External BFI Injection -- I am working with Retrotink 4K!

[Chief Blur Buster]

You will need to sacrifice 4K60 output to do 1080p240 output or 1440p120 output with a Retrotink. The BFI is adjustable in 25% (4ms), 50% (8ms), 75% (12ms) persistence when outputting 1080p240.

Chief, is 240Hz output required to unlock all 3 options? With 60Hz input, that is.

I'm confused because you've said the RT4K's BFI has to output 120Hz for 60Hz input (MitM limitation), which gives me the impression it can't modify those 60 frames, thus increases the refresh to enable insertions. 60/180 = 1:0:0 etc...

But following this logic, 120, 180 and 240Hz would all be locked to their own duty cycles. Where have I gone wrong?!

No, the duty cycle is more variable the larger ratio you get.

60fps at 180Hz would have the option of 1:1:0 and 1:0:0 duty cycles
60fps at 240Hz would have the option of 1:1:1:0 and 1:1:0:0 and 1:0:0:0 duty cycles
See TestUFO Variable-Persistence BFI as an example.

One other question: is this BFI as flickery as BenQ's old Z monitors with single strobing (at least aggressive)? I'd love my Z's 120Hz flicker without double images for 60fps games.

The longer the visible frame time (duty cycle), the less flickery. The old Z did subrefresh strobing, so at 2ms you get more motion blur with these levels of monolithic software BFI but you get less flicker with monolithic software BFI. You can't get less motionblur than minimum refreshtime (e.g. 1/240sec for 240Hz) for software BFI, while hardware strobing can go sub-refresh.

But, brute can help, e.g. 480Hz at 1:0:0:0:0:0:0:0 duty cycle would result in 2ms MPRT for 60Hz, which can outperform Z default settings for strobe. More Hz, the more duty cycle flexibility adjustment.

[roginthemachine]

Hey Chief, I just had a question about getting a particular feature of the Retrotink 4K with respect to motion clarity:

The Tink has a unique de-interlacing setting for 480i content called 'CRT Sim', which tries to recreate the interlaced look by inserting black scanlines in the gaps between the alternating fields, rather than simply filling in those gaps with a view to create convincing 480p like image.

I was curious whether enabling this 'CRT Sim' setting creates a benefit in motion clarity for 480i content, compared to say, the equivalent 480p like image, given that, I suppose, the persistence of each alternating field is cut in half.

In practice, I don't see much benefit - but perhaps it could be increasing the clarity by a small factor e.g. 25%?

Thanks for your help!

[RealNC]

Hey Chief, I just had a question about getting a particular feature of the Retrotink 4K with respect to motion clarity:

The Tink has a unique de-interlacing setting for 480i content called 'CRT Sim', which tries to recreate the interlaced look by inserting black scanlines in the gaps between the alternating fields, rather than simply filling in those gaps with a view to create convincing 480p like image.

I was curious whether enabling this 'CRT Sim' setting creates a benefit in motion clarity for 480i content, compared to say, the equivalent 480p like image, given that, I suppose, the persistence of each alternating field is cut in half.

In practice, I don't see much benefit - but perhaps it could be increasing the clarity by a small factor e.g. 25%?

Thanks for your help!

Don't know what the tink does, but on CRT TVs, interlacing would result each scanned out line filling parts of the above and below scanlines as well. So the end effect of a 480i signal is basically a line-doubled 240p signal with a progressive frame rate than equals the interlaced field rate. But because it's alternating the two frames, higher detail is being perceived. Because of the double-scanned appearance due to the bleed of each scanline into its neighbors, there are no interlacing artifacts.

Once you use a high quality CRT (like a VGA monitor) that doesn't have scanline bleed, it starts looking like a modern flat panel and interlacing artifacts become visible, and the illusion of full frames disappears.

Now I'm curious at what the tink does.

[Chief Blur Buster]

There are always 60 passes per second of 240 scanlines, regardless of 240p or 480i. So you've got 60 temporal images per second.

The problem is the pick-poison effect of faking information (remove the interlacing by deinterlacing algorithms) or preserving the interlacing. The two approaches can have different effects on motion resolution, e.g. muddy up or improve the 60fps look of 480i.

You can get the original 480p images if it's 30fps or less, like via the existing 2:2 or 3:2 deinterlace algorithm (inverse telecine) that I helped Mike of Retrotink 4K implement.

As a rule of thumb, for deinterlacing 480i to 480p, you have these options:
- For 60fps content in a 480i conduit, some information is necessarily spatially interpolated via a deinterlace algorithm, as the original information requires two refresh cycles to have full information delivered. You try to do your best, but the original motion resolution isn't always able to come back flawlessly because of the smudge factor of the vertical spatial interpolation. Whether by bicubic scaling, bilinear scaling, simple mathing, motion compensated mathing, or in other chips (other than Retrotink4K which has no AI) smart guessing via AI, etc.
- For 24fps and 30fps content in a 480i conduit, the inverse telecine algorithm will find the original fields and merge them (even and odd scanlines merged into one original 480p image).

There are pros/cons of the various approaches.

[NeonPizza]

Is it true that there's the occasional image artifacts when using TINK4K's triple strobe BFI for 24-30fps content(Movies & TV shows) when paired with a QD-OLED? I know there's less visible flicker(Thank god) than 60hz OLED BFI which can become unbearable on whites, along with a 1080p cap/limitation, which I'm 100% fine with since 1080p looks good enough on a 65" TV IMO.

But now there's image artifacts that can randomly show up on the screen? Maybe it's because the TINK4K is overworking itself?
Still, I'd deal. I need the 58%(144hz) motion blur reduction AND film judder reduction, which the Chief mentioned, gets lower than both plasma and CRT. Without either, watching movies & TV, as is, on QD-OLED/OLED is almost unbearable with 100% motion blur and what seems to be around 3x the amount of film judder Vs Plasma.

[BFI]

Just before I order one, can anyone please confirm the RT4K supports 1080p240 output with CRT beam emulation for 1080p60 inputs (ie HDMI consoles set to that res)?

I've read conflicting reports about this, with some people telling me the output limit for 1080p60 sources at 240Hz is 720p, rather than the 720p360 I read Chief mention as a possibility once.

I know HDMI 2.0 bandwidth supports 1080p240 or 1440p120, but not if the HDMI input contributes towards the pixel clock. All the examples I've seen are retrogames upscaled to 1080p240 with 25% persistence.

EDIT

I've been told the RT4K doesn't support overclocking or reduced chroma as I'd hoped, so can't output 1080p240 from a 1080p60 input. Shame.

[NeonPizza]

I'm thinking about picking up a 32" 240hz QD-OLED Monitor(MSI UPX), specifically for retro gaming, using a TINK4K.

Just wondering about the total amount of latency.

2ms > (TINK4K)
4ms > (TINK4K 240hz BFI)

But what about the latency coming from the monitor itself? Do I go by how much latency it has running an 8-16 bit retro game at 60fps, or do I go by how much latency it produces at 240hz with the 240hz TINK4K BFI? These are the latency numbers according to RTNGS >

Native Resolution @ Max 240Hz - 2.8 ms
Native Resolution @ 120Hz - 5.1 ms
Native Resolution @ 60Hz - 13.5 ms

if i have to add the 13.5ms from 60hz game mode, than that's around 20ms total which is a no-go for me. But if it is in fact just 2.8ms, than that combined with 2ms & 4ms is about 9ms which is fantastic. If not, back to a 27" Sony WEGA CRT.

[BFI]

Thanks for fixing this thread for me Chief (403 error). Is there any way to output 1080p180 from a 1080p60 input on the Tink?

The RAM load sits around 63% for 1080p120 output, rising to 92% at 180Hz, but I get something like a SBS 3D output using the CVT-RB modeline from https://guspaz.github.io/video_timings_calculator

1920, 48, 32, 160, 1, 1080, 3, 5, 98, 0, 179.98, "1920x1080p179.98"

The pixel clock for that mode is 440.60MHz, and I get a mostly black output with CVT-RBv2 and custom. So it looks like there's an unseen bottleneck, but I don't know if there's anything I could try because I don't know how the porch, sync, blank and polarity values manifest. The wiki tells me 179.98 is ignored because I'm in Frame Lock mode for BFI.

The documentation just says that 1080p inputs with outputs above 120Hz are known to be problematic. With CVT-RB it's like two images trying to overwrite each other, if you see what I mean. It looks like a double horizontal stretch with overlap.

Two more short questions: is Frame Lock or Gen Lock best for BFI, and am I correct to use LCD Saver (1 min) for my XG270 but not my DLP projector? I've never heard of a lamp suffering burn in, but possibly?! Cheers.