The new display does pretty darn good colors. I think it’s hard to compare a matte screen to a glossy. Both have their pros and cons.
In general, a matte screen is very good at dealing with ambient light and glare. Glossy screens have that “pop” and, in good lighting conditions, deeper blacks.
Before my Framework laptop I had a Lenovo with a really awesome glossy screen. It took a while but i became used to the matte display and really appreciate it. It might be psychological, but I think it’s easier on my eyes too.
I ran a battery test today. I jumped on YouTube and looped a video all day long starting at 100% charge until the battery died. I picked the video Baelin’s Route - An Epic NPC Man Adventure (hilarious!).
Test Conditions
Video resolution set to 1080p
Sound was muted
I have the larger 61Wh battery
Brightness set to about 30%
120Hz
Fedora 40
Firefox browser
Balanced power mode
The laptop ran for 6.73 hours. About 14.8% battery per hour.
I’ll order my Framwork 13 in the next weeks and I’m still hesitant about the screen. I read the posts about the bad scaling of the old screen, but I’m not sure if that is enough for me.
In my case, I’ll use the laptop for programming and browsing the Internet. I do not do hard gaming (an eventual point an click adventure, nothing that requires 120Hz) and barely watch any YouTube video. In other words: command line, a browser (Brave, nowadays) and VS Code, all in Linux
Is it still that big improvement for a use case like mine? I prefer more battery life than spectacular colors.
I reran the YouTube test again but set the display refresh rate to 60Hz to see if there’s any difference. Same conditions as the 120Hz test.
60Hz results: The laptop ran for 7.10 hours. About 13.7% battery per hour.
120Hz results: The laptop ran for 6.73 hours. About 14.8% battery per hour. (from the previous post)
Congrats on getting a Framework laptop! I really enjoy mine. The screen choice is definitely something to think about.
I would say the colors and contrast are about the same (as a human looking at a screen) between the two matte displays.
If you are a coder who likes tiny/small font, I think 100% scaling with the 2256x1504 panel could work. I think it’s too small to code myself but that’s my personal preference. DPI (dots per inch) is much better with 200% scaling on the new 2880x1920 panel. The font looks crisper and clearer which is more important to me than screen real estate. My personal opinion is if you need more screen real estate, workspaces and/or an external monitor (say a 15.4" portable USB-C monitor) works better than trying to cram windows next to each other on a 13.5" screen.
To give you an idea, I opened two terminals (on the 2880x1920 at 200% scaling) and put them side by side. In each terminal I edited two files using vim and I was able to get 71 (edit: off by one error, it’s actually 70) characters horizontal before wrapping and 37 rows before it started to scroll vertically. I used monospace font at size 12.
The 120Hz is really nice just working inside the operating system. Moving windows, cursor movement, scrolling, etc is more enjoyable at 120Hz. (again my opinion here)
In the terminal you might not see much difference between 60Hz vs 120Hz, but in VSCode (try VSCodium for the same app but without the MS tracking!), browser, and even lightweight games, I definitely see and feel a difference between 60Hz and 120Hz.
Maybe I’ll plug the 2256x1504 screen back in and do another battery test to see if there’s much difference.
Looks like you’ve swap the laptops around (left ↔ right) between the two photos.
The older / glossy panel’s white point is towards red, the newer panel’s white point seems to be slightly towards green.
You ought to have your own colourimeter because the panels do age (specifically the sub-pixels will age). Typical professional monitors in a colour grading environment would be required to be re-calibrated every 2 weeks or so. Those monitors would also record & show the powered-on hours of the panel.
Any downloaded ICC profile from Notebookcheck were calibrated against new panels at that point in time, by that particular colourimeter at that point in time. (There’s also a matter / question of colouritmeter accuracy aging as well)
I’m a coder that can’t work with tiny/small fonts anymore . My current laptop is an XPS 13, 1920x1080 with a PPI of 165, I need to use a monospace font of size 14 to feel comfortable. Given the bigger 200 PPI of the 2256x1504 I’m not sure I’ll manage to work comfortably at 100% scale.
Your test with two Vim sessions in a command line is really helpful for me, thanks for that.
Yes the display is definitely better, even when I set it to 60Hz to save battery it is smoother than the old one because of the reduced ghosting.
I was surprised that the panel is a good amount heavier and I really notice the difference to my framework from work (that still has the old panel).
Also the hinge works a bit worse with the more weight (when carrying the laptop around it more often just opens up 180 degrees through the movement).
Does anyone know how much the GtG response time actually is?
I’m seeing some people state >20ms, but a few months ago I asked on nrp on HN and got the response of something around ~12ms!
I agree that the 60Hz feels smoother than on the 2256x1504 panel.
You might check with Framework on the hinge. I think there were some older hinges that were not as strong as the current hinges. Mine does not behave the way you describe yours. It is strong enough to hold up the screen without any issues (at any angle) and it does not open up when I carry the laptop around.
I tried out 150% (scale res of 1920x1280) and 175% (scale res of 1645ish x 1097ish) screen scaling. At 120Hz the jerky screen that I saw on the 2256x1504 screen is less noticeable. 200% is smoother but I think both 150% and 175% are OK. I think 150% is too small for me. I could feel the eye strain. 175% I think would be OK. Eye strain was more than 200% but not bad for me. I gave the “vim” test again (two terminals side-by-side with vim running).
200% @ 14 point - 64x32
200% @ 12 point - 70x37 (I was off by one in my previous post)
Once I get my display in September (currently pre ordered about a week ago for Batch 5) I’ll post results from FreeBSD 14.1-STABLE. I’m a gen 1, batch 6 user from Dec 2021. I’ve done various upgrades though including the 61 Wh battery, and I still have the original glossy screen which I don’t like and had to pretty much immediately buy a matte film from Photodon, which massively helped the past few years… so I’m looking forward to finally having a natively matte display, and excited about the bump to 2.8k and 120 hz.
I reinstalled the 2256x1504 matte screen. I ran the same YouTube test. Same conditions and setup as before except I set the resolution scaling to 150% (what I used) instead of 200%.
60Hz (2256x1504) results: The laptop ran for 8.00 hours. About 12.5% battery per hour.
60Hz (2880x1920) results: The laptop ran for 7.10 hours. About 14.1% battery per hour. (from previous post; percentage calculated from time 100 / 7.10)
120Hz (2880x1920) results: The laptop ran for 6.73 hours. About 14.9% battery per hour. (from previous post; percentage calculated from time 100 / 6.73)
I remember reading on this forum that 100% and 200% scaling inherently consumed less power than 150% and 175%, but that was a while ago, and I’m not sure whether it still applies, and what other factors it might depend on (like compositor, firmware, etc.).
think so too been running mine for a few days. Updated from gen 11 with film. will repurpose older display on a project. Now they need to sell me the new camera…
I decided to run the test one more time with the 2256x1504 60Hz monitor but this time with 200% scaling. I was interested to see if there was any difference in battery life with whole number scaling vs fractional scaling. Here’s what I found:
Very interesting and surprising results! 200% consuming more power than 150% seems counterintuitive, simply because there should be less pixels to render the video on… But, considering a relatively small difference, perhaps the scaling simply plays no role, and they are all equal within the margin of error?
In any case, thanks a lot for running the tests, and collecting the data!
In Windows. Old screen was set to 30% brightness (obviously 60Hz), but with the light sensor enabled. New screen is set to 15% brightness, with the light sensor enabled, 120Hz.
I lose roughly 2 or so hours on average from the old screen this way. It’s rough.
I would say that 150% vs 200% is a tie. The difference is probably within error (6 minutes over about 8 hours is about 1.25%).
After doing the YouTube tests, there was a loss of about 1.25 hours, 16%, when going from the original 60Hz panel to the new 120Hz panel. That shrinks to about 54 minutes, 11%, if the new panel is set to 60Hz.
. My current laptop is an XPS 13, 1920x1080 with a PPI of 165, I need to use a monospace font of size 14 to feel comfortable. Given the bigger 200 PPI of the 2256x1504 I’m not sure I’ll manage to work comfortably at 100% scale.