madVR - high quality video renderer (GPU assisted)

[tp4tissue]

1. Force max GPU clocks in Nvidia control panel for mpc-hc64.exe. Will probably make your idle temps and power consumption go up to unacceptable levels though (it does on my 3060 so I don't use this one...for now). Ironically 4k60 files play smoother than 1080p60 because GPU is loaded higher by 4k60 thus preventing GPU downclock. Wish there was some way to set a hard lower limit to downclocking.




Also disable core-parking while you're at it. Makes a huge difference for games and general system snappiness.

[flossy_cake]

Also disable core-parking while you're at it. Makes a huge difference for games and general system snappiness.

Thanks, forgot all about that one.

Can't seem to get it to stick though:

https://i.imgur.com/u9ACQan.png

Should I start messing around in the BIOS with C-states and such?

[flossy_cake]

Huh that's weird, when I'm actually playing video it appears to be running entirely on e-cores (CPU 12-15)

And CPU 0 (first p-core) fluctuates between parked and nonparked status randomly during playback - probably not good for frame pacing :(

https://i.imgur.com/QreAu88.jpeg

It's a 12600KF btw

[flossy_cake]

Seem to have solved it with a BIOS setting

https://i.imgur.com/XguXFl7.png

All I did was set EIST to disabled in the BIOS.

EIST = intel speedstep and is different to intel speedshift

Apparently speedstep is where the OS tells the CPU what P-state to use and speedshift is where the CPU decides for itself

I have left speedshift at Auto in the BIOS

If I leave the PC idle at the desktop it doesn't seem to downclock, and power consumption and heat at idle seems to still be low, so I think C-states are still working it just it no longer parks and downclocks

edit: actually it may still be parking internally, just not reporting it to the OS? That's what AI is saying, so might need to disable c-states after all.

[flossy_cake]

DWM related keys possibly affecting frame pacing

HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Windows\DWM\Schedule\FrameRateMin=200 (decimal DWORD 32-bit)
HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Windows\DWM\OverlayTestMode=5 (decimal DWORD 32-bit)
HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Windows\DWM\ForceDisableFrameBuffering=1 (decimal DWORD 32-bit)

edit: it seems there is also

HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Windows NT\CurrentVersion\Multimedia\SystemProfile\EnableMMCSS=1 (decimal DWORD 32-bit) (enables multimedia class scheduler used by "SystemResponsiveness" below)
HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Windows NT\CurrentVersion\Multimedia\SystemProfile\SystemResponsiveness=0 (decimal DWORD 32-bit)
HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Windows NT\CurrentVersion\Multimedia\SystemProfile\AlwaysUseDirectFlip=1 (decimal DWORD 32-bit) (bypasses composition when in full screen windowed)
HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Windows NT\CurrentVersion\Multimedia\SystemProfile\AllowTearing=1 (decimal DWORD 32-bit) (disables forced vsync when going through compositor)


Cause I feel like the issues are with DWM because if I go FSE mode in MadVR + "disable fullscreen optimisations" in media player's .exe properties, seems to give a good result.

There's some stuff about Edge browser having better frame pacing as well, not sure it's true as it's still going through DWM in fullscreen mode and I've seen dropped frames but maybe after about 30 seconds of playback it seems to settle down a bit?

[tp4tissue]

Core parking can be disabled in registry, or use corepark app, it just clicks the registry setting for you.

[flossy_cake]

Yeah, I dont know, I think the whole rendering pipeline is broken on my system. I'm seeing all this weird behaviour like when mouse hovering over the progress bar, something about the drawing of the progress bar interrupts the rendering and cause massive stuttering. Same with anything else that gets "overlayed" into the video like youtube subtitles - every time a subtitle appears on screen it causes the video to stutter. In MPC-HC I tried MPC video renderer and it just stutters horribly on 60fps videos, tried all combination of settings, nothing resolves it. It's like the system has no clock or something, it's just spitting out frames whenever and its bouncing either side of the vsync so sometimes a frame is shown twice, other times its shown once. It's just total rubbish, no care has been taken. I tried HPET clock that just brought the whole OS to a crawl. Maybe it's the 3060's fault, maybe NVIDIA just "has no drivers". Because adaptive half rate vsync solves the stutters so that tells me its probably GPU driver issue. Maybe I need to try an AMD.

[Sunspark]

Have you tried using older drivers instead of current drivers?

[tp4tissue]

Have you tried using older drivers instead of current drivers?

Also, Team AMD. :D

[clsid]

Try disabling multiplane overlay (MPO), HAGS, and Virtualization Based Security (VBS).

With NVIDIA Profile Inspector you can set application specific settings, and more than available globally.

[flossy_cake]

Try disabling multiplane overlay (MPO)

Yep:


HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Windows\DWM\OverlayTestMode=5 (decimal DWORD 32-bit)



HAGS

Yep:

Windows Graphics Settings -> Hardware-accelerated GPU scheduling



Virtualization Based Security (VBS)

This one was already disabled by default.



With NVIDIA Profile Inspector you can set application specific settings, and more than available globally.

Ok thanks, will try.

[Sunspark]

Still think you should try an older driver. New drivers (and OSes) can and do introduce bugs all the time.

Use DDU to strip the current video driver out cleanly, try using 531.79, 537.42, or 546.33 and see how it goes.

[flossy_cake]

Still think you should try an older driver. New drivers (and OSes) can and do introduce bugs all the time.

Use DDU to strip the current video driver out cleanly, try using 531.79, 537.42, or 546.33 and see how it goes.

Sorry yes, I absolutely will be trying that first. Is there any reason you'd recommend those versions specifically - did you have success with them?

I was going to try spreading it out over a longer timeline eg. 465.89, 497.29 , 537.34. NVidia dont provide these anymore, it seems guru3d (https://www.guru3d.com/files/category/videocards-nvidia-geforce-windows-7-8-10/) is the only place hosting these older drivers. This Nvidia page here (https://www.nvidia.com/Download/Find.aspx/) doesn't work for me , the dropdown lists just do nothing and I can't download anything. I guess NVidia doesn't want us using old drivers?

In the meantime I have investigated GPU Resizeable BAR on/off (no observed difference) and Message Signaled Interrupts (https://forums.guru3d.com/threads/windows-line-based-vs-message-signaled-based-interrupts-msi-tool.378044/) , both of which were already enabled by default

I seem to be able to get good frame pacing if I use DX9 FSE mode and NVidia "prefer max performance" power profile which keeps the clocks at max. The power consumption is only 7 watts higher on my system because when MadVR is idle its not really idle it's still rendering a frame 120 times/sec so the savings aren't going to be as much. Whereas a web browser stops rendering entirely when you stop scrolling and clocks go down much lower and save about 18watts. So for madvr, max perf mode is an acceptable solution to me.

I also get much better frame pacing in Brave and Edge with max GPU clocks, but when subtitles appear on youtube videos those are rendered through dwm.exe compositor (made visible with brave://flags tint-composited-content=enabled which tints dwm components of the screen red, and when the subtitle appears it is the only thing on screen tinted red and induces a massive JERK both when it appears and disappears).

I get those same kind of jerks in madvr when hovering the mouse along the bottom edge of the frame to make the progress bar appear/disappear in windowed mode. I guess that progress bar part of the screen is being rendered by the dwm compositor and is causing some kind of synchronisation issue when it tries to "overlay" the dwm part onto the non-dwm part. This is what that regkey OverlayTestMode=5 is supposed to help with but it doesn't fix it for me. However this seems to be a different root cause than what is happening with the youtube subtitles as switching to MPC video renderer doesnt exhibit that jerking behaviour when making the progress bar appear/disappear by mouse hovering.

FSE mode doesn't use dwm at all it replaces that progress bar with its own purple and grey coloured graphic along the bottom and making that appear/disappear doesn't cause the jerks. So my issues seem entirely related to DWM synchronisation issues and GPU downclocking.

[Sunspark]

No particular reason. It's been a lot of years since I had Nvidia, but I remembered the other year there was a complaint about a new version having an issue with something.

The drivers are still there, just not searchable.

https://www.nvidia.com/download/driverResults.aspx/172143/en-us/1000/
https://www.nvidia.com/download/driverResults.aspx/184718/en-us/1000/
https://www.nvidia.com/download/driverResults.aspx/211637/en-us/USE/

You can also get drivers from the Windows Update Catalog, 3060 seems to go back to 2021. https://www.catalog.update.microsoft.com/Search.aspx?q=3060&p=14&scol=DateComputed&sdir=asc

You can also search in there by the hardware ID in the driver properties.

You need to unpack the cab file, and then in the device manager right click on your display adapter, select update driver and navigate to the folder you unpacked.

[flossy_cake]

It turns out that OverlayTestMode=5 is the culprit of my subtitle stuttering issues. The reason I didn't detect it sooner was because GPU clocks needed to be max to expose the difference. So OverlayTestMode=0 (or deleted, apparently) + GPU max clocks = no jerks when subtitles appear/disappear in youtube videos in Chromium browsers like Brave and Edge.

[flossy_cake]

Oh it might be possible to do a thing where you set eg. mpc-hc.exe and webbrowser.exe to max clocks in NVCP and then solve the heat/power issue by having MSI afterburner automatically apply eg. "power limit ~75%" when "2D" applications are running so it won't affect game performance. Not sure if this would work, will have to do some testing with my wattmeter but it sounds good in theory

edit: nope, my wattmeter is implying that Afterburner's "power limit" doesn't come into effect when idling at max clocks. So if I want the smooth video playback I have to pay a +15 watt price penalty at idle (idling in browser for instance). Kind of disappointing , but my whole system power draw is only 40 watts at idle with max clocks enabled. With downclocking (power profile "Normal" in NVCP, which according to Inspector is actually "Optimal"), 25 watts - even better, I would like that but seems like it's not possible if you want smooth video playback.

[flossy_cake]

Wow Afterburner's "power limit" feature seems to be really great for reducing power and heat under load. In Heaven benchmark for example, with the camera in a fixed position, at full bore I'm getting 140fps with 125 watts system draw on the 3060. Then if I apply power limit 80%, it drops to 105 watts and I still have 132fps. That's a 20 watt saving for only a 5% reduction in performance - hell yes I'll take that.

[flossy_cake]

Possibly the reason why my weak old r9 380 isn't dropping frames on 60 FPS scrolling videos is because it's weaker so it doesn't down clock as much on a light load. Kind of like how a smaller engine has to rev harder to do the same amount of work — those higher revs stop it from missing vsync intervals. What we really need is a "for this application don't allow the GPU clock to go below, say, 500mhz". MSI afterburner would be the ideal candidate but doesn't seem to have such functionality.

[huhn]

it doesn't. there is no vsync miss or stuff like this the issue is the DWM been hands down bad. and that old GPU doesn't support any of that newer DWM version so it still works.

you can not miss a v sync interval you would get tearing you will not get presentation time spikes from low or high gpu usage.

presentation takes like 0.15 ms that's nothing and it already has like 38 frames it can present just wating.

the issue is microsoft there are ways to fix madVR but general microsoft is a different beast.

the things you try may fix so good luck but the core issue is not GPU clocks.

[flossy_cake]

the things you try may fix so good luck but the core issue is not GPU clocks.

But then how come forcing max GPU clock in NVCP can solve it?

But you're right it's a DWM issue too. It's both a DWM a downclocking issue.

The DWM is putting its own layer of vsync on top of the application and the driver. So there are actually three different vsyncs: 1. Application vysnc, 2. Driver (NVCP vsync settings), 3. DWM. Probably all fighting with each other and when GPU clocks get low they go out of sync or something, I don't really know.

[huhn]

no they don't get out of sync because of work it's just a windows issue. again there is no work to be done. windows displays frames to early too...

the work has already been done the GPU is still sending the same number of frames no matter the clock speed. it already has the frames it wants to present it just going wrong.

yes if you push your GPU to far and the GPU downclock and drops frames because it can not render them in time with madVR that an issue but not presentation and the OSD will tell you but that is not the issue.
it still presents a frame and the work to do frame is the same.

windows is just fundamentally broken why else would every browser struggle? because of GPU clock even through the lowest is already total over kill?

also there is only one vsync that is the output buffer the rest doesn't work like that the output buffer is just read at the time of presentation that's why tearing and so on exist the rest is just buffer they can be off by a lot and it does not matter.

they don't need to fix that the acg user does not care.
also forcing max clock doesn't fix it in general it's also not even doing so it is just lying to you.

[flossy_cake]

also forcing max clock doesn't fix it in general it's also not even doing so it is just lying to you.

Well I am measuring with a wattmeter at idle and under light load and heavy load so I know whether it's really running at higher clocks and afterburner reports clock values consistent with that

But you're right the max clocks can't fully fix dwm frame pacing, it seems I need FSE + max clocks to achieve that.

It might not even be the down clocking on its own, it could be the fluctuations because when playing a video I can see it bouncing back and forth between 200 megahertz and 500 megahertz, and then for a bit it goes up to max clocks and then back to 200. All those changes in timing might be confusing things in the pipeline, I'm just speculating here.

[clsid]

Instead of forcing max clocks, you could try if raising the minimum clockspeed helps.
https://www.youtube.com/watch?v=WZH4nKEkJXA

[huhn]

kill the explorer (this will restart the DWM) and it may go away from a short while.

yes these option do something but the GPU changes the clock rate so fast and so rapidly the stuff shown to us is very very summarized AVG at best.

fixing madVR completely at least temporary is still possible.

[flossy_cake]

Instead of forcing max clocks, you could try if raising the minimum clockspeed helps.
https://www.youtube.com/watch?v=WZH4nKEkJXA

Oh my god that is exactly what I wanted thank you so much :thanks:

[Sunspark]

Good call. Higher minimums have been needed on other systems because if it is too low then they become more sluggish responding to input.

As for power consumption, I don't think it really worth caring about.. if it's 10 cents for 1000 watts for an hour, then does it really make a difference if it's at 100 vs 125?

[huhn]

for 4 hours a day that is:
4hx365=1460 hours
at 100 watt that is 146kwh a year
at 25 that's 3.65 extra

btw the avg kwh on earth is easily 25 cents in many develop countries it is 0.35 euro.

a GPU that is used will cost more money in terms of the powerbill than the buying price.

[flossy_cake]

Good call. Higher minimums have been needed on other systems because if it is too low then they become more sluggish responding to input.

As for power consumption, I don't think it really worth caring about.. if it's 10 cents for 1000 watts for an hour, then does it really make a difference if it's at 100 vs 125?

10c/kWh? Man what country are you living in :o

But even at 40c/kwh I guess the power bill is not the issue given it's not running for many hours a day under full load. But the temp reduces by about 6C and fan noise reduction too (crappy 2 fan MSI ventus cooler)

[clsid]

Given how often "high performance mode" fixes things, and given how many people are using that mode as a result of problems, NVIDIA should really give us more power management options. Simply increasing the threshold for downclocking could solve so many issues. It should downclock only if usage is low for a long(er) period, instead of bouncing up and down constantly in certain situations.

[huhn]

yes that was a huge issue with 1000 and 2000 series cards which defaulted to optimal power saving which was to aggressive with madVR resulting is massive rendertimes spikes.

my 4000 series has no such issue. optimal is gone it only has normal and maximum power.

[Siso]

yes that was a huge issue with 1000 and 2000 series cards which defaulted to optimal power saving which was to aggressive with madVR resulting is massive rendertimes spikes.

my 4000 series has no such issue. optimal is gone it only has normal and maximum power.

On normal it has also rendering spikes with 4000 series.

[Sunspark]

There is another place to look.. in the Windows Advanced Power Management options, there is a setting for "PCI Express Link State Power Management". I have that off on my system. Having it on may or may not cause issues with PCI Express devices which includes video cards.

[Sunspark]

10c/kWh? Man what country are you living in :o

I'm in Canada, using tiered instead of time of use. It's $0.12/kWh currently for the first 1000 kWh then becomes $0.142/kWh. Even cheaper in the province next door at $0.06905.

I hear you about the Euro pricing.. I was just in Greece, and the amount of crying over turning on the heat to the hot water tank was annoying. They are constantly turning the power off to it at the electrical panel.

[flossy_cake]

optimal is gone it only has normal and maximum power.

Normal appears to be Optimal according to Profile Inspector.

Also some apps (executables) have the power management mode removed from the NVCP, eg. firefox.exe. Obviously the developers said to nvidia "we want to save the planet and nobody should be running our browser in max clocks" (yeah...unless you watch youtube in it and want smooth frame pacing lol)

So firefox.exe can only be forced to max clocks in Inspector. And I saw that brave.exe has forced "ultra low latency" and prerender frame queue=1 in Inspector, and that was not visible in NVCP. Whereas Edge browser doesn't have that. But edge is a bloated nagging mess but at least it doesnt use that app manifest v3 so you can still install ublock origin

[flossy_cake]

Ugh, the nvidia-smi setting is good and I can already see that setting min clock to 650 or a third of max clocks is preventing DWM stutters when subtitles appear, but it's still frustratingly not quite a solution because I leave my PC on for many hours a day and when it's just idling at the desktop doing nothing like playing music with the screensaver on I want it to downclock all the way to 200 and the nvidia-smi is a global setting. I need to set it per process, so I'll have to find some way to automate it.

edit: looks like process lasoo would do the trick...

https://bitsum.com/wp-content/gallery/lasso-featured/watchdog-exec.png


edit: ugh the watchdog feature is not free

[flossy_cake]

Huh that's weird, I saw that when forcing max gpu clock with nvidia-smi "C:\windows\system32\nvidia-smi.exe -lgc 2000,2000" my idle power consumption and temp barely increases. Afterburner reports the gpu clock is indeed maxed. Then I tried forcing max mem clocks too and got this result

NVCP power mode nvidia-smi.exe idle power draw
----------------------------------------------------
Normal untouched 25w
Normal max gpu 29w
Normal max gpu+mem 40w
Max performance untouched 40w

I wonder though about the relationship between gpu and mem clock whether running max gpu clock with low mem clock would cause timing issues. Because normally the driver will scale them in tandem

edit: for reference


https://docs.nvidia.com/deploy/nvidia-smi/index.html

C:\windows\system32\nvidia-smi.exe -lgc 650,2000 # force min,max gpu clocks globally
C:\windows\system32\nvidia-smi.exe -rgc # reset min,max gpu clocks to driver default

C:\windows\system32\nvidia-smi.exe -lmc 2500,7500 # force min,max memory clocks globally
C:\windows\system32\nvidia-smi.exe -rmc # reset memory clocks to driver default



edit: oh and thankfully the afterburner "power limit" is a separate setting that doesn't get affected or reset by nvidia-smi. I tested that with Heaven benchmark and got the same result

[flossy_cake]

The sweet spot for me appears to be leaving the memory clock alone and just forcing a min GPU clock of about 66% of the maximum (1333mhz for my 3060). This gives me the exact same idle wattage and temps as letting the GPU downclock all the way to 200mhz at idle, and for variable loads the 1333->2000 spike isn't aren't as big as say 650->2000 so hopefully that will help with frame pacing too.

I am a bit concerned what effect this will have on longevity over many years by running the card 8 hours a day at 1333mhz instead of 200mhz.

Also the screensaver must be disabled like ribbons and stuff that animates cause the GPU has to draw all that , about 8 watts extra just for the screensaver. But it's more the idle temps and longevity I'm worried about than 8 watts.

[clsid]

You could also just put two bat files on your desktop to toggle it on and off.

But if power draw is low, then no need to worry about longevity.

[flossy_cake]

You could also just put two bat files on your desktop to toggle it on and off.

Yep that's what I'm doing now, it works well :thanks:

edit: and that shortcut .lnk file that runs the .bat can be set to run at logon in Task Scheduler and there is no flash of a cmd window on logon either so that's nice and seamless, doesn't feel like I'm doing anything janky. I'm using the official nvidia smi interface so no third party subscription software like process lasso is needed


But if power draw is low, then no need to worry about longevity.

Yeah I think it's probably fine, it's only a 3060 anyway

But my main longevity concern now is with temp stability because apparently its bad for electronics to be constantly cycled through +/-20C swings every few minutes. Better to have it consistently at one temp, even if that temp is hotter. Difficult to achieve with the way this fan curve works and the different load between say a 60fps file and a 24fps file.

The 60fps file is a higher sustained load so the fan runs at minimum speed through the whole file and temp is stable the whole time. But the 24fps file is lower load so eventually the fan goes to 0 RPM as it passes the hysteresis level at 40C, but then it cant maintain this equilibrium so after about 10 minutes of playback it warms up over the hysteresis to 60C, then cools it all the way back down to 40C again, repeat. This repetitive heat cycling of 40C -> 60C -> 40C can't be good for the electronics I would say.

[Sunspark]

Leaving the clock high won't hurt anything.. on a reasonable chip that is. I've left my iGPU's clock at 100% for years, but that's easy to do when it was already 15W maximum. My Steam Deck is also 15W maximum as well. In the good old days everything was always 100%, though you could force a downclock by pressing the turbo button.

Try the pci express link state power management setting!

[flossy_cake]

Try the pci express link state power management setting!

That one was disabled from the get go, but there appears to be a setting for it in the BIOS as well - although it shouldn't matter since it's disabled in OS anyway, but you never know....

Leaving the clock high won't hurt anything.. on a reasonable chip that is. I've left my iGPU's clock at 100% for years, but that's easy to do when it was already 15W maximum. My Steam Deck is also 15W maximum as well. In the good old days everything was always 100%, though you could force a downclock by pressing the turbo button.

I don't really understand the relationship between clock speed and idle power consumption. If I enable downclocking for both the CPU and GPU, I don't see any reduction on the wattmeter - I remain at 25 watts entire system draw. So it's not the core clock, it's something else. For the CPU I reckon its something to do with C-states (which I've got enabled in the BIOS even though it's not downclocking or parking at idle) and for the GPU it's something to do with the memory clock.

The method I am using for setting min GPU clock is to measure idle wattage at the desktop, then keep increasing the min GPU clock until the wattage starts to increase. On my 3060 that is around 1333-1500MHz, which I thought is surprisingly high.

Anyway I'm quite pleased with this system now, just need to double check everything with the 4k120 HDMI output as I'm currently only testing at 1440p120 over Displayport.

[YxP]

Gentlemen, I understand you are talking about 60fps videos and such, which might be beyond my "solution", but I just upgraded to W11 and had to deal with sudden random sync drops with 23, 24 and 25p videos when I was sending them to my TV with the same frame rate. I messed around with settings quite a bit and found, that at least for me, setting the present frames in advance to 1 in MadVR (as is the default in Nvidia control panel) would keep the the sync nicely. I haven't done any more testing, but thought maybe you could try it as well.

[flossy_cake]

I messed around with settings quite a bit and found, that at least for me, setting the present frames in advance to 1 in MadVR (as is the default in Nvidia control panel) would keep the the sync nicely. I haven't done any more testing, but thought maybe you could try it as well.

Cheers I'll add that to my list.

It might be synergistic to combine it with Nvidia Profile Inspector "maximum prerendered frames = 1" and the following DWM regkeys which seem like they might limit the frame queue:

HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Windows\Dwm\ForceDisableFrameBuffering=1
HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Windows NT\CurrentVersion\Multimedia\SystemProfile\AlwaysUseDirectFlip=1
HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Windows NT\CurrentVersion\Multimedia\SystemProfile\AllowTearing=1

(32-bit DWORDs)

edit: i suppose an interesting test on your system might be to try only those regkeys instead of MadVR's setting. If that also solves the issue, then DWM is the culprit and those keys aren't placebo

edit2 : but have you tried forcing max GPU clock for the media player executable in NVCP? Because that one has been the most impactful on my system. I've since reduced the min clock to 1333 using that nvidia tool and that seems to be enough for smooth performance without increasing the power and heat - it's a free lunch!

[YxP]

I'll do some testing later for sure, since I still have the problem of not getting the sync first time every time. Also pausing the video breaks it quite often. But at least without touching anything it seems to work well.

[flossy_cake]

I am a bit concerned what effect this will have on longevity over many years by running the card 8 hours a day at 1333mhz instead of 200mhz.


Oh I had an idea - make some script that sets the fully lowered GPU clock when the screensaver triggers, and restores the elevated min on screensaver exit. AHK script seems to be capable of that

[clsid]

You could also create a scheduled task that triggers on eventid (https://learn.microsoft.com/en-us/previous-versions/windows/it-pro/windows-10/security/threat-protection/auditing/event-4803) (100/101/4802/4803). That can be set to run as admin without UAC prompt and requires no third party tools.

[flossy_cake]

You could also create a scheduled task that triggers on eventid (https://learn.microsoft.com/en-us/previous-versions/windows/it-pro/windows-10/security/threat-protection/auditing/event-4803) (100/101/4802/4803). That can be set to run as admin without UAC prompt and requires no third party tools.

Excellent - even better :thanks:

[Celio1080p]

270 nits setting is for madvr because it's the peak brightness of your TV RU7100 in it's Warm 2 setting.

In Madvr HDR, use these settings as base. https://imgur.com/dCMmibb

Under Madvr's calibration tab, select this display is calibrated for rec709, gamma 2.2

On TV,

Go to Picture Mode -> Movie (If the movie mode looks desaturated (clamping gamut), then leave it on standard).

Go to Expert Setting -> Backlight Max, Brightness 0, Contrast 45, Sharpness 0, Color 25, Tint 0, Color Tone Warm2

You do the expert settings part whether you're in standard or movie.

Hey there tp4tissue. I'm so sorry for the extremely late reply. My fault here.

Ahem, I can't do those things you asked for, most of them. I use my TV 65 Samsung RU7100 on my PC, with a RTX 5060 Ti 16GB through HDMI. On PC mode, and game mode, of course.

https://i.imgur.com/A1FkKfc.png

Anyway, guys, I'm using the wrong version of MadVR? Why those options aren't available for me?


color tweaks for fire & explosions

highlight recovery strength


Please help?
I use Windows 11 Pro.
Should I try the latest beta version? Can someone send a link, please?

Sorry my english.

[huhn]

https://www.videohelp.com/software/madVR
copy the files of 208 over the normal install.

[Celio1080p]

https://www.videohelp.com/software/madVR
copy the files of 208 over the normal install.

https://i.imgur.com/wg6SlIh.jpeg

Hey there. Uh, you were talking to me? xD

So, I did what you said, I even redownloaded MadVR, reseted all settings, just to be sure.

Those two options seems not be available still. But many others showed up.

Things I noticed. For some reason, it seems to be misunderstanding my resolution. What is 3840x4320 ? That's not my resolution.

My native resolution is 3840x2160.

Other thing is, madVR is crashing a lot if I try to keep advancing the video, and...

The movie keeps flickering sometimes? I'm not sure why.

Any help is appreciate it. Sorry my english.

I am using MPC-BE 1.8.8 x64