[X] My Assistant

Borrowed a well-off friend's colorimeter; now my laptop screen finally doesn't show grays way too warm; it also fixed the washed-out gamma (it's calibrated at 2.2 now).

Of course, that doesn't fix it only covering around ~56% of sRGB, but baby steps. At least it doesn't look quite as terrible next to my PowerBook's anymore.

My desktop monitor's much closer to 100% (Dell U2412M - expecting around 95%), so I doubt I'll see quite as much change when I do that next.

Did you have to trade anything off? For my current screen when it was calibrated it covers more of the color spectrum accurately but I lost some peak brightness. Theoretically gained contrast.

Did you have to trade anything off? For my current screen when it was calibrated it covers more of the color spectrum accurately but I lost some peak brightness. Theoretically gained contrast.

Possibly a little bit of visibility of very dark (near black) grays.

They're still technically visible but you'd better be at full brightness. They were probably a bit too bright before, though, so it doesn't bother me much. I might want to look at a PLUGE (from colour bars) and see how those look now.
My laptop's LCD is nominally a 300 nit display, so if I lost any brightness it wasn't major.
Of course I don't 100% know what I'm doing so it's possible I did something wrong, too. Set a correction factor for PFS phosphor WLED screens (which I hope was correct; I know it's WLED but no idea beyond that). Told it to use the measured white point (planning to try 6500°K tomorrow and see what happens).

Those colours it can display should be more accurate, at least. So far, I'm happy with it. Even if it's not got quite the range (especially noticeable in purples/pinks like in my profile pic), things still look much closer on average to what I expect to see on my desktop's screen.

This is a Hydis (Hynix) FFS-type panel, btw, so the contrast ratio was already pretty good (compared to a TN panel).

This post has been edited by dragontamer8740: Jul 25 2021, 04:33

Update: when calibrating to D65 I do get a slight luminance drop on my desktop. It's fine though, my LCD's got a ~300nit backlight which is a bit excessive anyway.

It's sort of funny how far the 6500°K preset on the screen is from what the colorimeter says is 6500°K. Hope that doesn't mean that I am feeding it bad information.

Wish I could keep this colorimeter a few more days to play with it more. But I'm afraid that if I did that I'd end up keeping it forever :|

This post has been edited by dragontamer8740: Jul 25 2021, 21:19

There is an 850 SFX psu now from cooler master. My 750w Corsair unit is doing well but it gets hot, I am probably easily using 650w+ when gaming.

I mean it hasn't blown up, but it does get toasty.

There is an 850 SFX psu now from cooler master. My 750w Corsair unit is doing well but it gets hot, I am probably easily using 650w+ when gaming.

I mean it hasn't blown up, but it does get toasty.

I think my desktop tops out at around 130 watts. I use some shitty thermaltake PSU (think it's rated 430W or something? Doesn't matter) and a dual ball bearing replacement for the original sleeve bearing fan.

Edit: Actually, since I am using five spinning 7200rpm HDD's, maaaybe 150-170 or something if they're all thrashing away for some reason. I don't think I realistically have ever drawn full power from both the CPU and GPU simultaneously for more than a couple seconds on account of synchronicity and the like.

I still think it's funny that my Amiga is the only computer I have right now with 10k RPM drives.

My friend had his ryzen CPU fail, and he got new stuff, so I just got a free Gigabyte Aorus ATX motherboard (AX370 Gaming K7).
I think we mis-flashed his bios rom's while trying to figure out what was wrong, so I'll need a 1.8V SPI flash programmer if I want to get it working, but before I do that I'd need some DDR4 and a working CPU anyway so it's just a nice future project for now. Might need a stronger PSU as well but I need to look at specs. All I know is my friend was using water cooling so he probably stressed his setup a lot harder than I do. That or he just had bunches of money to throw around. Maybe both.

Think I might try something stupid like using voltage drop from diodes to program its roms with a TTL device. But that's for another time.

This post has been edited by dragontamer8740: Jul 26 2021, 05:08

My friend had his ryzen CPU fail, and he got new stuff, so I just got a free Gigabyte Aorus ATX motherboard (AX370 Gaming K7).
I think we mis-flashed his bios rom's while trying to figure out what was wrong, so I'll need a 1.8V SPI flash programmer if I want to get it working, but before I do that I'd need some DDR4 and a working CPU anyway so it's just a nice future project for now. Might need a stronger PSU as well but I need to look at specs. All I know is my friend was using water cooling so he probably stressed his setup a lot harder than I do. That or he just had bunches of money to throw around. Maybe both.

Think I might try something stupid like using voltage drop from diodes to program its roms with a TTL device. But that's for another time.

The bios updates are a bit of shit show for Ryzen. I am surprised, I am still on a bios from last year as I am waiting for a stable bios release.

Some the new bios have great features, such as less voltage for more speed but there have been loads of stability issues. The premium mobo I got lacks dual bios and easy bios flash back for some god forsaken reason. I have to remove the power cables and short mobo pins if there is an issue, I think the last time I had to do that on a comp was 2006.

They released a new stable bios today, but I want to wait a week to see if it bricks any Redditor's PCs first.

Update: tried a different (probably more correct) backlight type for my laptop and it's definitely darker now.

They don't have any options for my panel manufacturer, but since there's only one generic white backlight correction type that isn't high gamut (and since this laptop screen is only rated to cover about 60% of sRGB) I think I picked the right one this time.

I think it looks pretty good still; I'm still gonna keep both profiles even if one's wrong just because it corrects the color temperature somewhat without dramatically changing my gamma point.

Now this screen does look a bit closer to the sort of look I'd expect from my desktop monitor or powerbook screens.

Also I never overclock because it's just asking for trouble.

This post has been edited by dragontamer8740: Jul 26 2021, 07:26

Also ruqqus has no plain HTML/no-javascript method of using it, which is sad but not surprising.

Is it causing bad browser performance like "new" reddit & deviantart?

Is it causing bad browser performance like "new" reddit & deviantart?

It's causing "no browser performance" in Netsurf, which is my real issue.

From the ten minutes I spent on it the other day, my 10 year old laptop with HDD's didn't grind to a halt or heat up significantly, so it's probably better than new Reddit/DeviantArt.

I'm sure my Powerbook G4 will struggle, though.

Anyway, point is, new website designs suck. Even if you have to use JS there are better ways to do so than a billion frameworks and libraries now.

[dragontamer8740.gitlab.io] Have a skeleton for a text based game (sort of derived from the general Corruption of Champions concept, but not written in Flash) that I made in HTML/CSS/JavaScript a while back.

This one is performant everywhere with a non-IE, non-EdgeHTML-based-Edge browser (I think), and uses zero external Javascript libraries. I tested it in modern WebKit (WebKitGTK), Blink (Chromium), and Seamonkey/Firefox (~FF57) when I was writing this three years ago or so.

Back when it was still developed I did report the rendering bugs in EdgeHTML, but they never fixed them before they switched to Blink.

It also uses zero canvas elements (the stat bars are actually dynamic, but since it's just a skeletal engine you have to change variables in a JS console to see that in action. The bars themselves are SVG elements that I manipulate in Javascript). It acts as a "state machine" and can serialize/deserialize every part of its state (to save and restore data).

One evil thing I did was making the buttons on the bottom dynamic and bound to functions called via 'eval' since I couldn't find another way to do it. But it works and is part of the save-able state of the game. If I were writing it in C/C++ I'd have used function pointers.

Object.keys(player.stats) will get you a list of player stats that you can edit. You can do 'player.stats.def=55' to change the defense stat to 55, and notice how the stat bar automatically updates. I used JS getters and setters to do this, which made serialization/deserialization a little annoying but works well. The 'privData' sub-object contains the raw values which the getters/setters access.

BTW it also has a Javascript shell built-in which you access with the button on the side or the F9 key. I made that for testing on mobile devices.

This post has been edited by dragontamer8740: Jul 26 2021, 17:44

California, Hawaii, Oregon, Vermont and Washington banned gaming computers because they use too much power.

Dell had to change alienware prebuilt computers to use shitty low power CPUs for those states but it wasn't enough. The 30 series GPUs use too much power. Nvidia 1650 is the most powerful card allowed.

They are claiming the ban will save a gigantic amount of power, its almost like they don't know how PSUs work. A 1200w psu does not draw that full amount 100% of the time. Did some idiot do their calculations?

They are claiming they are going to save a huge amount by targeting idle power...

Why not just mandate gold and platinum PSUs....

Modern PCs and monitors have been far more efficient especially when idle since 6-8 years ago?

The legislation was sponsored by the energy companies that are having supply issues and donate to democrats. Rather than fix the issues, they have focused on a pretty nebulous issue.

Their calculations are that by banning gaming PCs they will save 1.5% to 100%?? of residential power usage.

Phase 3 is putting energy efficiency ratings on games and encouraging people not to play them wow.


Reddit is weird

I was reading about Reddit users complaining and being elitist about Ram overclocking. They claim that the DRAM calculator tool for Ryzen is pointless because people should learn to manually adjust ram and noobs using the calculator can run into problems when they input the incorrect data into the tool or don't understand the settings they are inputting.

I found the tool useful as it allowed me to dial in the ram in 30 mins in what would normally take a month of tweaking and testing. The extreme setting did have some stability issues but that is normal.

Reddit:
1. Noobs can not work out what ram they have so they should stick to XMP
2. Noobs should learn the very involved knowledge required to accurately adjust all the subtimings and voltages
3. Noobs that do not know how to recover their comp from a bad overclock should not touch the bios or this tool.

Three might be reasonable but its weird elitism. I found the Dram calculator's safe calculations to be better than what I would have bothered to dial in manually in the past and way better than the XMP settings.

The reddit users were complaining that the tool did not dial in everything perfectly and the extreme settings and voltages could be too high. They complained that they had problems (but they are elite) so nobody should use the tool.

The extreme/fast setting was only meant to be a guideline for where you could further dial in the ram for advanced users. I found the safe setting gave 95% of the performance gains and was stable compared to the extreme settings. The tool should just have more pop ups warnings for the extreme settings, in my opinion. I found the extreme settings assumed a good mobo and cool ram eg <40oC, however it did not explicitly state that.

edit: reading other threads, they are complaining when they buy cheap ram Dimms that the manufacturers tricked them by not explicitly saying that they were not as good as the expensive kits. They seem to have this belief that if they buy cheap kits that they will luck out and always get some god tier cheap but awesome silicon lottery kit.

Reddit: "I bought a cheap $100 dollar ram kit, why does it not perform the same as youtubers using $500 dollar kits, the manufacturers are evil. Someone else on reddit bragged that they bought the cheap kit and it performed the same as the expensive one but it only crashed for me."

It doesn't help that manufacturers don't tell you explicitly what is in each kit and many look similar (except for price.). However people should know how aggressive silicon is binned these days. Its not like 10-15 years ago where the silicon lottery could really work in your favor.

This post has been edited by EsotericSatire: Jul 27 2021, 02:13

California was a mistake.

laughs with low-power machine that would be even lower power if the PSU wasn't a terrible design
Glad I'm not impacted (even if I'd lived there).

Also, the Huion tablet (H1060P) I ordered a few days ago came in the mail today.

This is a 10×6 inch (12" diagonal) drawing area.

I got it sealed (in a damaged box) for ~$42 after shipping/tax on ebay, which is far less than the $100 I'd normally be spending on one.

First impressions are excellent.

The impetus for this purchase was misplacing my Wacom's stylus for the millionth time, since the strap no longer holds the pen in place. After an hour of searching, I decided to spend the additional ~$12 for a bigger better tablet instead of just buying a new pen that I already wasn't too happy with. It was prone to leaving 'dots' of high pressure when the pen first came in contact with the surface if I was writing or drawing something quickly.

A massive step up from the dinky little Wacom Intuos Draw tablet (model CTL490DW) that I got from a Best Buy in ~2015 that I was using before (…which was around $100). For comparison, that had a 6×3.7 inch (7" diagonal) drawing area, half the pressure sensitivity, and no tilt sensitivity.

Using the "[github.com] digimend" drivers, I got everything working with it, tilt support included, after specifying to use the 'wacom' driver for it in my xorg.conf. Apparently they have it use the 'libinput' driver unless they've verified a specific model works fine with the wacom one. This one does but was defaulting to libinput.

This pen also has a nice rubberized grip on it, a good shape/size, and the tablet came with this neat little two-fingered glove in the box to reduce friction from your hand dragging across the tablet surface.

This tablet also seems to have a smoother surface than the Intuos, which I actually like more. People claim the surface on the Draw is supposed to feel like a pencil on paper, but to me it felt more like drawing sandpaper than anything else.

I'm also digging the larger surface area; it's pretty much exactly the same width and height as the screen of my laptop (12" diagonal 8:5 ratio), and a much better fit for my desktop's 8:5 ratio 24" screen than the 7" one was.

My only concern is that from what I can see Huion is a bit aggressive when it comes to screwing people who don't buy from them directly and have proof of purchase. Apparently to get a firmware update (which mine already has thankfully) you need to send a serial number, proof of purchase, etc. to them. Which is kind of annoying. I'm sure they have had some sort of counterfeiting problem at some point, but come on.

At the price, still happy though.

This post has been edited by dragontamer8740: Jul 27 2021, 05:34

California, Hawaii, Oregon, Vermont and Washington banned gaming computers because they use too much power.

What heresy is this? They're blaming PC gaming for "high" power usage? As opposed to cryptomining farms?

I typically use an ultra-low-power netbook that uses something like 2-5W total. I've never wanted a "gaming PC". Until now.

I typically use an ultra-low-power netbook that uses something like 2-5W total. I've never wanted a "gaming PC". Until now.

Do they go that low? I thought 5W+ was a kind of wall for netbooks with the screen wifi etc.

I'll have to put the digital power meter on it instead of going by somebody's post. I think 2W is for "Topping off the battery, while off", which technically it spends time doing.

Why not just mandate gold and platinum PSUs....

That doesn't work because it has to be a government standard such as EnergyStar.
Such a mandate will only cause sellers to change to chinese brands that slap on the CSA TUV logo even though their products don't meet the power stability standard after their changes. I've bought chinese models from the early runs and then much later bought the same model because I thought the first one was good. The latter one was the same exact model number and the same FCC identification registration mark among other things. But the insides were quite different and now had "fewer" parts too on that small circuit board. There's no way such changes would have used the same registration codes as it was essentially a completely redone circuit board. Basically it was a different product that masqueraded in the same external chassis container using the same external connectors. Therefore, mandating Gold or Platinum efficiency from chinese brand power supplies is probably not a reliable approach for PS designs that end up with open circuit faults if they go bad which protects the mobo and hard drives from a power spike burnout. I bet the cheap gold/platinum ones will send a power spike before dying due to their "too few parts is better design" approach.

For lack of a better thread, why do people seem to think science changing in light of new data disproves it rather than strengthens it?

More strictly on tech, git another new HT radio I didn't rightly need. Need to make sure I have the proper CPS to program it

For lack of a better thread, why do people seem to think science changing in light of new data disproves it rather than strengthens it?

It is not disbelief in science but in the scientist. If you get new data that is contrary to the first couple of times you did run it, it means you suck at experiment design and stats - which should be the two most important qualities.

Unfortunately "journalist" think that "experiment design" is the power point presentation used to present results. So they make the bad even worse.

been a hot minute since posting again, that aside I come to this thread today with a riddle of something I picked up for €2 at my local thrift store:
what has a 5 pin din, a switch, a button, 2 neon indicators and runs off mains?


some hints:
the blue transformer on the 'board' creates a 6v:
the red indicator on the front is just to indicate mains power, the switch on the back just toggles mains power.
the 'board' its on is just a piece of plastic, its a glued/taped ratsnest under it.

regretably the din plugged devices predate me a bit, they seem to have had a wide range of uses so I can imagine this being some tester where the second neon indicator on the front would light on button press if the connected whatever would work correctly. but thats as much as I can think of.

quickly measuring on the din:
going from the top pin as ground all others look very 'floaty' nothing of great use.
between the outer pins is ~8v which goes to ~7v on button press
this is probably unrelated nonsense without the intended device connected

any thoughts or ideas on what it could possibly be?

This post has been edited by cate_chan: Jul 29 2021, 21:17