madVR - high quality video renderer (GPU assisted)

[mzso]

Originally Posted by dansrfe

1. No
2. Yes
3. All Spline @ 3 taps *
4. It's better to *


* These answers are my opinions

For some reason I couldn't get the red jaggediness unnoticable. Even when I used 100 soft cubic for everything. The simple yv12 shader of potplayer did a better job for some reason.

The reset settings .bat file didn't work too...

[Xello]

Does anyone have any experience playing japanese BD sources using madVR?

Apparently they use TV 16-235 rather than PC levels. This can be fixed when using EVR-CP by enabling "16-235 -> 0-255 [SD] [HD]" under Shaders, does madVR have something to sort this? I've tried switching between tv and pc levels in madVR menu and both give different but wrong results.

[nevcairiel]

All Blu-rays (and like 99.99% of all commercial content) use 16-235. madVR handles this automatically.

[Xello]

It's not handling this though, here is my testing with the file:

Played with EVR-CP = too bright (can see it in the black borders immediately)
Played with EVR-CP with the 16-235 -> 0-255 [SD] [HD] shader option enabled = perfect
Played with madVR set at normal 0-255 = too bright, identical to EVR-CP
Played with madVR set at tv 16-235 = even brighter the previous "too bright" result

Any suggestions ?

[Qaq]

Enable gamma processing in madVR. 2.45-2.60.

[mzso]

Originally Posted by nevcairiel

All Blu-rays (and like 99.99% of all commercial content) use 16-235. madVR handles this automatically.

Well, that's another illusion shattered.
It always turns out that things are suckier than I think they are...

[mzso]

Didn't anyone manage to try this little test file: https://docs.google.com/open?id=0Byf...kI4YjVHNndqRVE
16-235 ->0-255 conversion doesn't seem to be working for this one here.

[Xello]

Originally Posted by Qaq

Enable gamma processing in madVR. 2.45-2.60.

Tried this but the borders still aren't black even with 2.6. I guess I will be using EVR-CP for this file.

[mzso]

I ran "restore default settings.bat" I ran "uninstall.bat", deleted the single registry key which had madVR in it at this point. Deleted the files from the folder reinstalled madVR. Yet the settings remain, as is my problems. I'm at a total loss now on what happens here...

[nevcairiel]

Originally Posted by mzso

Didn't anyone manage to try this little test file: https://docs.google.com/open?id=0Byf...kI4YjVHNndqRVE
16-235 ->0-255 conversion doesn't seem to be working for this one here.

That file is flagged as full-range content in its headers, which means madVR won't perform range expansion.
You can temporarily fix it by pressing Ctrl-Alt-Shift-I a number of times to select TV as input range, but its really just a badly encoded file.

[Qaq]

@Xello
If your display is TV you can try to set Gamma=High and Black Level=Low.

[mzso]

Originally Posted by nevcairiel

That file is flagged as full-range content in its headers, which means madVR won't perform range expansion.
You can temporarily fix it by pressing Ctrl-Alt-Shift-I a number of times to select TV as input range, but its really just a badly encoded file.

Thanks for the info.

[Xello]

Originally Posted by Qaq

@Xello
If your display is TV you can try to set Gamma=High and Black Level=Low.

I could probably try to adjust it with my TV settings but i'm not going to, I have the tv calibrated for rec709 and d65 with optimum black and white settings and it's normal for every file so far but this one. It could be something wrong with the code/flags in the file itself, I will try to get the US version of the bluray maybe (it's 3:10 to Yuma japanese BD)

[cyberbeing]

I was re-creating my custom CRT refresh rates from scratch on a new Ivy Bridge + NVIDIA computer, and I noticed the following equation resulted in optimal refresh rate as reported by madVR's drop frame interval.

(VideoFPS x RefreshMutliple)(ClockDeviation% + 1) = OptimalRefreshRate

(23.976023976 x 4)(0.00163% + 1) = OptimalRefreshRate
(95.904095904)(1.0000163) = 95.90565914076324

(23.976023976 x 5)(0.00163% + 1) = OptimalRefreshRate
(119.88011988)(1.0000163) = 119.882073925954

This is the first time I've ever realized the direct mathematical correlation between refresh rate and clock deviation, so I just found it interesting. I tested this when fine-tuning the two refresh rates above, and it seemed to hold true for both. The closer I got to OptimalRefreshRate as calculated above, the more infrequent madVR's calculated drop frame interval was. It will certainly save me some time as I add more refresh rates that can be used without ReClock. The only problem with not using Reclock when fine-tuning refresh rates like this, is it depends on having a clock deviation which stabilizes quickly to the same value, each time you playback a video.

[chros]

Originally Posted by nevcairiel

That file is flagged as full-range content in its headers, which means madVR won't perform range expansion.
You can temporarily fix it by pressing Ctrl-Alt-Shift-I a number of times to select TV as input range, but its really just a badly encoded file.

Thanks nev, I forgot this option.
And in this case the picture will be the same as it was encoded properly?

[nevcairiel]

Originally Posted by chros

Thanks nev, I forgot this option.
And in this case the picture will be the same as it was encoded properly?

Yes, it'll be fine.

[konok]

Does anyone know how to get Infinite Pin Tee filter or Inftee SDK sample to work with madVR? I'm trying to show a video in two displays... This works with EVR, but when I tried using madVR, it freezes with either Tee filter.

I've tried something simple like this in GraphStudio as a test, but it still freezes:

source -> ffdshow -> Infinite Pin Tee -> madVR
(only one tee output)

madVR works perfectly without Infinite Tee.

[dansrfe]

Originally Posted by cyberbeing

(VideoFPS x RefreshMutliple)(ClockDeviation% + 1) = OptimalRefreshRate

(23.976023976 x 4)(0.00163% + 1) = OptimalRefreshRate
(95.904095904)(1.0000163) = 95.90565914076324

(23.976023976 x 5)(0.00163% + 1) = OptimalRefreshRate
(119.88011988)(1.0000163) = 119.882073925954

This is the first time I've ever realized the direct mathematical correlation between refresh rate and clock deviation, so I just found it interesting. I tested this when fine-tuning the two refresh rates above, and it seemed to hold true for both. The closer I got to OptimalRefreshRate as calculated above, the more infrequent madVR's calculated drop frame interval was. It will certainly save me some time as I add more refresh rates that can be used without ReClock. The only problem with not using Reclock when fine-tuning refresh rates like this, is it depends on having a clock deviation which stabilizes quickly to the same value, each time you playback a video.

This is a very interesting find indeed. I will try this with my nvidia card and tv today. Out of curiosity, do you have some sort of calculation for determining what to put into the fields for custom resolutions? I have tried creating a custom resolution for 23.976 and I can never get it spot on since it ends up around 23.978. To be honest I'm not exactly sure of the functions of these various options either.

[namaiki]

Originally Posted by dansrfe

Out of curiosity, do you have some sort of calculation for determining what to put into the fields for custom resolutions? I have tried creating a custom resolution for 23.976 and I can never get it spot on since it ends up around 23.978.

For me, I usually select "Manual" and then type in the refresh rate to 3 decimal points and then put it back to Automatic before clicking Test for best results (for me).

I'm surprised (and glad) that my Sony laptop was able to push 72Hz on it's inbuilt display.

[Nya_Su]

Hello everyone,

As I was stuck with my problem despite every settings I tried, I decided to give "fullscreen exclusive mode" another chance.

First, I dicovered that while it looked like there were a lot of dropped frames, it was in fact a lot of presentation glitches. Queues were stables. So I tried some various flush setting, but queues were dropping toward zero and dropped frames appeared. Changing CPU/GPU queue size values didn't change anything, so I went back to default (CPU 12, GPU 8).

Then I tried some other options. For now, I think I found a good combination for the exclusive mode setting : first three options are checked, 16, all flush to "flush & wait (loop)", and only "limit redering times to avoid glitches" is checked in the last three options (it looks like that in my case, it is the most important option...). Worst case, queues stabilizes to:

decoder queue 11-12/12
upload queue  7-8/8
render queue  3-4/8
present queue 3-5/16

And now, obviously except when exclusive mode activates itself, I have no dropped frames, no delayed frames and no presentation glitches.

Well, what do you know? I may end up loving exclusive fullscreen mode after all!