madVR - high quality video renderer (GPU assisted) [Archive] - Page 1326

Sunspark

2nd February 2026, 18:34

I moved my PC to the old 1366x768 32" PVA panel recently acquired TV which runs at 1360x768 on a PC. I wanted to watch some shows on a larger screen in a softer chair. It's not bad since it has 1:1 pixel, chroma 4:4:4 and full range blacks at 1360. I wish the backlight had a higher PWM rate (phone can see the flashing when you drop the backlight below 90%), maybe I will get used to it, even at a 100% backlight setting it still seems to have a shimmer. Testing displays is a must. This thing absolutely despises RGB input, it bands. You need to use YCbCr. Curiously, if you feed it a 1920x1080 input it doesn't care if it's RGB and the scaler chip will use it without banding or hotspots but overall it probably still prefers YCbCr.

What's been interesting to notice is that on a screen like this you can really see the difference between the various downscaling algorithm choices. On a monitor that is already normal sized or small, you don't really notice that much, but on a larger screen with larger pixels, it's a lot more visible! Leaving chroma upscaling the same, you can see large differences changing the algorithm when downscaling the image from 1920x1080 to 1360x768. Some are sharper, some are smoother, some are grainier, colours are affected (shadows, sunbeams), etc. I'm somewhat surprised how noticeable it is.

Sunspark

3rd February 2026, 20:18

Saw this post on another forum from a decade ago.

"AMD's DXVA2 algorithms are the most primitive one, Blinear, for all three. NVIDIA image upscaling algorithm differs depending on the source content's resolutions: for SD, it's Lanczos with some edge sharpening filter; for HD and full HD contents, it's Bilinear with with some edge sharpening filter. AMD's and NVIDIA's Chroma Bilinear upscaling is slightly different from madVR's.

Intel's are the best among the three. Chroma upscaling is close to Bicubic 50, image upscaling is Lanczos+AR, image downscaling is a bit difficult to identify, it looks more like Catmull-Rom+AR+unsharp mask than Catmull-Rom+AR+LL. If you are going to use Intel HD Graphics (in any SNB / IVB desktop CPU), madVR DXVA2 is the best choice. When there is no image scaling (mainly full HD contents in a full HD display), you have to choose an chroma upscaling algorithm and it depends on the number of EUs (Execution Units)."

I found this interesting because in part, I saw a comment from huhn awhile ago where he said he doesn't like bicubic150 for downscaling because of motion artifacts or something and prefers catmull-rom/bicubic 50. The dxva downscaling in the older Intel generations (mine is Broadwell) is not to be slept on.. it does appear equally as good or better than a number of the software algos for downscaling.

The info is still relevant for me since this is old hardware I am working with. I wonder if in the years since that was written if AMD or Nvidia have improved things, or if it's still the same as it was back then?

huhn

4th February 2026, 01:53

no it has not improved and intel has also utterly changed.

BTW. the old intel kernel is utterly broken and that's why they are the worst.
madVR bicubic beats all of them and it is really not good. DXVA has as said before a broken kernel matric tons of banding issues with chroma position color error and so much more it is the worst. d3d11 api is such an upgrade.

and while it is true that i do not like bicubic 150 in motion and think it should be fixed jinc and ssim are save and better then bicubic. bicbubic 150 is save for upscaled content at least as far as i know there are just other scaler more worth testing.

Sunspark

9th February 2026, 16:16

DXVA has as said before a broken kernel matric tons of banding issues with chroma position color error and so much more it is the worst.

This might be generation-specific. I just don't have the chroma issue.. we looked at a test image awhile ago and the shift is not present.

The downside of using that on this system is that yes, while it gives the system a big break on system load, it also means Intel had to be conservative with settings so it wouldn't break on certain content. So it is just average at everything and does have more grain.

It wouldn't surprise me if it was broken today, since everything from the big corporations is vibe-coded slop now.

and while it is true that i do not like bicubic 150 in motion and think it should be fixed jinc and ssim are save and better then bicubic. bicbubic 150 is save for upscaled content at least as far as i know there are just other scaler more worth testing.

I agree. I tested again and decided while it's OK, it could be better. Ssim and jinc provide good visual results but I can't use them right now due to thermal runaway causing the gpu load to ramp up to 100% due to throttling. Waiting for some putty to arrive and then I will replace the fan and all compounds/pads and see if the same thing happens. It's way overdue.

I read somewhere that spline36 (called spline 3-tap in madvr) is a good algorithm to use for non-integer factors when downscaling. When comparing it against lanczos, I agree. Lanczos is subtly sharper, but the problem is you don't want the output to be too sharp because then the eye catches on the edges/grain so spline is better here.

So it's a variable worth considering.. is any given algorithm recommendation better for integer or non-integer scale factors?

4K to 1080p is an integer but 1080p to 768p is not.

huhn

11th February 2026, 11:40

ingetar doesn't matter. madVR has some optimization that make it faster for integer but that's it.

yes people like talking about spline doesn't chance that it isn't impressive what so ever since we got AA.

it is irrelevant to talk about the sharpness difference of spline and lanczos when bicubic 150 is on the table a scaler magnitudes sharper then spline and lanczos 3 will ever be. the problem of lanczos was ringing which ahs been fixed.

DXVA scaler are usually utterly free on AMD and nvidia both can be done at idle and the lowerest power state. 480 to 8k or 8k to 480 it doesn't matter.

on older intel cards even the decoder puts load on the GPU which is just sad...

clsid

11th February 2026, 19:18

New beta 210 (https://www.avsforum.com/threads/improving-madvr-hdr-to-sdr-mapping-for-projector-no-support-questions.2954506/page-983#post-64465710) of madVR.

Schwartz

11th February 2026, 21:53

Cheers. So all the betas expire in 2027? How about the last stable, does that have a time bomb in it too? Just wondering.

dbcooper

11th February 2026, 22:03

Also from Madshi at that forum:

Some good news from Nvidia. I reported the HAGS issue to them with detailed instructions how to reproduce it, and they just told me they were able to reproduce it and will have a deeper look at it.

https://www.avsforum.com/threads/improving-madvr-hdr-to-sdr-mapping-for-projector-no-support-questions.2954506/page-983#post-64463606

huhn

11th February 2026, 23:40

they all have time bombs.

spectmajor

12th February 2026, 14:20

Can someone explain what is the issue with HAGS and how to reproduce it? I never disabled it, but I don't see any issue with it.

strumf666

12th February 2026, 16:35

Me neither, I don't think it affects all users.

huhn

12th February 2026, 16:46

the reported frame rate and composition rate changes after roughly 10-30 mins of playback.

i'm not affected it is usually reported with 23p mode which i do not use i use 119p.

pirlouy

12th February 2026, 19:06

It has been mentionned here (https://forum.doom9.org/showthread.php?p=2025704#post2025704) by... myself, but also in this thread.
Except that I had it with a Radeon card, not nVidia, and madshi reported it to nVidia and not AMD (from what I read). And I noticed it only in 23p mode, because 90% of what I watch is 23,976 Hz so I use 23,976 Hz only.

Sunspark

16th February 2026, 06:21

Finally got around to opening up the NUC to replace the fan and thermal compounds. One thing that surprised me, there was a thick dust mat between the blower and the heatsink fins, effectively shutting off airflow. No wonder. It wasn't visible from the outside.

Now that everything's replaced, instead of bouncing off 105C, turning on a bunch of the stronger algorithms only put it up to 75C.. that's a 30C degree difference in temperatures under load, wow and that's with the fan at 3800rpm instead of 8000+. Usually around 3300rpm now.

purplebaby

17th February 2026, 16:36

flossy_cake

17th February 2026, 17:16

I can't get HDR mode to enable when I play a hdr file on my LG CX (connected to PC via HDMI from RTX 5090 FE). Latest madvr version. Windows 11 LTSC. The video is displayed with SDR tonemapped colors. The TV doesn't switch to HDR mode.

I've set the hdr mode to 'pass through hdr to display' and 'send hdr metadata' on.

I kind of had that issue on my Windows 10 22H2 system with my 3060 outputting 4k120 RGB HDR over HDMI to my Samsung S90D

Firstly directX 9 doesn't support HDR at all so you will need to use DX11 in madvr settings

Secondly i was having an issue where even in Dx11 mode it would randomly not switch to HDR when in windowed mode but would reliably switch when going fullscreen exclusive, so try full screen exclusive

Lastly right click the media player executable and "disable fullscreen optimisations"

But there is some issue with madvr not setting HDR mode in quite the orthodox way or something like that, try asking AI to give more details there could be something else i forgot

huhn

18th February 2026, 02:40

yes ask the ai...

madVR tries to use the AMD and nvidia HDR api. and there is nothing wrong with that. a very old one in the AMD case amd_ags_x64.dll from 23 03 2017 5.0.6.0.

the alternative is to change the windows from SDR to HDR which Microsoft doesn't want you to do. so do it as mircosoft wants you to do it manual.

purplebaby

19th February 2026, 19:56

I had to remove the display (TV) profile, reinstall madvr, and then have mpc-hc/madvr's first exposure to HDR content be with only my TV connected, so that this became the primary display.

Now HDR works on the TV.

But now I can't get it to NOT use HDR mode on my monitor (which isn't suited for HDR)...

So it's either all SDR or all HDR. Lol.

huhn

19th February 2026, 23:56

install the newest GPU driver over the current install if it comes from windows update it may not have all features shipped like the nvidia HDR api which would explain this.

purplebaby

20th February 2026, 10:39

install the newest GPU driver over the current install if it comes from windows update it may not have all features shipped like the nvidia HDR api which would explain this.

I installed 591.74 as soon as I set up this PC.

flossy_cake

20th February 2026, 21:33

But now I can't get it to NOT use HDR mode on my monitor (which isn't suited for HDR)...

So it's either all SDR or all HDR. Lol.

Windows has it's own toggle setting for HDR mode ("Windows HD Color Settings") so ensure that is turned off as well

Also remember to set your HDMI output to 10-bit in the GPU control panel. Actually now I remember: setting the output to to 8-bit caused the HDR to not flake out in windowed mode. So that affects it too, so that's another knob you can try twiddling to sidestep the bugs. Eventually you will find the magical combination of settings that allows it to work lol

huhn

21st February 2026, 07:34

the nvidia api only care if HDR is technically possible nothing else bit deep or what ever does not matter it isn't amd.

go to the windows folder usually on C and search for nvapi64.dll
if it is missing created a restore point.
https://drive.google.com/file/d/1DrlFDB5FaJxaz5MZGc-74yUh6F7JQ_JA/view?usp=sharing
do what you must.

Sunspark

24th February 2026, 17:44

I won't be able to test again until later tonight because it's too bright in here, but I'm starting to think that downscaling may have issues.

I have multiple players and setups because it's interesting to compare.

I observed last night that on BE with the EVR-CP renderer using bicubic -0.6 and floating point I had great sharpness downscaling and good black levels/rich colour depth. With HC using madVR the blacks were lighter and the image was blurrier. So now I need to begin a series of testing.. switch to software decoding to see what changes with blacks/colours and then check again with all AR and LL turned off, and then again with both chroma and luma on bicubic 60 to see if it can match the clarity of the shader that EVR-CP is using.

Almost all testing most do has been with upscaling. I suppose nowadays there might be more downscaling again because of 4K videos being downscaled to 1080p. In this case I am downscaling 1080p to 768p.

huhn

24th February 2026, 17:59

fhere are two parameter in bicubic -0.6 and 60 mean the same but the second is usually not mention and madVR uses something there.

Sunspark

24th February 2026, 19:37

So I looked at the code.. strange.. the values are not apples to apples.. https://github.com/Aleksoid1978/MPC-BE/blob/master/src/Shaders/Resizers/resizer_bicubic_base_x.hlsl

I'm having trouble mapping these out. different AIs helping me with the math here keep giving me different values for -0.6, -0.8 and -1.0 when trying to convert it into B and C.

Sunspark

26th February 2026, 19:10

Woke up from a nap last night and went to watch a show and everything looked like it was moving faster than I expected and "glassy". Still the same in the morning. Wondered if I had smooth motion on, but no. Playback speed was 1.0 and display is 60.015 Hz. What changed was that I was experimenting with D3D11 Native paired with D3D9 overlay using a 24 fps (not 23.976) source so the timings must have been really good. Changing it to copy-back and windowed mode made it look more like what I am used to.

Turning smooth motion on makes it even more glassy like ice skating. You guys like that? Why?

huhn

26th February 2026, 19:31

your screen is at 60 HZ meaning it is doing 3:2 that is massive stutter and not the creators intent.

SM will show something very very similar to framerate matching with 24p in that case while still been at 60 hz.

what ever you mean with glassy i do not know. if you forced SM at 24 with 24p in you did something horrible right now.
if your device does any type of motion trash the result with SM will be terrible.
with SM 120 or more is highly recommended.

why would copyback or native have anything todo with that?

Sunspark

27th February 2026, 00:09

By glassy I mean it seemed like it had more frames than it should had have, or interpolated ones or something. AI said I am just not used to something that skips some abstraction layers like DWM which affects timing. I don't have perfect optical processing speed though so I'll put this aside for now.

So, now I'm contemplating which is the better choice for this TV.

With a 1:1 pixel mapping for the desktop, I am locked at either 60.015 Hz (disp/comp) or 59.91327/59.955 (disk/comp, slight mismatch) and locked to YCbCr 4:4:4 (no RGB, colour will distort--no idea why). Downscaling is done in software from 1920x1080 to 1360x768. madvr smooth motion is needed due to 3:2 pulldown. For whatever reason, it won't let me create a custom resolution of a lower refresh rate.

The TV however does accept 23.976, 24 and 25 CEA-861 signals, with a catch.. they're at a signal resolution of 1920x1080 which is not the native panel resolution. It does however in that situation accept both YCbCr and RGB without colour distortion (I suspect it still prefers YCbCr). Downscaling is done in hardware from 1920x1080 to 1366x768 ("Native" but not 1:1 so the chroma 4:4:4 test pattern breaks) using an unknown algorithm. madvr smooth motion does not appear to be needed.

Looking at the mainboard, it has a Zoran ZR39748 SupraHD chip that does everything, no separate scaler chip. Scaling is a polyphase FIR scaler which puts it between bicubic and lanczos in quality, so I guess that's good enough for luma scaling.

So I guess I have my answer there.. use the desktop at the lower 1:1 resolution and have madvr auto-switch to 1920x1080 and have the tv scale to panel and hopefully the refresh rate is true.. it doesn't seem to be stuttering. Advantage is not making the PC have the system load of both smooth motion and luma scaling. Flexibility to play with the added load of super-xbr if I want to.

Tip: If your TV is 2-prong power, get your multimeter out and identify continuity points between it and the PC. Then add a grounding wire to it. I did. It helped a lot with the static charge buildup, but still charges me up.. not as bad as before I added the ground wire though. Wire needs to be added to the TV. PCs have too many insulated standoffs and anodized paint on casing won't conduct properly whereas this unit has continuity between 2 of the 4 VESA screws, 1 of the case screws and all the ports. Boy, do I hate 2-prong devices. So lazy and cheap to not have them be grounded from the start.

huhn

27th February 2026, 10:25

why would you care about display and composition rate they both don't mean anything relative to each other.

annovif

27th February 2026, 10:39

brazen1

27th February 2026, 16:24

Hi everyone, I'd like to replace my video card but I don't want to waste my money, so I'm asking you about my concerns.
I have a new HP Pro G11 laptop with an Intel Ultra 5 processor and an NVIDIA RTX 2050 mobile graphics card, which can handle the same MAdVR filters as my old AMD RX580, even though on paper it should be superior. So it's true that MAdVR is optimized for NVIDIA graphics cards, so at the same cost NVIDIA will always be superior to AMD. I'm thinking for example of the AMD 9070xt versus the NVIDIA 5700. What do you think about it ?
Fabio

An Nvidia 5700 is 23yrs old. Video cards do a lot more than simply render madVR. They are the workhorse required to run a lot of other software too. Most prefer Nvidia exclusively because it works better than any other GPU's.

So, ask yourself, do you want to spend a bunch of money just to use higher settings in MadVR for a slight return on your investment or use it for other things too?

If simply madVR, get the least expensive Nvidia card you can - A GTX 960 sufficed for me with MadVR for a decade and so should your 2060.

If there's a chance you might use other GPU strength dependent software in the future, you can pull double duty and invest in an Nvidia card that will deliver more bang for your buck.

In that case, speed costs money... how fast you wanna' go?

Asmodian

27th February 2026, 16:30

So it's true that MAdVR is optimized for NVIDIA graphics cards, so at the same cost NVIDIA will always be superior to AMD.

Your one anecdote is not enough to base a conclusion on.

If madVR was optimized for Nvidia, it would probably be optimized for Fermi or something. That said, since Envy is running on Nvidia GPUs (The Envy Core MK2 and Pro MK3 use a 5060), any new optimizations that might make it into madVR from that work would be optimized for Nvidia. So, if you are using the new HDR stuff, that could be a factor.

I'm thinking for example of the AMD 9070xt versus the NVIDIA 5700. What do you think about it

I am assuming you meant the 5070, a 9070xt (2025) would be a lot faster than a 5700 (2003). :p

For raw performance in madVR I would probably go with the 9070xt, it is a much faster card for raster than the 5070. The 9070xt is competitive with the 5070 Ti.

The 9070xt has a theoretical texture rate of 760 GTexel/s v.s. the 5070's 482 GTexel/s, and game benchmarks are about 20% faster on the 9070xt, but I would still try to find some benchmarks to judge efficiency and real performance in madVR.

Or go with the 5070 even though the 9070xt is theoretically faster, AMD seems to have some persistent driver issues that do not play nice with some of madVR's features and Nvidia's market share makes any bugs more likely to be addressed.

Sunspark

27th February 2026, 16:35

Asmodian

27th February 2026, 17:27

but I still find it disquieting when I see that they're not both in sync.

I think we get jitter in whatever is getting composited, mouse movement being the obvious one. This is not noticeable for me at high refresh rates (100+), but not ideal either.

However, it is not a concern for video in madVR.

Siso

27th February 2026, 17:36

Out of curiosity when using smooth motion on a 120hz display, the first option in madvr's smooth motion doesn't work, only the last 2 work. Is there any difference using smooth motion "always" option, or the previous "if refresh rate is multiple of etc."

Sunspark

27th February 2026, 17:59

If you have 120, you don't need smooth motion as you have integer divisors. You would just switch between 120p or 119p refresh.

120/2=60
120/4=30
120/5=24
119.88/2=59.94
119.88/4=29.97
119.88/5=23.976

The only one that would be an issue is 25.

Siso

27th February 2026, 18:35

huhn

27th February 2026, 19:32

@annovif What you have is fine, but if it'll make you happy to get a new video card, first you must defeat Sam Altman.

@huhn I figure if both are in sync it works better. I have seen glitches in the past where the refresh would be 60 but the composition rate would be 30 and it definitely affected mouse movement, etc. These days those gross mismatches don't happen anymore, but I still find it disquieting when I see that they're not both in sync.

yes but a composition rate mismatch has nothing todo with the screen. it's a OS thing

Out of curiosity when using smooth motion on a 120hz display, the first option in madvr's smooth motion doesn't work, only the last 2 work. Is there any difference using smooth motion "always" option, or the previous "if refresh rate is multiple of etc."

not really if you would end up at 23p and play a 120 or 60 fps file you would see hell with always but that'S about it.

If you have 120, you don't need smooth motion as you have integer divisors. You would just switch between 120p or 119p refresh.

120/2=60
120/4=30
120/5=24
119.88/2=59.94
119.88/4=29.97
119.88/5=23.976

The only one that would be an issue is 25.
it will fix any and all sync issues.
there is a reason it is possible to use it exactly for that.

distino

27th February 2026, 19:54

Hello.

For playing 30 fps 1080p videos on 1080p 60hz display with MPC HC and madVR, are there any specific settings to improve experience?

I record game footage on my PC in 30 fps and then I'd like to watch those videos on my laptop. This laptop only has 60 and 48hz refresh rates available.
I tried smooth motion in madVR and while it did make motion a tiny bit smoother, the blur or the "soap opera effect” was too annoying.

Sunspark

27th February 2026, 22:13

distino

28th February 2026, 02:10

If the video is 30 fps and the display is 60 Hz then you don't need to do anything as it's an integer multiple.

ctrl-j to see the stats screen. The top will show the display rate and you will be able to see if the display is actually 60 Hz and if the video is actually 30 fps or 29.97 fps.

Yes pretty sure it's exactly 30 frames. I use 30.00 cap in both framecap software and recording software.

Here are the stats
https://i.imgur.com/KKkfd1n.png

I know about this, as I spent a lot of time fine tuning the settings for visual aspect. Rendering time is around 21 ms without stats displayed(dropped frames are only due to startup lag for the first 1 or 2 seconds when I start video).

This laptop is rather old and it can't handle any kind of frame interpolation and madVR at the same time. Though since those are only 1080p 30 fps videos and with smooth motion turned off, I'm able to use Jinc for chroma, YV12 Chroma Upsampling as pre-size shader, SharpenFlou as post-size shader and have pretty much everything unchecked in "trade quality for performance" tab except the last one "run custom pixel shaders in the video levels instead of PC levels".

So the image quality is very good, pretty much 1:1 like playing videos on my PC(that's like 10 times more powerful than this laptop).

I was just wondering if there are any specific settings in either madVR or MPC-HC that I should use when playing 30 fps videos on 60hz display. For example, does it matter if I force v-sync for the MPC in graphics control panel?

Sunspark

28th February 2026, 08:15

You don't need to do that. If you have "present a frame" checked in madvr next to direct3d 11 then that is your v-sync.

If everything is unchecked, then it is d3d9 windowed mode and still v-sync because it uses the Windows dwm compositor.

huhn

28th February 2026, 09:45

settings in madVR do not matter you always have vsync with madVR only external tool can try and disable that.

on the other hand are you sure you didn't record tearing?

distino

28th February 2026, 16:31

settings in madVR do not matter you always have vsync with madVR only external tool can try and disable that.

on the other hand are you sure you didn't record tearing?

You mean in-game screen tearing when you have v-sync off and framerate and refresh rate don't match? I have v-sync enabled for that game and there's no tearing. Besides, I don't think it would be even visible if you're recording directly from your PC? I mean screen tearing is only visible on your monitor and while in game. It's not visible while playing a video file, unless you record your monitor with your phone or something.

The above screenshot looks like this, because I did take a photo of the laptop's screen with my phone, rather than doing print screen.

huhn

28th February 2026, 17:55

depending on what buffer is read there shouldn't be tearing but it can be recorded with it.

getting tearing on win 11 is quite a bit of work.

Sunspark

28th February 2026, 18:56

I think I am done twiddling settings for a bit, but here are some observations I made.

I observed that on this LCD TV receiving full-range YCbCr output from the GPU, setting "BT.709/601 curve" did look a tiny bit better (slightly darker shadows) than setting "pure power curve" which is pretty much what everyone uses on monitors. I guess it makes sense because they would have tried(?) to tune the screen for DVDs, etc.

I think I like the sharpness of 1:1 pixel mapping resolution better and having the downscaling be performed by the gpu instead of the display doing the downscaling. To that end, a nice downscaling combination is mitchell-netravali chroma upscaling, jinc downscaling. Just that, no checkboxes. Good blacks, colour, clarity. Other combinations worked well too, e.g. super-xbr chroma but with the display doing the downscaling.

Now for timings, this is interesting and a little strange because it's not consistent. With various settings my eyes were getting tired because things seemed a little "off" with motion. With a baseline of "disable fullscreen optimizations" checked on the player properties and fullscreen exclusive checked along with direct3d 11 presentation+present frame checked along with d3d11 presentation unchecked I observed the following timing difference on the stats screen.

With fullscreen exclusive checked and direct3d 11+frame checked, stats screen reports display refresh rate 59.940 Hz and composition rate 60.000 Hz.
With fullscreen exclusive unchecked putting it in D3D11 Windowed, it then instead reports 59.913 Hz for display refresh rate and 59.955 Hz for composition rate.
Fullscreen exclusive checked and D3d11 unchecked putting it in D3D9 Exclusive, it also reports the same 59.913/59.955 timing.

I find it strange that the D3D9 and D3D11 exclusive modes don't both report the same identical #s. I find it strange that one particular combination reports the refresh rate of the display as 59.940.. that's what it is supposed to be, but why don't the other display modes (overlay and all the rest) report that? Could it be that this particular combination of settings reporting 59.940 is actually the most accurate reading?

Update: super-xbr & jinc look very nice together too (as does M-N & spline 3-tap--popular downscaling choice), but interestingly with this super-xbr/jinc combo the video feels like it is playing at a subtly slower speed even though the counters are not changing, must have something to do with the pipeline because I can see present queue reporting 1-3 with this combo.
I have been using evr-cp too a lot more lately as well, and honestly, this renderer is totally fine with 8-bit content. When set to exclusive mode (fso off on player) the refresh rate seemed fine, presentation is just as good once you select the right options and it doesn't bog down your cpu/gpu like madvr does. Good to have choices these days. There are other choices to experiment with as well. mpcvr, mpv, vlc, and more.

The speed thing is interesting really.. certain algos in madvr it seemed like the frames were moving subtly too fast, others subtly slower and it has partly to do with load. The counters were not changing but there was a perceptual difference.

huhn

1st March 2026, 23:50

you are just making stuff up...

madVR presents on vsync that vsync happens no matter what at the exact same time with absolutely zero cares about madVR.

also audio video sync... if it would be running any faster or slower it would run async it doesn't.

Sunspark

2nd March 2026, 01:37

I swear I am not.. that one setup has the stats screen report the numbers I want to see and the others do not.

As for the perceptual differences in playback, AI says it is not my imagination, that depending on the timing of the pipeline how it appears on the screen is affected even if the framerate remains the same. Older hardware with shallower buffer depths, etc. is probably more sensitive to that than fancy overpowered stuff. Put another way, the way smooth motion option works is by blending frames right? you are still receiving the same number of frames, the percentage of appearance is adjusted.. so why can't that also happen with a rendering pipeline?

We take it on faith that the clocks for the entire process are always in sync and accurate but even the display plays a role.