Welcome to Doom9's Forum, THE in-place to be for everyone interested in DVD conversion.

Before you start posting please read the forum rules. By posting to this forum you agree to abide by the rules.

 

Go Back   Doom9's Forum > Hardware & Software > Software players

Reply
 
Thread Tools Search this Thread Display Modes
Old 22nd December 2021, 21:07   #62561  |  Link
flossy_cake
Registered User
 
Join Date: Aug 2016
Posts: 635
I seem to find the "bilateral old" chroma upscaling algorithm produces by far the cleanest result with a poor quality 4:2:0 SD source.

I suspect it might be an issue with chroma alignment of some SD sources where the chroma sample position was incorrectly encoded in the source, eg. certain seasons of Family Guy. In these cases the bilateral filter seems to work wonders and smooths it out beautifully without any visible blurring, aliasing or bleeding.

Curious to know how it works.
flossy_cake is offline   Reply With Quote
Old 22nd December 2021, 22:06   #62562  |  Link
tp4tissue
Registered User
 
tp4tissue's Avatar
 
Join Date: May 2013
Posts: 753
Quote:
Originally Posted by ryrynz View Post
Never? Don't the best laser projectors hit 20,000 lumens already? Wouldn't that reach approx 1000 nits on a 100" screen?
The absolute best projectors in theaters are hitting 100nits. That's Dolby Vision super duper $Millions laser projector.

The average theater, ~50nits.

From 100 to 1000 ? not going to happen, Even if it did, it'd be worthless because contrast ratio would be trash.
__________________
Ghetto | 2500k 5Ghz
tp4tissue is offline   Reply With Quote
Old 22nd December 2021, 22:09   #62563  |  Link
tp4tissue
Registered User
 
tp4tissue's Avatar
 
Join Date: May 2013
Posts: 753
Quote:
Originally Posted by huhn View Post
he only said for JVC auto cal.
do you know how meter aware this software is?
or how high the ANSI CR of these devices is if the auto software even cares about low light corrections at all...
These probes are software agnostic. The spyder 4 and 5 dark read problem is INHERENT in the hardware of the probe.

The CR of most projectors might be terribad enough to be in the range of the probe, but that is not uniformly the case, especially with dynamic iris.
__________________
Ghetto | 2500k 5Ghz
tp4tissue is offline   Reply With Quote
Old 22nd December 2021, 22:12   #62564  |  Link
aleksanderdev
Registered User
 
Join Date: Dec 2021
Posts: 1
Hi! Running rtx 3080 and intel 12900k.

I'm a n00b at madvr and settings.
Is there anyone with similar setup that could recommend settings to use?

I will be playing on a 1080p projector until i can upgrade to Epson TW9400.
So I need hdr colormapping and upscale/enhancing 1080p videos.
I prefer sharp details over more soft blur.
aleksanderdev is offline   Reply With Quote
Old 23rd December 2021, 00:03   #62565  |  Link
SamuriHL
Registered User
 
SamuriHL's Avatar
 
Join Date: May 2004
Posts: 5,372
https://www.avsforum.com/threads/off.../post-61286239

For those who keep asking if madvr on PC is dead....an answer from the man himself.
__________________
HTPC: Windows 11, AMD 5900X, RTX 3080, Pioneer Elite VSX-LX303, LG G2 77" OLED
SamuriHL is offline   Reply With Quote
Old 23rd December 2021, 00:06   #62566  |  Link
LordX2
Registered User
 
Join Date: Oct 2008
Posts: 168
Nice to see that he isn't giving up on the PC!
LordX2 is offline   Reply With Quote
Old 23rd December 2021, 00:44   #62567  |  Link
Manni
Registered User
 
Join Date: Jul 2014
Posts: 958
Quote:
Originally Posted by tp4tissue View Post
The absolute best projectors in theaters are hitting 100nits. That's Dolby Vision super duper $Millions laser projector.

The average theater, ~50nits.

From 100 to 1000 ? not going to happen, Even if it did, it'd be worthless because contrast ratio would be trash.
It would be really nice if you stopped making definitive statements as if you knew what you’re talking about. My JVC RS2000 does over 200nits in high lamp (new bulb), with the P3 filter on. The native contrast ratio is over 30,000:1 at that brightness level. There are many, many projectors (especially laser ones) that produce much higher output.

Quote:
Originally Posted by tp4tissue View Post
These probes are software agnostic. The spyder 4 and 5 dark read problem is INHERENT in the hardware of the probe.

The CR of most projectors might be terribad enough to be in the range of the probe, but that is not uniformly the case, especially with dynamic iris.
Again, you have no idea what you are talking about. Agreed, the spyders can be a crap shot for Color, and they don’t age well, but their ability to correct gamma remains mostly unaffected. This is because the JVC Autocal software does up to 33 point reads per Color (R, G, B). Because it’s entirely relative for gamma calibration, the absolute Color accuracy of the meter is of no importance.

However, you do need a better meter, preferably a Spectro, to correct the RGB balance at 100%, which will automatically correct the rest of the greyscale. Given that the spectro can’t correct gamma easily, a poor Spyder is a great complement, even if it can’t be used on its own for anything but gamma calibration with the JVC Autocal.

Regarding black levels, you can use some of the Spyder models facing the projector Lens, getting up to 5,000 nits for 100% white, which means that the “black” readings are well within the sensitivity range of these meters, which isn’t always the case when used facing the screen or in a flat panel.

FYI the spyders are not only meters reporting “fake” dark readings, but it’s up to the software to either report trash, or a fixed “fake” value that looks more correct. The i1display Pro, which is a much better meter than the Spyders, also does this. The value near black isn’t the same depending on the software used.

Again, you have a limited knowledge in calibration and I can’t spend my time correcting you from posting misinformation, so please do some research before posting so much incorrect information. Thanks!
__________________
Win11 Pro x64 b23H2
Ryzen 5950X@4.5Ghz 32Gb@3600 Zotac 3090 24Gb 546.65
JRVR/LAV/jRiver/MyMovies/CMC
Denon X8500HA>HD Fury VRRoom>Samsung 77S90C
Manni is offline   Reply With Quote
Old 23rd December 2021, 12:55   #62568  |  Link
mclingo
Registered User
 
Join Date: Aug 2016
Posts: 1,354
Quote:
Originally Posted by SamuriHL View Post
https://www.avsforum.com/threads/off.../post-61286239

For those who keep asking if madvr on PC is dead....an answer from the man himself.
Personally i'm not convinced, if he told everyone PC is dead then he might lose a lot of his testing base.

Of course i'd like this to be true but sadly these are just words until we see some actual action,. or at least an official announcement, he must have some sort of rough timeframe for v1.0 but he cant even give us that.

Again though, he owes us nothing and we've had years of MADVR FOC, so if this sounds ungrateful its not.
__________________
LG OLED EF950-YAM RX-V685-RYZEN 3600 - 16GBRAM - WIN10 RX 5700 - https://www.videohelp.com/software/madVR/old-versions

Last edited by mclingo; 23rd December 2021 at 12:57.
mclingo is offline   Reply With Quote
Old 23rd December 2021, 17:57   #62569  |  Link
ajp_anton
Registered User
 
ajp_anton's Avatar
 
Join Date: Aug 2006
Location: Stockholm/Helsinki
Posts: 807
I'm having problems with a HEVC 2160p 10bit stream. Using LAV to decode, it outputs P010 in all cases below.

DXVA native works with EVR, but "kind of" freezes madVR. MPC-HC continues to play, but with a black screen. Exiting closes the window but the program stays on and needs to be killed via task manager.

Software works with both renderers, except the CPU is just a hair too slow.
DXVA copyback works with both renderers, but drops over half of the frames.

Skylake i3-6100T, HD 530, Windows 10.

(Works fine on my laptop's i7-1185G7, W10.)

Is this a known problem and do workarounds exist?
ajp_anton is offline   Reply With Quote
Old 23rd December 2021, 19:00   #62570  |  Link
Perenista
Registered User
 
Join Date: Oct 2013
Posts: 208
What could account for random white dots in the video while viewing only H.265 content (H.264 is OK)? I thought this was my i5 11400 CPU having some sort of driver issue (in my case it's the onboard video being used, so no GPU), but if it's only affecting H.265 content, then something else is happening. I am using madVR and MPC-BE.

Oddly I activated HDR from Windows and the physical buttons of this 4K monitor and this only happened after I did both. But a dead pixel test from a website didn't reveal (so far) anything wrong. I suspect further filters need to be installed or something about D3D or anything else needs to be changed for MPC Video Decoder. Perhaps I should download LAV?

Note: check these posts, they are probably about the same issue I had:

https://www.reddit.com/r/htpc/commen...ite_artifacts/

http://forum.doom9.net/showpost.php?...ostcount=51024

** https://www.avsforum.com/threads/mad...880609/page-13

** says:
Randomly appearing white dots could be a bad HDMI cable, or a GPU or VRAM hardware issue (or overheating).

Last edited by Perenista; 23rd December 2021 at 19:05.
Perenista is offline   Reply With Quote
Old 23rd December 2021, 19:03   #62571  |  Link
clsid
*****
 
Join Date: Feb 2005
Posts: 5,664
@ajp_anton
Try with D3D11 hardware decoding instead of DXVA2 Native.

Otherwise, use DXVA2 Copyback and in LAV Video Decoder disable all 10/16-bit output formats. That reduced the amount of data that needs to be copied to GPU memory. This can solve performance issues (on Intel) that are caused by bandwidth bottlenecks.
__________________
MPC-HC 2.3.0
clsid is offline   Reply With Quote
Old 23rd December 2021, 19:34   #62572  |  Link
Perenista
Registered User
 
Join Date: Oct 2013
Posts: 208
Without even reading the last post I decided to try LAV. It fixed the problem right away. Before: old settings: MPC-BE 1.5.8.6302 - Filters currently loaded:
- Default DirectSound Device
- madVR Renderer
- Audio Switcher
- MPC Video Decoder
- MPC Audio Decoder
- MPC MP4/MOV Source

This is what I was getting: https://i.postimg.cc/N01S7z4R/BUG.jpg (random white dots) - An issue that never happened before.

https://i.postimg.cc/Ssk5ZfT3/MPC-VIDEO-DECODER.png (MPC video decoder settings)

All of this in a small 1080p H.265 video. So I downloaded LAV from here: https://www.videohelp.com/software/LAV-Filters

Once installed, I configured LAV to be the default filter for everything (we need to add as an external filter in MPC-BE), and have set in MPC-BE the "merit" high enough to use it. madVR still being loaded, along with:

MPC-BE 1.5.8.6302 - Filters currently loaded:
- Default DirectSound Device
- madVR Renderer
- Audio Switcher
- LAV Video Decoder
- MPC Audio Decoder
- MPC MP4/MOV Source

No more white specks. What I don't understand is why after restarting Windows 2 times and loading an insignificant small file (not CPU or GPU demanding) I still saw the dots. Unless said bottleneck lasts far more than a few minutes after it happens for the first time.

All of this indicates that even with a modern CPU and its onboard video (yes, there is no GPU here at this moment) you need to do further changes to prevent something like this from happening. Probably there was some overload due to HDR being used, something I never done for so long since I bought this monitor.
Perenista is offline   Reply With Quote
Old 23rd December 2021, 19:54   #62573  |  Link
Klaus1189
Registered User
 
Join Date: Feb 2015
Location: Bavaria
Posts: 1,684
I never saw such white little dots. Try another HDMI Cable, preferrable a short one.
Klaus1189 is offline   Reply With Quote
Old 23rd December 2021, 21:41   #62574  |  Link
clsid
*****
 
Join Date: Feb 2005
Posts: 5,664
My reply was for Anton.

White dot artifacts are not related to decoder or renderer! Usually an issue with cable to monitor not being good enough. Since you enabled HDR mode, the GPU has to output more data through the cable.

https://www.google.com/search?q=gpu+...rtifacts+cable
__________________
MPC-HC 2.3.0

Last edited by clsid; 23rd December 2021 at 21:43.
clsid is offline   Reply With Quote
Old 24th December 2021, 02:40   #62575  |  Link
tp4tissue
Registered User
 
tp4tissue's Avatar
 
Join Date: May 2013
Posts: 753
Quote:
Originally Posted by Manni View Post
Wrong Garbage

You're talking about on off contrast. , I'm referring to Ansi contrast.

Your point is irrelevant. The numbers I'm quoting is from dolby themselves. their ultimate projector does 100nits.

Sure scale that down it could be brighter, I'm only quoting a data point. It's neither an assumption nor a broad stroke.

JVC autocal is worthless, it doesn't matter how many points it reads. The reason is because it makes assumptions on linearity, while the degradation of the probe's reading/ response is non-linear. The native gamma response of the projector is also non-linear over time. Correcting the rgb balance at 100% does NOT fix the rest of the line.

You can say it's "better" in relative terms, but it's still junk, because it's being multiplied by 0 as the filters are gone.

The spyder can not be relied upon for any calibration purpose where low reads are necessary, and it certainly can't be used once aged.

I don't know where you're pulling your information from, it's full of misinterpretations and half truths. I also don't understand why you insist upon perpetuating misinformation.

Spyder4s and 5s cause gamut clipping WHEN BRAND NEW.. what good are they when they're so terrible out of the box. Their entire software suite is a chest of cheats to make up for how terrible their device is.
__________________
Ghetto | 2500k 5Ghz

Last edited by tp4tissue; 24th December 2021 at 02:47.
tp4tissue is offline   Reply With Quote
Old 24th December 2021, 10:56   #62576  |  Link
Perenista
Registered User
 
Join Date: Oct 2013
Posts: 208
Quote:
Originally Posted by clsid View Post
My reply was for Anton.

White dot artifacts are not related to decoder or renderer! Usually an issue with cable to monitor not being good enough. Since you enabled HDR mode, the GPU has to output more data through the cable.

https://www.google.com/search?q=gpu+...rtifacts+cable
The minute I turn on hardware acceleration from here, for HEVC:

https://i.postimg.cc/nrNJnwWF/PROB.jpg

It's the exact moment the white dots reappear, and this in a small 1080p file. I did this checking after turning LAV off (removing from external filters from MPC-BE) and using MPC Video decoder back again.

What can account for white dots in the image only when HW acceleration is enabled? Also, please note HDR is completely OFF, from Windows 11 and also the physical buttons of this monitor.

If I turn HW acc off the image is 100% fine. MEDIAINFO from the video: https://pastebin.com/KBzuViup

Quote:
Originally Posted by Klaus1189 View Post
I never saw such white little dots. Try another HDMI Cable, preferrable a short one.
The white dots are very subtle, but they are there. They can pollute the entire image or only a few appear. I don't think this is hardware related in my case. It has to be some Windows 11, driver or player wrongly configured.

The first sign of issue was H.264 that had to be disabled from MPC Video Decoder. Now HEVC. What does that tell you?

If I put any GPU here every one of those issues would be gone, for good? I replaced the PC and even with the integrated graphics 4K content almost reaches 100% from "GPU" resources when the file is being played fullscreen. I guess I underestimated how much we need a GPU, not just a CPU, for that.

P.S. Same issue from another user:

https://arstechnica.com/civis/viewto...f=15&t=1195703

Last edited by Perenista; 24th December 2021 at 14:10.
Perenista is offline   Reply With Quote
Old 24th December 2021, 11:54   #62577  |  Link
Manni
Registered User
 
Join Date: Jul 2014
Posts: 958
Quote:
Originally Posted by tp4tissue View Post
You're talking about on off contrast. , I'm referring to Ansi contrast.

Your point is irrelevant. The numbers I'm quoting is from dolby themselves. their ultimate projector does 100nits.

Sure scale that down it could be brighter, I'm only quoting a data point. It's neither an assumption nor a broad stroke.

JVC autocal is worthless, it doesn't matter how many points it reads. The reason is because it makes assumptions on linearity, while the degradation of the probe's reading/ response is non-linear. The native gamma response of the projector is also non-linear over time. Correcting the rgb balance at 100% does NOT fix the rest of the line.

You can say it's "better" in relative terms, but it's still junk, because it's being multiplied by 0 as the filters are gone.

The spyder can not be relied upon for any calibration purpose where low reads are necessary, and it certainly can't be used once aged.

I don't know where you're pulling your information from, it's full of misinterpretations and half truths. I also don't understand why you insist upon perpetuating misinformation.

Spyder4s and 5s cause gamut clipping WHEN BRAND NEW.. what good are they when they're so terrible out of the box. Their entire software suite is a chest of cheats to make up for how terrible their device is.
I was hoping you'd do a bit of research before replying, but clearly you still believe you are actually knowledgeable about calibration in general and the JVC Autocal specifically.

Okay.

I work in the film industry. FYI the Dolby Cinema standard is 108 nits for peak brightness (that's what they get with two Christie 4K laser projectors), not 100nits. It's also entirely irrelevant to consumer content as it's using a different grade to what we get on UHD Bluray or UHD TV, so let's forget about it for this discussion.

ANSI contrast is pointless for film content, it's an extreme case that's only relevant for business projectors displaying Excel spreadsheets. On/off contrast is a lot more useful, because contrast only goes down from there. If you want to look at contrast beyond that that would be relevant to film content, then look at lower ADL contrast, from 1% to 10% (that's where most film content lies) maybe up to 25-30% max for extremely bright content.

Anything above that (such as ANSI, which is 50% ADL) is mostly irrelevant for film content. There is no film content at 50% ADL, which is what ANSI measures. Even the Art of Flight (snowboarding documentary, so the brightest possible picture you could think of) tops at something like 32% ADL. The only interest in ANSI contrast is that it gives you the other end of the scale. So knowing on/off contrast and ANSI contrast, you know the best and the worst contrast the display is capable of, but you have to know what to do with these measurements in order to interpret them for film content. On/off is essential on fade to black (which is present in film content) and is crucial to get better ADL contrast in the 1%-10% range, where most of the actual content for the vast majority of movies/video content lie.

I don't think you're interested in learning anything, but if you are this is one of the best threads on the subject: http://projectiondream.com/en/movie-...-measurements/

The other issue with ANSI constrast is how difficult it is to measure it if you don't know how to do it, as the room has a huge impact on it, contrary to on/off contrast (provided there is no ambient light).

Regarding the JVC Autocal, I have thousands of hours of experience with it. This is the last thread I created on the subject: https://discuss.avscience.com/index.php?topic=1850.0 it has more than 100,000 views. I have stopped contributing to it because I am not using Autocal as much since JVC finally gave us profile off with the filter, which allows to create large 3D LUTs without posterization or banding from the native, uncorrected wide gamut. So the thread isn't up to date (quite a few things happened since), but it is a great piece of software.

You are still entirely missing the point about the spyders (I've never used the Datacolor software suite, so I can't comment on that, I use JVC Autocal, Colorspace Pro and Calman Ultimate, I have many spyders 4/5/X, a BasiCColor Discus, a Spectracal C6 and an i1pro 2).

If you are incompetent enough to use the Spyders for gamut work without checking them first against a reference meter (preferably a spectro), you can do lots of bad things with them. You can clip the gamut, you can also lose 10-15% brightness due to overcorrection of the RGB balance at 100% white. Before the JVC Autocal started supporting the i1pro/i1pro2, I rejected ALL the Spyder 5 I received (I think I tested five of them). They were all rubbish for color work and it was a real pain to correct them before we found out about the correction matrix in the JVC Autocal settings (more on this below).

Now that JVC Autocal supports the i1pro2 or its cheaper clones (which should only be used for color unless you're prepared to zoom your screen down and use a piece of cardboard or screen material closer to the lens if you can't move the PJ closer to the screen), the color accuracy of the Spyder is 100% irrelevant if you only use it for gamma. You can even use them for color correction if you have access to a reference meter because the JVC Autocal allows you to enter a correction matrix in its settings, so any Spyder can do color and gamma and be just as accurate as the i1pro (or a reference spectro if you have access to one) provided the filters are not too degraded, which takes at least a few years if you store the Spyder properly (peli-case with dessicant).

Finally, you are also entirely missing the point I made about the way these meters are used with projectors. The Spyder 4 and 5 are used with JVC Autocal FACING THE LENS. You set them up so that 100% white is around 5,000nits. So the "black" or near black measures well within the limits of the meter and you can measure up to 33 point for each RGB color very accurately. The Spyder X is to be used facing the screen, which means that for some it might not give results as good as a Spyder 5 or 4 facing the lens, which is fine as you can use almost any Spyder with any software version using unofficial .dlls (see link below).

The JVC Autocal is in fact using a small internal LUT and is the best built-in calibration tool on any consumer projector (when used properly, which clearly isn't the way you've used it). You can use it with unofficially supported meters such as the i1display Pro (a much better low cost meter than the Spyders, far more accurate OOTB at least for bulb LCOS) or even the Klein K10a if you want to (or any meter supported by Colourspace, which means you can use probe matching and use a Jeti as reference without having to manually enter the correction matrix in the JVC Autocal settings). See https://www.avsforum.com/threads/usi...#post-61289806 for more information about that.

Do you really think that the best pro calibrators would be using the JVC Autocal with a matrix corrected Spyder (if they don't want to or are not aware they can use better unofficially supported meters with it) if the results were not good, when they usually have a K10a profiled to a Jeti or CR spectro or better to check results over 1,000 points? I happen to share information privately with a few of them, and I can tell you you are badly mistaken. It's calibrators who calibrate a JVC without using the JVC Autocal first (when there is no external VP / LUT box available) who show their incompetence. You simply can't get optimal performance out of a JVC without doing a JVC Autocal first, unless you're using high bright / profile off with a 3D LUT. Otherwise, you are throwing away contrast and brightness, showing your incompetence, because that contrast and brightness can't be retrieved without correcting the factory tables. The only way to do that is to use the JVC Autocal, before any manual calibration.

Sure, it doesn't replace a proper calibration with an external LUT box / VP and a large 3D LUT, but given the price of the software (free), it does a great job to get the user 90% there when they don't want to invest $$$$$ in an Envy or a Lumagen (especially given that the DTM of the recent native 4K JVCs -- NX and NZ models -- is also very good, unlike any other consumer projector on the market).

I don't have the time to educate you. I was hoping you'd want to do a bit of research and learn about this instead of posting misinformation, but sadly it doesn't seem to be the case.

I won't reply to any further post from you on the subject, unless you acknowledge that you were posting misinformation. If you want to keep posting that JVC projectors are awful, that JVC Autocal is a piece of junk and that Spyder meters are useless with it, that's your right. Hopefully people will be able to make an informed judgement about your claims when they read this thread.

It would be better though if you realised that you're not as knowledgeable on the subject as you think your are, because you could do some research, get JVC Autocal to work properly and stop posting nonsense about it, the JVC projectors or the Spyders.

Anyway, apologies everyone about this long and mostly off-topic post, and Happy Holiday!
__________________
Win11 Pro x64 b23H2
Ryzen 5950X@4.5Ghz 32Gb@3600 Zotac 3090 24Gb 546.65
JRVR/LAV/jRiver/MyMovies/CMC
Denon X8500HA>HD Fury VRRoom>Samsung 77S90C

Last edited by Manni; 24th December 2021 at 15:41.
Manni is offline   Reply With Quote
Old 24th December 2021, 15:12   #62578  |  Link
clsid
*****
 
Join Date: Feb 2005
Posts: 5,664
Quote:
Originally Posted by Perenista View Post
If I put any GPU here every one of those issues would be gone, for good? I replaced the PC and even with the integrated graphics 4K content almost reaches 100% from "GPU" resources when the file is being played fullscreen. I guess I underestimated how much we need a GPU, not just a CPU, for that.
You already have a GPU. Just not a powerful one. So consider using MPC Video Renderer instead of madVR.

Disable "Contrast Enhancement" in the Intel driver settings. That is known to cause a whole variety of issues, including artifacts and stuttering.
__________________
MPC-HC 2.3.0
clsid is offline   Reply With Quote
Old 24th December 2021, 16:03   #62579  |  Link
tp4tissue
Registered User
 
tp4tissue's Avatar
 
Join Date: May 2013
Posts: 753
Quote:
Originally Posted by Manni View Post
Egotistical rant

Again, I've said NONE of the things you assume I've said.

I said JVC autocal is worthless, WITH a spyder 5 or spyder x.

This is in fact the case for the majority of users who WILL NOT have a spectro or any of the knowledge base to do your "Cardboard" trick, or LUT corrections.

I never said, we can't use JVC autocal. Everything I've mentioned is focused on why the S5/x will not work for almost all end-users, and their rapid degredation (relative). Therefore, why buy them.


The other half of this conversation is just You misinterpreting , talking down to people, and being an ass. What with your shallow republican gotcha's oh look dolby said 108 nits instead of 100. All the while their own promo trailer has the engineer quoting 100nits. But that's besides the point.
__________________
Ghetto | 2500k 5Ghz
tp4tissue is offline   Reply With Quote
Old 24th December 2021, 16:39   #62580  |  Link
ajp_anton
Registered User
 
ajp_anton's Avatar
 
Join Date: Aug 2006
Location: Stockholm/Helsinki
Posts: 807
Quote:
Originally Posted by clsid View Post
@ajp_anton
Try with D3D11 hardware decoding instead of DXVA2 Native.

Otherwise, use DXVA2 Copyback and in LAV Video Decoder disable all 10/16-bit output formats. That reduced the amount of data that needs to be copied to GPU memory. This can solve performance issues (on Intel) that are caused by bandwidth bottlenecks.
Thanks, totally forgot about the D3D11 option, and it works. However CPU utilization is about 50% and there's quite a lot of stuttering/jitter (no dropped frames though). DXVA scaling doesn't seem to work.

Also tried MPCVR, and the jitter situation is a little better, but still a lot. DXVA2 also "works" here, but it drops half of the frames.

Forcing copyback to 8bit still drops a lot of frames.

So I went back to DXVA2 with EVR-CP. Almost no CPU utilization, and less jitter than anything else. It's for my mom's computer, so "it should just work" is of higher priority than quality.
ajp_anton is offline   Reply With Quote
Reply

Tags
direct compute, dithering, error diffusion, madvr, ngu, nnedi3, quality, renderer, scaling, uhd upscaling, upsampling

Thread Tools Search this Thread
Search this Thread:

Advanced Search
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off

Forum Jump


All times are GMT +1. The time now is 04:38.


Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2024, vBulletin Solutions Inc.