madVR - high quality video renderer (GPU assisted)

[Siso]

No idea then, unless you're using more than one monitor, which can cause similar issues.

Only one monitor and that's all. I'm thinking it might be happening after shutdown of the pc, tried with restart again and it was ok, I'll test with shutdown again later.

[mclingo]

The current algos are great with light banding, no issues there are all, these can stay as they are. Howeve they are not suitable for near black macro blocking and other near black issues because the algo works on the entire picture so if you have to go heavy to deal with really bad near black issues it destroys the rest of the image.

We really need something that just works at near black, lighter areas of the picture dont get compressed as much and tend to look ok.

The final game of thrones seaon is a good example, it was totally unwatchable at times on OLED but it wasnt much better on high end LCD's either.

An example bluray is Snowpeircer, here is a clip, its all kinds of awful.

https://1drv.ms/v/s!AgvFafeelEBijd4jDEfXshvmKKfkWw?e=8p04yS

[huhn]

a blu ray encoded with 1 ref frame and 11 mbit combined with 32 kbit mono AAC and cropped?
why do i even get a feeling this is CBR encoded from someone that has ittle to no clue what they are doing which will make dark scenes even worse.

[mclingo]

yes its a quck and dirty rip, but hey, why not throw in a totally unnecessary insult , that always helps doesnt it....

its looks identical to the bluray which I also have.

[mclingo]

bluray

https://1drv.ms/u/s!AgvFafeelEBij5kHF5JUgm0sdK6NKg?e=BWDUzW


dirty rip

https://1drv.ms/u/s!AgvFafeelEBij5kGitESdrI2luDFtA?e=xBYcGQ

[huhn]

so having little to no clue what someone is doing in something like this is an insult?

an tool like mkvtoolkit can cut files without any reencoding and it should be faster too.

[huhn]

can you plz make sure it is the same frame. you can navigate per frame by using control + arrow key left right in mpc-HC.
both are still cropped and that's not allowed on a blu ray.

BTW. you don't have to provide a sample from a blu ray your reencode is fine too but you should specify that because if you don't the first question should be if the encoded is bad because it is not a blu ray so a problem with an reencode.

edit: plz lossless images like png we want to judge rendered images not images that are again compressed.

[mclingo]

you've pointed out a percieved inadequcy with a clip ive posted which turned out to be wrong, the clip was perfectly adequate for the intented purpose as you can see in the screen shots, therefore I knew exactly what I was doing when I made it several years ago, if I made it again now, it might be different but its totally fine.

But that isnt really the problem, its your abraisive manner, you could have simply said " are you sure that clip is a good enough representation of the bluray", but no, you want straight into accuse mode.

Do you even know when you are being abraisive, are you deliberately trying to wind people up or is there a language barrier here?

[huhn]

you just posted jpg of different frames.

so again you have no clue what you are doing good day sir.

[mclingo]

it should be clear to everyone that the compression artefacts are so bad that its irrelvant that they dont share the same frame, this was meant as a comparison shot to show that the clip was close enough to the bluray.

stop being so petty and puerile.

[ryrynz]

This feature request has been brought up multiple times, we're all aware of the limitations of some sets and the likelihood we'll see anything addressing those in madVR in the short term. Didn't really need it brought up again, by the time anything is likely implemented you'll have a new TV.

[Alexkral]

We really need something that just works at near black, lighter areas of the picture dont get compressed as much and tend to look ok.
The final game of thrones seaon is a good example, it was totally unwatchable at times on OLED but it wasnt much better on high end LCD's either.

My TV doesn't have a good enough contrast ratio, so I had to increase the brightness of your images a little to be able to see what you're talking about, and yet it doesn't seem too noticeable. So you say that these artifacts are very noticeable in an OLED in a not completely dark environment? In that case I guess that reducing the brightness a bit should help.

This feature request has been brought up multiple times, we're all aware of the limitations of some sets and the likelihood we'll see anything addressing those in madVR in the short term. Didn't really need it brought up again, by the time anything is likely implemented you'll have a new TV.

The point is that this is not due to a limitation but because of a feature such as the contrast ratio that is likely to continue to improve on future TVs.

[exuvo]

I recently added a new graphics card to my computer, a RX 5700 XT on PCIE3x8, in parallel with my older Radeon R9 290X on PCIE3x4. Both work fine with madVR except with the 5700 card it takes an additional 2 seconds to start video playback compared to the old card with exactly the same madVR settings. Seems unrelated to the file being played. I assumed this was shader cache related and tried clearing that from Radeon settings but nothing changed. I enabled "delay playback start until render queue is full" to avoid dropping frames at the start for now but would rather remove the delay.

Start time from when MPC-HC window appears to video playing:
On display connected to R9 290X: 1.3s
On display connected to RX 5700 XT: 3.1s

OS Windows 7 64-bit

Any ideas what could be causing it and how I could fix it?

I also seem to have been hit by coil whine with NGU (very high is very noticeable, others are slightly audible) but can almost completely (only audible if room is quiet or ear next to computer) resolve that by setting "don't flush after intermediate render steps". Combined with lowering GPU powerlimit to -50% (still plenty of performance left) it completely removes the coil whine (barely audible with ear next to GPU).

[j82k]

My TV doesn't have a good enough contrast ratio, so I had to increase the brightness of your images a little to be able to see what you're talking about, and yet it doesn't seem too noticeable. So you say that these artifacts are very noticeable in an OLED in a not completely dark environment? In that case I guess that reducing the brightness a bit should help.

I think the problem is that he has a 2015 Oled.
2015/2016 Oleds had terrible near-black gradation. I've had a LG C6 before and that TV would show any near-black artifacts exaggerated.

Recent Oleds generally handle near-blacks much better but there are still some issues...

[huhn]

My TV doesn't have a good enough contrast ratio, so I had to increase the brightness of your images a little to be able to see what you're talking about, and yet it doesn't seem too noticeable. So you say that these artifacts are very noticeable in an OLED in a not completely dark environment? In that case I guess that reducing the brightness a bit should help.
if you calibrate your system properly and in this case not clipping levels is enough to see compression artefacts
if an screen is unable to display this properly is just an completely different issue.

what so ever this file is not nice. deband high does a pretty good job on it but this is not fixable is just how much do you want to hide vs how much damage to the not dark part are fine to you. it's could be possible to trigger the dreaded black makro block flickering on these LG oleds: https://www.youtube.com/watch?v=Sl_t3UQ3gAk

mastering screen are usually IPS 1000:1 screen like the eizo coloredge series there are other screen that are not IPS yes there are but that's not the point.
This feature request has been brought up multiple times, we're all aware of the limitations of some sets and the likelihood we'll see anything addressing those in madVR in the short term. Didn't really need it brought up again, by the time anything is likely implemented you'll have a new TV.
lowing the bit deep for dark scenes using dithering was talked before for oled but i can't remember dynamic post processing.

[tony359]

Hi there

wondering if anybody could help on my last question I posted here (https://forum.doom9.org/showthread.php?p=1886695#post1886695)?

Thanks!

[huhn]

using HDR mode on projector is not recommended.

Since I'm here, another unrelated question: if I play some older videos (in my case a 720x480 video), picture is not upscaled to full screen. What am I missing?

that's a dvd resolution they don't have square pixels and can be used to displayed at 4:3 or 16:9 if this information is missing it will be displayed at 3:2 which is wrong if the file is for some reason supposed to be displayed like this it will have black boarders left and right.


you didn't by any chance added every resolution you could find under display mode didn't you?

[Furgiuele]

Hi. I'm Italian and new on this forum and I don't speak English very well.
I appreciate a lot the guide about MadVR and I'm trying to optimize it for my needs.
I have a htpc with MSI Z390A-pro mobo, I5-9600K cpu, 16 Gb ram, Nvidia 1660 gpu, bd drive, ssd and nas with a lot of concerts and movies ripped from br discs.
I have a Sony VPL-vw870es (calibrated BT2020) as projector and a Trinnov Altitude 16 as processor. I use Jriver Media Center 25 with MadVR, but I don't like to use more than one setup.
So I set MadVR with PC levels 0-255,10 bit, already calibrated BT2020 pure power curve 2.20, 2160p23, 24, 25, 29, 30, 50, 60, tone map HDR using pixel shaders, tp 100, BT2390, highlight medium, measure each frame's peak; if in doubt, deactivate deinterlacing, disable automatic source type, force video mode, only look at pixels in the frame center; chroma ups. NGU Antialias medium, image ups. luma NGU Sharp, luma qu. direct qudrupole very high, chroma high; "use Direct3D 11..." and " "present a frame for every VSync" flagged in general settings, smooth motion only if there would be motion judder; ordered dithering, no flag in trade for quality.
It works very fine with all 23.970 fps videos (4k hdr, fhd, etc.), without interlacing. But I've issue to watch 29.970 fps videos, with a lot of dropped frames, motion juddering and stuttering. I have a lot of concerts with this fps (Diana Krall, Carlos Santana, Miles Davis, ecc.).
Can someone help me?
Thanks.

[nsnhd]

Are there benefits or harms by setting madVR for 2160p23, p24 in comparison with setting for 2160p59, p60 only ? Could that help to prolong the time a frame will repeat in...when the display switches to around 23.970 hz by playing 23.970 fps movies ?

[huhn]

if you play a 23p file at 59p you are "repeating" 12 frames every second.

[3ll3d00d]

I have noticed inconsistent behaviour when upscaling DVD to 4k, it behaves something like it isn't upscaling for some period of time and then it corrects itself and starts upscaling again. This can happen at arbitrary times, e.g. within a single scene without having touched any playback controls or changed any settings.

An example to illustrate, I created these by just replaying a scene on a loop. It rendered it "correctly" v few times, it is mostly looking bad. Look at the guy on the wire for the obvious difference. The fact this is two different frames is irrelevant btw as the effect occurs regularly on different content, i.e. it's not specific to one scene or one particular source file.

Anyone seen something like this before? Any tips on how to resolve?

Bad - https://ibb.co/7YLkcg9
Good - https://ibb.co/P6VwpKD

[nevcairiel]

Your screenshots are very low resolution, making it hard to judge anything, or even read the madVR OSD at all. Perhaps upload the original screenshots to a better image host?

[Asmodian]

This is not an issue with upscaling. It is due to deinterlacing working well or not. madVR either uses the GPU hardware for deinterlacing or it has its own IVTC algorithm if you put it in film mode. This behavior should depend on the DVD itself, is it telecined or true interlaced?

On that note, you do not want to use 25 Hz for this DVD. It is being deinterlaced to 50 fps, as indicated by the 20ms frame time.

Edit:
Your rendering times are very high for a 20ms frame time, perhaps the deinterlacer is getting confused due to performance issues? Try turning down the NGU 4x to medium, or maybe only double with NGU high, for testing. (maybe a problem with paused OSD... I shouldn't pay too much attention to the rendering times when paused. :o)

Your screenshots are very low resolution, making it hard to judge anything, or even read the madVR OSD at all. Perhaps upload the original screenshots to a better image host?

On desktop (Chrome) I got a link on the image to download the full resolution, which did get a good image.

[nevcairiel]

Oh now I see it. Thats some bad UX for an image host.

[ryrynz]

Yeah definitely upscaling correctly, the obvious NGU upscaling can be seen throughout the image, definitely a deinterlacing issue.
You may wish to change between deinterlacing types to see what yields the best result.

[3ll3d00d]

This is not an issue with upscaling. It is due to deinterlacing working well or not. madVR either uses the GPU hardware for deinterlacing or it has its own IVTC algorithm if you put it in film mode. This behavior should depend on the DVD itself, is it telecined or true interlaced?

thanks

I'm not 100% sure how to tell if it is telecined or not, mediainfo reports


Frame rate mode : Constant
Frame rate : 25.000 FPS
Scan type : Interlaced
Scan order : Top Field First


with regards to the "deinterlacing on" or "deinterlacing off" message, am I correct in thinking this refers to whether madvr or upstream is doing the deinterlacing? i.e. if on then madvr is deinterlacing, if off then something upstream is doing it.

[el Filou]

if in doubt, deactivate deinterlacing, disable automatic source type, force video mode [...]
I've issue to watch 29.970 fps videos, with a lot of dropped frames, motion juddering and stuttering. I have a lot of concerts with this fps (Diana Krall, Carlos Santana, Miles Davis, ecc.).Based on your description, what could be happening is those 29,970 fps concerts are telecined, and as you are forcing video mode they are deinterlaced with very bad quality to 59,94 fps and you have 3:2 motion judder, but it's hard to be sure without knowing what the source is: Blu-ray? DVD? TV?
Try to enable film mode for those concerts (you can assign a keyboard shortcut key to it) and see if it is better.
with regards to the "deinterlacing on" or "deinterlacing off" message, am I correct in thinking this refers to whether madvr or upstream is doing the deinterlacing? i.e. if on then madvr is deinterlacing, if off then something upstream is doing it.No, it just informs on what is the source of madVR's decision to deinterlace or not. In this case the upstream filter, i.e. video decoder, says it's interlace so madVR deinterlaces.
This looks to be a PAL DVD you're playing. In most cases with films on PAL DVDs you can just force disable deinterlacing and you will get a 25 fps progressive image (I've got only one PAL DVD in my film collection where this isn't the case). You will then see 'deinterlacing off' and 'frame time 40 ms'

[Furgiuele]

Hi. I post a question yesterday. Please, can someone help me?

[Furgiuele]

Hi. I'm Italian and new on this forum and I don't speak English very well.
I appreciate a lot the guide about MadVR and I'm trying to optimize it for my needs.
I have a htpc with MSI Z390A-pro mobo, I5-9600K cpu, 16 Gb ram, Nvidia 1660 gpu, bd drive, ssd and nas with a lot of concerts and movies ripped from br discs.
I have a Sony VPL-vw870es (calibrated BT2020) as projector and a Trinnov Altitude 16 as processor. I use Jriver Media Center 25 with MadVR, but I don't like to use more than one setup.
So I set MadVR with PC levels 0-255,10 bit, already calibrated BT2020 pure power curve 2.20, 2160p23, 24, 25, 29, 30, 50, 60, tone map HDR using pixel shaders, tp 100, BT2390, highlight medium, measure each frame's peak; if in doubt, deactivate deinterlacing, disable automatic source type, force video mode, only look at pixels in the frame center; chroma ups. NGU Antialias medium, image ups. luma NGU Sharp, luma qu. direct qudrupole very high, chroma high; "use Direct3D 11..." and " "present a frame for every VSync" flagged in general settings, smooth motion only if there would be motion judder; ordered dithering, no flag in trade for quality.
It works very fine with all 23.970 fps videos (4k hdr, fhd, etc.), without interlacing. But I've issue to watch 29.970 fps videos, with a lot of dropped frames, motion juddering and stuttering. I have a lot of concerts with this fps (Diana Krall, Carlos Santana, Miles Davis, ecc.).
Can someone help me?
Thanks.

Please, can someone help me?:thanks:

[ashlar42]

Please, can someone help me?:thanks:You need to be patient. Nobody is payed here to provide assistance. Everybody does it in his/her free time. Be patient and somebody capable enough will try to help you, I'm sure.

[mclingo]

Please, can someone help me?:thanks:

have you created a custom display mode in MADVR for 29.970?

disply modes/custom modes/add

[Klaus1189]

@Furgiuele:
See post #547 in AMD, Intel and Nvidia driver issues and last recommended version (https://forum.doom9.org/showthread.php?p=1883743#post1883743)
Don't use 25Hz, 29Hz and 30Hz modes, madvr picks these three first, so delete them, if you want to play properly interlaced content which has double "motion resolution" by creating fields (top field and bottom field) and not frames and store two fields in one frame.
But use 2160p for your setup, if you have "4K" Equipment ;)

[huhn]

I have noticed inconsistent behaviour when upscaling DVD to 4k, it behaves something like it isn't upscaling for some period of time and then it corrects itself and starts upscaling again. This can happen at arbitrary times, e.g. within a single scene without having touched any playback controls or changed any settings.

An example to illustrate, I created these by just replaying a scene on a loop. It rendered it "correctly" v few times, it is mostly looking bad. Look at the guy on the wire for the obvious difference. The fact this is two different frames is irrelevant btw as the effect occurs regularly on different content, i.e. it's not specific to one scene or one particular source file.

Anyone seen something like this before? Any tips on how to resolve?

Bad - https://ibb.co/7YLkcg9
Good - https://ibb.co/P6VwpKD

you are deinterlacing 25i to 50p which is the correct thing to do for a deinterlacer but you are displaying it at 25 which creates all kind of problems.

the next issue you are dropping frames and quite a lot of them because of some render spikes it seems hard to judge screens are from the paused state.

hardware deinteralcing needs to get a constant stream of frames not feeding it one frame my already break it so try a different player/decoder and it needs multiply frames to kick in so you can skip into the file and look at the frame it needs like 3 or frames to even start working.

[Siso]

Can't remember which one is better for creating a custom resolution - nvidia control panel or cru 1.4.1.

[3ll3d00d]

you are deinterlacing 25i to 50p which is the correct thing to do for a deinterlacer but you are displaying it at 25 which creates all kind of problems.
what sort of problems?

I don't watch many DVDs admittedly but I haven't noticed any issues (apart from the deinterlacing one). I'm bandwidth constrained (due to HDBaseT) til I can run a new hdmi cable anyway.

the next issue you are dropping frames and quite a lot of them because of some render spikes it seems hard to judge screens are from the paused state.
there always seem to be a few dropped frames when playback starts, the stats shown are from pausing and jumping back and forth so aren't indicative of normal behaviour. i.e. it isn't regularly dropping frames in normal playback.

hardware deinteralcing needs to get a constant stream of frames not feeding it one frame my already break it so try a different player/decoder and it needs multiply frames to kick in so you can skip into the file and look at the frame it needs like 3 or frames to even start working.
It would stick in that state for a prolonged period of time as opposed to flickering in and out. As far as I can see, switching the config in LAV so madvr is presented with an already deinterlaced stream seems to produce a better result. I'm not sure exactly what is changing in this case though as I'm comparing my LAV config to the jriver packaged version (which they don't expose config for)

[Klaus1189]

@3ll3d00d
Take a look at my previous post, same issue here.

See post #547 in AMD, Intel and Nvidia driver issues and last recommended version (https://forum.doom9.org/showthread.php?p=1883743#post1883743)
Don't use 25Hz, 29Hz and 30Hz modes, madvr picks these three first, so delete them, if you want to play properly interlaced content which has double "motion resolution" by creating fields (top field and bottom field) and not frames and store two fields in one frame.
But use 2160p for your setup, if you have "4K" Equipment ;)

[Furgiuele]

You need to be patient. Nobody is payed here to provide assistance. Everybody does it in his/her free time. Be patient and somebody capable enough will try to help you, I'm sure.

Sorry! And thanks to all of you.

[Furgiuele]

[QUOTE=el Filou;1888560]Based on your description, what could be happening is those 29,970 fps concerts are telecined, and as you are forcing video mode they are deinterlaced with very bad quality to 59,94 fps and you have 3:2 motion judder, but it's hard to be sure without knowing what the source is: Blu-ray? DVD? TV?
Try to enable film mode for those concerts (you can assign a keyboard shortcut key to it) and see if it is better.

:goodpost: Thanks!!! It solves my issue. I disabled "automatically activate deinterlacing..." and I flagged "disable automatic source... force film mode... only look at pixels...". Now no dropped frames, no stutter, no judder"! Can I leave this settings for all my videos? I have all blu ray discs ripped (fhd and uhd)...

[Furgiuele]

Sorry, I have another question. I have a Sony vpl-vw870es prjector (laser lamp with 2200 lumen). MadVR v0.92.17. In devices, my device, hdr, I flagged "tone map HDR using pixel shaders", "measures each frame's peak..." and "output video in HDR format". How many nits I've to wright in "target peak nits"? Thanks.

[Asmodian]

Can I leave this settings for all my videos? I have all blu ray discs ripped (fhd and uhd)...

Anything that is actually interlaced will look bad but you might not have any real interlaced content.

Also, once you disabled automatically activating deinterlacing the other options don't do anything.

[huhn]

what sort of problems?
well you need rendertiems below 20 ms but oyu are not using these iamges true interlanced content can't be displayed correctly
I don't watch many DVDs admittedly but I haven't noticed any issues (apart from the deinterlacing one). I'm bandwidth constrained (due to HDBaseT) til I can run a new hdmi cable anyway.
cables have nothing todo with this every cable that has the bandwidth of hdmi 1.3 can do 50 hz at 4:2:0 but i have no clue what HDBaseT is.

there always seem to be a few dropped frames when playback starts, the stats shown are from pausing and jumping back and forth so aren't indicative of normal behaviour. i.e. it isn't regularly dropping frames in normal playback.

your OSD already shows 19 ms from the max of 20 and spikes with 50+
It would stick in that state for a prolonged period of time as opposed to flickering in and out. As far as I can see, switching the config in LAV so madvr is presented with an already deinterlaced stream seems to produce a better result. I'm not sure exactly what is changing in this case though as I'm comparing my LAV config to the jriver packaged version (which they don't expose config for)
maybe madVr is able to drop frames before they are rendered which could stop deinterlacing.

Sorry, I have another question. I have a Sony vpl-vw870es prjector (laser lamp with 2200 lumen). MadVR v0.92.17. In devices, my device, hdr, I flagged "tone map HDR using pixel shaders", "measures each frame's peak..." and "output video in HDR format". How many nits I've to wright in "target peak nits"? Thanks.

is not recommended to output HDR as HDR on projector thanks to the very basic HDR tone mapping.
i can't give you a nit number to many changes on a topic i don't care yet.

on an old madVR build 2x output nits where ok no clue about projector that would be in the 100-300 range.

[Furgiuele]

is not recommended to output HDR as HDR on projector thanks to the very basic HDR tone mapping.
i can't give you a nit number to many changes on a topic i don't care yet.

on an old madVR build 2x output nits where ok no clue about projector that would be in the 100-300 range.

Many thanks.
Anyway I can't flag "use direct 3D 11 for presentation", else I've random skip in motion (altough without dropped frames).
Do you think I've to change my gpu (GTX1660) with RTX 2060 or AMD RX 590 to have a better rendering?:thanks:

[Fandangos]

Can you guys tell me if users with Ryzen integrated gpus have any kind of problem outputing the correct color space to the display using MadVR renderer?

HDR metadata is correctly triggered but color space remains in bt.709?
Or is this a non existent problem?

[Klaus1189]

What driver version do you use? What display/projector do you use? What player do you use? MPC-BE, MPC-HC, ...

[huhn]

Many thanks.
Anyway I can't flag "use direct 3D 11 for presentation", else I've random skip in motion (altough without dropped frames).
Do you think I've to change my gpu (GTX1660) with RTX 2060 or AMD RX 590 to have a better rendering?:thanks:

can't be judged with these informations.
there are a lot of possible setting combination and a 1660 can clearly do d3d11.

the equivalent to a 2060 is the rx 5700 not the rx 590 i can't recommend them.

[Warner306]

Can you guys tell me if users with Ryzen integrated gpus have any kind of problem outputing the correct color space to the display using MadVR renderer?

HDR metadata is correctly triggered but color space remains in bt.709?
Or is this a non existent problem?

There have been at least two Ryzen users who were failing to get anything other than BT.709 from HDR passthrough. The Ryzen drivers may have an issue identifying the color coordinates sent by madVR and change the color space.

[DMU]

There have been at least two Ryzen users who were failing to get anything other than BT.709 from HDR passthrough. The Ryzen drivers may have an issue identifying the color coordinates sent by madVR and change the color space.

There may be an issue with the entire AMD Vega GPU architecture, which are used in the iGPU embedded in the Ryzen processors. This can be verified by users of Vega 56/64 and Radeon VII.

[Furgiuele]

can't be judged with these informations.
there are a lot of possible setting combination and a 1660 can clearly do d3d11.

the equivalent to a 2060 is the rx 5700 not the rx 590 i can't recommend them.
I can't use D3D11! Don't know why. On Nvidia setup I use 3840 x 2160 60hz res, Nvidia color curve Ycbcr444, 8bit depth and I can't flag anything else.
I have a Sony vpl-vw870es projector (actually the best of Sony). Now I order from Amazon a Gigabyte RTX 2060 mini ITX OC6G and next week I'll see if I can use D3D11 without skip in motion. But I don't know if I can keep it because my psu is 350 watt 80 plus gold.
I hope you can give me a good suggestion...
Thanks

[Owyn]

yesterday madvr was working fine but today mpc-hc videoplayer just closes itself after long "opening file" when I try to play anythig with madvr selected as renderer
I didn't update anything (perhaps windows did? (thought after checking update log it shows nothing for at least two weeks))

here's the debug log after activating the debug mode: https://pastebin.com/82X68xn4

edit: after todays cumulative update for .Net it started working again yay

[wolfman2791]

Just wanted to chime in here... a lot of the issues have to do with Windows updates. I finally had to refresh my windows 10 install and it fixed everything with madvr (i run jriver). If you're like me and you tweak your windows, it creates problems.

But before you do that, uninstall the latest windows updates one by one and see where the break occurs. That's how i noticed it was a windows issue.