OSSC Pro

[fernan1234]

Just out of curiosity, are you judging the motion clarity of a Sony PVM from that article or have you purchased one?

I have/had a couple of different models, all of which have panels that vary by grade but are the same in terms of the scan driving. I think the article significantly underrates it (it suggests that it is 1/4 the motion clarity of CRT, which doesn't make sense if that's also supposed to be the clarify of frame-matched BFI, but maybe the latter would be 1/8 instead).

[marqs]

It has 5V barrel plug input like before, but practically all ADCs and logic use 3.3V or below so the input is not fed directly to any sensitive parts. I would still recommend using the 2.5A PSU that we plan to include with the board - you are likely to just ask for trouble if you use an underpowered USB hub or a thinny USB cable.

Hmmm I saw you using LP5912 LDO which won't be able to filter out any high frequency noise. PSRR is rated at very very low frequency of say 1-50 khz, most SMPS will be a lot more than this, 1 MHz being very common and sometimes more, especially if it is small size (= small size inductors). Looks like LP5912 noise rejection figure is only good up to 10 khz if not less. with 100 khz it is at 40 db.

What is the noise and ripple generated by your supplied wall PSU? at what frequency does it operate at? you must see it in oscilloscope to verify but it is known that LDOs will not filter the noise\ripple at higher frequencies.

It doesn't matter if your sensitive parts use 5v supply or not since 3.3v is generated from it directly without proper filtering (LC, Pi filters, etc...). I think you know all this and took it in consideration, I just want to point it out. I am really interested about the PSU you are going to supply with the device in terms of price, origin, and most importantly its noise\ripple specs. Chinese stuff are almost always very bad in this regard.

If you still take suggestions, then I suggest getting an AC-DC power module from meanwell which will be < 10$ and will act at a frequency less than 100 khz (maybe 65 khz) which is easily decimated by good LDO of yours.

I am really pumped for the pro this year!

My comment was more about output voltage tolerance in general rather than high frequency noise which is a potential concern as you noted. The PSU candidate which I've tested measured <80mV ripple at 20-40kHz so it's not too bad but not great either. Meanwell also has a few potential PSU models (e.g. GE18I05-P1J or GST18A05-P1J) which could be considered.

[H6rdc0re]

Sounds like you don't like using "scanlines", but for people that do need to use them especially at 100% it's basically a deal-breaker because that's an additional 50% brightness loss on top of the 50% loss from 60Hz BFI which is what you need for most optimal motion clarity, and how close that gets you to "comparable to CRT" seems to vary by perception. For me it's not that close, and according to measurements it's basically 1/4 as clear as CRT which is not bad but still ways to go IMO.

I do agree that for many people commercial OLEDs + modern scalers can be quite a satisfactory alternative to CRTs.

OLEDs will get brighter in the near future. Samsung will enter the OLED market next year and will surely combine the crystal layer (quantum dot) to their OLED TVs. This will force LG to follow suit. With higher brightness OLEDs you can combine BFI with scanlines. Now you’ll have to choose. I like scanlines but like perfect motion more so I choose perfect motion.

Seems to me you have never seen a LG 2020/2021 OLED with the right settings. When running matched content (60fps @60Hz and 120fps @120Hz) with OLED Motion Pro @high is amazing. When running with the wrong settings motion is just the usual blur. I have 5 CRTs myself so I know what perfect motion should look like.

[fernan1234]

Seems to me you have never seen a LG 2020/2021 OLED with the right settings. When running matched content (60fps @60Hz and 120fps @120Hz) with OLED Motion Pro @high is amazing. When running with the wrong settings motion is just the usual blur. I have 5 CRTs myself so I know what perfect motion should look like.

I've seen the CX and C1 models, and yes matched content at the high BFI setting is very good (though it was equally good with matched content on the older 60Hz panels, though obviously they couldn't match 120fps content), but like I've been saying I don't consider it comparable to CRT, but definitely a step in the right direction. It's just disappointing that it's not as good as another persistence-blur-clearing technique that has been used on Sony's pro OLED panels since 10 years ago. And even that is not the same as CRT obviously, but gets closer, twice as closer to my eyes.

As to the brightness issue, Samsung's QD-OLED is expected to be less bright, at least at first and in terms of peak brightness. But what I think will be more important is whether it can handle full screen brightness better, with less ABL than LG's panels for SDR signals (which is what the OSSC Pro will output). Current LG WOLEDs can get very dim in SDR, which is compounded further with BFI + scanlines double-whopper brightness halvings.

[VEGETA]

My comment was more about output voltage tolerance in general rather than high frequency noise which is a potential concern as you noted. The PSU candidate which I've tested measured <80mV ripple at 20-40kHz so it's not too bad but not great either. Meanwell also has a few potential PSU models (e.g. GE18I05-P1J or GST18A05-P1J) which could be considered.

80mV is bad. However, if it is at low frequencies such as 20-40 khz, you could get away with it by using LDOs. Either way you should have put filters such as LC or Pi filters using big inductor and capacitors at the very input to your circuit. Maybe 2 stages of 10uH with 1000uF caps before and after the L... could be a bit less capacitance but 1000uF elec. caps are not pricey. then feed this into everything else and you won't get anything near the 80mV.

You also have a buck converter for 1.1v which is going to generate a lot of high frequency noise regardless of its input supply, so more multistage LC\Pi filters are a must after it. With all that, you must put a ferrite bead at the input of all your ICs power pins followed by regular decoupling caps which you put.

How much price will your PSU from China cost? Meanwell is not cheap at all and yet they still have +100mv_p-p ripple, fortunately at low frequency.

EDIT: I found suppliers on Alibaba selling 12v 3A and similar interchangeable wall mount PSUs for 1-3$!! what a bargain! I assume these are good choice even if they say it is 200 mVp-p noise which is very bad. You can use it to supply 12v or 5v, then wack-in multistages LC\Pi filters which is going to help a lot especially if this PSU is running on low frequency. I hope you post some measurements of both before and after filtering on loaded condition so we can see.

[H6rdc0re]

I've seen the CX and C1 models, and yes matched content at the high BFI setting is very good (though it was equally good with matched content on the older 60Hz panels, though obviously they couldn't match 120fps content), but like I've been saying I don't consider it comparable to CRT, but definitely a step in the right direction. It's just disappointing that it's not as good as another persistence-blur-clearing technique that has been used on Sony's pro OLED panels since 10 years ago. And even that is not the same as CRT obviously, but gets closer, twice as closer to my eyes.

As to the brightness issue, Samsung's QD-OLED is expected to be less bright, at least at first and in terms of peak brightness. But what I think will be more important is whether it can handle full screen brightness better, with less ABL than LG's panels for SDR signals (which is what the OSSC Pro will output). Current LG WOLEDs can get very dim in SDR, which is compounded further with BFI + scanlines double-whopper brightness halvings.

There is no blur what so ever on slow, medium or fast pans. I don’t understand what is worse about the motion performance according to you. Like I said motion is no worse than on any of my CRTs.

Things should get interesting next year. We’ll see how Samsung comes out, perhaps they won’t use sample-and-hold technique. Doubt it though.

SDR gets plenty bright even with ABL at about 150cd/m2 on a 100% sustained window and much higher than the 100cd/m2 or 120cd/m2 picture standard. Throwing in both BFI and scanlines is too much but either BFI or scanlines is possible.

I’ll stop now because I’ve derailed this topic enough.

[fernan1234]

There is no blur what so ever on slow, medium or fast pans. I don’t understand what is worse about the motion performance according to you. Like I said motion is no worse than on any of my CRTs.

You can use this to see the differences across different display technologies and flickering/strobing techniques more easily: https://www.testufo.com/blackframes

Persistence blur is measurable in MPRT (moving picture response time). If CRT has a persistence blur of 1ms (60fps content), it's known that BFI at its best (duty cycle matching content FPS) is 4x the persistence blur of CRT (4ms). That is objectively worse than CRT and should be noticeable to any attentive eye, but certainly an improvement over the 16ms persistence blur without any flicker/strobing at all.

Things should get interesting next year. We’ll see how Samsung comes out, perhaps they won’t use sample-and-hold technique. Doubt it though.

Maybe a bit of confusion going on here, but sample-and-hold is the display type, not a technique for addressing persistence blur. All sample-and-hold displays (LCD and OLED) have 16ms persistence blur and need some kind of flicker, strobing, scanning, etc. technique to improve it. This article is pretty good at giving some perspective on what it takes to get to actual CRT motion clarity for all cases (basically need 1000Hz refresh rates, so we still have a long way to go): https://blurbusters.com/blur-busters-la ... -and-hold/

And no worries, I don't think this is that off-topic, since the OSSC Pro and such are meant to provide CRT alternatives, and pairing them with OLED displays is the best option available for the next few years, but expectations should be set accurately.

[jalyst]

Fascinated in this product, only just started getting into the whole Mister FPGA scene.
May I please ask; are there still plans to sell & ship it worldwide?
And is it fair to say that it's the best device of its kind, on the planet?

Thank you very much / BR.

[marqs]

My comment was more about output voltage tolerance in general rather than high frequency noise which is a potential concern as you noted. The PSU candidate which I've tested measured <80mV ripple at 20-40kHz so it's not too bad but not great either. Meanwell also has a few potential PSU models (e.g. GE18I05-P1J or GST18A05-P1J) which could be considered.

80mV is bad. However, if it is at low frequencies such as 20-40 khz, you could get away with it by using LDOs. Either way you should have put filters such as LC or Pi filters using big inductor and capacitors at the very input to your circuit. Maybe 2 stages of 10uH with 1000uF caps before and after the L... could be a bit less capacitance but 1000uF elec. caps are not pricey. then feed this into everything else and you won't get anything near the 80mV.

You also have a buck converter for 1.1v which is going to generate a lot of high frequency noise regardless of its input supply, so more multistage LC\Pi filters are a must after it. With all that, you must put a ferrite bead at the input of all your ICs power pins followed by regular decoupling caps which you put.

The Meanwell brand you recommended has 50/60mV ripple spec for the PSUs I listed so I don't see 80mV so bad. After -50dB from the LDO ripple amplitude is less than a millivolt. Sure, pi filters on analog supplies could have been used as an extra safety. The on-board bucks are not directly driving any analog circuits so I'm not too concerned with them aside from potential noise generated upstream. 1.1v is for FPGA digital logic and Intel's recommendation doesn't include a filter on that line - perhaps because the inductor would need to be fairly large in order not to cause notable voltage drop with 2-3A current.

How much price will your PSU from China cost? Meanwell is not cheap at all and yet they still have +100mv_p-p ripple, fortunately at low frequency.

EDIT: I found suppliers on Alibaba selling 12v 3A and similar interchangeable wall mount PSUs for 1-3$!! what a bargain! I assume these are good choice even if they say it is 200 mVp-p noise which is very bad. You can use it to supply 12v or 5v, then wack-in multistages LC\Pi filters which is going to help a lot especially if this PSU is running on low frequency. I hope you post some measurements of both before and after filtering on loaded condition so we can see.

I'm not thinking of getting random cheapo PSUs from Alibaba/Aliexpress. Even if you scattered the board with filters to eliminate ripple, there are other issues like safety concerns, quality of caps etc.

[VEGETA]

Sure, pi filters on analog supplies could have been used as an extra safety.

yes, this is always best.

The on-board bucks are not directly driving any analog circuits so I'm not too concerned with them aside from potential noise generated upstream. 1.1v is for FPGA digital logic and Intel's recommendation doesn't include a filter on that line - perhaps because the inductor would need to be fairly large in order not to cause notable voltage drop with 2-3A current.

I know they are not doing any analog stuff but still noise and ripple can and will find its way in ground path even if you have perfect layout loop. Yes, it is good to look for recommendation from manufacturers but those are not mentioning the optimal settings for your design. DC-DC converter datasheets do not tell you how to design Pi\LC filters for your application but rather how to use their part.

I finished designing a dreamcast power supply by using TPS62913 new part from TI which promises <1mV ripple and noise, they precisely mention that you should get this without any external filters... because their IC includes ferrite bead inside feedback loop itself which is not how it should be usually. However, other TI parts just shows you the regular schematic without any external filters and leaves it to you to decide.

I'm not thinking of getting random cheapo PSUs from Alibaba/Aliexpress. Even if you scattered the board with filters to eliminate ripple, there are other issues like safety concerns, quality of caps etc.

caps quality are over-exaggerated by designers very often. Getting cheap elec. caps and ceramic caps from LCSC is ok unless you need very specific low ESR caps and so on. I don't care about how Rubicon elec. cap is good since it costs a fortune for very little improvement if any... except that it can last 25 years instead of 20 years or so. this is my humble opinion.

As for cheap psus from alibaba, my original OSSC from Aliexpress came with one of them and it is doing great till now. I even have an extremely low quality Chinese laptop adjustable PSU (5~24v) but despite being noisy and so on, its safety features are ok. short circuit makes it shutdown for example.

I guess, you can rely on such cheap PSUs in terms of short circuit protection..etc but not in terms of noise and ripple.

What is the price and model you wish to have as default option?

[vol.2]

caps quality are over-exaggerated by designers very often. Getting cheap elec. caps and ceramic caps from LCSC is ok unless you need very specific low ESR caps and so on. I don't care about how Rubicon elec. cap is good since it costs a fortune for very little improvement if any... except that it can last 25 years instead of 20 years or so. this is my humble opinion.

that's a terrible idea. buying cheapo caps is like playing russian roulette with the device. if you don't have problems with them, you're just lucky.

[VEGETA]

caps quality are over-exaggerated by designers very often. Getting cheap elec. caps and ceramic caps from LCSC is ok unless you need very specific low ESR caps and so on. I don't care about how Rubicon elec. cap is good since it costs a fortune for very little improvement if any... except that it can last 25 years instead of 20 years or so. this is my humble opinion.

that's a terrible idea. buying cheapo caps is like playing russian roulette with the device. if you don't have problems with them, you're just lucky.

You can't be lucky all the time. putting multiples in parallel like you should always helps. I see designers using cheap caps and resistors more and more than before.

what real issue you are concerned with? forget about low esr caps or similar stuff which requires high quality caps... I am talking about PSU caps, decoupling, bypassing, LC filters, etc... assuming heat is not an issue, which should be the case.

There is a case where I don't recommend using them, it is when you design +1000$ instruments like those very precise lab PSUs where such caps can make a difference, as small as it is. As someone says "in precise lab psus your job is to get rubicon caps" or something similar.

Top brands are Rubycon, Nichicon, Panasonic, Chemi-Con... and stuff like Kemet, TDK, Vishay, etc come after those but still good quality. Another step below is Lenon which is not considered low quality but not high quality.

Funny thing is, you can get TDK, Kemet, Murata, etc.. very very cheaply in LCSC these days. the only costly stuff are big elec. caps.

All and all, if you take care of heat and maybe ripple current, it is probably gonna be fine.

___

Example: GRM155R71C104KA88D from Murata which is 0.1uF 16v 0402. Price for 1000 quantity is 1.6$ from LCSC and 5.38$ from Digikey. I think it is cheap enough while being high quality. I don't think LCSC has bad parts here or fake ones, it is not Aliexpress. You DON"T buy good brands from Aliexpress!!

this is what I meant previously, you can get very cheap parts from LCSC especially in bulk while being high quality reputable brands as well.

[vol.2]

I'm not even going to justify that by arguing with you about it. There's no way Marqs is going to put rubbish caps in such a premium device to save some pennies, and people wouldn't accept it, no matter what you believe.

[VEGETA]

I'm not even going to justify that by arguing with you about it. There's no way Marqs is going to put rubbish caps in such a premium device to save some pennies, and people wouldn't accept it, no matter what you believe.

I didn't tell him to put crappy caps but rather cheaper ones to save some cost for non-sensitive components. For OSSC PRO which is not a business, you can do whatever you want but when you sell for business it is typical to optimize for cost. Get the cheapest good Capacitor you can get, not just cheap. It has to be good.

[underage]

marqs, could the OSSC Pro adapt a signal to make it FreeSync compatible? For use with a standard Freesync capable monitor or TV, like the recent OLED TVs.

[maxtherabbit]

except that it can last 25 years instead of 20 years or so

an additional 25% service life is a pretty big deal

[mikechi2]

except that it can last 25 years instead of 20 years or so

an additional 25% service life is a pretty big deal

There's no good reason, imo., for using electrolytic caps (they're also useless for high frequency decoupling) in a modern low-voltage system. Unless the device is price sensitive, spend an extra dollar on ceramic caps and one less problem to worry about. And spare the poor chump that would otherwise have to recap the PCB in 25 years, lol.

[VEGETA]

except that it can last 25 years instead of 20 years or so

an additional 25% service life is a pretty big deal

There's no good reason, imo., for using electrolytic caps (they're also useless for high frequency decoupling) in a modern low-voltage system. Unless the device is price sensitive, spend an extra dollar on ceramic caps and one less problem to worry about. And spare the poor chump that would otherwise have to recap the PCB in 25 years, lol.

Elec. caps are still good despite no designer loves them. bulk capacitance cannot be done without them, how can you add say 1000uF using ceramics? They are good for having a lot of filtering for DC-DC converters especially those with low to mid frequency ripple, combined with inductors and ceramic caps of lower values you will get decent results.

I don't have enough data to judge that we don't need elec. caps anymore TBH, but I know for a fact they are useful in power supply filtering circuits. I mean, no one use them to decouple video ICs xD.

But as you said, they are not useful in high frequency. For that you would need ceramics and ferrites.

[mikechi2]

No offense, but if you're designing something like a video converter that needs 1 mF of decoupling, you're doing something wrong.

[vol.2]

there are also applications where ceramics are a bad choice because of the piezo effect. i deal with circuits that are noise sensitive to the point where that limits me. i generally can't use electrolytics though because of the size. mostly end up using tantalum or film $$$

[VEGETA]

No offense, but if you're designing something like a video converter that needs 1 mF of decoupling, you're doing something wrong.

you cannot guarantee the input supply, the ac to dc converter. unless you designed it yourself which is not possible, at least in small designs. It also depends on your standards, like how much ripple and noise can you tolerate. elec. caps exist for a reason, you don't need 1mF all the time but if you are designing a Pi or good LC filter you are going to need some capacitance there. if your supply is kinda noisy or you want very little noise\ripple you will need to put some good filtering circuits. this is my opinion.

I saw people considering 100 mVp-p noise and ripple to be good, but I am definitely disagree.

Hope this clears it. Thanks

[mikechi2]

No offense, but if you're designing something like a video converter that needs 1 mF of decoupling, you're doing something wrong.

you cannot guarantee the input supply, the ac to dc converter. unless you designed it yourself which is not possible, at least in small designs. It also depends on your standards, like how much ripple and noise can you tolerate. elec. caps exist for a reason, you don't need 1mF all the time but if you are designing a Pi or good LC filter you are going to need some capacitance there. if your supply is kinda noisy or you want very little noise\ripple you will need to put some good filtering circuits. this is my opinion.

I saw people considering 100 mVp-p noise and ripple to be good, but I am definitely disagree.

Hope this clears it. Thanks

If you're relying on 1 mF of electrolytic capacitance (or really any electrolytic cap at all in a non-price sensitive design) to clean up a noisy supply on a low-voltage design, something isn't right. Not trying to be argumentative, just trying to help you with your design process.

[VEGETA]

No offense, but if you're designing something like a video converter that needs 1 mF of decoupling, you're doing something wrong.

you cannot guarantee the input supply, the ac to dc converter. unless you designed it yourself which is not possible, at least in small designs. It also depends on your standards, like how much ripple and noise can you tolerate. elec. caps exist for a reason, you don't need 1mF all the time but if you are designing a Pi or good LC filter you are going to need some capacitance there. if your supply is kinda noisy or you want very little noise\ripple you will need to put some good filtering circuits. this is my opinion.

I saw people considering 100 mVp-p noise and ripple to be good, but I am definitely disagree.

Hope this clears it. Thanks

If you're relying on 1 mF of electrolytic capacitance to clean up a noisy supply on a low-voltage design, something isn't right. Not trying to be argumentative, just trying to help you with your design process.

ok, what solution do you propose?

I usually prefer doing at least one stage LC filter with on elec. cap at the end (as big as I can afford) then finally a ferrite. this is typical of the least you can do to ensure low noise on any supply source.

I don't put 1000uF always, rather a 47u or something similar. However, if I have the space and the cost isn't an issue nor the design (not coming after LDO = doesn't affect its feedback response) then going for bigger cap is good.

I use ceramics everywhere else, including the LC or Pi filter itself... multiples with multiple values. I just like to put one big filtering elec. cap to ensure best circumstances to fight low frequency ripple and noise.

I've done a design of very low Dreamcast power supply (similar shape to DreamPSU) which I used only 1 47uF elec cap for each rail besides the ceramics. I could have used Ceramics only.

However, since the input is 12v from any PSU you have = quality is not guaranteed... I had to put good filtering circuit for the 12v rail and input. Eventually, all worked very well!

I know this example is not similar to low voltage video design, but still won't hurt to do it properly. I mean, what if I had a crappy supply only? this way all noise will go in if no proper filtering is done.

I really wish there was a thread dedicated to such discussions and solutions from engineers who build and test such stuff... it will help all designers.

[underage]

Seems to me you have never seen a LG 2020/2021 OLED with the right settings. When running matched content (60fps @60Hz and 120fps @120Hz) with OLED Motion Pro @high is amazing. When running with the wrong settings motion is just the usual blur. I have 5 CRTs myself so I know what perfect motion should look like.

Sorry, but you got me curious.
Is it even advisable to use Motion Pro @high with 120fps content? I thought the 'high' setting corresponds to 60hz BFI, so wouldn't that just fully blackout every 2nd frame with 120fps content?

[ldeveraux]

I really wish there was a thread dedicated to such discussions and solutions from engineers who build and test such stuff... it will help all designers.

Then start one so we don't clog up this thread with side-questions.

[H6rdc0re]

You can use this to see the differences across different display technologies and flickering/strobing techniques more easily: https://www.testufo.com/blackframes

Persistence blur is measurable in MPRT (moving picture response time). If CRT has a persistence blur of 1ms (60fps content), it's known that BFI at its best (duty cycle matching content FPS) is 4x the persistence blur of CRT (4ms). That is objectively worse than CRT and should be noticeable to any attentive eye, but certainly an improvement over the 16ms persistence blur without any flicker/strobing at all.

I made a picture of my LG CX running that image. The image with 60fps @60Hz is absolutely perfect in motion. See the picture in the link. https://ibb.co/9GJ7JGw I don't see how you could improve on that.

Maybe a bit of confusion going on here, but sample-and-hold is the display type, not a technique for addressing persistence blur. All sample-and-hold displays (LCD and OLED) have 16ms persistence blur and need some kind of flicker, strobing, scanning, etc. technique to improve it. This article is pretty good at giving some perspective on what it takes to get to actual CRT motion clarity for all cases (basically need 1000Hz refresh rates, so we still have a long way to go): https://blurbusters.com/blur-busters-la ... -and-hold/

And no worries, I don't think this is that off-topic, since the OSSC Pro and such are meant to provide CRT alternatives, and pairing them with OLED displays is the best option available for the next few years, but expectations should be set accurately.

I know the LCDs and OLEDs are sample-and-hold but in time manufacturers could emulate a CRT lightgun seeing how OLED is self-emitting and the response time is quicker than CRT @0.001ms compared to 0.01ms. Seeing how BFI already performs I doubt something like that is needed. Further development in light output should be more than enough.

Sorry, but you got me curious.
Is it even advisable to use Motion Pro @high with 120fps content? I thought the 'high' setting corresponds to 60hz BFI, so wouldn't that just fully blackout every 2nd frame with 120fps content?

You can use Motion Pro @high on both 60fps and 120fps. Just be sure to use matched refreshrates. I've added a link to show how BFI works in both 60Hz and 120Hz. https://i.rtings.com/assets/products/Jz ... -large.jpg

[dojima]

I made a picture of my LG CX running that image. The image with 60fps @60Hz is absolutely perfect in motion. See the picture in the link. https://ibb.co/9GJ7JGw I don't see how you could improve on that.

Anyone could take a picture with a still camera at like 1/1000s and show zero blur on a television. In order to really show what the TV can do, you'd have to use a methodology like RTINGS uses here. I agree with you that BFI on OLEDs works fantastically and that motion resolution rivals CRTs, but your image doesn't prove anything.

[H6rdc0re]

I made a picture of my LG CX running that image. The image with 60fps @60Hz is absolutely perfect in motion. See the picture in the link. https://ibb.co/9GJ7JGw I don't see how you could improve on that.

Anyone could take a picture with a still camera at like 1/1000s and show zero blur on a television. In order to really show what the TV can do, you'd have to use a methodology like RTINGS uses here. I agree with you that BFI on OLEDs works fantastically and that motion resolution rivals CRTs, but your image doesn't prove anything.

Seriously!? This picture is taken with my Iphone 6, nothing special or any trickery.

[fernan1234]

Anyone could take a picture with a still camera at like 1/1000s and show zero blur on a television. In order to really show what the TV can do, you'd have to use a methodology like RTINGS uses here. I agree with you that BFI on OLEDs works fantastically and that motion resolution rivals CRTs, but your image doesn't prove anything.

Yeah it's not accurate to try to compare persistence blur using still pictures. The only ways to compare are with your own eyes and with measurements, which have already been taken for BFI vs CRT (4ms vs 1ms).

Personally I always have to question statements like this because I'm not sure what "rivals" here exactly means, and when you see this BFI next to an actual CRT the difference is significant and reflects the objective measurements, which correspond to the fact that BFI has 4 times the persistence blur of CRT. That IMO doesn't "rival" CRT. But, it does get you much closer than any other consumer-level solution, and destroys the 16X persistence blur you'd get on a sample-and-hold display without it.

I don't see how you could improve on that.

When you see a rolling scan/bar flicker/strobe method you'll see how it can be improved (I estimate about 2x improved, but don't have measurements to back that up). You can learn more about how this would be an improvement and why it's been a requested feature for RA here: https://github.com/libretro/RetroArch/issues/10757

[H6rdc0re]

When you see a rolling scan/bar flicker/strobe method you'll see how it can be improved (I estimate about 2x improved, but don't have measurements to back that up). You can learn more about how this would be an improvement and why it's been a requested feature for RA here: https://github.com/libretro/RetroArch/issues/10757

What are you talking about? Every test pattern is completely sharp with no blur what so ever. Are you ignoring my photo? Here's 2 more also taken with the same Iphone.
https://ibb.co/nc1dGNC
https://ibb.co/Ntc3V2g

Perhaps you haven't seen a 2020/2021 OLED with the right settings. When using the wrong settings, it's blur city.