PCB if there are buttons at the bottom of the monitor.
And that can only happen if you use way too much water really.
And most current monitors don't even have anything there.
This screen has burn mark. Litterally a little crater in the screen. Plastic/glass gone.
Happened a lot to Macbook retina screens..
Here another example.
oh many degrees : yes
That's why we pay 2K for a Macbook 
did they ever fix the FSR 3 implementation? https://www.youtube.com/watch?v=J2FMQfEiBOA
Get 30% discount with my "SKAG" code!
Windows 11 Pro ($22): https://www.gvgmall.com/gvg/SKAG1
Windows 10 Pro ($15): https://www.gvgmall.com/gvg/SKAG
Windows 10 Home ($14): https://www.gvgmall.com/gvg/SKAGH
Windows 11 Home($20): https://www.gvgmall.com/gvg/SKAG1H
Office 2019 ($46): https://www.gvgmall.com/gvg/SKAG9
Hello Dear Viewers :D
T...
i found a way to enable FSR upscaling in ark ascended and im curious to see if anyone can spot the difference between TSR and FSR both running at 58% screen percentage
The weird stuff you have to build in Mechwarrior 5: Mercenaries for specific mission type...
You are standing still. I notice (atleast in some games) the difference while moving.
Especially in SF, FSR and moving objects look so smeared (like falling leaves).
That UI is ... questionable. Doesn't look like a game for me.
in SF i find FSR at balnced to look better than TAA
Mods to give tabletop level of customization instead of the original very limited ones.
mostly because it has better image stability
Baldur, you explained once to me why in SF things look smeared for FSR.
Like the leaves falling down from the sky, they had ghost effect.
Originally you can only switch weapons, and need to be same size and type.
Anything without motion vectors will smear badly.
And in SF, lot of the GPU fakery things don't have those.
It doesn't happen for other modes.
the only thing i found in satisfactory that doesnt have motion vectors are reflections
Belt items, machine animations.
leaves...
i think the added motion vectors to those in 1.0
Pretty sure I am using Intel ExSS atm.
And it doesn't have that problem ๐คทโโ๏ธ
neither does FSR at performance
But maybe it's not a FSR problem, but just SF specific.
SF is updating, I check after it.
or at least i cant see it
here is XESS VS FSR at 1080p performance (compression might have ruined it)
it ABSULUTLY did
even though im recording at 10000 Kbps
weirdly, i get much more ghosting on xess than fsr
thats honestly kinda cool
though i had to disable antialiasing and enable fsr2 for that, which is bonkers to do
in SF i only see ghosting with FG
streamable.com/qle0d5 That's a video. FSR / No downscalling.
unless you have 1-wire setup and you just need to fuck around and find out
where are you because i cant find those very much if at all
in game of course
new save in the grasslands
especially at the end of the video
Arrow is velocity
Look how smeared that is.
they do look blury but im not sure at all that XESS would do a better job
I was mistaken. I use TSR ๐
Ooooh
TSR THE WORST upscaler
Well, it's the one that doesn't do that.
And I have good FPS. No Downscalling.
Not the best FPS though, I noticed that. But I am happy with how the game looks with it.
im going to chect that out myself because i almost never see anything that ghosts so much
especially with FSR
only if you have other options
i went there and i did see it with 60% but not with 100%
mhm, well you can clearly see it in the video.
and I just took screenshot of my settings, so not sure what's going on ๐คทโโ๏ธ
wait no, i was using XESS
๐คฆโโ๏ธ
like an absolute dumb***
now that im actually using FSR i can see it
but i wouldnt call that something that makes me want to use XESS and give up on FG
I was mistaken, like I said, I use TSR.
And I don't want to blame FSR either. Can be just SF specific. But I noticed it and it annoyed me so I switched.
the only thing i found that made xess look better was its fine detail reconsrtuction (1st FSR 2nd XESS)
both at 50%
i think ive spent like hours looking for things that one does better than the other
but almost every time i find something that looks bad on one of them i then go to the other to see if it looks just as bad and often i keep second guessing myself in terms of which looks better
different time of day W FSR at 100%
and what should i be looking at?
The shadows look fine.
The details look fine. I don't understand why it's a bad upscaler. It sure looks better than FSR.
turn it down to 60% and look at decently far away foliage (AKA:trees)
oh, and disable lumen
They both look ok. But FSR looks weird with moving particles, like I showed.
so does TSR
Check, the little flies: https:// streamable.com/tdqaz0
didnt work
there is a space in there, else the bot gets angry
got it
and just checked, TSR doesn't have it that badly.
I don't know, it's like black smears
but FSR does have more ghosting here
Wait ... what ๐ฎ The bot sleeping? Or it's whitelisted. Good to know.
But yeah, that video is TSR.
you just need to remove the https:// portion, lol
can be copied and pasted into the browser without issues and the bots don't detect it either
You can clearly see the little details better. In my opinion.
The little butterflies and whatnot.
Yeah I know, but I didn't know streamable was whitelisted.
i have tried it myself and seen it but for me that still doesnt out weight the extreme flickering on the trees
I don't have flickering at 100%. But I can see it being an issue.
at 100% the only problems you can see about an upscaler is ghosting
or bugs
like this one https://youtu.be/to2MuvbZ_zs
a bug that occures when using FSR with screen space reflections that makes the reflections extreemly grainy
that's not what I'm referring to anyway.
by splitting the two portions of the link, it was no longer copy-paste-able (hence why you needed to give that extra piece of clarification)
had you just removed the https:// part without putting a space in between, this wouldn't've happened
Oh yeah, I do that. But this time I was lazy ๐
I though they would notice
for which i also have a fix for https://youtu.be/Gw-MumKIbnw
this is what you have to insert into the engie.INI file:
[/Script/FFXFSR3Settings.FFXFSR3Settings]
r.FidelityFX.FSR3.UseSSRExperimentalDenoiser=true
r.FidelityFX.FSR3.CreateReactiveMask=true
i do when looking at these trees
I haven't noticed that.
But yeah, just play with what you feel comfortable with. That's why there are options. But most upscalers have drawbacks.
true
but i will end it with this comparison: https://youtu.be/8rLtskFJcm4
youtube compression makes it useless though
got a fun one: whenever i let my monitor go to sleep, but every time it wakes back up AMD Adrenaline is open
Oh, my pc was doing that every time the pc itself went to sleep
I don't remember what the fix was
This is a known bug for them.
good to know, thought it might've been my system being weird
Rambler doing video that is only little over 7 minutes long!
https://www.youtube.com/watch?v=qY0kEB-1MIc
Patreon: https://www.patreon.com/buildzoid
AHOC Shirts: https://actually-hardcore-overclocking.creator-spring.com/
Bandcamp: https://machineforscreams.bandcamp.com/music
My FPS game: https://buildzoid.itch.io/hv-shooter
Twitch: https://www.twitch.tv/buildzoid
Twitter: https://twitter.com/Buildzoid1
Facebook: https://www.facebook.com/actuallyhar...
He also expected that the two dumbasses were installing the CPUs with MB already installed in a case and case upright...
I've never even thought of mounting a CPU while the mb is vertical
Like, you'd start thinking "hmmm, the CPU doesnt stay in place, lemme rotate my case so gravity works with me" after a bit, right?
just an avarage day playing ark ascended....
I did it maybe twice.... once with an old 80386 (they still had very sturdy and not too many pins on the underside, and no cooling system to attach so fractions of mm could kill something because of the amount of pressure the cooler applies.... and then with a, I think Pentium II or Pentium III -- there were 1-2 generations where the CPUs came on or were little logic boards that went into the mainboard...
yes, I'm old ๐
Yes, those were the times: https://en.wikipedia.org/wiki/Pentium_II
or at least maintain the cpu in place permanently by grabbing it from the frame hole
the code that keeps ark together is amazing
first time i got the hotspot temps over 90c
I wish there wasa built into OS (UI) way to concatenate contents of files with same name from different folders in bulk
WTF
one day I should really make a software with detailed synthetic performance mesurement
so one could take the data and use an application profile to accurately guess the app specific performance / scaling
Heh
So someone is planning to use old system for dedi, and someone else then tries to get him to buy crappy new system?
And CPU performance both doesn't and does matter
Which is very interesting
And if you wanted to make new system for SF Dedi right now, 5700X3D would probably be the cheap option...
Which would run circles around that 13500 for that.
The "newer is always better" rhetoric with hardware amazes me
Because often the old hardware you have is better than the new hardware you don't
Baldr please dont make me regret the forward, I know for same reasons as for twitter quote it's stupid how it pulls it out of context
but so here the original topic was on selecting the right OLD system cheap stuff second hand enteprise ewaste to get
and these only use intel shit
Ok
But somehow I don't expect 13th gen Intel to be in the enterprise ewaste category yet either.
Because usually that is 3 years old and older.
yepp, pats his i7-9700KF livingroom vr pc
yea i think there probably still going to last even if not as long
my 13700K hassent shown a single sign of wear and i havent been on the default preset for the entire past year turns out
I always make my then "old" gaming PC the new home-server, that way I get about 8-10 years out of the hardware. Only the past iteration wasn't like that because I took my work PC with me when being let go 5 years ago...
No, because the normal way for 3 year old systems to end up from companies to used market is because the warrant & support often runs out at that 3 year mark.
Anything newer isn't usually being sold out or otherwise gotten rid off.
So perfectly working systems are replaced to keep full warranty & support running all the time.
honestly i dont see a problem with that if the deals are good and they arent absolutely totaled by the end of the 3 years
But point was, 13500 isn't old enough to be in that 3 year or older category.
So there shouldn't be much being sold.
Only from possible bankrupsies etc.
Launch Date Q1'23
you can be lucky and get an almost brand new "defective" laptop, where the user complained the webcam is kaputt while they actually had used a privacy sticker. We had that TWICE within a few months at my job... and it's not a large company
but I'm glad they're not using refurbished former leasing hardware anymore... that was... funky...
Dang Baldur new role
anyone heard of a way to cluster timeseries?
for context: some way to group together signals that go like bottom ones, regardless when they began, and the top ones (rising, or rather, not falling in this particular case). curious how many similar behvaiours are there
Elitebook 845 g7 for like 300usd a couple of years ago. Ryzen 7 4750u + 16gb ddr4 3200
Quite happy about it
Meanwhile I bought a thinkpad p53 with 9750h and rtx5000
With a bad screen and trackpad which I can fix for $75
not bad not bad
the thinkpad cost me like 480 too
oooo really not bad
ok i thought that the rtx 5000 was ampere for some reason, still good tho
this was scary ๐
and saddest thing - i have to do it again since i forgot 1 tiny important column
honestly wonder why pandas takes so much space while CSV export makes just about 9GB file
gota check their 'lowe memory' feature
Turns out there were 100 nuclear explosions of magnitude about 6 (i say it like that because it's a bit weird how data is filtered...). there was a pause from 1995, but last one is from 2017 - it was registered by 2000 seismic stations
oh it's just for reading CSV. sad
the 2017 one was NK i assume
exactly
A customer came in with a bent CPU and chipped substrate in one corner. Asked him how he got that much damage, he told me he forgot to put the processor while the motherboard was out of the casing and didn't want to take it out because it was a pain installing it vertically with the IO shield... Asked him why he didn't tilt the casing horizontally, told me he didn't want to scratch the casing. 1000$CAD damage not to scratch a 70$ casing. People are stupid. Good analysis.
And lot of other people also telling about their fuckups in installing CPU while case was vertical in that BuildZoid video I posted yesterday...
I've been guilty of installing a CPU vertically before, actually to be more precise, I had to change my AIO, and in the process involve having to remove the standard Intel locking mechanism, opted to be lazy and not take out the whole motherboard, I tried doing it vertically with the MB still in the case. I noticed an issue with my GPU after the install as it wasn't running at PCIE 4.0 x 16, but instead it was doing it at x 4, after days of troubleshooting, ended up finding out that I bent my CPU corner right where it was responsible for PCIE lanes, because the CPU likely moved as I was mounting it vertically. Damaged pins and CPU, had to replace both the MB and the CPU.
yea, there's basically nothing that saves you time or money when moving a system to do a horizontal cpu swapout
hell, it'd be cheaper to replace an entire CLC than the cpu
i vertically installed on lga775 once, do not recommend
Good thing you could just swap the outside parts to the latest case as the actual frame is still the same ๐คฃ
oh i think it's dead dead
but yes, pretty sure it's the same chassis as every other
there's always an option to put pc case on something soft, like sofa or bed. noscratches.
Or just towel or like on desk etc.
also never fully assemble case before first boot
oh hell no ๐คฃ
asking for trouble
the only serious mistake i've made when assembling my systems was when i completely forgot to plug in CPU power & my mobo yelled at me
that stupid plug really likes to hide in the darkness
learned that the hard way, too.... the first mainboard where you had to screw a counter-plate to the bottom for mounting the fan... and the case of course didn't have a hole in that place. noticed after mounting and attaching the hard to reach places with cables etc.
all of these results had atleast 1 autosave during them so they have to be retested
Starship IFT-6 planned for next week.
No earlier than Monday morning local time, correct?
Slipped one day very recently.
So Tuesday 2024-11-19 22:00 UTC as the current targeted opening of 30 minute window.
https://www.spacex.com/launches/mission/?missionId=starship-flight-6
๐
every thing has been retested!
jesus fuckin christ
how's raptor lake with a regular SSD faster than arrow lake with optane
ask intel
they surely know
you would hope
i actually find these very interesting:
lumen off 100% resolution scale
TAA 77
FXAA 77
NONE 78
TSR 63
FSR 69
XESS 67
upscaling performance at 75% (TAA for comparison)
TAA 89
TSR 79
FSR 83
XESS 78 (thats surprising)
How do you test different upscaling methods objectively?
Because, you can measure FPS gain objectively. But you cannot measure "how it looks". That's more opinion based I think? And different upscaling methods look different. So how can we test these methods?
you dont... the end
best you can do is create side by sides with ICAT and then grab a magnifying glass
But what about unexpected behavior, like the flickering trees.
mhm, though so.
depends on your definition of unexpected behavoiur
Well, TSR, the trees flicker at 60% in the distance. Other upscalers don't have that. So to me, that's unexpected behavior.
so things that one upscaler does but the others dont
What's the proverb? Comparing apples to oranges?
you guys are riding on the red line of this being satis talk and not off topic
mh, that's because we had previous talk about it. But the fact remains, different upscalers, give different visual results.
People telling me that upscaler is worse than another made me think about it.
how would this be satis talk specificly?
Like, only thing you can say, that one upscaler gives more frames.
but at what cost.
because you are using data colected from the game
but either way im interested in the results
results?
well id love to see a comparion of each one
lightning put in some work comparing them
but that was very quick with screenshots.
but that was purely visually. So without FPS gain in mind.
Fps gains right here
This video set it all off ๐
Me too, hence my question how do you do it properly.
no one responded to that one
id say zoom in on something from a decent distance away with each one
Measure frames per second. And then take different visual elements and compare them; shadow, object far away, moving objects, edges, etc.
been there done that
and then for a moving one maybe stand on a belt asuming a SF context
Well, it got me interested in the topic ๐
Also, how much impact do games have on the results of upscalers?
I mean, I assume it has to be implemented by the developers of the game. How much can do optimize?
probably a decent impact
Because maybe one upscaler is really good in a particular game, but sucks in another.
because diffrent contexes will allow something to work better
like frame generation would work like hell on something that is consistently very similar frame to frame
judging from the amount of console commands available for FSR in SF and ark
not that much
console commands isn't the same as the actual code of implementing upscalers.
and the fact cyberpunk added FSR 3 clearly means it needs some implementation work
well, I am not so sure about that.
fsr2, xess and dlss2 are all forms of temporal AA; they all need the same inputs and are interchangeable
like motion vectors and a depth buffer
So, motion vectors and depth buffer is provided by developers?
yis
but once you support one, you support them all
this is also why it's a pain to mod them into games that do not already support one
and what about FSR 3?
frame generation needs more data, hence FSR3 and DLSS frame gen not being drop-in
but the upscaling part of FSR3 is identical to FSR2
using mods you can literally swap a .dll file in a game that supports DLSS to make it use FSR2 instead
and vice versa
very handy for developers.
microsoft is standardising this with directSR, an API that developers target, where windows then chooses the best upscaler
so; game supports directSR, your PC can do dlss, it runs the latest dlss provided by your graphics driver
one day we won't need developers to lift a finger
it doesn't have frame generation in it yet but that'll be easy
wonder how much more they can improve it
improve what?
upscalers/frame generation
well FSR can definitly get a BIG upgrade (AI)
DLSS: i dont know if there is room for improvment there
FSR4 is using ML
i think DLSS has peaked already, there's only so much information to be extracted
ye thats what i was eluding to
DLSS ray reconstruction looks shit ass honestly, it hallucinates detail like chatgpt
i cant try that
reflections added that aren't already there, detail smushed out of existence
but i heard good things about it
it changes the image
Looks like RX 6800 non-XT deals have finally dried up in the USA, that thing was an absolute steal when it was around $340.
shame
Looks good to me.
340??!!?!?!?!??!?!
nah
nvidia's showcase they are going to make it look the best it can
what
cup babied out of life
where did my copper wire go???????
i cant even build my power poles no more maan๐ญ
oof, you right watching another video and some weird stuffis happening.
A walkthrough in "Hogwarts Legacy" with maxed out settings, Raytracing set to "Ultra", DLSS Quality + Frame Generation and improved visuals by "Ultra Plus - Ray Reconstruction - Performance and Hotfixes". ๐ฎ
Ultra+ Mod - Now with Ray Reconstruction, and function keys for Lighting Presets and toggle other features, like f.e. fog or different glob...
Huh. PCP just tripping, Amazon has the 6800 for $360.
it invents super shiny reflections that should not be there
it's straight up AI hallucination bs
at the lake, you can see the lake just switching out differently
it has some sort of fog, and that just dissapears instantly. Nvm, they disable the fog ๐
Still these reflections don't look good.
it's like there is also a fog on the floor.
that just looks like the game is trash
im scrolling trough the video and most of this dosent even seem to be primarely raytracings fault
keep in mind that version of RR is modded
a few moments later it looks better:
really zoom in on this, it looks like AI generated images with the weird patterns out of nowhere
This stuff?
When it's off.
i grabbed some footage of each upscaler in SF at 60% and im going to put them into ICAT and im going to look for flaws
even at potato quality you can see the ghosting
well thats why its called potato quality
maaan i would love it if AMD encoders didnt SUCK
use av1
i cant
much of the reflection on the left of the puddle is entirely wrong
ah heck
RX 6000
doesnt have AV1
well this way the only 1 comparison that is usable is the stationary one
you could use CPU encoding
cpus can encode btw
or jack the bitrate up
its already at 12K
lmao
or atleast AV1 does
try h264 cpu
That is true. I guess it's WIP
it's over a year old
a well done cubemap would look better
I think the idea is, that it's low effort for developer.
the ML does the work for the developer... but at what cost.
instead of denoising raytracing then denoising TAA in DLSS, it's meant to denoise both in one step
but it clearly hallucinates detail
it made up a whole building in the reflection
they're asking the ML to do too much
with a ๐ด sticking out of the building.
lmao i didn't see that
I didn't notice it as first too.
But I mean, you really have to pay attention for that no?
i'll try x264
so i guess if you don't pay attention to the picture it's great
it has room for improvement for sure
OH GOD ITS SO MUCH WORSE
i fear a 5600g isn't strong enough
lets try with 12K bitrate
dont worry it can handle it
maybe*
its better now
and it did handle it
the palm tree also is the wrong color in the reflection
yeah, the more I look at it, the more I notice
but im not satisfied yet.. how about 22K bitrate?
and the sky reflection seems a bit bright considering that the brighter horizon is caused by artificial lighting, the actual sky should be pretty dark
its just not good
i think that is just the ground being un ailiased
lets try this
for what i need it for there can be NO compression artifacts
then you need to use CPU encode in smth like DNxHR
even nvenc has compression artifacts
if you're using OBS doesn't that have a lossless mode
finally i think that it looks as good as what i see in game
doesn't discord compress?
probably does
use a video editor to nearest neighbour upscale it, zooming in on an area that shows off any artifacts you want to show off; this will help it endure the discord compression
at first im just going to put them in to ICAT to campare them
1st FSR loses to XESS in terms of pixelation
don't you need the exact same image?
I find it hard to compare those two side by side
they both have grass
and were stopped
yeah, but first one looks like camera is closer to the grass
I am nitpicking, but if you want to be precise ๐
It does look like first one has more pixels.
Power cables between power towers would be interesting. They look awful in motion with DLSS, even when they're contrasting only against static sky.
these 2 are the same distance from the camera
FSR almost makes them disapaer at range
I'd rather they were just gone altogether than the flicker in and out I see with DLSS.
what %?
for me its 60% for all my tests
Out and about, can't recall off the top of my head, but toggling performance/quality didn't seem to make a different specifically with cables.
this is what they look like W FSR
but i will have better tests for this later
that waterfall in the background ๐
thats just how the game is
next TSR VS FSR pixelation test
Pretty bad in motion too I'm guessing? Screenshots only tell half the story, even when they're good/bad.
But like here, it's way harder to spot the difference.
yeah
TSR looks extremly blurry here
but less pixelated
which looks better here
If Blackwell and RDNA4 end up being on TSMC N4, we all need to brace for screenshots from Nvidia and AMD showing off great upscaled performance... in cherry picked screenshots. And no 1% low or latency info.
Because that's the only way they'll be able to portray big gains. Unless die sizes go way up.
i vote FSR (its sharper)
marketing 
bigger number better
or put fancy acronyms on the box.
next XESS VS TSR
Coming up at CES, the next iteration of the Lenovo Legion 7 laptop: the Legion 7 AI X Pro XT Max GTA 6 Nvidia Geforce RTX Edition.
have my money
you had me at AI.
XESS has better clearity in motion and is less blury (IMO)
Honestly, I would love for companies to paper launch special editions like that just to hear tech press say it.
next FSR VS XESS image stability test
its hard to show but XESS has some instability with the trees and the the water extractors in the back
and FSR is sharper
FSR has pixelated rocks oof.
they are oversharpened
for sure
FSR's sharpening pass is known to be more agressive than XESS
heres the same scene but with FSR's sharpening pass lessened
but i forgot to turn the fog off
๐ญ
FSR sharpening set to 0 (but it has fog)
next TSR VS XESS
i always noticed XESS had shitass stability with lumen on
i noticed the opposite
oh tsr is not bringing its best
TSR also has issues with the trees an water extractors
also has this weird effect (the random pixels in the sky)
the fuck
like XESS also has issues with these trees
is 60% considered balanced?
just more
or peroformance??
in SF it is performance
depends who you ask
Yeah, TSR doesn't look good at all at 60%.
other games:66% is quality 58%balanced 50%performance and i think ultra performance is 33%
btw, another good game to benchmark is RD2.
I had some fun finding the right upscaling in that game ๐
Fair, it's a good game too.
Maybe for steam sales in a month? If you really want it ๐
Anyway, it's used a lot for benchmarking.
im not realy into games like that
TSR still with these random pixels maaam
same here
its meant to be a stress test to see all upscalers at their worst
Yeah, fair, but still.
I don't know which I want when I look at both of them.
How noticable are the pixels zoomed out?
extremly
Oh hell no.
it's really sharp though
you can turn the sharpening pass off
but out of the box it is more sharp than xess
this is how r.fidelityFX.FSR3.sharpness 0
XESS also has more brighter spots.
you mean like on the pipe?
could just be time of day
i personaly like the sharpnes
moving on
next is the shimmer test (XESS VS FSR)
accidentally selected TSR
so in terms of ghosting XESS wins by a mile (or i dont even know what artifact im looking at with FSR)
here too
these trees are more stabil with FSR
in fact everything is more stabil with FSR
next TSR VS FSR
I mean at some point we're going into the territory of pixel peeping
this is all at motion isn't it?
the newer stuff is in motion
yeah so isn't it mostly irrelevant to actual gaming?
I mean yeah for cs2 and stuff like that sure but you're not playing that with FSR/XESS/TSR
do you mean that because this is in motion it isnt relevent to actual gaming?
half/half
The thing is, pixels are the difference. Look at those two.
maybe
From what i've seen, TSR is the weakest at 60%.
XESS and FSR look alike. So I guess it's up to FPS.
especially because of its halucinations
though from what ive seen FSR is sharper than TSR but has more ghosting (in motion)
I cannot handle the ghosting of FSR.
your preference
And I mean that in this specific game.
Don't want to say FSR has the same issue in other games.
TSR VS XESS is next
I was watching other games that did the same comparison. And I didn't notice much ghosting.
FSR 3.1 has more ghosting than previous versions but has more image stability
you were probably playing a game with FSR 2.2 or FSR 3.0
FSR runs better
yeah saw it
DLSS is the upscaling king for a reason
im going to grab some 100% and maybe 75% footage
also that must be a prety old video
2 years
tbf, the trustworthiness of that source is highly questionable.
fwiw, it could be another UBM guy trying to manipulate supposedly fair comparisons against AMD
i've posted selfmade videos showing ghosting with FSR.
So, I take the video more serious since I experienced it myself.
i can also confirm that FSR has ghosting present in certain scenarious
Same but oh no a little bit of ghosting for 50fps more ๐ฑ
im going to have TAA in for 100% res as well
we are aware of that
but I rather have not the extra FPS and no ghosting. win-win.
I assume most people are using the quality purest for a few more fps. Ghosting is a non issue there
other upscalers don't have it as bad
oooooof TAA and that power line are not friends
and you still save more frames. Not more than FSR, but more than no upscaling.
what powerline? ๐
OMG TSR halucinates at 100%too
I don't notice it.
FSR and XESS are perfectly stabil everywhere here
maybe has something to do with lumen
or the rare AMD driver bug
TAA and TSR arent stabil in these places:
TSR is better than TAA though
at 75% FSR is still perfectly stabil
XESS is slightly unstabil here
and here
big screenshot dump
also...
in this particular frame TSR was faster than XESS
here is what each of them ran like in the 60% setting
imo xess wins here
Faced with some clearance issues, we needed a printed circuit board, but with no electronics experience, how would we go about that? Well, we figured we might just be able to print our own! Here's what went into making the PCB in Episode 39 of Project Binky, and why it took so long to get that episode to you!
Find us on Twitter @badobsessionms...
The Mad
How To Not Get Self-Designed PCB
PCB prototype production services exist for a reason.
Ryzens?
Pretty old at this point to get.
If you already own, depending on use, might be good idea to get 5700X3D to replace.
(Yes, I know you probably meant Nvidia GeForces, but you didn't specify anything specific ๐)
Sandy Bridge ๐ถโ๐ซ๏ธ
say.
your.
name.
Why would I?
i think combination of 3 phrases is just proper good
I like this idea but any normal scammer will immediately notice he is talking to AI https://youtu.be/RV_SdCfZ-0s?si=VTCbybrdtofGmp2y
O2 has created human-like Daisy, an AI โGrannyโ to answer calls in real time from fraudsters, keeping them on the phone and away from customers for as long as possible.
With Daisy revealing how youโre not always speaking to the person you think you are, O2 is urging customers to remain vigilant, leave โscambaitingโ to the AI experts and report...
Also, scammers use different numbers and sometimes clone numbers so I am not really sure if this will help
Go to https://surfshark.com/redshirtjeff for 4 extra months of Surfshark at an unbeatable price!
The Pi 5 can now game at 4K Ultra, with Ray Tracing: ON.
Check out my blog post for the complete guide to setting up your own AMD 7000-series graphics card on a Pi:
https://www.jeffgeerling.com/blog/2024/amd-radeon-pro-w7700-running-on-raspberry-p...
Playing Doom Eternal in 4K on raspberry Pi 
i wish there were more radeon pro laptops
thats long fixed, no?
should be
i decided to compare FSR to TSR in ark ascended (left side is TSR right side is FSR)
and both are at 60%
mission imposssible:
UNNOTICABLE COMING
here is the overall image
cannot notice anything
we have to cook|| cat gun pfp images ||
in the overall it IS difficult to see but when you play the game you can see TSR flickering all over the place
oh
also when i start moving TSR applies a soft blur
(TSR is on the right this time)
for some reason FSR just isnt rendering those little peices of grass
and they both have this weird blocky thing around the trees
I guess they get lost in the sand
here is the overall
Seems like a texture issue then
FSR is more flickery in motion (this doesnt demonstrate it very well)
the TSR blur is strong with this one
FSR has the tieniest amount more ghosting
and i think they have about the same amount of pixelation here
the water looks pixelated with FSR
at 100% they look extremly similar
except for the place where FSR doesnt render this little bit of foliage
and FSR makes the water look pixelated
but TSR still has more flickering
once again the TSR blur is real
but so is the FSR pixelation
the timing was a bit mis aligned (so i took 2 screenshots)
TSR still with its blur
and then FSR is a little pixelated
HUB continues to talk about CPU limited gaming while sticking to the same limited subset of games, only throwing in a quoted GN Tweet about Stellaris. ๐คฆ
are 3000 series good
or are 2000 series good
just wana know the comparison
(nvidia geforces)
Yes.
Ask better questions, get better answers.
Any given 2000 or 3000 series card will run any game. Some better or worse than others.
you make google sound informiatve
none of them because none of them are going to fit in your $0.50 budget
Example of a better question: what's the best card for [budget] in [region] if I play [list of example games].
a 2080 ti is faster than a 3050, therefore 2000 better
thought really you should get a radeon 7600 because it's a bigger number
Xs make everything faster. This is why the XFX RX 7900 XTX is the fastes GPU to date
ah yes of course, but then we gotta bow down to the XFX RX 580 GTS XXX
thats my favorite gpu
I will be amazed if next gen GPUs launch without some sort of AI branding.
honestly, i wait more for when AMD enters market of server AI-training units
(well technically they have stuff like MI200, but i didn't hear anyone being amased by it)
The pointless AI naming in Ryzen mobile is hopefully just an aberration, but I think it's more likely to spread.
i suspect we'll see it until there's some seriously bad press around the term, then it'll stop
Well, not that pointless right now.
As Win11 24H2 has some features that are only available if your CPU has 40 TOPS or higher performance NPU.
And those specific CPUs are the only AMD ones that fulfill the requirement.
2070 = 3060
3060 Ti < 2080 Ti < 3070 โ Titan RTX
This is using 3dcenter's GPU performance comparison chart for 1080p
If you want a better answer, you gotta be more specific with your questions.
"Is XYZ good" is a pointless question to ask without giving any info on what you consider good or bad
Aristotle discusses the Forms of Good in critical terms several times in both of his major surviving ethical works, the Eudemian and Nicomachean Ethics. Aristotle argues that Plato's Form of the Good does not apply to the physical world, for Plato does not assign "goodness" to anything in the existing world. Because Plato's Form of the Good does not explain events in the physical world, humans have no reason to believe that the Form of the Good exists and the Form of the Good thereby, is irrelevant to human ethics.
I would guess we will see more of the NPU stuff we already have in Smartphone processors in "normal CPUs"...
Microsoft mandated that computer/laptop must have 40+ TOPS NPU to be able to be "CoPilot+", and to be able to use all the 24H2 features on the "AI" front.
Baldur, I have a port on my switch, that remains powered on after turning my PC off. I though it was magic package, but I disabled it on my PC side. What else can it be? My NAS ports on the switch and my macbook port don't have this issue.
Because your PC keeps its LAN powered with MB standby power.
Most MB don't even have setting to disable that in BIOS, it just is.
No!
Those magic packet etc. settings are just what is allowed to wake the PC over that limited link.
PC no power = switch port powered on.
I disable magic package in BIOS. And Windows.
And if you check, you will see that the PCs LAN port lights are also on while the PC is off.
even if MB doesn't have power, the port is on low power mode?
MB always has power, when the PSU has power.
Like I take out the power chord of my PC, and on my switch the port is still active.
5V standby line.
Ok, that is then weird.
If it still has link while the PSU is fully off.
What happens if you change the network cord to another switch port?
Let me check.
Might stay on even without LAN cord, which would tell that the control for that LED on the switch side is broken.
Those LEDs are software controlled by the switches microcontroller/CPU.
And then that causes a LED driver to change the state.
And fault in either that LED driver, or the CPU code or like etc. can cause one to get stuck always on.
So, it also happens to other ports. But if my NAS switches off, those ports go out. And same for macbook pro. It's only with my PC it's happening atm.
I took out the chord of my PC for 20 seconds again, the port remains orange.
I will check switch side to see what's going on.
Normally that "turn PC off" means just shutting it down, but keeping the PSU connected, and in that case the MB still has that 5V standby power running to it.
You only specified fully that the PSU input is disabled much later.
Do the PC side LAN port light also stay on?
PC no power = switch port powered on.
No
Try turning the computer on with PSU disconnected.
The MB and PSU capacitors can hold charge for long time.
Because the PC has no power ๐
That was inconclusive in my head, because for most consumers, that would still have PSU on with power cable connected.
Because most just don't know that PC off isn't power disconnected.
That try to turn the computer on without PSU getting power will discharge all the stuff fast, to real off state for the MB.
Otherwise I have seen MB status lights staying on for over a minute after removing power cord while in turned off mode.
still powered on, orange. even after holding button to discharge
out of curiosity I switched the lan cable from port while PC was fully turned off. And then the light doesn't go on.
And that changed between switch ports?
And disappears when you disconnect the LAN cable from the computer?
What switch
I think my switch keeps the port active for some reason, netgear S305 if I remember correctly.
So that would point to a firmware bug in the switch.
That it doesn't properly detect the loss of link with the PC and turn the LED off.
GS308T sorry
Until the cable is disconnected physically.
Ok, will see if I can reset switch or update it.
If it was something on the PC end, the LED would turn back on once the cable is reconnected.
I noticed it last night when I went to bed, didn't trust it, so I pulled out cable to be sure ๐
Yeah, and the lan port would also have power on the PC. I would assume.
Unless power can travel through the PC lan port, without going through the led, which I don't think is possible.
There is no power in normal LAN links.
And the LEDs are not connected to the actual link pins at all, they are purely controlled by the LAN chips on either ends via separate pins.
Have to go for now. Will keep you posted. Thanks for input Baldur ๐
Lame.
Got to love when tools to assist just fail because of shitty implementation.
Use jira to create branch, branch name too long. Refuses to truncate name...
Ok, so I am tweaking around in my switch interface. I enabled EEE mode and Auto Power Down Mode for all my ports to test it out.
Now if I power of my PC (take button back at the PC) the port switches off.
If I turn back power to PSU (back button at PC), the port switches on. So now it's clearly something my PC is keep alive and I don't know why.
I disabled magic packet and PCIE in bios.
PCIE Device Power -> Disabled
Windows side:
Overlooked one.
That was it.
I didn't expect windows to take priority over BIOS. But I think it's a race condition, Windows shuts down, tells switch to standby for package for WOL, switch receives and keeps port active.
Then PC turns off, motherboard turns off and there is no WOL active at my PC side, but my switch doesn't know that, since , my PC told my switch there is WOL (Windows).
Yeah, on Windows side there is that one toggle on many of the LAN cards that puts the basic tech on or off.
And then various "Wake on ..." options to select which methods are supported.
Then another under that Power Management tab if Windows will react to wake signal from the card or not.
But anything under that Advanced tab is LAN driver specific.
None of it is standard Windows functionality, but drivers options that it exposes.
Damn... Didn't check the dashcams old manual, it only support 128GB SD cards as max size and I got 256GB one to replace old 64GB.
Because that 128GB size max is just weird, but it was real.
The camera correctly formatted the card, but cannot use it and just asks to format it again and again.
happens to the best of us
what kind of black magic that is?
Still, didn't expect this behavior. I assumed turning it off in BIOS takes priority over anything driver/OS side.
That specific setting in BIOS based on the name only manages if UEFI looks for that Magic Packet and starts the computer on it.
Even if OS hasn't done anything.
But it doesn't control the full wake on lan subsystem at all.
In UEFI land the OS can influence BIOS settings.
But time to go see if doing just 128GB FAT32 partition onto the 256GB card will allow the dashcam to use it.
i honestly hate that Windows changes clock in bios every dam time
Dualboot? Tell one OS to not do it
Both Linux and Windows do that.
But Linux defaults to having the HW clock in UTC, and Windows in local time.
But both allow configuring to the other.
[HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\TimeZoneInformation]
"RealTimeIsUniversal"=dword:00000001
Yep...
PCIE Devices Power On
[Enabled] Select this item to allow the system to be waked up by a PCIE device and
enable wake on LAN.
source: download.asrock.com/Manual/Software/AMD B650/Software_BIOS Setup Guide_English.pdf
That's all information the manual gives ๐ฆ
Windows does not tell the switch anything. It tells the NIC to stay awake for WoL. The switch just sees that something on the other end is still there and keeps the port active
ok turn out i should hate Ubuntu for doing that automatically. but i suport having clock in UTC, so i will choose to contniue disliking Windows for having me to re-set it's clock every time
No. That was NOT the case in the beginning.
I could fully power of my PC, chord and everything.
The switch port stayed active. That might have been a bug.
In this case the Switch thought it had another end when the NIC was completely powered down too.
With power cord removed from the computer.
But seems that the link light in this case works by just the NIC saying "I'm going down, but will still listen, keep the light on" or something like that.
That's what I was saying...
Yeah, just rewriting.
Thanks ๐
Normally that's the case. Exceptions, like in your case, exist.
I have a mini PC that won't power down the NIC period. (As long as it has power)
I enabled EEE mode and Auto Power Down Mode for troubleshooting (switch-side), I put them back to default and it all works perfectly now. So it was the toggle "Shutdown Wake-On-Lane" I missed causing the confussion.
Maybe I am too paranoid, I don't know. Don't like ports enabled when devices powered off ๐
But didn't work.
The camera detected the partition size correctly, but still wanted to format.
Might be doable with some extra tweaking, but I'll just use the 256GB card in Pi for now.
And have two 64GB cards in the car, and get 128GB card later.
Got this set in 2018 with GPS and hardwiring addons: https://dashboardcamerareviews.com/mini-0906/
Would need to look at replacement at some point.
A stealthy, 2x1080p front-and-rear dash cam with above average video quality, at an accessible price. Comes with GPS, parking mode, capacitor, CPL filter.
About 100e for full set back then.
But not for sale anymore.
The thing is, I once asked a local police officer if I can use my dashcam to report infrictions. And they told me, it had little legal ground for police cases.
Maybe it has more value for court cases.
Because the main reason I want it, is because if some dumbo crashes into my car, I have more proof I did nothing wrong.
Wanted bigger card, as that 64GB one can only store about 4 hours worth for both cameras at best quality.
Yes.
Was mainly about being able to get specific part of long trip after the trip has ended.
Where nothing real happened but wanted something smaller.
And didn't want to stop to mark specific part as something not to be overwritten.
So, you want it more as footage of the trip itself?
Yes.
The point of the second card (had crap 16GB card before as backup) is so you can switch the card if something happened.
Don't people use a Go-Pro for that because of the image quality?
Just for something that happened during.
Not as expected before starting the trip recording.
But wasn't crash or like, where the recording would actually be needed.
And wanted after the fact recheck what had happened, but couldn't as the recording was already overwritten in one case at least etc.
Don't currently use the parking guard, as that still uses almost 0.5W of power, almost same power usage as full recording.
Even if only set to take one frame per 15 seconds always, or full speed with motion detection, or both.
The hardwiring kit has battery low monitor and cutoff, but if on, will get to that level in few days max, depending on what voltage level you set as the cutoff.
Added physical switch I can use to switch between the power delivery versions.
If the HW PSU gets full time power or just when the car is on.
0.5W doesn't sound bad
Ah, sorry, 0.5A.
400-500mA depending on settings etc.
But that might have been the 5V amperage after the kit.
So less on the 12V side.
But when normal car batteries have 20-50Ah capacity in total, IIRC.
Edit: Ok, Go up to around 100Ah for normal car batteries, and higher for larger diesels etc.
Almost miniscule usage while the car is running, but very high for constantly on load while the car is shut down.
But with lead-acid batteries, only pretty small percentage of that is actually usable in constant use, somewhere around 25% IIRC?
And also need to keep enough charge to actually start the car too ๐
My 3D scanner (3DMakerpro Lynx) arrived and this is how about 2h self training results looks like this with 2 combined scans:
To me looks really good
How's the dimensional accuracy?
Looking at how it picked up the cast numbers and cat logo, its less than 1mm but will do some measuring tomorrow
I wonder if it's worth the trouble and will actually be faster to call a process pool executor to compress and save 20MB string
Certainly acceptable
why do i feel like this is a scam
Trying to get people hooked on the extras.
The caliper measure is around 145,60 mm if my eyes still serve me right?
Yes
But neither seems to have been trying to get that accurate result.
And if the mismatch is same in all directions, you can just calibrate.
Learning how to read calipers back in elementary school paid off 
Etc.
The result doesn't need to be exact on the measurements without any work, if it can be easily made to match.
Although it will be harder to read as your vision gets worse
I got that too ๐
got my dad a digital one a few years ago for that reason ๐
I am considering getting a digital caliper too because reading classic (vernier) calipers hurts my eyes 
But anyways, now I got two sets of double-shot PBT keycaps \o/
To replace the original painted ABS keycaps, even when they haven't lost any of the paint yet in multiple years of use.
its never free... it costs your credit card info
"Nice"... Bought Seagate HDD from one of the retailers on Seagates own list for Finland.
And the drive has sticker saying the warranty is only valid in China.
But on other hand, the Seagate site where you can input country and serial said that the warranty was valid...
I bought a 16pm from one of my parts vendors bc they had one new in box for somehow cheaper than retail
New phone time
Before anyone says stolen or whatever, I thought that too but it activated just fine
So yeah idk must be a B2B thing
And my plans for cooling the drive in the enclosure just went out of the door.
4TB drive is not full height one, only 20.20mm high and not 26.11mm high.
Which the datasheet told, but I didn't think to look.
Well, will need to place the thermal pad material between the spindle on the drive underside, and not on the top cover.
With full height drive, there would have been about 3mm gap between the top and the case, but now there is 9mm gap.
Fortunately the average operating power is only 3.7W for this lower platter count version, when it is 5.0W for the 6TB model that is the first full-height model, and 8.7W for the 8TB model
And fortunately that spindle motor is where lot of the power usage comes from directly.
Unfortunately the usable underside area is much lower than the topside area:
Different model with different sticker, but same metal part:
In the end, needed 3mm of Arctic TP-3 on the spindle underside, and then put 7mm of the stuff on the other side to keep the underside pressed against the outer casing.
1mm thick sheet turned over three times and then one layer removed when the 8mm stack couldn't be inserted once the drive was in position.
Time to connect it to the computer and put self-test on, and monitor drive temperatures.
And mistake.
The enclosure didn't come with any kind of USB cable, and the only USB 3.0 Full-B cable I have is very short...
No, I'm just stupid.
There was cable inside cardboard "spacer"
Asrock is going to bring PSU's on the market.
And it has optional feature to prevent voltage drop by providing 5,2V instead of 5V...
and ntc-sensors in the power cable of the GPU that can measure the temperature and shut down incase of overheating
Couldn't easily start the SMART self test, so just went with full formatting.
Writing "only" at 37.5MB/s with spikes to little above, with 500ms average response time.
But as the important thing, the drive temperatures is still just 34C.
70C listed as the maximum operating temperature for the drive reported temperature.
No, that was for the bigger model, 65C for this one.
So the specific USB dock is significantly limiting the transfer speeds, but not enough for it to be problem in the intended usage.
Wonder if there is some kind of connection problem and it dropped to USB2 speed connection?
No, just the sync writing Windows was doing by default.
Changed the USB drive mode from Quick removal to Better performance and now writing at 186MB/s.
216MB/s now.
Up to 41 C drive temp.
How come Quick Removal has less performance?
Because that controls if everything is done synchronously (no data loss on removal), or if all the normal ways to make writes faster are in use, but that can lead to data loss if the drive is disconnected or loses connection in middle of a write.
I guess IO operations are throttled or something when doing quick removal? Or there is extra mechanism.
Because people have been stupid and just take the USB drives out without using the "eject" mechanic.
So the default was changed long time ago.
When not using quick remove, writes are cached
So they appear to happen more quickly but they're actually cached on your machine and finished in the background
And fed to the drive in large transactions, not in minimal steps.
Not really relevant to this, when I'm doing 4TB write.
Just that synchronous with minimal steps vs. asynchronous with large transfers.
Well, almost synchronous.
Because this is just format doing full drive 0-byte-write.
So the write caching itself doesn't do much in this case.
When the one write string will take multiple hours.
Only that write size difference.
When the Quick Removal needs to do things in inefficient way to keep the amount of data in flight to minimum.
And Better Performance can just hammer the drive with as much of max size transfers it can.
Including using the drives own write caching to do continous writing, instead of those more controlled writes that cause the drive to do reseeks etc.
Not just the OS side write caching and combining.
There are so many interesting things happening with I/O that most common people don't even know. Like when you delete a file. The data is not overwritten instantly. Else it would be 2 writes. One write to overwrite it with 0. Then another write when new data is stored.
Instead the reference in the master file table is deleted (so like a pointer). And thus there is only 1 write. Since there is no reference, the data will be overwritten somewhere in the future.
And that's why you have secure erase etc. But who knows stuff like this?
With HDDs, or SSDs without TRIM/DISCARD.
SSD with TRIM, there are garbage collection passes that actually delete the data once large enough part of the erase area is marked for TRIM.
As changing bits in one direction can be done one-by-one, in other direction only in large chunks.
Where that TRIM/DISCARD allows the OS to inform the drive that the specific LBA isn't valid data anymore and can be safely reused.
What we do for performance/optimization
And that's without accounting for the lower-level firmware/software or event delving into the hardware side of things, which is just as, if not more, complex
For sure.
And really a deep understanding of what's going on with the hardware is what even most people with advanced knowledge are missing
Very true
It's a big pet peeve of mine when people that work with software try to brush away the hardware side of things and pretend they don't really matter because thier python script is platform-agnostic
Oh yes. You can get quite the speed ups if you choose data structures that benefit you access patterns. But for that the programmer needs to understand how the data structure is stored/accessed
42C, while on wooden desk in 23C room, so seems the heat transfer is good enough.
The aluminum outer case is warm to touch, but not excessively so, but it is right way around, so I cannot touch the bottom where the best heat transfer should be.
True, same with cars. A lot of people can drive and use a car, but who really knows whats going on under the hood. Especially in these days with all the new technology.
Modern cars require just as dep an understanding of the tech side as the mechanical side
Yes.
Take throttle for example. And this explenation:
One type has a mechanical cable to open the throttle, and the other has electrical cable.
The electrical one uses an adjustable resistor (potentiometer) on the pedal side and an electrical motor on the throttle side. By using an electrical system, you avoid the transfer of noise and vibration, and you also eliminate friction in the throttle cable.
Throttle-by-wire has been around for ages, thankfully
The first one, looks to me, more simple than our modern solution. Although our modern one is better, just a tiny bit more complex.
Since you have adjustable resistor at pedal side and a electrical motor on throttle side.
The issue is that you already need a throttle position sensor even with mechanical throttle
Just as a simple example ๐
And mechanical throttle is actually more prone to sticking
So really you move the throttle position sensor to the pedal instead of the throttle body and then put a closed loop actuator on the throttle body
This looks and sounds simpler. Not saying it is. But my instincts say this is much simpler ๐
Only when you don't have cruise control or like etc.
And doing that kind of cable run without it causing issues isn't that easy to route etc.
But that's kind the point I made, PC and cars got more and more complex because we keep adding features and keep optimizing.
And other things. Like a photo camera used to be simple.
Camera Obscura
Modern day camera:
Yep, more and more complex at the cost of more and more capability
For me, it's so hard to keep up with all these technologies.
My issues with modern vehicles is that they're adding complexity without benefitting capability
Example?
It's my full-time hobby and even then you can really only get a surface level understanding if you want to cover a broad selection of fields
And you have to cherry pick?
I don't assume you know every part of cars AND DSLR camera's AND PC's.
Digital climate controls
cherry pick = specialize.
I know those three fields much, much better than most
Cost cutting.
Well, start naming the parts and what they do ๐
But at the cost of increased complexity
And the new paradigm of more complex climate control systems really isn't neccecary and is far more expensive down teh road
Reduced complexity for the manufacturing line.
That is true
But yeah, I keep it more high level ๐ I know what some parts are for a DSLR/Car/PC. But I am no expert.
- Rear lcd
- Sensor
- Memory card
- Battery
- Integrated flash
- Mounting flange
- lens body
๐
Fortunately that one got finally pushback from EU regulators.
Haha, yeah those, but I mean every little chip etc ๐
Yeah, I know what you mean
I just chose to be a pedant
IIRC any new model from next few years onwards has to have certain controls, including some of the climate control ones, back to separate controls, and cannot just be on a screen.
And the best part is, slowly there is awereness around repairability of tech. And I think they need to come up with a solution. A bicycle, someone can repair themselves probably. But to repair a laptop is different. My point is, with old technology, it was simpler and one could repair a lot themselves. Now that's no longer the case.
With later date when everything needs to comply, including older models still being sold.
I believe one of those was drive selector as well, right?
Didn't pay attention, and cannot find source right now.
So while things get more and more complex, goverments are demanding those complex tech, to be repairable more easy.
Yep
One of my primary life goals is to be capable on pretty much the broadest terms I can
cough cough **apple ** cough cough
Ah, NCAP (safety testing) is having some specific ones as needing to be not on screen, in pushback to Model 3 type idiocy too.
So it wasn't regulation, just the safety testing scoring getting changed.
And I very passionately make an effort to do taht
I have never in my life paid for someone else to repair something, even if it takes me far longer than it's worth to do it myself
you don't buy Apple macbook
There was also some movement on that climate control etc. controls front.
Yes, I intentionally buy things that I know are more repairable
I still drive a 20yo vehicle
Because I'm confident enough to attempt most any repair on it
With testing having been done on how much worse those are for safety already.
The amount of "heads down" time needed for basic actions was just horrible with those.
I hope they (the law) never force you to buy a modern car.
I am pretty confident with working on those as well
The primary issue is that many repairs are locked behind proprietary software and hardware tools on the electrical side
Diagnosing issues is more complicated, but electrically they're still very simple compared to devices and mechanically they're rather similar
(compared to older vehicles)
Yep
Lots of sensors
Not really, you need to go into 90s cars to get less for most part.
Do you know Mat Armstrong?
A Youtuber, who buys cars that are totalled. And then fixes them.
He and his dad don't do anything else than that. And when I see all the cables and technology of a stripped modern car. It's insane.
It is
The most important thing with cars is that they're broken down into discrete, simple systems
And it's a very useful skill to be able to isolate those
Can't find the video, he has so many lol.
But it was just one big pile of cables, cables, cables, all electronic ofc.
From one car ๐
I don't understand how they can put the cars back together, how they remember what is what.
It's a testament to the momentum of legacy systems that the sheer amount of electrical crap in cars hasn't pushed them to 48v just to cut the literal weight and space of wires.
And that pile is probably smaller in todays car than one from 2005-2010.
eh, I think it was from a VERY fancy car.
Yeah, but move the CANBUS etc. lowered the amount of wiring significantly.
And the fancy cars in that time frame already had most of the sensor shit, outside of the semi-self-driving ones.
so, expensive car, with a lot of features, needs lots of cables.
You would be suprised how much wiring there is already in 90s basic car.
Can lowered wiring requirements in a lot of ways but also led to a more distributed compute architecture which meant it really wasn't that much lower
Even just stereo and power locks and windows adds up to a lot
When the main runs are size of your wrist already in some specific places.
Wiring of ENGINE of a Audi RS6 GT3