31KHz CRT: 240p120 vs. 240p120 w/BFI vs. 480p60 w/ Scanlines

The place for all discussion on gaming hardware
SuperSpongo
Posts: 315
Joined: Sat Mar 17, 2018 2:49 pm
Location: Germany

Re: 31KHz CRT: 240p120 vs. 240p120 w/BFI vs. 480p60 w/ Scanl

Post by SuperSpongo »

Another minor disadvantage of the potential OSSC Pro 240p120 mode is that it introduces at least half a frame of lag. Don't know, if this has been mentioned here yet.
User avatar
Kez
Posts: 818
Joined: Thu Jul 20, 2017 7:09 am

Re: 31KHz CRT: 240p120 vs. 240p120 w/BFI vs. 480p60 w/ Scanl

Post by Kez »

Josh128 wrote:From what I read, the 240p120 w/BFI is supposed to work on 120Hz LCDs and people already use it with RA, but I cant confirm personally if its 240p or some higher resolution at 120Hz.
I can't imagine that looking very good even if you can get a monitor to display it.
User avatar
Josh128
Posts: 2123
Joined: Thu Jan 16, 2014 9:01 am

Re: 31KHz CRT: 240p120 vs. 240p120 w/BFI vs. 480p60 w/ Scanl

Post by Josh128 »

SuperSpongo wrote:Another minor disadvantage of the potential OSSC Pro 240p120 mode is that it introduces at least half a frame of lag. Don't know, if this has been mentioned here yet.
Yeah Marqs mentioned it in the OSSC Pro thread. He also mentioned that adding motion interpolation would add some additional lag on top of the 8ms, but he didnt specify how much. He also didnt mention that if BFI were implemented in 120Hz mode if it would cause additional lag or the same 8ms as the native 240p120 mode.
Last edited by Josh128 on Thu Feb 13, 2020 4:37 pm, edited 1 time in total.
User avatar
Xer Xian
Posts: 881
Joined: Sun Feb 06, 2005 3:23 pm
Location: Italy

Re: 31KHz CRT: 240p120 vs. 240p120 w/BFI vs. 480p60 w/ Scanl

Post by Xer Xian »

Kez wrote:At 240p, the gun is drawing a scanline every frame and it takes around 1/240th (-h blank) of a frame to move from right to left.

At 480p + scanlines, in the same period the gun draws two scanlines.. the actual image and then a "blank" line. It is spending half as long actually lighting up phosphors, so the image is darker.
Your analysis is correct, but as you said yourself, when displaying 480p the beam is moving at double the frequency, therefore the end result will be practically the same - after all, either at 240p or at 480p w/ scanlines, there are still 240 active lines being drawn every 1/60 of a second..

About your initial suggestion of using higher line multiples than 2x - of course they would look brighter (960p w/ 100% scanlines in particular being as bright as 240p120), and if the CRT no longer shows any space in between the scanlines (ie no longer shows 'natural' blanklines) at the resulting total vertical resolution and the multiple is an even number then there should not be any visible shortcoming that I can think of.. the look overall might be different than at 240p120 though. Scanlines and artificial blanklines would have nearly the same width (scanlines might still be thicker due to the effect described by Unseen in the first page of the thread). If anyone thinks something is wrong about my reasoning - and isn't yet fed up to death of this debate - please speak up :)
User avatar
Kez
Posts: 818
Joined: Thu Jul 20, 2017 7:09 am

Re: 31KHz CRT: 240p120 vs. 240p120 w/BFI vs. 480p60 w/ Scanl

Post by Kez »

Xer Xian wrote:Your analysis is correct, but as you said yourself, when displaying 480p the beam is moving at double the frequency, therefore the end result will be practically the same - after all, either at 240p or at 480p w/ scanlines, there are still 240 active lines being drawn every 1/60 of a second..
240 lines every 1/60 of a second, yes.. but the proportion of that time spent actually emitting light is effectively halved at 480p w/ scanlines.

I don't have a CRT PC monitor but I do have a BVM D20 and an OSSC, at some point I will get it hooked up to do some experiments with the different line multipliers and scanlines. I will say that I have tried my N64Advanced at both 240p and 480p w/scanlines just out of curiosity, and the latter was disappointing.
User avatar
Josh128
Posts: 2123
Joined: Thu Jan 16, 2014 9:01 am

Re: 31KHz CRT: 240p120 vs. 240p120 w/BFI vs. 480p60 w/ Scanl

Post by Josh128 »

Kez wrote:
Xer Xian wrote:Your analysis is correct, but as you said yourself, when displaying 480p the beam is moving at double the frequency, therefore the end result will be practically the same - after all, either at 240p or at 480p w/ scanlines, there are still 240 active lines being drawn every 1/60 of a second..
240 lines every 1/60 of a second, yes.. but the proportion of that time spent actually emitting light is effectively halved at 480p w/ scanlines.
Precisely. At 240p120 native, the electron beam never turns off, except during blanking when retracing the from the end of one porch to the beginning of the other. Same as all native resolutions without artificial scanlines. At 480p60 w/ artificial scanlines, the color guns are all going 0 or turning off every other horizontal scan. Assuming the power, in watts, to the electron beam remains constant regardless of the resolution or refresh rate, 480p60 w/ artificial scanlines is effectively transferring half of the power or energy to the phosphor substrate that it does at the 240p120 native.
User avatar
Xer Xian
Posts: 881
Joined: Sun Feb 06, 2005 3:23 pm
Location: Italy

Re: 31KHz CRT: 240p120 vs. 240p120 w/BFI vs. 480p60 w/ Scanl

Post by Xer Xian »

Kez wrote:240 lines every 1/60 of a second, yes.. but the proportion of that time spent actually emitting light is effectively halved at 480p w/ scanlines.
I think that would be significant only if the phosphors had an exceptionally short persistence. In general, the phosphors stay lit for at least one or two milliseconds while the lines are scanned in the matter of microseconds, so the time 'lost' in scanning blank lines should have a negligible impact.

I should test this on my multisync monitor too, though.
User avatar
Josh128
Posts: 2123
Joined: Thu Jan 16, 2014 9:01 am

Re: 31KHz CRT: 240p120 vs. 240p120 w/BFI vs. 480p60 w/ Scanl

Post by Josh128 »

Xer Xian wrote:
Kez wrote:240 lines every 1/60 of a second, yes.. but the proportion of that time spent actually emitting light is effectively halved at 480p w/ scanlines.
I think that would be significant only if the phosphors had an exceptionally short persistence. In general, the phosphors stay lit for at least one or two milliseconds while the lines are scanned in the matter of microseconds, so the time 'lost' in scanning blank lines should have a negligible impact.

I should test this on my multisync monitor too, though.
The reason for the much brighter image is very simple- you are scanning the same area with twice the electron beam power per second.
User avatar
Xer Xian
Posts: 881
Joined: Sun Feb 06, 2005 3:23 pm
Location: Italy

Re: 31KHz CRT: 240p120 vs. 240p120 w/BFI vs. 480p60 w/ Scanl

Post by Xer Xian »

Yeah, but addressing the question asked by Shelcoof, Kez and I were talking about the brightness level of 480p w/ scanlines as opposed to standard 240p60, not 240p120.. OT closed here on my part.
User avatar
Josh128
Posts: 2123
Joined: Thu Jan 16, 2014 9:01 am

Re: 31KHz CRT: 240p120 vs. 240p120 w/BFI vs. 480p60 w/ Scanl

Post by Josh128 »

I see. I could test 480p vs 960p w/ scanlines if I had an emulator or program that would blank every other line on the 960p. Not sure if I can get that. It would be interesting, I would expect them to be the same.

BTW, heres a 1/4000 sec shot of the scanline and phosphor persistence of my monitor at 240p120 vs 480p60 w/ scanlines. From these images its very easy to see why the 240p120 is so much brighter.

480p60 w/ scanlines
Full Size - https://i.imgur.com/E43l7z8.jpg
Image

240p120
Full Size - https://i.imgur.com/3rM3BaN.jpg
Image
User avatar
Shelcoof
Posts: 1520
Joined: Mon Nov 03, 2008 9:36 pm
Location: Canada

Re: 31KHz CRT: 240p120 vs. 240p120 w/BFI vs. 480p60 w/ Scanl

Post by Shelcoof »

Love the whole discussion. Unfortunately all I can do is stand by and read everything :lol:

You know... this all stems from my desire to use a PC monitor to play most of my consoles with scanlines. Its the closest thing to a PVM/BVM for me. I get a great crips picture with a PC CRT monitor but if I add in 100% scanlines I get a dark picture. Without scanlines its bright enough but is missing the details gained from the scanlines.

Best alternative so far is to run 25% scanlines for 480p and it gives me enough of a 240p look while being bright enough. Would love those 100% thick scanlines but if the image is too dark forget about it.
SuperSpongo
Posts: 315
Joined: Sat Mar 17, 2018 2:49 pm
Location: Germany

Re: 31KHz CRT: 240p120 vs. 240p120 w/BFI vs. 480p60 w/ Scanl

Post by SuperSpongo »

You should try the OSSC's hybrid scanlines then :)
User avatar
Josh128
Posts: 2123
Joined: Thu Jan 16, 2014 9:01 am

Re: 31KHz CRT: 240p120 vs. 240p120 w/BFI vs. 480p60 w/ Scanl

Post by Josh128 »

Shelcoof wrote:Love the whole discussion...

You know... this all stems from my desire to use a PC monitor to play most of my consoles with scanlines. Its the closest thing to a PVM/BVM for me. I get a great crips picture with a PC CRT monitor but if I add in 100% scanlines I get a dark picture. Without scanlines its bright enough but is missing the details gained from the scanlines.

Best alternative so far is to run 25% scanlines for 480p and it gives me enough of a 240p look while being bright enough. Would love those 100% thick scanlines but if the image is too dark forget about it.
You would love the 240p120 image. BVM+ quality on fire sale pricing CRTs. If the double strobing illusion on bright, fast moving objects doesnt bother you, you'd be set. If it does, and motion interpolation is an option that could be implemented without adding too much additional latency, there you go. Done deal.

Honestly, for me, most games I've tried are fine with the native mode. I can see the double strobe effect at times if I look for it on some games, some games not at all. If you do see it, sometimes its hard to "unsee" it and you find your eyes going back to it. I find its more visible on bright, fast scenes and less visible on darker scenes regardless of speed. The effect is exactly the same as what you see in any 32 bit generation 30fps game, except less pronounced because its on 60fps vs 30fps. Some people may find it more objectionable than others, fine, understood.

For me personally, the amount of total latency Im willing to accept for ANY mode on a CRT from a console is 2 frames or 32ms. Marqs has already stated that 240p120 native will require 1/2 frame, or 8ms latency. If the motion interpolation could be accomplished in an additional frame to frame and a half, total latency would be 24ms to 32ms. Being that the CRT will not add anything at all to that, it would be acceptable to me. Anything over 32ms, IMO, just wouldnt be worth the effort because its too much of a tradeoff.

Regardless, if Marqs and the devs can make 240p120 native happen, they should be considered heroes. I'd be great with that. Anything on top of it, like interpolation, is just icing on the cake to me. Props to all the devs working to make this happen.
User avatar
Shelcoof
Posts: 1520
Joined: Mon Nov 03, 2008 9:36 pm
Location: Canada

Re: 31KHz CRT: 240p120 vs. 240p120 w/BFI vs. 480p60 w/ Scanl

Post by Shelcoof »

SuperSpongo wrote:You should try the OSSC's hybrid scanlines then :)
I have played around with it and it just adds more darkness. I'll stick to the regular Horizontal scanlines.


Josh128 wrote:
Shelcoof wrote:Love the whole discussion...

You know... this all stems from my desire to use a PC monitor to play most of my consoles with scanlines. Its the closest thing to a PVM/BVM for me. I get a great crips picture with a PC CRT monitor but if I add in 100% scanlines I get a dark picture. Without scanlines its bright enough but is missing the details gained from the scanlines.

Best alternative so far is to run 25% scanlines for 480p and it gives me enough of a 240p look while being bright enough. Would love those 100% thick scanlines but if the image is too dark forget about it.
You would love the 240p120 image. BVM+ quality on fire sale pricing CRTs. If the double strobing illusion on bright, fast moving objects doesnt bother you, you'd be set. If it does, and motion interpolation is an option that could be implemented without adding too much additional latency, there you go. Done deal.

Honestly, for me, most games I've tried are fine with the native mode. I can see the double strobe effect at times if I look for it on some games, some games not at all. If you do see it, sometimes its hard to "unsee" it and you find your eyes going back to it. I find its more visible on bright, fast scenes and less visible on darker scenes regardless of speed. The effect is exactly the same as what you see in any 32 bit generation 30fps game, except less pronounced because its on 60fps vs 30fps. Some people may find it more objectionable than others, fine, understood.

For me personally, the amount of total latency Im willing to accept for ANY mode on a CRT from a console is 2 frames or 32ms. Marqs has already stated that 240p120 native will require 1/2 frame, or 8ms latency. If the motion interpolation could be accomplished in an additional frame to frame and a half, total latency would be 24ms to 32ms. Being that the CRT will not add anything at all to that, it would be acceptable to me. Anything over 32ms, IMO, just wouldnt be worth the effort because its too much of a tradeoff.

Regardless, if Marqs and the devs can make 240p120 native happen, they should be considered heroes. I'd be great with that. Anything on top of it, like interpolation, is just icing on the cake to me. Props to all the devs working to make this happen.

I'm not sure if my consoles or set-up would allow 240p120. I play mostly on original hardware so I assume this could only be achieved via emulation on PC.

If the strobing is anything like motion blur or pixel response time on an LCD it would totally bother me. The biggest reason why I enjoy playing on a CRT is because I find the blur affect from motion really bothering on LCD. Example would be playing the original Sonic games only to have a blurry screen the entire time.

I've always been bothered by this... it drives me nuts lol
User avatar
orange808
Posts: 3196
Joined: Sat Aug 20, 2016 5:43 am

Re: 31KHz CRT: 240p120 vs. 240p120 w/BFI vs. 480p60 w/ Scanl

Post by orange808 »

Shelcoof wrote:
SuperSpongo wrote:You should try the OSSC's hybrid scanlines then :)
I have played around with it and it just adds more darkness. I'll stick to the regular Horizontal scanlines.
Are you certain?

Increasing the hybrid scanlines (in either multiplication and subtraction mode) should increase brightness on any display. I've never heard of any display getting darker and I can't think of any reasonable scenarios where it would occur.

Maybe you have vertical scanlines and hybrid scanlines confused?
We apologise for the inconvenience
User avatar
Josh128
Posts: 2123
Joined: Thu Jan 16, 2014 9:01 am

Re: 31KHz CRT: 240p120 vs. 240p120 w/BFI vs. 480p60 w/ Scanl

Post by Josh128 »

Shelcoof wrote:
I'm not sure if my consoles or set-up would allow 240p120. I play mostly on original hardware so I assume this could only be achieved via emulation on PC.

If the strobing is anything like motion blur or pixel response time on an LCD it would totally bother me. The biggest reason why I enjoy playing on a CRT is because I find the blur affect from motion really bothering on LCD. Example would be playing the original Sonic games only to have a blurry screen the entire time.

I've always been bothered by this... it drives me nuts lol
Thats where the OSSC Pro unit thats currently being developed comes in. This thread stemmed from a question/request I had on the OSSC Pro thread. The unit will be able to take a standard SNES that has been RGB modded, and convert it to 240p120 to be displayed directly on an SVGA CRT monitor.

ALSO-- should you choose to stick with LCD, the OSSC Pro could also potentially put out a 120Hz BFI mode, which would completely eliminate the blurring that bothers you on LCD. I'm like you, I cant stand the blur of retro systems 2D games and for that very reason I only game on plasma or CRT. I still use LCD IPS on my PC, and I find its acceptable for modern 3D PC games like Doom 2016 and such, but yeah, this new OSSC will theoretically be capable of doing both these things and more. I'm very excited about it.
User avatar
Shelcoof
Posts: 1520
Joined: Mon Nov 03, 2008 9:36 pm
Location: Canada

Re: 31KHz CRT: 240p120 vs. 240p120 w/BFI vs. 480p60 w/ Scanl

Post by Shelcoof »

orange808 wrote:
Shelcoof wrote:
SuperSpongo wrote:You should try the OSSC's hybrid scanlines then :)
I have played around with it and it just adds more darkness. I'll stick to the regular Horizontal scanlines.
Are you certain?

Increasing the hybrid scanlines (in either multiplication and subtraction mode) should increase brightness on any display. I've never heard of any display getting darker and I can't think of any reasonable scenarios where it would occur.

Maybe you have vertical scanlines and hybrid scanlines confused?
Oh yhaa possibly lol.

However though when I do play around with what you are talking about I find it messes up the color and it looks off. Maybe I'm not setting it up right?






Josh128 wrote:
Shelcoof wrote:
I'm not sure if my consoles or set-up would allow 240p120. I play mostly on original hardware so I assume this could only be achieved via emulation on PC.

If the strobing is anything like motion blur or pixel response time on an LCD it would totally bother me. The biggest reason why I enjoy playing on a CRT is because I find the blur affect from motion really bothering on LCD. Example would be playing the original Sonic games only to have a blurry screen the entire time.

I've always been bothered by this... it drives me nuts lol
Thats where the OSSC Pro unit thats currently being developed comes in. This thread stemmed from a question/request I had on the OSSC Pro thread. The unit will be able to take a standard SNES that has been RGB modded, and convert it to 240p120 to be displayed directly on an SVGA CRT monitor.

ALSO-- should you choose to stick with LCD, the OSSC Pro could also potentially put out a 120Hz BFI mode, which would completely eliminate the blurring that bothers you on LCD. I'm like you, I cant stand the blur of retro systems 2D games and for that very reason I only game on plasma or CRT. I still use LCD IPS on my PC, and I find its acceptable for modern 3D PC games like Doom 2016 and such, but yeah, this new OSSC will theoretically be capable of doing both these things and more. I'm very excited about it.
I have a very fast 1ms response time on my BenQ @60hz and I have a Panasonic Plasma as well and both I find blurry.

I thought the blur was due to the technology of the liquid crystals inability to change fast enough which causes blur on fast moving scenes?

The one thing I love about an LCD Panel is the crystal clear "Bright" image and colors have gotten so much better as well. I find scanlines don't really affect the brightness on a flat panel either.

If what your telling me is true I might finally move on to flat panel displays.
User avatar
Josh128
Posts: 2123
Joined: Thu Jan 16, 2014 9:01 am

Re: 31KHz CRT: 240p120 vs. 240p120 w/BFI vs. 480p60 w/ Scanl

Post by Josh128 »

Shelcoof wrote:

I have a very fast 1ms response time on my BenQ @60hz and I have a Panasonic Plasma as well and both I find blurry.

I thought the blur was due to the technology of the liquid crystals inability to change fast enough which causes blur on fast moving scenes?
For LCD, its due to sample and hold method of updating the screen rather than pulsing the screen as in CRT. Plasma is kind of a hybrid of the two methods. It "blurs" a bit more than CRT but definitely less than even a 1ms LCD. However, if the BFI technique is used with the 1ms LCD, the result can better the plasma and come very close to CRT. A true 120Hz or better capable LCD is required for this. Some newer panels also have built in backlight strobing to aid with blurring as well, but Im not sure how configurable it is and if it can be used effectively on game console content.

Shelcoof wrote: The one thing I love about an LCD Panel is the crystal clear "Bright" image and colors have gotten so much better as well. I find scanlines don't really affect the brightness on a flat panel either.

If what your telling me is true I might finally move on to flat panel displays.
For brightness, LCDs are unmatched. They are easily the brightest displays by far, and their light is produced completely independent from how they produce their color image and this is why you can get full artificial scanlines with an acceptably bright picture easily on most LCDs. They are a transmissive display in regards to the colors you see rather than an emissive display such as OLED, CRT, or plasma. Which is also why their contrast in dark rooms is still not nearly as good as those three types.

But yes, its going to be possible on the new OSSC Pro, its just a matter of devs implementing it.
User avatar
Shelcoof
Posts: 1520
Joined: Mon Nov 03, 2008 9:36 pm
Location: Canada

Re: 31KHz CRT: 240p120 vs. 240p120 w/BFI vs. 480p60 w/ Scanl

Post by Shelcoof »

Josh128 wrote: But yes, its going to be possible on the new OSSC Pro, its just a matter of devs implementing it.
I wasn't too hyped about the OSSC Pro but I sort of am now. I didn't think it could get any better than the original OSSC so I'm getting excited.



orange808 wrote:
I've played with the Hybrid scanline option again and it does help increase the brightness a bit. I'm starting to fine tune each settings.

I find the ratio of 2:1 is good example 50% to 25%hybrid scanline or 100% to 50% hybrid scanline.

I find this ratio gives you enough scanline details without sacrificing brightness. I think this is a better option than 25% scanline because you get more details.

I have a few CRT PC Monitors that has a brightness boost and that helps a ton as well. Feeling happier :D
SuperSpongo
Posts: 315
Joined: Sat Mar 17, 2018 2:49 pm
Location: Germany

Re: 31KHz CRT: 240p120 vs. 240p120 w/BFI vs. 480p60 w/ Scanl

Post by SuperSpongo »

There you go :)

Concerning the colors: I think Borti mentioned that the substraction method will mess them up which is why the multiplication method was introduced. I don't remember the details, but he took a couple of comparison shots. They should be somewhere in the original OSSC thread.
User avatar
kitty666cats
Posts: 1270
Joined: Tue Nov 05, 2019 2:03 am
Location: Massachusetts, USA

Re: 31KHz CRT: 240p120 vs. 240p120 w/BFI vs. 480p60 w/ Scanl

Post by kitty666cats »

Those concerned about brightness in scanline mode should definitely consider the X-Vue Gamma Boost, I randomly got one on eBay for cheap and it helps quite a lot with the darkness when generating scanlines
User avatar
Shelcoof
Posts: 1520
Joined: Mon Nov 03, 2008 9:36 pm
Location: Canada

Re: 31KHz CRT: 240p120 vs. 240p120 w/BFI vs. 480p60 w/ Scanl

Post by Shelcoof »

kitty666cats wrote:Those concerned about brightness in scanline mode should definitely consider the X-Vue Gamma Boost, I randomly got one on eBay for cheap and it helps quite a lot with the darkness when generating scanlines
I asked you in the other thread but would you be able to do up some comparison shots?
Or let us know if the gamma boost affects color accuracy or not? I'm having a hard time finding the correct terms but when brightness increases I find colors get washed out and you get less detail.

This happens on my HD Fury Nano when I increase the gamma brightness to max. Its not terrible by any means but I'd rather play without scanlines than ruin the colors.
User avatar
kitty666cats
Posts: 1270
Joined: Tue Nov 05, 2019 2:03 am
Location: Massachusetts, USA

Re: 31KHz CRT: 240p120 vs. 240p120 w/BFI vs. 480p60 w/ Scanl

Post by kitty666cats »

Shelcoof wrote:
kitty666cats wrote:Those concerned about brightness in scanline mode should definitely consider the X-Vue Gamma Boost, I randomly got one on eBay for cheap and it helps quite a lot with the darkness when generating scanlines
I asked you in the other thread but would you be able to do up some comparison shots?
Or let us know if the gamma boost affects color accuracy or not? I'm having a hard time finding the correct terms but when brightness increases I find colors get washed out and you get less detail.

This happens on my HD Fury Nano when I increase the gamma brightness to max. Its not terrible by any means but I'd rather play without scanlines than ruin the colors.
I returned mine because it seemed to have a defective PSU input. Plugged it into several different 5V power supplies that fit its port (that awkward TINYYY barrel connector used by old Nokia phones) and the power indicator would not light up at all. The only time I got it to light up and work is when connected to the VGA output of my Portta HDMI to VGA.

I found a detailed picture of the packaging for NIB Gamma Boosts and it says they don't accept YPbPr signals into or out of it, but I am real curious if it'd work out plugging one into the dsub15 port of my Kramer FC-14...

Anyhoo, there's a few cheap ($20 USD, shipping included) NIB sealed old stock Gamma Boosts available on eBay, and I am probably gonna scoop one up and see how a brand new unit treats me. I can snap some pictures once I get another one, will probably order tmrw :D
strayan
Posts: 671
Joined: Sun Mar 19, 2017 8:33 pm

Re: 31KHz CRT: 240p120 vs. 240p120 w/BFI vs. 480p60 w/ Scanl

Post by strayan »

Josh128 wrote:You would love the 240p120 image. BVM+ quality on fire sale pricing CRTs. If the double strobing illusion on bright, fast moving objects doesnt bother you, you'd be set. If it does, and motion interpolation is an option that could be implemented without adding too much additional latency, there you go. Done deal.
I finally got around to testing 240p120 on the cheapest PC CRT I could find and by Jove, Josh is right, it looks extremely PVM-like https://i.imgur.com/4IjsmD3.jpg.

The double strobe does bother me but only on some games, exactly like he says.

I too will be able to deal with the input latency added by BFI when I've got a PVM quality picture on a junk yard 31KHz CRT.
User avatar
Josh128
Posts: 2123
Joined: Thu Jan 16, 2014 9:01 am

Re: 31KHz CRT: 240p120 vs. 240p120 w/BFI vs. 480p60 w/ Scanl

Post by Josh128 »

^^

lol Looks amazing doesnt it? :mrgreen: The standard 240p120 without BFI has the most beautiful image quality at the cost of the double strobing illusion on fast moving graphics. BFI will fix the strobing illusion at the cost of brightness. Perhaps an interpolation mode from the OSSC Pro could give us the standard 240p120 image quality with no strobing at the cost of a bit of lag. Only question is how much? Depending, that mode could be the "holy grail".

Im currently looking at the same technique for displaying medium res 25KHz CGA arcade games.
Spoiler
Its looking very promising!
Blur Busters
Posts: 2
Joined: Sat May 30, 2020 11:21 pm

Re: 31KHz CRT: 240p120 vs. 240p120 w/BFI vs. 480p60 w/ Scanl

Post by Blur Busters »

Chief Blur Buster. Founder of TestUFO here!

As the black frame insertion authority, we're happy to answer BFI-related questions with definitivity. In fact, I custom modified a TestUFO Black Frame test in order to "show and tell" to micdrop my answers here. Currently, Blur Busters is known as the famous "refresh rate micdropping factory" so we mythbust this shit all the time, whether it be "Human Eyes Cannot Tell 30fps vs 60fps" or any refresh rate myths.

Here's a double-image effect in software-based black frame insertion. View this TestUFO test on a non-flickering LCD!

TestUFO Animation: Software-Based Double Image BFI Demo

This works on all 60Hz LCDs / OLEDs!
- Runs 15fps on 60Hz LCDs
- Runs 30fps on 120Hz LCDs (best emulation of CRT 30fps at 60Hz)
- Runs 60fps on 240Hz LCDs

There are many adjustable settings at the top of the above TestUFO test, play with them, to get variable BFI effects. Even triple-image effect (like CRT 20fps at 60Hz) and variable-length BFI for 240Hz (e.g. 75%:25% duty cycle, 50%:50% duty cycle, 25%:75% duty cycle).

How did we make it work?

It's a 4-frame software-based 2-repeat sequence of 1 visible, 1 black frame, using a cadence of 100%,0%,100%,0% (of the same static frame) per 1/15sec on a 60Hz monitor. That generates a software-based double-image effect on nearly all display technologies (LCD, OLED, DLP, etc). The motion of human eye-tracking is actually the main cause of LCD motion blur, clearly demonstrated in animation demo http://www.testufo.com/eyetracking ...GtG pixel response is NOT the main cause anymore, it's static pixel visibility time from sample-and-hold. Common LCD motion blur is mathematically directly proportional to pixel visibility time (excluding GtG time), as explained in Blur Busters Law: The Amazing Journey To Future 1000Hz Displays as well as Pixel Response FAQ: GtG versus MPRT.

Terminologically, all impulse-based technologies are essentially a form of per-pixel BFI
For the purposes of this post, terminologically we will treat BFI and CRT impulse scanning as an overlapping venn diagram. CRT is terminologically a "phosphor-trailed rolling-scan BFI" as BFI doesn't need to be engineered as full frames, or even a 50%:50% refresh-cycle-granularity BFI. Normally we don't call CRT as BFI, because it's a natural inclusion. But we will, for the purposes of this post (from a human vision perspective), treat CRT a BFI behaviour that is essentially defacto 1%:99% ON:OFF per-pixel-based BFI. This depends on how slow/fast the CRT phosphor is, but you get the idea, brief pixel flash, lots of black time.

We don't necessarily think of BFI as integers, actually, and BFI can also work on VRR, and BFI can also work with rolling scans too. We also have a "Temporal HLSL" idea proposal at the GroovyMAME forum, to emulate a CRT electron gun via the sheer refresh rate of ultra-high refresh rate monitors (such as the upcoming DELL 360Hz IPS-panel monitor this year, which could theoretically do a 6-segment software-based rolling-scan BFI for a 60Hz emulator!)
Josh128 wrote:The trade-off of that is the double-image effect (and a handful of milliseconds of lag)

Fact: Not Necessarily

If programmed properly, there is no lag for 240p120 software-based BFI, as long as you display the visible frame on the first refresh cycle of the 2-frame sequence (100%,0%).
Xer Xian wrote:The double-image effect of repeated frames on a CRT cannot be alleviated with BFI.
Fact: FALSE

Yes, it can. View this on a 120Hz CRT: http://www.testufo.com/blackframes and you'll see single-image 60fps!

Detail: You have a single-image effect at 60fps on a 120Hz CRT when you combine software BFI with hardware BFI. Combining software BFI (software flicker) and hardware BFI (CRT flicker) to use software hide every other hardware flicker. This is the technique used with LightBoost displays, using software BFI to block every other 120Hz strobe. The problem is manufacturers don't want to manufacture 60Hz strobe backlights because of painful flicker, but it's actually a 1-line firmware change, as seen in this thread, Dear ViewSonic: Please Add 60 Hz Single-Strobe for PureXP on XG270. So we often have to combine software BFI with hardware BFI, to workaround a manufacturer limitation of a too-high-impulse-rate. For excess refresh rate, the software-based BFI only responsibility is only merely just simply to black-out unwanted refresh cycles. That's it!! When someone thinks simpler (E=mc^2 instead of complex calculus) and BFI becomes easier to understand. Play with those parameters at the very top of http://www.testufo.com/blackframes#easteregg=1 (especially on a 120Hz+ display), combined with a little study of recent Blur Busters display research and you may become a BFI expert within an hour.
Xer Xian wrote:thus losing information (even if redundant).
Fact: FALSE

Eliminating a repeat-refresh isn't loss of information. It actually adds information because of reduced display motion blur, because motion blur (LCD) + duplicate images (fps double Hz) is caused by eye-tracking across unwanted extra persistence. It's mathematically beautifully simple if you study up.
Xer Xian wrote:If you feed 240p120 to a modern display and activate 120hz BFI you will substitute one kind of vision artifact (image persistence) with another (double-image effect).
FACT: Sometimes true, sometimes false

Just fiddle with the parameters of the TestUFO test to make it appear and disappear on the same 120Hz+ LCD: http://www.testufo.com/blackframes#easteregg=1

Here are some great BFI demos:
TestUFO Animation: Variable BFI -- Demos variable amounts of blur against full-frame-rate
TestUFO Animation: Double Images -- Demos double image effect
TestUFO Animation: Triple Images -- Demos triple image effect (test at 120Hz+, flickers much 60Hz)
TestUFO Animation: Blue Frame Insertion -- Demos that blank frames don't need to be black

NOTE: For best TestUFO animation performances, make sure to run this in a browser capable of running framerate=Hz, though multimonitor sometimes interferes. To minimize stutters during this demo on some computers, you may need to close all tabs and run only on primary monitor. May need to test a different browser, FireFox vs Chrome.
Blur Busters
Posts: 2
Joined: Sat May 30, 2020 11:21 pm

Re: 31KHz CRT: 240p120 vs. 240p120 w/BFI vs. 480p60 w/ Scanl

Post by Blur Busters »

Let me know if you have any questions. Thanks!
Post Reply