Great questions. This is a very "Chief Blur Buster" quality question, as I prefer to talk about blur busting than other topics. I rarely have time to reply to the forums, but when I do -- this is the correct question to ask if my eyeballs are in the right place at the right time.
Guffman wrote: ↑08 Jul 2026, 12:48
1.) Is 144Hz worth running over 120Hz? Is there greater compatibility and/or stability with 120Hz/60fps BFI versus 144Hz/72fps? I've always been under the impression that the more frames you have to throw at motion clarity techniques, the better the results, but maybe sticking with 120hz final output with 60fps BFI would be better or more stable math-wise?
It's recommended to use 120Hz over 144Hz if you're doing external 60fps content. Much better looking.
While CRT simulator theoretically can run at non-divisible refresh rates, it works better at large ratios (e.g. 60:280 for 60fps at 280Hz, than 60:144 for 60fps at 144Hz). You have temporal nyquist artifacts (aka flicker), much like the spatial blur of scaling 640x480 -> 650x490 instead of 640x480 -> 1920x1080. Brute spatial/temporal scaling for the win! Same metaphor, same problem. Time dimension instead of space dimension, really.
Guffman wrote: ↑08 Jul 2026, 12:48
2.) Is any of this even viable given my limit to 144Hz? I've used CRTs and am lucky/proud enough to keep an FW900 on my computer desk, so most of the point has been to replicate that glory with a big, modern TV in a home theater kind of setup rather than being confined to a desktop. Should I give up until 240Hz OLEDs and other more premium display technologies become cheaper and more mature, whenever that might happen?
Motion clarity progressively becomes clearer with maximum BFI:
60fps + (max BFI and/or framegen) at 120Hz = 50% less motion blur
60fps + (max BFI and/or framegen) at 240Hz = 75% less motion blur
60fps + (max BFI and/or framegen) at 480Hz = 87.5% less motion blur
Also, 60fps->120fps interpolation has less lag when output to a 240Hz desktop display than a 120Hz-144Hz desktop display.
Being that said, good framegen all the way (e.g. NVIDIA DLSS 4.5 doing 240fps in CP2077 on a 240Hz OLED) looks better to my eyes more than 90% of the time than doing BFI. So if I had a choice of fantastic framegen (superior to interpolation) or compromises-filled BFI, in a non-latency-critical situation, it is now tempting to use framegen instead. But if you're using VSYNC locked framerates (VSYNC ON 60fps CP2077), that will definitely look clearer motion on your FW900. But some interpolation software (and older DLSS/FSR modes) doesn't let you interpolate 60fps->240fps, and you can only interpolate 60fps->120fps. That's where BFI can help get motion clarity all the way to 240fps. See my answer to the question on combining interpolation and BFI by budgetting your maxHz.
So you don't *need* maxFPS to benefit from maxHz; it can simply help reduce your 60fps lag in the first place too!
If you have the holy grail of a FW900 on your desk, you'll definitely appreciate this. Alas, I understand OLED TVs don't have 240Hz yet, but you can at least get a 240Hz monitor and begin playing/tweaking with it, so your computer is ready when 240Hz TVs arrive.
Guffman wrote: ↑08 Jul 2026, 12:48
3.) If I did get a "dumb" BFI to work where every other frame is strobed black, could it still work with the frame interpolation in any way? Lossless Scaling uses DXGI or WGC for capture options, and it can also apply a flat frame multiplier to take a 60Hz source to 120, 180, etc. I've tried a few times now to run things like Shaderbeam on top of Lossless Scaling's output, but have had mixed results regarding compatibility/functionality, so I'm wondering if I ultimately have to compromise between either the frame gen working or the BFI working. Maybe they interfere with each other...?
Yes, you can combine BFI/CRT with framegen, but you must do it staged properly.
Use interpolation to go 60fps->120fps.
Use BFI to go 120fps->240fps.
It looks amazing. 75% less motion blur with just 2x framegen, due 50% blurbust via framegen layering on 50% blurbust via BFI.
Needless to say, you need more Hz-room for a staged approach. Treat your refresh rate as a blur busting budget. Split carefully between blackframe-based approach and framegen-based approach.
Educate yourself on BFI duty cycles versus motion blur:
www.testufo.com/blackframes "try on your 144 before buy 240 or 480"
Guffman wrote: ↑08 Jul 2026, 12:48
4.) I've seen conflicting info on Vsync, especially VRR - should it be enabled or not? I've seen posts saying VRR should always be enabled to ensure a proper frametime sync, but sometimes it seems to be referring to Vsync in general and not VRR, while other posts say to never leave VRR activated since you want the shaders/effects to easily sync up with the monitor's basic or maximum refresh rate. Then there are in-app settings; if I want to activate Vsync across the board on a driver level (Nvidia Control Panel/App), should I also toggle it on in various applcations like Reshade, ShaderGlass/Beam, Lossless Scaling, etc.?
It's possible to do VRR+BFI but it is MUCH harder (100x harder) to get stable if you're using software BFI. A 0.8ms variance on 8ms frametimes via software BFI = 10% flicker. Software is not perfectly accurate since VRR keeps Present()-to-photons as exact-time-relative as possible, so software jitter = flicker.
Retrotink 4K has BFI+VRR perfectly stable because it's using an FPGA and it can time the VRR frametimes perfectly zero-variance.
Instead, you need help from the one-dimensional grid snapping effect of VSYNC, much more forgiving (8.3ms safety margin for VSYNC ON at 120Hz, instead of being vulnerable flicker from sub-millisecond software jitter). VSYNC ON is much more forgiving on de-flickering software BFI.
Use high CPU/GPU priority for ShaderBeam, low CPU/GPU priority for game
But even so, you still need to raise ShaderBeam priority and lower game priority.
Framedrop in game = looks like a normal framedrop on CRT
Framedrop in CRT simulator = looks like a malfunctioning CRT
Therefore CPU/GPU priority on game = lower than ShaderBeam priority.
Even better: Use two GPUs. Dust off your old GPU in your closet, put that as your 2nd ShaderBeam-only GPU in your computer, and call it a day. You can mix GPUs (AMD, NVIDIA, Intel), give the more powerful GPU to the game, and less powerful GPU to ShaderBeam. Make sure you have enough bus bandwidth between the slots for the frame blitting every refresh cycle, tho: ShaderBeam is bus-bandwidth-heavy.
Guffman wrote: ↑08 Jul 2026, 12:48
That aside, just wanted to make a note that I don't post here often, but this website is a goldmine to me since I find all this fascinating and I love learning about all of it. Thank you very much for maintaining and updating this place, as well as keeping it a bastion of good conversation for all our blur-busting needs!
You are welcome.
1. Fix your integer divisor. Use 120Hz, not 144Hz, for external 60fps content.
2. Fix your CPU/GPU process priorities for ShaderBeam and your game. If something framedrops, make it the game not ShaderBeam
3. If you manage to stabilize it, you'll probably get another good upgrade if you buy a 240Hz OLED if you want more software blur busting options, or combining layered approaches. I see generic 240Hz desktop OLEDs on sale for cheap now.
Hope this helps your experimenting! Also, you can use utilities like ToastyX or the old NVIDIA Control Panel to create custom refresh rate that are not supported by your TV.