The frequency converter here tells me that 1000 Hz is a flicker every 1ms
https://www.unitjuggler.com/convert-frequency-from-Hz-to-ms(p).html?val=1000
Help wanted to measure flicker and/or test a flicker-free monitor
- Edited
GregAtkinson i ordered one yesterday! Good to explore ebay as a way to solve the logistics problem of getting the tech in front of more of us. Reagan
reaganry Nice! Thank you for giving it a shot. I put it in the mail this morning.
GregAtkinson I read a couple places that this monitor supports 167 colors. I assume that means 167 out of a possible 256.
Whoops… I just realized that 167 means 16.7 million… (And according to Display Specifications, it's 6 bits + FRC.)
- Edited
I've been thinking about what a few people have been saying about refresh rates, temporal dithering/FRC, screen size, etc.
I had tested about 100 monitors and I took screen shots of the ones that had OK flicker. There ended up being 19 monitors with OK flicker. In other words, if you're willing to put up with a bit more flicker for a faster refresh rate or a true 8-bit panel, then this list could be useful.
| Manufacturer | Model # | Refresh | Color depth | Resolution | Screen size | Response time |
|---|---|---|---|---|---|---|
| Samsung | U28R55 | 60Hz | 8-bit + FRC | 3840 x 2160 | 28 | 4 ms |
| HP | M27h FHD | 75Hz | 6 bit+ FRC | 1920 x 1080 | 27 | 5 ms |
| LG | 27UP600 | 60Hz | 8-bit + FRC | 3840 x 2160 | 27 | 5 ms |
| Samsung | S27BG40 | 240Hz | 8-bit | 1920 x 1080 | 27 | 1 ms |
| HP | M27h | 75Hz | 6 bit+ FRC | 1920 x 1080 | 27 | 5 ms |
| HP | M24h | 75Hz | 6 bit+ FRC | 1920 x 1080 | 24 | 5 ms |
| Dell | s2422hz | 75Hz | 6 bit+ FRC | 1920 x 1080 | 24 | 4 ms |
| LG | 24mq450-B | 75Hz | 6 bit+ FRC | 1920 x 1080 | 24 | 5 ms |
| LG | 27mq450-B | 75Hz | 6 bit+ FRC | 1920 x 1080 | 27 | 5 ms |
| LG | 29wp50s | 75Hz | 6 bit+ FRC | 2560 x 1080 | 29 | 5 ms |
| LG | 32up83a | 60Hz | 8-bit + FRC | 3840 x 2160 | 32 | 5 ms |
| Philips | 45b1u6900ch | 75Hz | 8-Bit | 5120 x 1440 | 45 (curved) | 4 ms |
| MSI | g271cp | 165Hz | 8-Bit | 1920 x 1080 | 27 | 1 ms |
| MSI | g321cu | 144Hz | 8-bit + FRC | 3840 x 2160 | 32 (curved) | 4 ms |
| Lenovo | g32qc-30 | 165Hz | 8-Bit | 2560 x 1440 | 32 | 1 ms |
| Lenovo | q27h-10 | 75Hz | 8-Bit | 2560 x 1440 | 27 | 4 ms |
| Acer | EI322QUR sbmiipphx | 165Hz | 8-Bit | 2560 x 1440 | 32 | 1 ms |
| AOC | Cq32g3su | 165Hz | 8-Bit | 2560 x 1440 | 32 | 1 ms |
| Dell | SE2422H | 75Hz | 6 bit+ FRC | 1920 x 1080 | 24 | 5 ms |
- Edited
qb74 You could also use Rtings' tests for monitor flicker, which is pretty reliable
https://www.rtings.com/monitor/tests/motion/image-flicker
Now that I look closer at the site, I shouldn't have dismissed it so quickly. If you really look at the waveform closely, you can see whether it flickers or not. Here's a good example that I took from their site that doesn't seem to flicker:
https://www.rtings.com/monitor/reviews/asus/rog-strix-xg27aq
I kinda got thrown off by the fact that nearly all of their monitors receive a rating of 10 for flicker (10 being the best). But if you ignore their numerical rating and look very closely for a perfectly flat, thin waveform, you could probably find a flicker-free monitor that way.
The RTings site has a suggestion board. I created a discussion about how they could provide more detailed flicker graphs.
Here's the link. If everyone comments and likes I bet we could get their attention.
https://www.rtings.com/discussions/bxSoNXRO-ChZGzW3/detailed-flicker-graphs
You bought an LG 27GL650F-B (8bit) monitor, what's next?
Want to know how "response time" affects pixel flickering?
The backlight will flicker if you turn on MBR (black frame insertion)
- Edited
My return window is closing soon so I'm going to return all unsold monitors on Friday morning. Order by EOD Thursday if you want one.
I bought a true 8-bit monitor just to see if I could tell a difference between that and 6-bits+FRC. It a fairly inexpensive monitor ($228). It's a Lenovo Q27h-10. 27", true 8-bit (according to https://www.displayspecifications.com/en/model/01a62201). 75hz max frequency.
The flicker is about 1.4% using the formula (max-min)/max. It's in the top 10 (out of about 100 monitors tested) in terms of low flicker.
This screen shot is 100x zoom, 100% brightness and 30% contrast.
I haven't used it yet (I have a bit of a headache today), but I'll report back in the next week or so.
GregAtkinson Another thing Rtings is good at is to look for response times charts. Certain trainsitions / colors may flicker as well.
Such as this VA panel during a rise transition (from darker to lighter shade)
This sawtooth wave u can see in the scope is very likely related to the scanline logic, a brightness dip happening every refresh cycle due to it (180hz = 5.5ms)
This is measured with a 12bit scope + thorlabs photodiode.
I wouldn't be surprised if this is the issue you're experiencing on the monitors you've tested.
qb74 I wouldn't be surprised if this is the issue you're experiencing on the monitors you've tested.
Yeah, I've seen a lot of waveforms like that. I zoom in 100x so the waveforms are a lot easier to see. That particular monitor would likely not work for me due to the noisy waveform and sawtooth pattern. Eyeballing it, it looks like the flicker is about 5% or so.
I received the LG 27MQ450-B from @GregAtkinson last weekend, and I set it up as my main monitor at work to give it a thorough test. Around the same time I also received the Opple Light Master 4 I had ordered, which I know has limited capabilities, but I thought it would at least provide an empirical measurement along with my subjective perceptions.
Leaving the display settings as they came (Brightness=100, Contrast=40), my very first impression was that the picture appeared somewhat unstable. I had a slight perception of the picture moving or flickering. I took a few readings of a white background with the Light Master, and all readings were something like this (I calculate 12.6% flicker using (max-min)/max):
I ultimately reset the brightness and contrast settings to their defaults (I believe ~75 for both), and the picture looked brighter and more stable, although still not as stable as my old Dell monitor from 2013. I initially thought this might be usable once I got accustomed to it. I took another Light Master reading of a white background, and got this (I calculate 7% flicker using (max-min)/max):
I then tried reading emails on the screen for an hour or so (my typical use pattern). Unfortunately, I got the symptoms I get with most modern displays: focus problems, blinking constantly, a feeling like a flashlight is shining in my eyes, lack of comprehension, etc. I forced myself to use it for a few days, and perhaps adapted to it slightly but not significantly.
I am open to thoughts on this, because this sounds contradictory to GregAtkinson's testing indicating very low flicker. My initial thought was that the AC power in my office may not be very clean, or that there could have been a GPU output issue (however, none of these changed from my old setup, which was perfectly comfortable). I was connected via DisplayPort, just like previously (well, previously it was a DisplayPort to DVI adapter), so I wonder if I should try a different cable or try connecting via HDMI. I also am a novice user of the Light Master device, but the measurements do seem to correspond to my real-world discomfort.
For reference, this was a Light Master reading of my old Dell monitor from 2013, which I found perfectly comfortable (the PWM is quite evident, but I can use this all day long):
macsforme I took a few readings of a white background with the Light Master, and all readings were something like this (I calculate 12.6% flicker using (max-min)/max):
You and I get pretty similar results with the Opple, FWIW. Here's what I got with my monitor. I think I'd probably turned the brightness down to 50%, hence the lower absolute values. But the percentage flicker is roughly the same.
I think the bigger issue is FRC, maybe that's what is causing those symptoms.
macsforme I ultimately reset the brightness and contrast settings to their defaults (I believe 75 for both), and the picture looked brighter and more stable
Given that increasing the contrast was helpful, I'm wondering if your issue is related to the clarity of text. I'm curious if you've tried a display with high density pixels, like a Retina display or 3840 by 2160 pixels on a relatively small (24-inch or smaller) panel. What size is your 2013 Dell (screen size and pixel dimensions)? Similarly, have you tried making the text bigger? And experimented with anti-aliasing/font smoothing/etc.?
macsforme I was connected via DisplayPort, just like previously (well, previously it was a DisplayPort to DVI adapter), so I wonder if I should try a different cable or try connecting via HDMI
FWIW, I have been using HDMI (though TBH it kind of sounds like a long shot).
Your old Dell would give me a headache instantly with its PWM, so it seems likely that we have different issues.
- Edited
Given the fact macsforme could not use the monitor that was bought from GregAtkinson as I thought it woul be, because the problem is more then flicker, it is FRC, and dithering. Dithering is depends on GPU used, OS used. So hunting for perfect monitor is no as simple as measuring flicker.
GregAtkinson You and I get pretty similar results with the Opple, FWIW.
Interesting. I am curious what you attribute the discrepancy to? The Opple seems like a much less sophisticated device than your photodetector + oscilloscope setup, but then I wonder what the Opple is sensing that causes it to show that rough waveform.
GregAtkinson Given that increasing the contrast was helpful, I'm wondering if your issue is related to the clarity of text. I'm curious if you've tried a display with high density pixels, like a Retina display or 3840 by 2160 pixels on a relatively small (24-inch or smaller) panel.
I have tried several high-DPI screens. I used original Retina series MacBook Pros (2015) for several years, mostly without issue (some issues lately, but I currently attribute those to macOS operating system/driver issues starting with Catalina or Big Sur, or too much overall screen time). Back when I had a 2020 MacBook Pro 14-inch M1 Pro (when my major issues started), I tried three different high-DPI monitors with it: LG 24MD4KL-B, LG 24UD58-B, and Dell P2415Q. In all cases, it looked like the picture was constantly moving under my eyes. That issue was probably mostly attributable to the MacBook, but the LG 24MD4KL-B and Dell P2415Q both felt like the backlight had a harsh red undertone (I assume for wide gamut capabilities). Ultimately, I took the Dell P2415Q to work and tried it with my work desktop (as I did with the new one last week), and while 2X high-DPI scaling was beautiful, after a few hours I felt like it was piercing into my eyes and I could not use it. The Dell P2415Q backlight uses DC dimming, so I could not attribute that discomfort to PWM.
There are posts here about the backlight LED + phosphor changes in monitors after about 2015 (for better color reproduction and/or energy savings), and currently this is where I attribute my main sensitivity.
GregAtkinson What size is your 2013 Dell (screen size and pixel dimensions)? Similarly, have you tried making the text bigger? And experimented with anti-aliasing/font smoothing/etc.?
My older screen that I use daily at work is a Dell P2213 (22", 1680x1050, TN panel). This new one (27"/1080p) has only a slightly lesser PPI (82 versus 90). The UI looks comfortably readable on either, but after some time, my symptoms begin with the LG.
smilem Given the fact macsforme could not use the monitor that was bought from GregAtkinson as I thought it woul be, because the problem is more then flicker, it is FRC, and dithering.
There is definitely something going on. As I stated earlier, the first time I looked at the new LG display the picture looked unstable to me. And this was on the exact same computer, operating system, etc. that I was using earlier without issues.
macsforme Interesting. I am curious what you attribute the discrepancy to?
tl;dr: the zoom on my device filters out the ultra high-frequency flicker.
I was curious about that too. I figured it out. Here's a graph at 0x zoom. Note the X axis is 1 us (one millionth of a second) per div, so the entire screen shot is 1/100,000 of a second.
Here's the same graph at 3.5x zoom. Note the graph is actually smoother though you'd expect it to be bumpier given that it's zoomed in.
For me, I do not seem to be sensitive to flickering above 100khz, so I wasn't terribly worried about this. For instance, old CCFL monitors worked fine for me, and they tend to flicker quite a bit at 100khz (give or take).
Here's a graph at 0x zoom which shows a fair amount of flicker. It's not too hard to imagine that it's +- 1.5mv, so about 10-15% flicker:
Here's a graph that's quite a bit smoother at 3.5x zoom, since the ultra high frequency flicker is smoothed out:
23 days later
Since you have the most sensitive gear for flicker testing please get to nearest ZEPTER distributor, and test a pair of
Fullerene C60 doped glasses lenses. Make sure to choose indoor lighter color version.
https://shop.zepter.com/Medical/smart-glasses/HYPERLIGHT-OPTICS-SET,-INDOOR-LENSES
I wonder if they remove the flicker? The price is 180Eur if you manage to get good discount. But only if it works as claimed. Perhaps these could remove the dithering too.