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. |
21st October 2015, 15:28 | #33801 | Link | ||||||||
Registered User
Join Date: May 2009
Posts: 120
|
first of a bugreport!
in MadVR. devices - sony vm500es - properties - green text right of "test projector" says (JVC) until i press "FIND"/"TEST" again, even if the ip is correct. Quote:
Quote:
Quote:
EDIT: i now tried to make a script. Is this correct? on a "1.85:1" movie it still uses "2.40:1" my profiles are called "1.77:1", "1.85:1", "2.40:1" Code:
if (AR = 1.85:1) "1.85:1" else if (AR = 2.35:1) "2.40:1" else if (AR = 2.40:1) "2.40:1" else "1.77:1" Quote:
Quote:
Quote:
Quote:
Quote:
__________________
Gigabyte G1.Sniper Z87, I7 4770K (3.5ghz), Gigabyte Radeon 280x, Win8.1 x64, ATI 15.7.1, MPC-HC BE 1.4.6 1478, MadVR v0.90.21, LAV Filters 0.68.1, XySubFilter 3.1.0.746 Last edited by DigitalLF; 21st October 2015 at 16:55. |
||||||||
21st October 2015, 18:08 | #33803 | Link |
Registered User
Join Date: Aug 2004
Location: Canada
Posts: 860
|
Is there some reason why I can't get the madvr test patterns (ytp files) to load? It's been like this for 6 months or more. Just says cannot render file. I'm on 88.21 and MPC 1.7.9, both 64-bit. Does this maybe have a 64-bit conflict? Windows 7 SP1
|
21st October 2015, 18:29 | #33804 | Link | |
Registered User
Join Date: Apr 2011
Posts: 54
|
Quote:
Regards!
__________________
Intel Core i7-4790 CPU @ 3.60GHz, RAM 32 GB Dual-Channel DDR3 @ 665MHz (9-9-9-24), Panasonic TX-P42G20E, NVIDIA GeForce GTX 970, Win 10 Pro x64, PotPlayer 1.7.16291 64-bit, madVR v0.92.17 |
|
21st October 2015, 20:46 | #33805 | Link | |
Registered User
Join Date: Jan 2014
Location: France
Posts: 76
|
Quote:
You can find below the settings and results with Linear Light On & Off : OSD MPC-HC & GPU-Z with Linear Light OFF : OSD MPC-HC & GPU-Z with Linear Light ON : madVR - General Settings : madVR - Output : madVR - Trade Quality for Performance : madVR - Downscale : madVR - Chroma Upscaling : madVR - Double Luma : madVR - Upscaling Refinement : madVR - Dithering : madVR - Smooth Motion : madVR - Image Enhancements : madVR - Deband : Hope it helps |
|
21st October 2015, 21:09 | #33806 | Link |
Registered User
Join Date: Nov 2012
Posts: 11
|
I'm using madVR with MPC-HC.
I noticed that if I take a screenshot with the window resized below the actual video size the image is taken at the window resolution, with aspect ratio correction (if any) applied. This is an interesting behaviour and I would like to ask if it's possible (and how) to:
|
21st October 2015, 22:13 | #33807 | Link | |
Registered User
Join Date: Oct 2008
Posts: 187
|
Quote:
If the JVC is off or it's been turned on locally (JVC remote control or power button) everything works as expected using your program. If the JVC has been turned on remotely by me (either by RS232 or TCPIP) then your program will still find it, but always reports power state as standby and NONE of the other commands work. My software does not issue IP (or rs232) control commands once power up is completed until power down. Note that I close the socket after every control command (and that the JVC would close it if I don't). Thus it seems that once the JVC has been turned on by a control command from one source, it will no longer accept control commands from any other source (TCPIP or RS232) until a power off command is issued from the same source that issued the power on. |
|
21st October 2015, 22:14 | #33808 | Link | |
Registered Developer
Join Date: Mar 2010
Location: Hamburg/Germany
Posts: 10,344
|
Quote:
Maybe a period of quiet and collecting and fixing issues is a good thing.
__________________
LAV Filters - open source ffmpeg based media splitter and decoders Last edited by nevcairiel; 22nd October 2015 at 00:29. |
|
22nd October 2015, 00:54 | #33810 | Link | |
Registered User
Join Date: Nov 2012
Posts: 138
|
Quote:
i am mostly just using Jinc3AR with AdaptiveSharpen 0.7 pre upscale Last edited by truexfan81; 22nd October 2015 at 00:57. |
|
22nd October 2015, 07:07 | #33814 | Link |
Registered User
Join Date: May 2012
Posts: 57
|
Hi guys,
I really need your help in understanding the effect of some changes I have made in madVR that allowed me to have a smooth playback again after hours of fighting with a nasty stuttering. Sorry if this seems stupid but I just need to know whether I have finally found the magic solution or is just something circumstantial. First let me list my rig and the initial settings: - Intel NUC D34010WYKH, intel Core I3 4010U (r (1.7 GHz Dual-Core processor with 3 MB smart cache), Intel HD 4400 GPU, Crucial 2x4GB = 8GB DDR3 1600MHz, 60GB OCZ Agility SSD - LAVFIlters: DXVA Copy-Back - madVR: * chroma-up: BCB75AR * image-up: DXVA * image doubling: None * image-down: DXVA * enable fullscreen exclusive mode * dithering: ordered * smooth motion: off - SVP: default 1g profile, 59.94Hz - TV; Sony KDL26EX320, 60Hz, 1366x768 - ReClock: WASAPI exclusive Asus Xonar U3, 24 bit padded into 32, resolution of 59.94 input manually first time the video plays - movies: Blu-Rays: 1920x1080 / 800, 23,976fps, DVDs: 720x576, 25,000fps (PAL) , Samsung S4: 1920x1080, 29,911fps. With the above settings and everything else left as default in madVR I had a terrible stuttering even if madVR stats showed 1 frame drop every 6-7 hours. Please see print screen below: But then I have just spotted that my "present queue" was showing 5-7/8 instead of 7-8/8 so I decided to drop the frames presented in advance from 8 to 4 in fullscreen mode and from 8 to 3 in windowed mode. Also in the tab trade quality for performance I've checked the first 4 options up to "store custom pixel shader ..." ... and voila my "present queue" was now 2-3/3 (windowed) and 3-4/4 (fullscreen) and the stutter was almost non percievable. Then I remember an article on Anandtech where they tested the NUC (the Core I5 version) with madVR and used "CPU and GPU queue sizes of 128 and 24". I said why not try, did that and no more stutter visible, madVR shows 1 drop frame every 1.02 days and SVP performance graphs show CPU average at 67% with 1g profile. Indeed the guys at Anandtech also increased the frames presented in advance to 16 ... and this is where I'm a bit lost. They have practically maxed everything. http://www.anandtech.com/show/7566/i...ff-pc-review/5 Can you please tell me in which way did my PC benefited from the above and whether maxing out CPU and GPU queues size is a right thing to do in my case? Are there any other settings that would give me a better bang for the buck? Last edited by ionutm80; 22nd October 2015 at 07:37. |
22nd October 2015, 10:45 | #33815 | Link |
Registered User
Join Date: Jun 2007
Posts: 652
|
mashi, I fully admit that I'm ignorant about this, and you might have already given an answer for this. Considering your experience with dedicated scalers (Lumagen, if I remember correctly from our Panasonic discussions for 1:1 pixel mappings), is Non Linear Stretch for 4:3 material something you might be considering for the future?
|
22nd October 2015, 10:55 | #33816 | Link | |||||||||||||||
Registered Developer
Join Date: Sep 2006
Posts: 9,140
|
Quote:
Quote:
Quote:
Quote:
Quote:
Yep, seems to be a GPU RAM size issue. When enabling linear light, I'm using 32bit floating point textures for some processing steps, which consume twice as much RAM as the default format. I'm a bit surprised GPU RAM size seems to more than double, though, that doesn't make a lot of sense, to be honest. Anyway, you can probably work around the issue by reducing the size of the GPU queue a bit. Quote:
Quote:
http://madshi.net/madHcCtrlSony.rar Does that fix the problem for you? It might also make lens memories work for your Sony. Quote:
Code:
if (AR > 1.9) "2.40:1" else if (AR > 1.8) "1.85:1" else "1.77:1" Quote:
I'm aware of that it's pretty hard to say when linear light should be enabled and when not. So maybe I should remove the option. But which option should I set it to then? Especially with Anime content the difference between linear light and gamma light can be *huge*. Black lines are very noticably fatter when using linear light. Quote:
Quote:
1) projector is turned on by rs232 How is this possible?? At least my JVC X35 has a setup menu option to either use rs232 or ip control - I have to choose! Does your projector not have such a menu option? If it does have such an option, have you set it to rs232 or ip control? Does your own ip control software work if you turned the projector on via rs232, while my ip control test tool fails? 2) projector is turned on by your ip control software In this situation my ip control test tool fails, too? Does your own ip control software manage to switch lens memories or stuff in this situation, while my test tool fails? 3) projector is turned on physically (remote or project power switch) In this situation my ip control test tool works perfectly? 4) projector is turned on by my ip control test tool What happens in this case? Does my test tool work? Does your ip control software work? Quote:
Anyway, I think the biggest stability problems were caused by adding support for rendering in paused&stopped graph state, followed by changes needed to make black bar detection work. All those changes were rather invasive, and thus came with a high risk of new stability issues. However, I think most of the other new features planned for the near future are less risky, so stability should trend up now, regardless of whether I add new features or not. Do you have any specific problems with v0.89.12? My impression from feedback so far is that it's working pretty well and stable. Quote:
Quote:
Quote:
|
|||||||||||||||
22nd October 2015, 10:59 | #33817 | Link |
Registered Developer
Join Date: Mar 2010
Location: Hamburg/Germany
Posts: 10,344
|
So far it seems pretty good, indeed. Will have to keep an eye on it.
__________________
LAV Filters - open source ffmpeg based media splitter and decoders |
22nd October 2015, 11:16 | #33818 | Link | |
Registered Developer
Join Date: Sep 2006
Posts: 9,140
|
Quote:
FWIW, the later v0.89.x builds have several stability bug fixes in them, which e.g. v0.88.x/v0.87.x and older builds are lacking. E.g. I've been working with CiNcH on live TV playback stability when switching TV channels. He says the later v0.89.x builds are more stable with DVBViewer than any previous build ever was. So I'm hopeful about stability, especially considering that no further *invasive* changes are planned for the near future. |
|
Tags |
direct compute, dithering, error diffusion, madvr, ngu, nnedi3, quality, renderer, scaling, uhd upscaling, upsampling |
Thread Tools | Search this Thread |
Display Modes | |
|
|