madVR - high quality video renderer (GPU assisted)

[ADude]

To answer my own question, I discovered that the complete change log can be found at:

http://www.videohelp.com/software/madVR/version-history#changelog

[sneaker_ger]

There's also a changelog.txt in every madVR package.

[huhn]

thanks.
What driver version is this?

In 350.12 the range settings are still in the "adjust desktop color settings" menu.
352.63

[Siso]

352.63

Will it work on gtx 550 ti?

[huhn]

Will it work on gtx 550 ti?

i don't know for sure it's a dx11 GPU should work.

[nevcairiel]

Thanks.
What driver version is this?

In 350.12 the Range settings are still in the "adjust desktop color settings" menu.

The options moved around in the windows 10 drivers, presumably because MS also improved support for this kind of thing on the desktop.
It is however not recommended to install these drivers on any earlier version of Windows.

[madshi]

Damn :( I wanted to sleep early tonight... now I will have to stay awake all night... refreshing my browser :)
FWIW, be aware that there are different time zones. I'm going to bed right now, and have a full day of work on madVR left tomorrow. So it will be at least be 12 hours, maybe 16 from now. Can't say for sure...

[toniash]

FWIW, be aware that there are different time zones. I'm going to bed right now, and have a full day of work on madVR left tomorrow. So it will be at least be 12 hours, maybe 16 from now. Can't say for sure...
Good morning & good work :):)

[James Freeman]

I wonder if 0.88.0 will be a bug fix release, or there will be new goodies inside.... can't wait.

[nevcairiel]

I wonder if 0.88.0 will be a bug fix release, or there will be new goodies inside.... can't wait.

Inrementing the first number usually has a specific meaning. Although madshi has had a few features in the minor releases as well!

[MokrySedeS]

My friend has a problem with madVR and here's what he says:I recently tried switching from PotPlayer (version 1.5.39659) to MPC-HC – latest nightly 1.7.8.162 (+madVR – 0.87.21) and I'm having some problems. I'm getting "madVR reports: creating direct3D failed 80070005" error on MPC-HC but it only happens when my TV switches to fullscreen from 60Hz to 24Hz. I probably should add that PotPlayer works fine with the same madVR settings including the following display modes "1080p23, 1080p24, 1080p29, 1080p30, 1080p50, 1080p59, 1080p60". If I delete my display settings from madVR then everything is fine.

I'm using two displays. One connected through dvi cable and tv is connected through HDMI. Graphic card GeForce GTX 770 Twin Froz 2GB – GPU version: 350.12, Release date: 4/13/2015.

Here's a log file: http://www74.zippyshare.com/v/r9ZZa1fS/file.html
I failed to help him unfortunately :(

[XMonarchY]

Something is wrong with madVR again. I made a Rec.601/SMPTE C 3DLUT with dispcalGUI 3.0.0 and ArgyllCMS 1.7.0, but when I loaded it into madVR on the appropriate SMPTE C line, I cannot press Accept because I get repeating error that I need to specify which external 3DLUT I want to use for that display, although SMPTE C 3DLUT was selected!

Also, madTPG stopped loading completely... The control settings in lower-right corner do load and I get madTPG in taskbar, but the actual window does not show up, even after a restart.

Finally, I have a question - if I select BOTH Rec.709 and SMPTE C 3DLUT's for appropriate line, how will madVR know which one to use? I know 3DLUT type can be selected in madTPG if it was working properly, but aside from madTPG, how would madVR know? Would it base it on content? Some content does not specify whether it uses Rec.709 or Rec.601/SMPTE C ...

[huhn]

Something is wrong with madVR again. I made a Rec.601/SMPTE C 3DLUT with dispcalGUI 3.0.0 and ArgyllCMS 1.7.0, but when I loaded it into madVR on the appropriate SMPTE C line, I cannot press Accept because I get repeating error that I need to specify which external 3DLUT I want to use for that display, although SMPTE C 3DLUT was selected!

Also, madTPG stopped loading completely... The control settings in lower-right corner do load and I get madTPG in taskbar, but the actual window does not show up, even after a restart.

Finally, I have a question - if I select BOTH Rec.709 and SMPTE C 3DLUT's for appropriate line, how will madVR know which one to use? I know 3DLUT type can be selected in madTPG if it was working properly, but aside from madTPG, how would madVR know? Would it base it on content? Some content does not specify whether it uses Rec.709 or Rec.601/SMPTE C ...

yeah that's a bug.

it is checked if the source is flagged with a colorspace and if this is not the case madVR uses the resolution of the video to choice that.
if you create a 720p video with bt 601 and didn't flag this in the stream you failed as a creator and bt 709 will be used.

[potomac]

Using Bicubic should work fine for image upscaling. For chroma you can also use Bicubic, if your GPU can handle that, or Bilinear otherwise. I think playback should work just fine with madVR for Blu-Ray, too. No need to use EVR for that.

thanks for the answer. everything works fine:
chroma: bicubic 75
upscaling: lanczos 3 taps
and some trade quality for performance
but to play this blu ray i need to disable deinterlacing in lav video:

Video #1
ID : 4113 (0x1011)
Menu ID : 1 (0x1)
Format : AVC
Format/Info : Advanced Video Codec
Format profile : High@L4.1
Format settings, CABAC : Yes
Format settings, ReFrames : 2 frames
Codec ID : 27
Duration : 3h 1mn
Bit rate mode : Variable
Maximum bit rate : 34.0 Mbps
Width : 1 920 pixels
Height : 1 080 pixels
Display aspect ratio : 16:9
Frame rate : 29.970 fps
Color space : YUV
Chroma subsampling : 4:2:0
Bit depth : 8 bits
Scan type : MBAFF
Scan order : Top Field First
Color primaries : BT.709
Transfer characteristics : BT.709
Matrix coefficients : BT.709
Color range : Limited
format_identifier : HDMV
Source : 00101.m2ts

[michkrol]

but to play this blu ray i need to disable deinterlacing in lav video

Nothing to worry about, disabling it means madVR does the deinterlacing. Software deinterlacing in LAV uses your CPU, while madVR uses your GPU to deinterlace and it's usually better to leave the deinterlacing to madVR.
To make sure everything works fine, check if the debug OSD (CTRL+J) says "deinterlacing on" (right over the queues in the OSD).

[bcec]

352.63

I see range setting show up for a split of a second and then it gets removed. Is there any particular thing I need to do to get to see those options in the first place? Using 350.12

EDIT: Oh, is this a windows 10 thing only?

[billqs]

My friend has a problem with madVR and here's what he says: Originally Posted by creativeopinion
I recently tried switching from PotPlayer (version 1.5.39659) to MPC-HC – latest nightly 1.7.8.162 (+madVR – 0.87.21) and I'm having some problems. I'm getting "madVR reports: creating direct3D failed 80070005" error on MPC-HC but it only happens when my TV switches to fullscreen from 60Hz to 24Hz. I probably should add that PotPlayer works fine with the same madVR settings including the following display modes "1080p23, 1080p24, 1080p29, 1080p30, 1080p50, 1080p59, 1080p60". If I delete my display settings from madVR then everything is fine.

I failed to help him unfortunately :(

Where are display settings stored in madvr? I've had a dickens of a time getting 60hz to work in 4k after switching from a R9280x to a GTX960.

I finally got it to work in MPC-HC by forcing the refresh rate change to 60hz from 23 hz with a cap on 30hz, but that shouldn't be necessary since the new card is capable of 60hz@4k (that and the HEVC hardware encoder are the only reasons I side-graded with the 960 in the first place.) I was looking for a config file in MPC-HC that might be limiting me, but if the above is true, I should be hunting for one in Madvr.

[billqs]

BTW, I wasn't overtaxing the GPU in getting the freezes, I was only getting them when NNEDI3 was engaged. All other combinations I tried worked. The reason it became a hard crash is that I was operating in "fullscreen exclusive mode" in madvr. When I used "windowed mode" I eventually got a freeze using NNEDI3, but because it didn't crash like before with exclusive use, I was able to get a cursor and get out of the program back to desktop. I haven't had a crash yet with the 960 card, but I believe the OpenCL regkey entry that was mentioned a few pages back got rewritten with the installation of the new card.

[madshi]

My friend has a problem with madVR and here's what he says:
I failed to help him unfortunately :(
Does he have the display modes setup only in madVR or also in MPC-HC? If MPC-HC and madVR are both trying to change display modes, that can cause problems.

Something is wrong with madVR again. I made a Rec.601/SMPTE C 3DLUT with dispcalGUI 3.0.0 and ArgyllCMS 1.7.0, but when I loaded it into madVR on the appropriate SMPTE C line, I cannot press Accept because I get repeating error that I need to specify which external 3DLUT I want to use for that display, although SMPTE C 3DLUT was selected!

Also, madTPG stopped loading completely... The control settings in lower-right corner do load and I get madTPG in taskbar, but the actual window does not show up, even after a restart.
Can you please create a bug entry on the madVR bug tracker with a detailed guide on how to reproduce this bug? Thanks.

Where are display settings stored in madvr? I've had a dickens of a time getting 60hz to work in 4k after switching from a R9280x to a GTX960.

I finally got it to work in MPC-HC by forcing the refresh rate change to 60hz from 23 hz with a cap on 30hz, but that shouldn't be necessary since the new card is capable of 60hz@4k (that and the HEVC hardware encoder are the only reasons I side-graded with the 960 in the first place.) I was looking for a config file in MPC-HC that might be limiting me, but if the above is true, I should be hunting for one in Madvr.
Just because one random user had a problem which he solved by deleting something from his madVR settings doesn't mean it will solve a totally different problem for you.

Right now I don't have enough information to really help. The first thing to check is whether the Windows desktop can be switched to 4kp60. If that works, it should work with madVR, too.

[MokrySedeS]

Does he have the display modes setup only in madVR or also in MPC-HC? If MPC-HC and madVR are both trying to change display modes, that can cause problems.

I'm pretty sure that he has default mpc-hc settings (except for using madVR instead of evr-cp of course), so we can rule that out.

[madshi]

I'm pretty sure that he has default mpc-hc settings (except for using madVR instead of evr-cp of course), so we can rule that out.
Maybe he can try installing a different/newer driver? I'm not sure what else to suggest. Most other users don't seem to have that specific problem. Well, he could also try to switch display modes with MPC-HC instead of madVR, but usually I recommend it the other way round...

[madshi]

madVR v0.88.0 released

http://madshi.net/madVR.zip

* added Direct3D 11 presentation path
* added support for native 10bit output
* added NEDI image doubling option
* added "image enhancements" settings page for source enhancements
* added "upscaling refinement" settings page for upscaling post processing
* added FineSharp to source enhancement and upscaling refinement
* added LumaSharpen to source enhancement and upscaling refinement
* added SuperRes to upscaling refinement
* added chroma upscaling option "activate SuperRes filter"
* added Shiandow's deband algorithm
* removed chroma image doubling options
* fixed some 64bit stability issues
* fixed crash when two madVR instances had the same parent (DVBViewer BiB)
* fixed one potential DXVA decoding freeze
* trade quality option "lose BTB and WTW..." is disabled by default now
* added SettingsSetBoolean("FinalizeOnWMClose") option
* (un)install.bat now complains if you don't run them as admin
* (un)install.bat now doesn't try to install madVR64 on a 32bit OS, anymore
Many many new features and options for you guys to play with - enjoy! :)

Some random comments:

1) The Direct3D11 presentation path seems to work very well for me, maybe even better than the D3D9 presentation path. Please give it a try and let me know how it works for you. I've noticed that on my PC the D3D11 path seems to react a bit faster when switching between windowed <-> exclusive mode. Also it has two advantages: First of all, it allows native 10bit output (see below). Secondly, in FSE mode D3D11 allows me to skip some VSyncs without having to present a frame, while still being able to present frames in advance. Which means that GPU performance in D3D11 FSE mode should be slightly better compared to D3D9, if you have a display refresh rate which is higher than the source frame rate.

2) Now native 10bit output is supported. Interestingly, my JVC projector looks bad with 10bit output, it looks much better with 8bit output. My LCD monitors look fine with 10bit output, though. So this may vary from display to display. AFAIK, most displays today do *not* have a native 10bit panel yet. So in many cases it is possible that you'll get a better image quality outputting dithered 8bit from madVR. But let your eyes be the judge.

3) I've decided to offer sharpening in two different places: In the "processing\image enhancements" page. This page applies sharpening to the source video, when it is still in its original resolution. And in the "scaling algorithms\upscaling refinement" page. The algorithms on this page can be run during/after upscaling. The general idea is that there are 2 different causes of blurred video: Either the original source is already soft/blurred. In that case it makes sense to sharpen the original sources. This will vary from source to source, though. The other cause of soft/blurred video is if you upscale. The larger the upscaling factor, the blurrier the result. So my thinking is that it makes sense to apply a certain amount of sharpening as part of the upscaling process to make sure that upscaling doesn't make the image blurrier than it originally was.

4) All the algorithms in the "upscaling refinement" page can be applied either just once, after upscaling is done. Or alternatively you can tell madVR to upscale the video in multiple steps, each one with a 2.0x scaling factor, and to apply upscaling refinement after each step. This makes sense because all the refinement algorithms work better that way. They are less effective if you run them after e.g. a 4.0x upscaling.

5) Please note that currently there are tons and TONS of options. This definitely is overkill and needs to be cleaned. This is a task we have to work on together. For each algorithm I'd like to define max 3 options (e.g. low, mid, high) and that must be sufficient. Otherwise the sheer amount of options will shy new users away.

-------

Now I need FEEDBACK from you guys. But since there are so many new features/options, it doesn't make sense to just collect feedback from everyone about everything. That just wouldn't work well. So please allow me to coordinate testing and feedback:

I would like to start with evaluating the new "DEBAND" algorithm written by Shiandow. Please could you guys try it with your various sources and let me know how it compares to madVR's other debanding algorithms in terms of quality and performance? Based on your feedback I will either remove the new algorithm again, or I will replace one or multiple of the old options with the new one. I want to end up with only one Deband option with max 3 different settings. Your feedback is very welcome - thanks!

[sneaker_ger]

1. cosmetic: settings still say "87.21"
2. Microsoft EMET goes crazy after closing MPC-HC. Like 20 error messages.

[madshi]

Settings say "madVR v0.88" on my PC. Please make sure your browser cache didn't give you an old file. Does the changelog include info for v0.88.0?

[Qaq]

@madshi
Thanks for the new version.
Seems its time to upgrade from Catalyst 13.12 to try 10 bit output. Can you recommend any newer Catalyst version?

[madshi]

I don't think you need to upgrade the driver. 13.12 should support 10bit output just fine.

[sneaker_ger]

Changelog does include 0.88.0 and I always go out of the way to delete old madVR folder completely. I will try again, though.

I also got crashes and a log on the desktop (will do proper reports later). Did you mix up debug and normal .ax (x64)?

[madshi]

Don't know what to say. Both release and debug builds are definitey stamped as v0.88.0 (I've downloaded and checked the zip from my own server again, just to be safe), and the settings dialog fetches this information from the ax files. Must be something strange going on on your PC. Maybe some other madVR installation from ZoomPlayer, J.River MC or some other media player?

[MartinW]

Don't know what to say. Both release and debug builds are definitey stamped as v0.88.0 (I've downloaded and checked the zip from my own server again, just to be safe), and the settings dialog fetches this information from the ax files. Must be something strange going on on your PC. Maybe some other madVR installation from ZoomPlayer, J.River MC or some other media player?

I also get v0.87.21 as version number. I noticed that the 64 bit version (madVR64.ax) is dated the first of May?

[Ver Greeneyes]

64-bit MPC-HC 1.7.8.162 crashes on exit with an access violation with this version (running on Windows 7). Stack:

fffff9600008e6d8() Unknown
user32.dll!UserCallWinProcCheckWow() Unknown
user32.dll!SetWindowTextW() Unknown
> mpc-hc64.exe!CWnd::SetWindowTextW(const wchar_t * lpszString) Line 246 C++
mpc-hc64.exe!CMainFrame::OpenSetupWindowTitle(bool reset) Line 11356 C++
mpc-hc64.exe!CMainFrame::OnFilePostClosemedia(bool bNextIsQueued) Line 3463 C++
mpc-hc64.exe!CMainFrame::CloseMedia(bool bNextIsQueued) Line 14604 C++
mpc-hc64.exe!CMainFrame::OnClose() Line 984 C++

Could be a bug in MPC-HC of course. Haven't tried the 32-bit version. Doesn't seem to matter if I use the DX11 presentation path or not, though I haven't tried resetting settings altogether.

Edit: Still crashes in the refreshed build.

[vomanci]

CTRL+J doesn't work, no OSD is displayed

[XMonarchY]

yeah that's a bug.

it is checked if the source is flagged with a colorspace and if this is not the case madVR uses the resolution of the video to choice that.
if you create a 720p video with bt 601 and didn't flag this in the stream you failed as a creator and bt 709 will be used.

OK, 2 questions then:

1. Can I just use Rec.601 3DLUT for both lines - Rec.709 & Rec.601 to force Rec.601 in whichever video I have?

2. The other bug is a bigger problem. I can no longer start madTPG and see madTPG window. When I try, I get a madTPG taskbar icon, and mad home cinema control icon in lower right corner, both of which also show up in Task Manager, but the actual madTPG window never shows up, even after I restart my PC! How do I go around that?

[madshi]

Ah, stupid me, I still need to get used to having 64bit support now. I totally forgot to recompile the 64bit version. Will do so in a minute...

[huhn]

OK, 2 questions then:

1. Can I just use Rec.601 3DLUT for both lines - Rec.709 & Rec.601 to force Rec.601 in whichever video I have?

2. The other bug is a bigger problem. I can no longer start madTPG and see madTPG window. When I try, I get a madTPG taskbar icon, and mad home cinema control icon in lower right corner, both of which also show up in Task Manager, but the actual madTPG window never shows up, even after I restart my PC! How do I go around that?

no don't do that.
use a rec 709 madVR does the rest. of cause you can add the 601 with the bt 709 but don't put a 601 in the 709 slot!

[madshi]

Ok, zip refreshed with a proper 64bit build inside now:

http://madshi.net/madVR.zip (v0.88.0)

32bit build unchanged.

[Devrim]

Ok, zip refreshed with a proper 64bit build inside now:

http://madshi.net/madVR.zip (v0.88.0)

32bit build unchanged.

:thanks:

OSD is working here again!

[luk008]

There are a lot of new options! Thank you for the great job, madshi! :thanks:

How do I know if my display support 10bit color output? My tv is an LG LB6500.

[madshi]

Use the madTestPattern test patterns (see link in the first post of this thread) to compare image quality with 8bit and 10bit output. The test patterns should be smooth and banding free.

[Sparktank]

My goodness, you've been busy. :cool:

Thanks for the mega update.

Going to remux Insidious, I know that movie suffers dearly from banding.

[MS-DOS]

Both 32 and 64-bit MPC-HC 1.7.8.162 crash on Ctrl+C or on exit now, no MadVR settings changed.

[ikarad]

Bug with madvr0.88 64bits.
When I quit mpc-hc with madvr, mpc-hc crash.

If I activate use directd11 for presentation in win7 64 there are lots of presentation glitches in fullscreen mode.

If I use 10bit, in windowed mode, image is very dark. In fullscreen mode, it's like as 8 bit.

[ikarad]

Use the madTestPattern test patterns (see link in the first post of this thread) to compare image quality with 8bit and 10bit output. The test patterns should be smooth and banding free.

When madtestpattern is installed. How use it?

[yok833]

Thanks Madshi for all the new options.. we have a lot to play with now :)

When I switch to the new D3D 11 presentation it become very laggy ..; however the queues are all full and the rendering stats are very good (around 19 ms as usual). Everything back to normal when i swith to D3D 9... any ideas?

update : As Ikarad said I have a lot of presentation glitches too and it only happen when I am in fullscreen - Win7 64

[luk008]

Why 10 bit only looks bad in windowed mode and okay in fullscreen mode? Is there any explanation for that?

[Moony349]

Also having crashing and presentation glitches in FSE mode with x64 MPC HC 1.7.8 build 162 on Windows 7.

Crashing on closing MPC HC occurs regardless.

Presentation glitches only occur when using DX11.

EDIT: Also crashing when moving between monitors.

[MS-DOS]

Render queue is stuck at 3-4\8 using D3D11 and presenting 4 frames in advance in windowed mode. Increasing presented frames changes nothing, while setting them to 1 fills up the render queue to its maximum and... causes hundreds of presentations glitches :D

[noee]

fyi, Appears to be the same crashing on MediaCenter20 as on mpc-hc, changing video or closing the player. Also on video stop (Ctrl+s) on MediaCenter20.

[tickled_pink]

Crashing on exit also occurs is MPC-BE.1.4.5.364.x86

[madshi]

My goodness, you've been busy. :cool:

Thanks for the mega update.

Going to remux Insidious, I know that movie suffers dearly from banding.
Yes, quite buys, it was a lot of work. Debanding feedback would be great, thanks.

If I activate use directd11 for presentation in win7 64 there are lots of presentation glitches in fullscreen mode.
Fullscreen exclusive mode or fullscreen windowed mode?

If I use 10bit, in windowed mode, image is very dark. In fullscreen mode, it's like as 8 bit.
Very strange. I don't see that here. Anybody else with the same problem?

update : As Ikarad said I have a lot of presentation glitches too and it only happen when I am in fullscreen - Win7 64
Also having crashing and presentation glitches in FSE mode with x64 MPC HC 1.7.8 build 162 on Windows 7.
Render queue is stuck at 3-4\8 using D3D11 and presenting 4 frames in advance in windowed mode. Increasing presented frames changes nothing, while setting them to 1 fills up the render queue to its maximum and... causes hundreds of presentations glitches :D
Seems to me the presentation glitch problems with D3D11 are probably all caused by Windows 7 behaving differently than Windows 8. Not totally sure, though. On my Windows 8.1 x64 PC everything runs well.

Why 10 bit only looks bad in windowed mode and okay in fullscreen mode? Is there any explanation for that?
Which OS are you using?

[madshi]

madVR v0.88.1 released

http://madshi.net/madVR.zip

* fixed media player shutdown crash/freeze