I run my LG Ultrafine 5k @ 5k with my late 2013 MBP, Thunderbolt 2 cable, and Apple T2-T3 adapter plugged into the display. It works great.
Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Studio Display with old Macs
- Thread starter bingeciren
- Start date
- Sort by reaction score
5K@60? I thought 20 GB/s wasn't fast enough for that. Maybe with some sort of compression?
Your 5K is running at 4K. Nowhere in macOS does it tell you the real output resolution - it is only showing you the framebuffer resolution which it scales down to the 4K output resolution. Instead, check the timing info in SwitchResX to see the output resolution.
If you want 5K then you need to create a custom timing that is 5K39Hz but the GPUs of Old Macs usually can't do more than 4K width without artifacts.
If you want 5K60, then you need an eGPU.
If it is, I honestly can’t tell the difference between the 4k and 5k native! The image fills up the screen, and macOS reports 5k60 in both resolution picker and about this Mac. If it is just the frame buffer, then the signal is being produced at 5k, downsampled to 4k for transport to the display, then back upsampled to 5k to produce the image on the panel. Do I have that right?
Maybe you haven't seen 5K before. Have you connected the Studio Display to a modern Mac? Try a test pattern with single pixel wide alternating horizontal and/or vertical lines of black and white. View the test pattern at 100% using a 5120x2880 non-HiDPI mode.
Like I said before, macOS doesn't report output resolution.
There's no 5K signal in your case. The framebuffer in memory is 5K. It is scaled down to 4K by the GPU for output over DisplayPort (HBR2 x4). The display scales 4K up to 5K.
I also have a 14" M1 MBP that I can use with the same LG Ultrafine 5k. I honestly can't tell the difference between both computers. The 5k/4k/5k upscaled looks perfectly good to me - do you recommend any specific test patterns I could use? And, since it's displaying a 5k image on the screen, how is the output any different than 5k using DSC?
I have an iPad Pro that when I connect to the LG Ultrafine runs in 4k mode. That's noticeable, and the iPad does report it is outputting in 4k. This looks better.
I'm so confused.
This website has a test pattern:
https://www.eizo.be/monitor-test/
You should be in 5120x2880 mode. The test will switch temporarily to full screen to display the test pattern.
The sharpness test may also be useful - it draws text in black & white with no antialiasing.
If you see something that looks gray in either of those tests (except in the grayscale gradient part) then it's because of the scaling.
Use SwitchResX. Go to the Current Resolutions Tab. Double click the currently selected mode. It will show the pixel clock and active pixels. For 4K60 it will be 3840x2160 ≈533MHz. For 5K60 it will be 5120x2880 ≈960MHz.
Sorry, busy morning and I'm just getting to this. I ran the test pattern you suggested at full screen - thank you! But, the image still looks like 5k to me? I'm attaching a picture taken with my phone so there are some moire artifacts - these don't exist with the mk1 eyeball. I see no banding, grey, or fuzziness with the vertical/horizontal bars.
But, using switchresX (also thanks) I do see that the computer is using the 522 MHz pixel clock and then upscaling to 5120x2880 internally. This is so bizarre - I've never seen another display that allows you to think you're outputting a higher resolution image than is actually happening. Why does MacOS hide this from the user??
The good news is I still can't really tell the difference in resolution between the 5k/4k/5k of the late 2013 and the "real" 5k scaling of the M1 MBP. The late 2013 scaled resolution looks good enough for me - it's much better than the 27" 1440p native display my wife uses for work.
Thanks again for the suggestions getting to the bottom of this.
The screenshot on the right appears to be HiDPI. Are you using 2560x1440 HiDPI mode?
You need to be using 5120x2880 non-HiDPI mode to run the test pattern.
Switch to 5120x2880 non-HiDPI mode using SwitchResX.
Run the test pattern.
Take a screenshot of the test pattern by pressing Command-Shift-3.
Open the screenshot in Preview.app
Zoom into the vertical lines 10 times (Command +). You should see that each vertical line is a single pixel wide. This is not true when you use a HiDPI mode - in that case you'll see a gradient of black to white that is 4 pixels wide.
View Actual Size (Command-0). Drag the window around. When you move horizontally, the vertical lines will appear to animate or flash. When you move vertically, the horizontal lines will appear to animate or flash.
My display is 4K so I can test perfect 4K and imperfect 5K.
I get flashing in either case. But there is a difference in quality if I look at Actual Size at 4K and Actual Size at 5K. When scaling isn't being use, there appears to be more lines (it's more clear), and dragging causes more flashing than animating. I guess that makes sense - moving vertical lines one pixel left or right is like inverting all the pixels (when scaling isn't being used).
In your case, you need to compare imperfect 5K on the 2013 MBP and perfect 5K on the M1 MBP (except M1 uses DSC so it's not perfect perfect but close enough since DSC is supposed to be visually lossless)
Thanks for catching that I was using 2560x1440 HiDPI mode. I re-did the test in 5120x2880 on both computers using switchresX. I took a screenshot, then zoomed in to confirm each vertical line was 1 pixel wide. It is in both cases. Then, I viewed the screenshot at actual size and dragged it around the desktop on both machines. There is flashing on both computers. The M1 MBP shows a 933 MHz pixel clock, but the 2013 displays a 533 MHz pixel clock. And, now that I am sitting in front of the monitor, A/B testing both computers, I still can't really tell the difference in practical use. The display is very sharp in either case.
Thanks for the pointers @joevt - I had no idea MacOS was being so sneaky behind my back by hiding the real (output) resolution, but practically I don't think it matters for my use. And this 2013 sits in the basement running my 3D printer now, so it's not like I'm spending long hours in front of the machine.