madVR - high quality video renderer (GPU assisted)

[xabregas]

xabregas, what driver are you running with your HD7770? Also, which decoder, player and what OS?

MPC HC + LAV filters (lav video descoder + lav audio decoder + lav splitter source) + reclock (wasapi)

windows 7 x86

driver is 12.10 official

going to install 13.1 now, FEAR :scared:

[noee]

I'm running 13.1 now on HD6570. MadVR with MPC-HC and jRiver MC18 is flawless on film SD and HD (bluray rips) material (jinc3/AR for both Chroma/Luma). Running a chessy old Athlon II...10bit x264-encoded sources....Encoding video with x264 at the same time as playback....

Anyway, are any of your madVR queues starved?

Try going into REclock advanced and "Clean up Video Timings Database", then start a video and let it get it's groove back on, see what happens.

[Dodgexander]

In "PAL" video streams (1080i/25 or 576i/25), there's basically only two types of video you can get: true interlaced content (1080i/25 or 576i/25), or progressive content (1080p/25 or 576p/25). With interlaced content you want to deinterlace using the best algorithm available (e.g. vector/motion adaptive) to get a progressive video suitable for display (1080p/50 or 576p/50). For progressive content, you want to simply merge pairs of fields together because they belong to a single frame with no motion (i.e. weave deinterlacing).

However, when you tell a GPU to perform deinterlacing, it should detect whether the video is actually progressive and thus weave needs to be used rather than vector/motion adaptive. Therefore, you can basically leave deinterlacing on all the time ("video mode") and it'd play back everything perfectly (25p content would be frame-doubled to 50p but this should have no impact on image quality).

Unfortunately, there are apparently some combinations of GPUs and progressive video content where this doesn't happen as intended. Therefore you need to force "film mode" for these to get proper weave deinterlacing so that quality isn't sacrificed by unnecessary vector/motion adaptive deinterlacing.

I don't believe I have any such content (or my GPUs aren't affected) so I can't test this.


Okay thats clear, but how is that related to my problem?

If a channel switches from 1080/25p too 1080/25i, if using dxva deinterlacing, shouldn't it detect it needs to go from "no deinterlacing, or "weave" as you describe from the progressive source. Too "vector" for the interlaced source?

Because in my case it doesn't. It always thinks the video is progressive, even when I change the deinterlacing in the AMD control panel to vector or motion, there is no deinterlacing, or just "weave" deinterlacing for the 1080/25i video.

So since this just happens with Madvr, is this a communication problem between the renderer and the amd driver?

Why doesn't it work to set the deinterlacing to vector manually when this occurs?

Why does it work forcing deinterlacing on in Madvr?

If it could detect the change, should it be sending the information to change the deinterlacing type too the driver?

[dansrfe]

Two questions:

1) Is it better to use SVP in conjunction with madVR or does it not make a difference?

2) If I were to use SVP with madVR does the need for ReClock still arise or does SVP eliminate the problem which ReClock is trying to solve as well?

[QBhd]

I tried SVP but could not stand the look of it... It made everything nice and crisp, but it looked cheap and sterile and looked like a soap opera. Not my cup of tea, but some people love it.

QB

[Qaq]

2) If I were to use SVP with madVR does the need for ReClock still arise or does SVP eliminate the problem which ReClock is trying to solve as well?As far as I remember SVP author himself used to use ReClock.

[AndreaMG]

Two questions:

1) Is it better to use SVP in conjunction with madVR or does it not make a difference?

2) If I were to use SVP with madVR does the need for ReClock still arise or does SVP eliminate the problem which ReClock is trying to solve as well?

1) SVP intepolates newly added frames and works fine with MadVR, it does makes a difference if you like (as me) watch videos at 50/59.94fps. It works via avisynth, it would be awsome (for stability and performance) if it could be integrated into the video renderer :)

2) Reclock is still needed unless you have perfect timing.

Please visit SVP forum. The authors are nice guys and the result SVP is able to give in my case are FAR BETTER than the frame interpolation of my TV set.;)

[madshi]

They all do better than this.
I'm not sure about that. First of all: See my reply after the next quote. Furthermore: Your screenshot of the Cheese Slice test pattern does not show how proper motion adaptive deinterlacing with diagonal filtering looks like. Proper motion adaptive deinterlacing looks *much* better than that! You're putting way too much stock into the Cheese Slice tests. Furthermore, AMD's and NVidia's implementation of motion adaptive deinterlacing seems to be far from optimal, as far as I can see...

That sounds about right for motion adaptive deinterlacing but modern GPUs use vector adaptive deinterlacing, which is far better at reproducing the "original" full resolution progressive video than motion adaptive deinterlacing. Both the GTS250 and GT430 do some form of vector adaptive deinterlacing.

These examples are from the Cheese Slice test
Those Cheese Slice test patterns were the reason that I said that NVidia and ATI *might* do some sort of motion compensation. But I've just checked the Cheese Slice test patterns again, and I have to *totally* change my opinion. Actually, it seems everybody has been misinterpreting the Cheese Slice tests! Because, I've just found out that they're actually encoded as *FILM* with a 2:2 cadence. Just try it out: Play those Cheese Slice tests with madVR film mode. madVR will detect a 2:2 cadence for both 50i and 60i Cheese Slice tests. In the moment when motion starts, madVR will take a moment to switch onto the correct cadence, but once it has switched, playback is perfect. After that I've looked at the separate fields in the video stream, and really, every 2 fields belong together to form a progressive frame. So the Cheese Slice tests must be totally disregarded as *video* mode deinterlacing tests. Motion compensation wouldn't work at all for this test pattern because motion compensation expects movement between every field, which this test pattern does not have.

Edit: Btw, you can now use the 1080i50 Cheese Slice test patterns to check whether your GPU IVTC algorithm can handle PAL film content well... :) Compare it to madVR film mode. Use frame stepping and compare pixel by pixel. Or use the 1080i60 pattern to check how well your GPU IVTC can handle 2:2 cadences in 60i content. FWIW, madVR takes too long to switch to the correct cadence here. That's something I'll work on in the future.

Edit2: Disable the option "only look at pixels in the frame center" in the madVR settings, and madVR film mode will play all 1080i Cheese Slice test patterns with identical quality to the original progressive source, which is much better than anything AMD and NVidia seem to be able to do.

And im paying for that already, this gpu drivers are making me crazy, madvr is not smooth at all, i disabled everything i found in CCC video section and still sometimes is good and smooth, others is not smooth...

Maybe is powerplay changing the clocks for power saving or other crap saving, how can i disable that waste of technology in CCC?? I think the problem is CCC, i can see the clocks go to 1000HZ and then drop to 300 and up and thats making madvr exclusive mode unstable :scared:
You see the clocks changing while video playback is running? That's pretty bad. I think there are some tweak tools which allow you to fix the clocks at a specific value, but I'm not sure which tools can do that. FWIW, I'm using a 7770 in my development PC and on win7 x64 playback is always smooth in fullscreen exclusive mode. So I'm not sure why you have problems.

only have pulldown detection checked to auto, forcing smooth playback and thats all...
Please disable "forcing smooth playback". Nobody even knows what that does exactly.

Ah, i have aero disabled. With aero enabled i only get errors, f&$/&$/ aero :mad:
You shouldn't get errors with Aero enabled. What kind of errors do you get?

Some suggestions:

(1) Try with software decoding (just in case you have DXVA decoding enabled).
(2) Try without Reclock.
(3) If you still have problems, make a screenshot of the OSD in the situation when the problem occurs. If we're talking about FSE mode, then maybe PrintScreen won't work? (Not sure). In that case write down numbers of all the queues in the moment when the problem occurs. Then we may be able to identify where the bottleneck is.
(4) The madVR OSD probably does report frame drops when the problem occurs?
(5) Are you playing progressive files or interlaced files? Try progressive files (e.g. Blu-Ray), because they're easier to play. Does the problem only occur with interlaced files, or with progressive files, too?

Okay thats clear, but how is that related to my problem?
Not at all. Nobody said it would... ;) Film vs video has nothing to do with your problem.

[DragonQ]

It's only recently that I've gained interest in interlaced material, but CUVID does a hell of a job deinterlacing 29.97fps video material to 59.94fps. Quite frankly, it looks fantastic to me. James Cameron keeps saying that frame rate is more important than resolution(tring to push for 1080/48p instead of 4K/24p) and it sure doesn't look 288p to me :o

Now how much resolution do we lose when watching a NTSC (video mode) DVD deinterlaced to 59.94fps? GPU's use fairly advanced algorithms IIRC.
It's less "resolution loss" and more "deinterlacing artefacts". But they're usually difficult to spot at normal viewing distances with most material, assuming your deinterlacer is good.

Okay thats clear, but how is that related to my problem?

If a channel switches from 1080/25p too 1080/25i, if using dxva deinterlacing, shouldn't it detect it needs to go from "no deinterlacing, or "weave" as you describe from the progressive source. Too "vector" for the interlaced source?

Because in my case it doesn't. It always thinks the video is progressive, even when I change the deinterlacing in the AMD control panel to vector or motion, there is no deinterlacing, or just "weave" deinterlacing for the 1080/25i video.

So since this just happens with Madvr, is this a communication problem between the renderer and the amd driver?

Why doesn't it work to set the deinterlacing to vector manually when this occurs?

Why does it work forcing deinterlacing on in Madvr?

If it could detect the change, should it be sending the information to change the deinterlacing type too the driver?
I believe this is a bug in MadVR that madshi mentioned a few pages back - basically it only checks whether the video is interlaced or not at the start. You can get around this by setting LAV Video Decoder to "aggressive" deinterlacing.

With this enabled, my MBAFF videos recorded from DVB-T2 correctly switch between progressive and interlaced.

I'm not sure about that. First of all: See my reply after the next quote. Furthermore: Your screenshot of the Cheese Slice test pattern does not show how proper motion adaptive deinterlacing with diagonal filtering looks like. Proper motion adaptive deinterlacing looks *much* better than that! You're putting way too much stock into the Cheese Slice tests. Furthermore, AMD's and NVidia's implementation of motion adaptive deinterlacing seems to be far from optimal, as far as I can see...
I've mentioned them once and haven't used them in years, so I disagree with that comment.

I know what crappy deinterlacing ("288p") looks like - for example, the UK Coupling DVDs all exhibit this (576p/25 made from 576i/25 but they've clearly just taken every other field and line doubled them). I do not see this with "progressive video in an interlaced stream" material on any of my PCs or TVs.

Those Cheese Slice test patterns were the reason that I said that NVidia and ATI *might* do some sort of motion compensation. But I've just checked the Cheese Slice test patterns again, and I have to *totally* change my opinion. Actually, it seems everybody has been misinterpreting the Cheese Slice tests! Because, I've just found out that they're actually encoded as *FILM* with a 2:2 cadence. Just try it out: Play those Cheese Slice tests with madVR film mode. madVR will detect a 2:2 cadence for both 50i and 60i Cheese Slice tests. In the moment when motion starts, madVR will take a moment to switch onto the correct cadence, but once it has switched, playback is perfect. After that I've looked at the separate fields in the video stream, and really, every 2 fields belong together to form a progressive frame. So the Cheese Slice tests must be totally disregarded as *video* mode deinterlacing tests. Motion compensation wouldn't work at all for this test pattern because motion compensation expects movement between every field, which this test pattern does not have.

Edit: Btw, you can now use the 1080i50 Cheese Slice test patterns to check whether your GPU IVTC algorithm can handle PAL film content well... :) Compare it to madVR film mode. Use frame stepping and compare pixel by pixel. Or use the 1080i60 pattern to check how well your GPU IVTC can handle 2:2 cadences in 60i content. FWIW, madVR takes too long to switch to the correct cadence here. That's something I'll work on in the future.

Edit2: Disable the option "only look at pixels in the frame center" in the madVR settings, and madVR film mode will play all 1080i Cheese Slice test patterns with identical quality to the original progressive source, which is much better than anything AMD and NVidia seem to be able to do.
This is really interesting, I'll have a look when I get home later. Doesn't change the fact that my GPUs seem to handle 2:2 cadence fine but maybe there is a better way of seeing the effects of different deinterlacing algorithms on native interlaced content.

Presumably the best way to do this would be to create a 1080p/50 video, then cut away half of each frame to get a 1080i/25 video, then compare the two with different algorithms.


EDIT: Hmm, for some reason forcing film or video mode doesn't do anything on my laptop (HD4000).
EDIT 2: Oh OK, forcing film mode only works for software decoding.

[mzso]

I got the ati, i read it too late :o

And im paying for that already, this gpu drivers are making me crazy, madvr is not smooth at all, i disabled everything i found in CCC video section and still sometimes is good and smooth, others is not smooth...

Don't get to bothered. It often sucks with nvidia too. Sometimes the video gets jerky no matter what renderer you use. Only a system or driver restart helps.

Last time it happened I noticed on the madvr osd that the clock deviation was like 20%, which is funny...

For a while it seemed the problem was fixed then it came back again.

[mzso]

I tried SVP but could not stand the look of it... It made everything nice and crisp, but it looked cheap and sterile and looked like a soap opera. Not my cup of tea, but some people love it.

QB

It still amazes me how people can fall in love with artifacts/noise and the sort.

As far as I remember SVP author himself used to use ReClock.

Speaking of SVP is it only possible to use it with ffdshow software filter? Since it runs on the cpu and without HW accel the decoding also my cpu is far from sufficient and the video card's acceleration capabilities are wasted.

[DragonQ]

Don't get to bothered. It often sucks with nvidia too. Sometimes the video gets jerky no matter what renderer you use. Only a system or driver restart helps.

Last time it happened I noticed on the madvr osd that the clock deviation was like 20%, which is funny...

For a while it seemed the problem was fixed then it came back again.
I remember trying MadVR on my HTPC when I first got it a year ago - couldn't avoid frame drops no matter what settings I used, despite GPU and CPU usage being nowhere near 100%. Might get around to trying it again now that MadVR, LAV Filters and MPC-HC have all matured a bit. :D

The only time I get dodgy playback in MediaPortal (EVR) is when the GPU and/or driver craps out. This happens rarely but inexplicably. Basically I just get frame drops all over the place for no apparent reason and I have to restart the PC. The other day it was only happening during certain scenes on certain channels - e.g. the adverts would all be fine, but the programme would be crappy (despite everything being interlaced) and another channel would be fine.

Oh, the joys of video playback. ;)

[madshi]

@DragonQ, a year ago the NVidia drivers had serious bugs when using madVR fullscreen exclusive mode. These bugs have been fixed in the meanwhile (at least when using win7 or win8 x64). For windowed mode you can now also use Overlay in madVR which could also potentially help avoiding frame drops.

[truexfan81]

@DragonQ, a year ago the NVidia drivers had serious bugs when using madVR fullscreen exclusive mode. These bugs have been fixed in the meanwhile (at least when using win7 or win8 x64). For windowed mode you can now also use Overlay in madVR which could also potentially help avoiding frame drops.

i use overlay, and right now 36minutes into a dvd i have
1 frame repeat every 29.94 seconds
dropped frames 11
delayed frames 3

all queues are full, i guess its the vsync causing the frame drops?

[AndreaMG]

Speaking of SVP is it only possible to use it with ffdshow software filter? Since it runs on the cpu and without HW accel the decoding also my cpu is far from sufficient and the video card's acceleration capabilities are wasted.
It runs under Avisynth so ffdshow raw video filter is needed. It runs on CPU+GPU (Open CL) or CPU only.

[Qaq]

Speaking of SVP is it only possible to use it with ffdshow software filter? Since it runs on the cpu and without HW accel the decoding also my cpu is far from sufficient and the video card's acceleration capabilities are wasted.
1. SVP - http://www.svp-team.com/
SVP allows you to watch any video on your PC file with frame interpolation (like you can watch it on high-end TVs and projectors). It increases frame rate by generating intermediate animation frames between existing ones to produce very smooth, fluid and clear motion.
SVP provides GPU acceleration and it's possible to watch FullHD 1080p-video recalculated to 60Hz in real-time with mid-range CPU and almost any GPU hardware.
2. Dmitri Render - http://dmitrirender.narod.ru/ Even more GPU assisted. Not freeware, free 30-day trial.

[Dodgexander]

I believe this is a bug in MadVR that madshi mentioned a few pages back - basically it only checks whether the video is interlaced or not at the start. You can get around this by setting LAV Video Decoder to "aggressive" deinterlacing.

With this enabled, my MBAFF videos recorded from DVB-T2 correctly switch between progressive and interlaced.


Brilliant thanks!! And a coupling fan? Even more brilliant, should be more famous than it is!

So if I want to use YADIF2 deinterlacing in lav for standard def dvb channels, is there anyway to use lav without it applying yadif to hd channels also?

At the moment the only solution I have is to uncheck lav to use h264, but then I have to use cyberlink decoder for 1080i instead and id like to use lav instead.

[DragonQ]

Is there a particular reason you want to use YADIF? In most cases hardware deinterlacing is better.

[xabregas]

You see the clocks changing while video playback is running? That's pretty bad. I think there are some tweak tools which allow you to fix the clocks at a specific value, but I'm not sure which tools can do that. FWIW, I'm using a 7770 in my development PC and on win7 x64 playback is always smooth in fullscreen exclusive mode. So I'm not sure why you have problems.

I solved the problem by creating a profile and changing the profile in notepad where i put the clock to 1000 in every powerplay default config (there were 3), in the memory i put 1125 and voltage to stock value which is 175, memory voltage is 0000 in all settings, so i guess it can`t be changed...

Now the clocks are locked :D


Please disable "forcing smooth playback". Nobody even knows what that does exactly.

Ok, will disable it :D


You shouldn't get errors with Aero enabled. What kind of errors do you get?

Some suggestions:

(1) Try with software decoding (just in case you have DXVA decoding enabled).
(2) Try without Reclock.
(3) If you still have problems, make a screenshot of the OSD in the situation when the problem occurs. If we're talking about FSE mode, then maybe PrintScreen won't work? (Not sure). In that case write down numbers of all the queues in the moment when the problem occurs. Then we may be able to identify where the bottleneck is.
(4) The madVR OSD probably does report frame drops when the problem occurs?
(5) Are you playing progressive files or interlaced files? Try progressive files (e.g. Blu-Ray), because they're easier to play. Does the problem only occur with interlaced files, or with progressive files, too?

with aero enabled i feel vsync is applied twice somehow, maybe is only in my head :rolleyes: DO i have to disable desktop composition in general settings when using exclusive mode with aero enabled or is fine with only exclusive mode checked?

(1) i always try with software decoding, tried with lav and coreavc, coreavc is much more smooth

(2) yes, guess what, madvr started being smooth after disabling reclock. sound is bad now, but i went to the loft and brought my old pci xfi xtrememusic and sound is good again, no need for reclock :rolleyes:

(3) I will put here my osd after playing a progressive video, with everything you said to disable, now with locked clocks and with ati brand new drivers 13.1 :scared:

Thanks ;)

[xabregas]

Don't get to bothered. It often sucks with nvidia too. Sometimes the video gets jerky no matter what renderer you use. Only a system or driver restart helps.

Last time it happened I noticed on the madvr osd that the clock deviation was like 20%, which is funny...

For a while it seemed the problem was fixed then it came back again.

damn :(

Problem is i think this gpu is very fast, the PQ is amazing with jinc and lanczo with AR enabled, but there is too many things screwing madvr in CCC amazing settings :sly:

[iSunrise]

However, I agree with DragonQ that "proper" video deinterlacing is much better than simple Bob. I don't know of any consumer hardware solution which uses motion compensation for video mode deinterlacing (not sure about what ATI and NVidia do, though, maybe they do some sort of motion compensation, I don't know). The deinterlacer hardware chips in TVs, Receivers, Blu-Ray players etc all use motion-adaptive deinterlacing with some sort of diagonal filtering. The key feature of motion-adaptive deinterlacing is that it detects static vs. moving image areas. For static image areas you can safely weave the fields together without getting artifacts and as a result for static image areas you get full progressive resolution. For moving parts most implementations use a Bob scaling algorithm with diagonal filtering, which produces results which are better than a Bob algorithm based on simple Bilinear scaling.

I would guess that if you used something like Jinc on every separate field (like 6233638 seems to suggest), you'd get better image quality for *moving* parts than what the typical CE deinterlacer chip does. However, if you don't weave static image areas then you're going to lose a lot of resolution during "quiet" scenes.
Interesting. I always wondered why some interlaced material looks overly soft after it got deinterlaced. I originally thought that deinterlacing halves the frame rate but at least puts the frames together as close to the original as possible. I didn´t know that there´s also filtering applied which explains some of the results I´m getting.

That diagonal filtering also seems to occur with NV GPUs a lot. At least I usually get a WAY sharper image (more fine details) in the interlaced picture compared to active deinterlacing with a lot of my DVDs (I have about 500-600 DVDs that are NTSC, about 100 are PAL). Also, for some reason, on some DVDs, I don´t need to deinterlace at all and have a perfect and pretty sharp picture.

After years of madVR use, I also noticed that sharp algorithms like Lanczos/Bicubic on luma amplify some of the interlacing artifacts a lot, to the point where it gets overly distracting, while SoftCubic makes some of these artifacts to entirely go away. Which probably is one of the reasons why you included it in the first place, because it can hide artifacts. I really hate it´s desaturation, though, because the picture not only looks very soft afterwards, but also a bit dull. If you have a movie with a lot of bright lights (traffic lights in a big city or dark areas with bright lights, like large city camera shots, which there happen to be a lot when you have a TV series that got produced in the US) you will reduce the dynamics a lot and the picture looks almost "dead" afterwards. I think that once we get to OLEDs, the desaturation effect on SoftCubic is going to get even more pronounced, since OLEDs have a much higher contrast. SoftCubic also looked horrible on my old CRT, because CRTs already had quite a soft picture (compared to an LCD), because of the way they produce the image.

If there were some kind of method we could apply that acts like SoftCubic but without the desaturation effect, SoftCubic would work a lot better. Is there any possibility to counter that desaturation either in the SoftCubic algorithm itself (which is probably a result of the interpolation it does) so that edges that get overly softened and desaturated at least get a boost in saturation again. Is it possible to calculate the amount of desaturation according to the percent on SoftCubic (50%-100%) we choose in the madVR scaling options?

[Qaq]

(2) yes, guess what, madvr started being smooth after disabling reclock.You just had to disable vsync in ReClock.

[Dodgexander]

Is there a particular reason you want to use YADIF? In most cases hardware deinterlacing is better.

I seem to be trying the impossible.

Trying to set things up to so both SD and HD look good. Problem is, the settings in CCC that make SD look good, make HD look odd.

So I thought if I use yadif, it disables the AMD CCC settings so I can leave the HD settings on without losing what I consider to be good SD picture quality.

I would purely use yadif, but my PC can't handle running 1080i with it on.

The other problem I am trying to combat is motion when watching sport at 1080i. At the moment I find watching football in HD is so blurry and slow compared to SD. Ive tried the dimitri filter and also SVP, but they make the ball and players blurry.

I think what I am looking for is the best way to deinterlace SD content compared to HD.

[DragonQ]

That diagonal filtering also seems to occur with NV GPUs a lot. At least I usually get a WAY sharper image (more fine details) in the interlaced picture compared to active deinterlacing with a lot of my DVDs (I have about 500-600 DVDs that are NTSC, about 100 are PAL). Also, for some reason, on some DVDs, I don´t need to deinterlace at all and have a perfect and pretty sharp picture.
Most DVDs are of films and are thus progressive in nature and don't need deinterlacing. They should be IVTCed instead but the current dispute is whether 2:2 cadence on PAL DVDs (i.e. the fields just need weaving together) is detected properly in different situations.

I seem to be trying the impossible.

Trying to set things up to so both SD and HD look good. Problem is, the settings in CCC that make SD look good, make HD look odd.

So I thought if I use yadif, it disables the AMD CCC settings so I can leave the HD settings on without losing what I consider to be good SD picture quality.

I would purely use yadif, but my PC can't handle running 1080i with it on.

The other problem I am trying to combat is motion when watching sport at 1080i. At the moment I find watching football in HD is so blurry and slow compared to SD. Ive tried the dimitri filter and also SVP, but they make the ball and players blurry.

I think what I am looking for is the best way to deinterlace SD content compared to HD.
Hmm. What settings are you fiddling with in the Catalyst Control Centre? I find most post-processing algorithms (offered by nVidia at least) do more harm than good. I'd just set deinterlacing to "vector adaptive" and leave the rest disabled.

Also, have you checked your CPU and GPU usage during these "blurry and slow" sequences? Could be just lack of processing power.

[xabregas]

You just had to disable vsync in ReClock.

It was disabled, somehow ATI HDMI audio + wasapi + reclock = fail

or better said:

ati hdmi audio + wasapi + reclock + ME = FAIL :o

[madshi]

Is there a particular reason you want to use YADIF? In most cases hardware deinterlacing is better.
Based on what evidence? You may be right, but I fear many people judged based on Cheese Slices and that is just not valid way to judge video deinterlacing quality. Or did you test with real interlaced video content to come to the conclusion that hardware deinterlacing is better than YADIF in most cases?

Don't get to bothered. It often sucks with nvidia too. Sometimes the video gets jerky no matter what renderer you use. Only a system or driver restart helps.
damn :(
Don't give up too early. What mzso reports is not normal. Most madVR users have no problems whatsoever, from what I can say. Maybe a bit of tweaking is needed sometimes to get things up and running nicely, but that's about it, in most cases (if your PC is fast enough).

i have to disable desktop composition in general settings when using exclusive mode with aero enabled or is fine with only exclusive mode checked?
You don't have to disable it. You can try if it works better that way.

(1) i always try with software decoding, tried with lav and coreavc, coreavc is much more smooth
That's weird, is your CPU old and slow? There should be no big difference between LAV and CoreAVC, unless your CPU is very old, then maybe CoreAVC is ever so slighty faster, but not much (if at all).

(2) yes, guess what, madvr started being smooth after disabling reclock. sound is bad now, but i went to the loft and brought my old pci xfi xtrememusic and sound is good again, no need for reclock :rolleyes:
Then I would guess that something is wrong with the Reclock configuration. You don't necessarily have to keep it off. Maybe clearing the timing database and playing with the Reclock settings could fix the issue? But I can't help there much, my Reclock experience is rather limited.

Interesting. I always wondered why some interlaced material looks overly soft after it got deinterlaced. I originally thought that deinterlacing halves the frame rate but at least puts the frames together as close to the original as possible. I didn´t know that there´s also filtering applied which explains some of the results I´m getting.

That diagonal filtering also seems to occur with NV GPUs a lot. At least I usually get a WAY sharper image (more fine details) in the interlaced picture compared to active deinterlacing with a lot of my DVDs (I have about 500-600 DVDs that are NTSC, about 100 are PAL). Also, for some reason, on some DVDs, I don´t need to deinterlace at all and have a perfect and pretty sharp picture.
The diagonal filtering I mentioned does not reduce sharpness, it just removes jagging artifacts in moving parts of the image. The diagonal filtering is all good, not bad. It's also only used for video mode deinterlacing, not for movie content (unless your GPU failed to detect that the source is a movie).

Is it possible to calculate the amount of desaturation according to the percent on SoftCubic (50%-100%) we choose in the madVR scaling options?
That won't work. SoftCubic does not desaturate by itself, it just has some blur included. Which means that small colored highlights get desaturated (because of the blur). But bigger colored areas do *not* get desaturated at all. So you can't just up the saturation control to fix things. Doing so would fix small colored highlights. But then big colored areas (e.g. sky, grass, water etc) would get too much saturation.

[Dodgexander]

Turns out I was comparing the settings with "dynamic contrast" on in CCC and I didn't like the difference it was making between Yadif and Dxva deinterlacing.

Ill stick with hardware deinterlacing for now then, unless there is some software alternatives of better quality I can try. I did like ffdshows ffmpeg deinterlacer that worked well for my football 1080i sample.

Hey fiddlings the game, HTPCs the name :D

[DragonQ]

Based on what evidence? You may be right, but I fear many people judged based on Cheese Slices and that is just not valid way to judge video deinterlacing quality. Or did you test with real interlaced video content to come to the conclusion that hardware deinterlacing is better than YADIF in most cases?
Well I tested this a long time ago myself but that's only for one family of GPUs. AMD could well be different. Intuitively it makes sense that hardware deinterlacing would be better but I guess more tests need to be done.

Maybe we should get some clips and get people to upload screenshots with specific settings?

[Dodgexander]

The cause of the crash is that your filter does implement "IFileSourceFilter" and "IFileSourceFilter.GetCurFile()" returns success, but the returned file name is NULL. madVR then crashes when later calling CoTaskMemFree() to properly release the file name. You can fix the crash by either returning a proper file name, or by letting GetCurFile() return an error, or by not supporting IFileSourceFilter. I'll add a check for a NULL file name to the next build. But I think you should implement the suggested change in your filter, anyway.

I saw this, I use NPVR and for some reason, although it seems to work. The OSD of NPVR doesn't dissapear.

Also for some reaons with Madvr as the render, if you launch in fullscreem mode it doesn't scale and just shows a window of the image, until you resize and make fullscreen again.

Any ideas??

[madshi]

Well I tested this a long time ago myself but that's only for one family of GPUs. AMD could well be different. Intuitively it makes sense that hardware deinterlacing would be better but I guess more tests need to be done.

Maybe we should get some clips and get people to upload screenshots with specific settings?
I think it would be interesting. I'd really like to know how YADIF compares to the best AMD and NVidia can do. But maybe a new thread should be started for this, because it's not really madVR related. In any case, the first step would be to collect samples which show clear differences between YADIF and DXVA.

BTW, for NVidia and ATI, hardware deinterlacing is really using the pixel shaders, not hard wired circuits. For Intel I think there's a dedicated circuit for deinterlacing, but what I've been told it's not as good as NVidia's and ATI's deinterlacing. JFWIW.

I saw this, I use NPVR and for some reason, although it seems to work. The OSD of NPVR doesn't dissapear.

Also for some reaons with Madvr as the render, if you launch in fullscreem mode it doesn't scale and just shows a window of the image, until you resize and make fullscreen again.

Any ideas??
You should ask the NPVR developer about this, not me.

[leeperry]

It's very convenient to have mVR automatically hiding the mouse cursor after a few secs in FSE mode, but do you really find it to be mandatory for windowed playback? I mean you could let the movie play as "always on top" in the center of the screen in a small window and do other stuff simultaneously, atm it's a major PITA as the mouse cursor keeps disappearing :o

And let's not mention ppl with dual screens setups.

[ThurstonX]

Please disable "forcing smooth playback". Nobody even knows what that does exactly.



Ok, will disable it :D


I'm not saying this is the definitive answer on ATI's/AMD's "Enforce Smooth Video Playback (ESVP)" feature, but it's an interesting read which includes feedback from AMD:

http://www.anandtech.com/show/3601/the-final-word-on-the-best-radeon-htpc-card

"The short answer is that it’s an “idiot button” for users, to keep them from using so many post-processing settings on low-end video cards that It causes frames to drop."

There's more detail at that link.

[Dodgexander]

Using Mad vr with Media Player Classic and FFDshow DXVA video decoder results in a crash of mpc when moving to a second screen, or setting media player classic to start playback on a secondry screen.

[Matching_Mole]

I coming back regarding my "lack of smoothness" of the IVTC due to my higher screen refresh frequency.

Indeed, I tried several player+video decoding+splitter but,as some one else has already mentioned, no combination solved the issue. I tried also to use Reclock "speed adaption" feature with the same result.

I have a question regarding Reclock and its use with IVTC. I mean when I used it with a progressive film, the clock deviation reported by madVr is reduced dramatically to almost 0, which is expected. I payed also attention to the clock deviation with a telecined film and with or without Reclock it was around -0.4% so I was not sure of the effect of Reclock. Though Reclock indicated that it try to adapt the speed to 28.8fps instead of 29.97.

Someone can help me to understand if Reclock see the fps change due to the madVr IVTC process or if use Reclock with madVr IVTC process is more hurting than help? And with the IVTC process used, what should be the correct clock deviation expected with a refresh rate of 48hz?

[xabregas]

I'm not saying this is the definitive answer on ATI's/AMD's "Enforce Smooth Video Playback (ESVP)" feature, but it's an interesting read which includes feedback from AMD:

http://www.anandtech.com/show/3601/the-final-word-on-the-best-radeon-htpc-card

"The short answer is that it’s an “idiot button” for users, to keep them from using so many post-processing settings on low-end video cards that It causes frames to drop."

There's more detail at that link.

nice catch ;)

now i know i can disable it because i disabled every default post processing on CCC :D

And do anyone tried amd steady video??

I checked once and with dxva decoding i saw the amd steady logo when video started, looked very smooth, so much smooth the audio went OOS after 5 min and after 10 the movie dropped idk how many frames :p

Anyway is interesting to see how amd techs find some new settings to help us screwing video playback :sly:

Fingers crossed now has i`m going to unninstall amd drivers (FEAR and panic), and then install 13.1 brand new drivers, so happy i am, maybe they have a surprise in this drivers for HD7770 owners, idk, some new amazing setting to delete movies after playback :p

[Omega13Lives]

Don't know if this is where to post this. The folks at ZoomPlayer suggested I contact the developer of madVR for my question. Here:

There is a cool program called SoftTH which allows you to use Direct3D to span multiple monitors for video games AND you can specify the bezel/border size so the picture looks consistent like you were watching (at a drive-in, if anyone knows what that is) through a car windshield with the frame blocking part of the view but the picture/video looks normal when you move or turn. With ZP Video Wall I can stretch over 3 monitors but the video gaps to the outside monitors. Is it possible to adjust for the gaps/bezels to span multiple monitors AND allow for the bezel? tx

I know this is not common, but thought I'd ask

Eric

[chuuey]

Ineteresting, I've tried using madvr on the same machine but in windows xp (i need it for work) and madvr won't even display video with 1080p content, 720p stutters like crazy lol, using lav filters, geforce 670, 6 core cpu... weird... standard renderers play with no issues o.o

[dansrfe]

I know this is something minor but would it be possible to have a mouse pointer cursor (with the pointing finger) in exclusive mode? It seems a bit more natural that way. Possibly something to add to the v1.0 list.

Thanks!

[6233638]

I know this is something minor but would it be possible to have a mouse pointer cursor (with the pointing finger) in exclusive mode? It seems a bit more natural that way. Possibly something to add to the v1.0 list.What player are you using? You should still have a mouse cursor.

[secvensor]

What video card is necessary from the last that it was possible to reproduce 1080p@60fps@10bit@100MHz at jinc 8 taps hroma upsc?
ati ore nv?
7850-7970 ore 660-680

[AndreaMG]

What video card is necessary from the last that it was possible to reproduce 1080p@60fps@10bit@100MHz at jinc 8 taps hroma upsc?

This one: http://www.olcf.ornl.gov/titan/ :p

[secvensor]

Too much demands energy.

[Plutotype]

Folks,
regarding upscaling 2k to 4k, please check this review of Sony VW1000ES 4k projector. In the middle of the page, reviewer David Mackenzie shows 2k to 4k upscaling results and compares 3 situations: Upscaling OFF, Sony Reality Creation ON and Oppo BDP-103 upscaling ON.
http://www.hdtvtest.co.uk/news/sony-vpl-vw1000es-201301162503.htm?page=Picture%20Quality

The ultimate question is: Can madVR using JINC (3/4/8) beat these techniques in terms of upscaling artifacts and clarity? ( supposing we have sufficient hardware in the HTPC ).
Thanks

[AndreaMG]

I'm also curious about the comparison regarding SD to 4K scaling, the pixel gap is sooo huge that it should be easy to tell the difference between a good and a less good scaler. That said, in Jinc I trust :D:devil:

[6233638]

Folks,
regarding upscaling 2k to 4k, please check this review of Sony VW1000ES 4k projector. In the middle of the page, reviewer David Mackenzie shows 2k to 4k upscaling results and compares 3 situations: Upscaling OFF, Sony Reality Creation ON and Oppo BDP-103 upscaling ON.
http://www.hdtvtest.co.uk/news/sony-vpl-vw1000es-201301162503.htm?page=Picture%20Quality

The ultimate question is: Can madVR using JINC (3/4/8) beat these techniques in terms of upscaling artifacts and clarity? ( supposing we have sufficient hardware in the HTPC ).
ThanksI would expect Jinc 3 AR to look as good or better than DRC, especially with actual video content. That's not to say it's perfect - there is still room for improvement.

The problem is that those are photographs, and LCoS is rather soft, which helps smooth things out, so you can't really compare those images with screenshots. I would definitely be interested in seeing a comparison on a 4K native display though.

I did try creating a similar pattern (http://www.abload.de/img/4k-greyglpnu.png) though, and this is the result with Jinc 3 AR:
http://www.abload.de/img/jinc69r3a.png

What I also noticed when doing this though, is that if the source image is not a multiple of 8, madVR does not seem to resize it correctly.

"Double Size" in MPC-HC with a 112x128 source:
http://www.abload.de/img/correct3rqni.png

"Double Size" in MPC-HC with my original 113x131 image:
http://www.abload.de/img/madvrkqqwi.png

[Lyris]

Hi guys,
I'm David Mackenzie.

Sorry I didn't get a chance to test madVR. But, here's the original pattern I made:
http://www.mediafire.com/?8pd9y18uc5qfcdr

[Duffy Moon]

What video card is necessary from the last that it was possible to reproduce 1080p@60fps@10bit@100MHz at jinc 8 taps hroma upsc?
ati ore nv?
7850-7970 ore 660-680

For 10 bit, I believe you need a pro. card. Maybe a Quadro?

[Thunderbolt8]

damn :(

Problem is i think this gpu is very fast, the PQ is amazing with jinc and lanczo with AR enabled, but there is too many things screwing madvr in CCC amazing settings :sly:
do you have any suggestions which setting better to disable or enable in CCC, starting from default values?

[6233638]

Hi guys,
I'm David Mackenzie.

Sorry I didn't get a chance to test madVR. But, here's the original pattern I made:
http://www.mediafire.com/?8pd9y18uc5qfcdrThanks David. The results look pretty good with Jinc 3 AR here, but obviously I can't see how that compares to the other processing without access to one of those projectors.
Jinc can look a little soft with very fine details, but is preferable to aliasing or sharpening artifacts in my opinion.

Sony's DRC maybe avoids ringing a little better (hard to tell from photos) but Jinc seems to look a lot more natural with real picture content.

http://www.abload.de/img/sqrlgjuce.jpg

What video card is necessary from the last that it was possible to reproduce 1080p@60fps@10bit@100MHz at jinc 8 taps hroma upsc?
ati ore nv?
7850-7970 ore 660-680Do you have a sample of this type of content?

For 10 bit, I believe you need a pro. card. Maybe a Quadro?madVR does not currently support greater than 8-bit output, so there's no point in buying a pro card for it. (at least not yet)

[AndreaMG]

Hi guys,
I'm David Mackenzie.

Sorry I didn't get a chance to test madVR
David, next time you review a 4k tv/projector do not forget to test MadVR jinc upscaling and report back pronto :) :thanks: