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 6th July 2020, 09:47   #59781  |  Link
VincAlastor
Registered User
 
Join Date: Sep 2009
Location: Berlin
Posts: 173
Quote:
Originally Posted by deado View Post
Fullscreen exclusive mode has stopped working for me (player is mpc-hc 1.9.6).
here the same!

periodically dropped frames in DX11 fullscreen mode since windows 10 2004 update (Nvidia driver 451.48). No difference with on/off gpu scheduling

has anyone already found a solution to the problem?
VincAlastor is offline   Reply With Quote
Old 6th July 2020, 10:54   #59782  |  Link
Klaus1189
Registered User
 
Join Date: Feb 2015
Location: Bavaria
Posts: 1,666
Quote:
Originally Posted by VincAlastor View Post
...
Please try older drivers: GRD 446.14 or SD 442.92. Does that eliminate the issue?
Klaus1189 is offline   Reply With Quote
Old 9th July 2020, 21:00   #59783  |  Link
VincAlastor
Registered User
 
Join Date: Sep 2009
Location: Berlin
Posts: 173
Quote:
Originally Posted by Klaus1189 View Post
Please try older drivers: GRD 446.14 or SD 442.92. Does that eliminate the issue?
Thank you very much.

GeForce 451.67 GRD WHQL also fixed the issue of frame drops.
VincAlastor is offline   Reply With Quote
Old 9th July 2020, 21:05   #59784  |  Link
Klaus1189
Registered User
 
Join Date: Feb 2015
Location: Bavaria
Posts: 1,666
You‘re welcome. See the driver thread for issues:
AMD, Intel and Nvidia driver issues and last recommended version

I am very active and update as fast as possible.
Klaus1189 is offline   Reply With Quote
Old 14th July 2020, 20:03   #59785  |  Link
orena
Registered User
 
Join Date: Dec 2009
Posts: 4
What is the best MadVR settings for High End HTPC with Nvidia RTX 2080TI GPU?
orena is offline   Reply With Quote
Old 14th July 2020, 22:29   #59786  |  Link
nussman
Registered User
 
Join Date: Nov 2010
Posts: 238
https://forum.doom9.org/forum-rules.htm
nussman is offline   Reply With Quote
Old 14th July 2020, 22:59   #59787  |  Link
mclingo
Registered User
 
Join Date: Aug 2016
Posts: 1,348
Quote:
Originally Posted by orena View Post
What is the best MadVR settings for High End HTPC with Nvidia RTX 2080TI GPU?
Thats like walking into a DIY store and asking a member of staff how to build the best house.

install it, play around a bit, do a bit of reading so you understand what each option does and then you'll be able to push each option a little to see how your PC performs.

There are no "best " options, everyone has different setups and requirements, this is why MADVR is so good, it caters from the noobs who want to just install and go to pro's and everyone in the middle.

I'm no longer a noob but a long way from a pro.
__________________
LG OLED EF950-YAM RX-V685-RYZEN 3600 - 16GBRAM - WIN10 RX 5700 - https://www.videohelp.com/software/madVR/old-versions
mclingo is offline   Reply With Quote
Old 15th July 2020, 03:28   #59788  |  Link
Asmodian
Registered User
 
Join Date: Feb 2002
Location: San Jose, California
Posts: 4,406
Quote:
Originally Posted by orena View Post
What are some good settings MadVR settings for High End HTPC with Nvidia RTX 2080TI GPU?
I have some settings.bin's linked at the end of the third post of the thread in my signature. They are what I use on a 2080 Ti plugged into a LG CX OLED TV. If you are willing to do a bit of reading that thread is also a good place to start understanding the settings so you can tweak to your preferences.

HDR is a bit troublesome now, since I use current Nvidia drivers, but I still use passthrough for HDR. It does require turning on Windows HDR manually but I don't watch that much HDR content.
__________________
madVR options explained
Asmodian is online now   Reply With Quote
Old 15th July 2020, 03:50   #59789  |  Link
VBB
Registered User
 
VBB's Avatar
 
Join Date: May 2016
Location: Long Beach, CA, USA
Posts: 620
As, did you replace your C9?
__________________
Henry | LG OLED65C7P | Denon AVR-X3500H | ELAC Uni-Fi x7 | ELAC Debut 2.0 SUB3030 x2 | NVIDIA SHIELD TV Pro 2019 | Plex
VBB is offline   Reply With Quote
Old 15th July 2020, 04:19   #59790  |  Link
Asmodian
Registered User
 
Join Date: Feb 2002
Location: San Jose, California
Posts: 4,406
Quote:
Originally Posted by VBB View Post
As, did you replace your C9?
Yes, it is still around but I have a small (48") CX now too. I no longer have a C7, just in time.
__________________
madVR options explained
Asmodian is online now   Reply With Quote
Old 18th July 2020, 08:13   #59791  |  Link
mzso
Registered User
 
Join Date: Oct 2009
Posts: 930
Hello!

Which AMD driver version is considered the best (or least bad) for madVR right now?
mzso is offline   Reply With Quote
Old 18th July 2020, 09:13   #59792  |  Link
Klaus1189
Registered User
 
Join Date: Feb 2015
Location: Bavaria
Posts: 1,666
Latest 20.7.2 is fine.
Klaus1189 is offline   Reply With Quote
Old 18th July 2020, 16:59   #59793  |  Link
mzso
Registered User
 
Join Date: Oct 2009
Posts: 930
Quote:
Originally Posted by Klaus1189 View Post
Latest 20.7.2 is fine.
With the plethora of problems I have doubts. (Though Windows 10 might be garbage as well.)
mzso is offline   Reply With Quote
Old 18th July 2020, 17:09   #59794  |  Link
Klaus1189
Registered User
 
Join Date: Feb 2015
Location: Bavaria
Posts: 1,666
Which one do you prefer then?
Klaus1189 is offline   Reply With Quote
Old 19th July 2020, 02:39   #59795  |  Link
TheDarkTemplar
Registered User
 
Join Date: Dec 2011
Posts: 23
Hello! I keep getting this crash in MPC-HC about 50% of the time I open and fullscreen a video:

https://drdump.com/UploadedReport.aspx?DumpID=49079987

It is apparently an issue in MadVR. Is there some workaround to this or will it be fixed at some point? It's not the worst issue ever since I can just relaunch again but it's a bit annoying.

I think it may be related to switching out of HDR since I am generally fullscreening non-HDR videos.

Last edited by TheDarkTemplar; 19th July 2020 at 02:42.
TheDarkTemplar is offline   Reply With Quote
Old 19th July 2020, 16:17   #59796  |  Link
mzso
Registered User
 
Join Date: Oct 2009
Posts: 930
Quote:
Originally Posted by Klaus1189 View Post
Which one do you prefer then?
Maybe all's garbage. And/or the OS is...

Starting to get fed up with everything always being sh*t.

Last edited by mzso; 19th July 2020 at 16:22.
mzso is offline   Reply With Quote
Old 19th July 2020, 16:32   #59797  |  Link
Klaus1189
Registered User
 
Join Date: Feb 2015
Location: Bavaria
Posts: 1,666
Quote:
Originally Posted by mzso View Post
Starting to get fed up with everything always being sh*t.
Developers of graphic drivers don't seem to care about HTPC usage, for many years. None of the three teams is perfect, none will work as you want it to if you try every little thing out.
It is like finding the smallest evil
Klaus1189 is offline   Reply With Quote
Old 20th July 2020, 13:22   #59798  |  Link
Charky
Registered User
 
Join Date: Apr 2018
Location: Paris, France
Posts: 89
Is madvr known for causing a/v sync issues ?

I created a short movie using Adobe Premiere Pro CC 2020 (Vimeo 1080p23 preset, H264/AAC, MP4 container). In Premiere, everything is in sync. However, on my HTPC (MPC-BE, LAV, madvr v113, custom resolutions), I noticed the picture is slightly behind the audio, like 50 ms. With VLC ou MPC-BE using EVR, the sync is perfect, so I'm pretty confident that the issue lies with madvr itself.

I actually never noticed that before with "regular" content (i.e. bluray remuxes) but a 50 ms delay is not much and I probably haven't been paying as much attention as with my own movie, because I spent a lot of time lipsyncing dubbed dialogues and creating foley sounds and making sure everything is in sync.

I solved the issue by setting a 50 ms audio delay in LAV audio decoder (regular content now seem to be in perfect sync too), but am i missing anything ?
__________________
Charky

"Rule #1 : If it works, don't change anything."

Last edited by Charky; 20th July 2020 at 13:55.
Charky is offline   Reply With Quote
Old 22nd July 2020, 07:10   #59799  |  Link
Shiroe
Registered User
 
Join Date: May 2019
Posts: 8
Maybe someone can assist me with a problem. In PotPlayer, when the video renderer is set to madVR, subtitles flicker when seeking.

Here is a demonstration:

https://e.pcloud.link/publink/show?c...aE6vCEmfIODk8X

In those videos, I am jumping by frames. Subtitles disappearing is not limited to jumping by frames. When I jump backward by seconds, sometimes subtitles that had appeared during normal playback do not appear when the video automatically resumes. In that case, I can jump backward again, pause the video, and then play the video to make the subtitles visible. I recently discovered MPC-BE suffers from that same problem, i.e., subtitles disappearing when jumping backward by seconds, when the video renderer is set to madVR.

If I recall correctly, those problems were not present in PotPlayer when I installed it via LAV Filters Megamix, and as far as I know, those problems are also not present in MPC-HC, which I installed via K-Lite Codec Pack Standard.
Shiroe is offline   Reply With Quote
Old 23rd July 2020, 14:36   #59800  |  Link
Alexkral
Registered User
 
Join Date: Oct 2018
Posts: 319
Hi guys, I'm finally done with the Super-resolution implementation in AviSynth+ that I've been working on, you can try it here. Now I'm doing some tests, and since there was some interest, I started with one including all the NGU settings. For this I have used Set14, consisting of 14 images, which with 16 NGU settings has forced me to capture 224 screenshots, so for larger datasets this will not be possible. The other algorithms tested are:

- The five new AiUpscale models.
- The three FSRCNNX models.
- Waifu2x-caffe (upconv_7_photo model).
- Lanczos3 (Matlab) for reference.

The FSRCNNX models were implemented in the same way as the AiUpscale models. The metrics used are:

- PSNR
- SSIM
- IFC
- PI
- SIS

PSNR and SSIM are well known. IFC (Information Fidelity Criterion) is a metric based on natural scene statistics that has been shown to have a high correlation with perceptual scores in the context of SR evaluation. PI (Perceptual Index) is a No Reference metric designed for the 2018 PIRM Challenge on Perceptual Image Super-resolution, which is also highly correlated with the ratings of human observers in a human-opinion study. SIS is a method specifically designed for assessing the visual quality of super-resolved images, which also has a very high correlation with a human-opinion study carried out with a database built for the same purpose.

Downsampling was done with Bicubic (Catmull-Rom settings). All the metrics were computed on the Y channel. NGU Anti-Alias introduces a half pixel offset that needed to be corrected, which slightly worsens the scores (in the case of PSNR I'd say only a few hundredths). For the PI less is better, and since it is a No Reference metric, the reference is also included.

Code:
Model                         PSNR                  Model                         SSIM                  Model                         IFC                   Model                         PI                    Model                         SIS
__________________________    ________________      __________________________    ________________      __________________________    ________________      __________________________    ________________      __________________________    ________________

AiUpscale HQ                  32.5251817903342      AiUpscale Medium              0.90942544644301      AiUpscale Medium              8.27967367353770      Reference                     3.40104937646879      AiUpscale Medium              0.92414368352855
AiUpscale Medium              32.4232206599772      AiUpscale HQ Sharp            0.90843988514755      AiUpscale HQ                  8.24574655763528      AiUpscale HQ Sharp            3.58668578180591      AiUpscale HQ Sharp            0.92382785147580
Waifu2x-caffe                 32.3473914060035      AiUpscale HQ                  0.90839669836945      AiUpscale Fast                8.16061456261743      AiUpscale Medium              3.65547901829712      AiUpscale HQ                  0.92338216249267
NGU Sharp (very high)         32.2938176168901      AiUpscale Fast                0.90693736664703      AiUpscale VDSR                7.90344864991210      Waifu2x-caffe                 3.67943591351007      AiUpscale Fast                0.92006283988828
AiUpscale VDSR                32.1939373807653      AiUpscale VDSR                0.90521242215438      AiUpscale HQ Sharp            7.73815420390795      AiUpscale Fast                3.71741685428603      AiUpscale VDSR                0.91742165939633
AiUpscale Fast                32.1863953769405      Waifu2x-caffe                 0.90495347031380      Waifu2x-caffe                 7.68228442626807      FSRCNNX_x2_8-0-4-1            3.81516969239280      Waifu2x-caffe                 0.91636248251022
FSRCNNX_x2_56-16-4-1          32.1759503587084      NGU Sharp (very high)         0.90460008282567      FSRCNNX_x2_8-0-4-1            7.66717466529441      AiUpscale HQ                  3.84336339556552      FSRCNNX_x2_8-0-4-1            0.91619538116915
FSRCNNX_x2_16-0-4-1           32.1484038971244      NGU Sharp (high)              0.90193258684294      FSRCNNX_x2_16-0-4-1           7.56245532653046      NGU Sharp (very high)         3.93987906364856      FSRCNNX_x2_16-0-4-1           0.91375067249656
AiUpscale HQ Sharp            32.1199740239966      FSRCNNX_x2_8-0-4-1            0.90107924389368      NGU Sharp (very high)         7.56185554352781      FSRCNNX_x2_56-16-4-1          3.96283639678717      NGU Sharp (very high)         0.91283152130317
NGU Sharp (high)              32.0745914126075      FSRCNNX_x2_16-0-4-1           0.90021250932189      FSRCNNX_x2_56-16-4-1          7.54238083322097      NGU Sharp (high)              3.96668297143171      FSRCNNX_x2_56-16-4-1          0.91250552777137
FSRCNNX_x2_8-0-4-1            32.0254211392229      NGU Sharp (med)               0.90008345962041      NGU Sharp (high)              7.10509959144021      FSRCNNX_x2_16-0-4-1           3.98720638993818      NGU Sharp (high)              0.90383627513313
NGU Sharp (med)               31.9361908894940      FSRCNNX_x2_56-16-4-1          0.89928737851361      NGU Sharp (med)               6.93150412700636      AiUpscale VDSR                4.04068349093325      NGU Sharp (low)               0.90309609902881
NGU Sharp (low)               31.7104151080959      NGU Sharp (low)               0.89500452165092      NGU Standard (very high)      6.79878287005572      NGU Sharp (med)               4.04345458415675      NGU Sharp (med)               0.90084060562877
NGU Standard (very high)      31.5974304026016      NGU Standard (very high)      0.89373264838815      Lanczos                       6.78807153422038      NGU Sharp (low)               4.20257164181289      NGU Standard (very high)      0.89213869604695
NGU Standard (high)           31.4416756290765      NGU Standard (high)           0.89223745931491      NGU Sharp (low)               6.76680062532443      NGU Standard (very high)      4.27329141474409      NGU Standard (high)           0.88663713794603
NGU Standard (med)            31.3266464934018      NGU Standard (med)            0.89079871365700      NGU Standard (high)           6.65548044824456      NGU Standard (high)           4.34545507861726      NGU Standard (med)            0.88214872829691
NGU Standard (low)            31.1795059486520      NGU Standard (low)            0.88724999653284      NGU Anti-Alias (very high)    6.54847171682001      NGU Standard (med)            4.34697051435912      NGU Standard (low)            0.87988322632132
NGU Anti-Alias (very high)    31.0032971499628      NGU Anti-Alias (very high)    0.88462531208344      NGU Standard (med)            6.52544719594284      NGU Standard (low)            4.47568646264313      NGU Anti-Alias (very high)    0.87742671984170
NGU Anti-Alias (high)         30.9569815543731      NGU Anti-Alias (high)         0.88321233882888      NGU Anti-Alias (high)         6.42962113230166      NGU Soft (very high)          4.87787622216863      Lanczos                       0.87560000704257
NGU Anti-Alias (med)          30.8970336747559      NGU Anti-Alias (low)          0.88065405839635      NGU Standard (low)            6.38911485598076      NGU Soft (med)                4.89969182841594      NGU Anti-Alias (high)         0.87458590111814
NGU Anti-Alias (low)          30.8339491643038      Lanczos                       0.88061992663747      NGU Anti-Alias (low)          6.30590655792457      NGU Anti-Alias (very high)    4.91336898065817      NGU Anti-Alias (med)          0.87280516539207
Lanczos                       30.6067648881095      NGU Anti-Alias (med)          0.88043102098649      NGU Anti-Alias (med)          6.18364967410367      Lanczos                       4.92084403030973      NGU Anti-Alias (low)          0.87012785173923
NGU Soft (very high)          30.6012993686409      NGU Soft (very high)          0.87854781224895      NGU Soft (very high)          5.98479762818895      NGU Soft (high)               4.92618718914022      NGU Soft (very high)          0.85343713547148
NGU Soft (high)               30.4664433204248      NGU Soft (high)               0.87680636052519      NGU Soft (high)               5.92007932481318      NGU Anti-Alias (low)          4.92710046176526      NGU Soft (high)               0.85054526901960
NGU Soft (med)                30.4483124249698      NGU Soft (med)                0.87621006480048      NGU Soft (med)                5.89496906342640      NGU Anti-Alias (high)         4.94724879619383      NGU Soft (med)                0.84712992882588
NGU Soft (low)                30.2594794628864      NGU Soft (low)                0.87162551537298      NGU Soft (low)                5.71276482218494      NGU Anti-Alias (med)          4.95971449743798      NGU Soft (low)                0.83820122012128
                                                                                                                                                            NGU Soft (low)                5.07866068664092
As you can see, the NGU scores are completely ordered (Sharp > Standard > Anti-Alias > Soft) especially in terms of PSNR and SSIM, with some differences in the other metrics (e.g. AA low performs better than AA med for three metrics). NGU Soft is consistently below Lanczos, same as AA for all metrics except PSNR and SSIM, and even Standard for IFC. FSRCNNX_x2_8-0-4-1 performs better than the other FSRCNNX for all metrics except PSNR. Waifu performs well, and needless to say, I am very satisfied with the AiUpscale results.

Last edited by Alexkral; 20th October 2020 at 00:48.
Alexkral 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 20:46.


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