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. |
12th October 2011, 23:10 | #10081 | Link |
Registered User
Join Date: Jan 2009
Posts: 1,210
|
Can anyone guide or link me to a basic to advanced calibration method for an LCD Screen, LCD LED Screen, and a Plasma screen? I have completely disabled driver software changes to the display (aka catalyst & nvidia control panel) and just have a windows 7 calibration done on all three of my monitors after putting them on default brightness/contrast levels (which seem pretty clear/normal).
What my question essentially boils down to is: What tools do I need to setup up the best possible calibration which I can do myself and not need to have a professional come in and do. I'm on Windows 7 x64 w/ latest madVR. It's not that I don't want to spend money on a professional calibration. I'm just interested in doing it myself and learning how to do it rather than have someone else do it, even if costs more in the end. Thanks! |
13th October 2011, 00:12 | #10082 | Link |
Registered User
Join Date: May 2011
Posts: 94
|
This site is a great source of info.
http://www.curtpalme.com/forum/viewtopic.php?t=10457 |
13th October 2011, 08:01 | #10083 | Link | ||||
Registered User
Join Date: Aug 2008
Location: the Netherlands
Posts: 851
|
Quote:
Will he need to do a re-calibration for my pj for the best result? EDIT: got a new reply from the IFS guy. I ll try to translate my best: Quote:
EDIT 2: And again a reply: Quote:
EDIT 3: Another reply from him: Quote:
Last edited by THX-UltraII; 13th October 2011 at 10:42. |
||||
13th October 2011, 11:57 | #10085 | Link | |||||
( ≖‿≖)
Join Date: Jul 2011
Location: BW, Germany
Posts: 380
|
Quote:
Quote:
Quote:
The point here is that you should get the projector as close as possible to BT.709 (he's done this), then *profile* it so you have a detailed report of all of the inaccuracies. You should do this over at least 2000 sample patches, but I'd recommend more. An optimal profiling should probably have been done overnight, as they can take many hours at several thousand samples. Then, you can use your very, very powerful (compared to the projector) HTPC to adjust for all of the tiny little differences. You'll probably be able to cut the delta E in half that way. I doubt you'll see a difference with the eye, but better is always better, right? Quote:
Quote:
Upload your attachment somewhere else, getting them approved on doom9 can take quite a while.
__________________
Forget about my old .3dlut stuff, just use mpv if you want accurate color management Last edited by nand chan; 13th October 2011 at 11:59. |
|||||
13th October 2011, 12:02 | #10086 | Link | |
Kid for Today
Join Date: Aug 2004
Posts: 3,477
|
Quote:
|
|
13th October 2011, 12:06 | #10087 | Link | |
( ≖‿≖)
Join Date: Jul 2011
Location: BW, Germany
Posts: 380
|
Quote:
Yes and madVR runs off 16-bit processing, and uses 16-bit LUTs. Even if the projector is capable of 12-bit internal processing, how many samples does it really store?
__________________
Forget about my old .3dlut stuff, just use mpv if you want accurate color management Last edited by nand chan; 13th October 2011 at 12:08. |
|
13th October 2011, 13:01 | #10088 | Link |
Registered User
Join Date: Aug 2008
Location: the Netherlands
Posts: 851
|
Thanks for your very extensive reply and time you take for my postings nand chan. The only problem is that the stuff you talk about is way out of my league and I just don t understand all these values/settings you discuss. I emailed all your replies directly to him so I hope he will take the time to reply to this again.
But is the conclusion that you are saying that I will only get some minor (maybe not seeable) improvements with the things you talk about? Just want good PQ with correct colours displayed. Right now my setup is: - latest MPC-HC with madVR as output renderer. - perfect video sync with 1-3 dropped frames/hour (reported my madVR) - LAV video decoder with all boxes checked in the output formats field and HQ video checked - ATI 5450 with latest ati driver and RGB Full as Pixel Format - 0-255 checked in madVR - projector set to RGB as colourspace and 'ENHANCED' as hdmi input |
13th October 2011, 13:12 | #10089 | Link | ||
( ≖‿≖)
Join Date: Jul 2011
Location: BW, Germany
Posts: 380
|
Quote:
Quote:
__________________
Forget about my old .3dlut stuff, just use mpv if you want accurate color management |
||
13th October 2011, 14:02 | #10092 | Link | |
( ≖‿≖)
Join Date: Jul 2011
Location: BW, Germany
Posts: 380
|
Quote:
Now the problem is, the eye (or rather the brain) doesn't process light linearly. If something is surrounded by a very dark color, it'll appear different than if it's surrounded by a very bright color. It's the reason those optical illusions work (A and B are the same color). So what we ended up doing is we started encoding images using a lower gamma value (1/2.2), which results in the function shown in red. However, if we decode this on a monitor that uses a higher gamma value (eg. 2.4, shown in the blue curve below), we get the result shown in purple, which is darker than a truly linear source. What basically is happening here is that colors are appearing darker on our monitors than they should. Since the monitor is surrounded by a dark environment, this effect cancels itself out in the brain (something surrounded by darkness will appear brighter), and we “perceptually” get the original colors back. However, on your monitor, you have a curve of 2.3 PPC used to decode, so the result will be closer to linear. This works fine for the average dimly lit playback environment, but if you're in a perfectly dark room, you will need to emulate a higher decoding value. What I recommend doing is going into your madVR options, check “enable gamma processing”, select “pure power curve” and choose something like 2.4 or 2.5. You'll probably have to play with the value somewhat, but this should result in a more optimal result for playback in a very dark environment. I personally use 2.5 at night, 2.4 in the evening and 2.3 during the day. Then, go to your calibration settings and choose “the display is already calibrated”, select “BT.709” for the primaries, “pure power curve” for the curve and 2.30 for the setting. I hope this clears things up a bit You're probably right. THX-UltraII, if you have any follow up to this post, send it to me via PM.
__________________
Forget about my old .3dlut stuff, just use mpv if you want accurate color management Last edited by nand chan; 13th October 2011 at 14:11. |
|
13th October 2011, 14:59 | #10093 | Link | |
Kid for Today
Join Date: Aug 2004
Posts: 3,477
|
Quote:
-I'm really not up to date on 10bit output, as I happen to be highly XP resilent...but from what I gathered, 10bit output doesn't work in mVR yet? So at the end of the day, your grossly oversized 16bit LUT is being encoded in RGB24 TMDS. And yes, 10/12(or maybe 14/16?) internal processing in his spanking new JVC pj will look better and less bandy/noisy. -The RS1 didn't have any gamma settings at all, so I would agree w/ your statement as this very much applies for the need of an ICC...but I believe they've come a long way since then? I've got no interest in LCD based pj's(due to problems such as misconvergence), but a DDP3010 Darkchip3 DLP combo works in 10bit and provides just the right amount of colorimetry settings to avoid highly lossy 8bit adjustments. -I might be wrong but all the stuff you like to nitpick about lately is only true if you're using a top of the line (freshly factory recalibrated) spectrophotometer and recalibrate your display at least on a weekly basis. I believe hardly anyone can DBT the same test pattern w/ a ΔE of 0/1 or 2. And many ppl can't be hassled to learn the calibration 101 and prefer to pay an ISF guy to come to their house....but we both know how fluctuant the colorimetery of a high pressure light bulb is, so one week down the road their calibration is pretty much fubar. Theory is nice and all, and of course everything should be done w/ the highest accuracy, but please don't overlook the real world consequences. And tbh, I rest my case that this kind of discussion has little to do in the mVR thread. I've already helped THX-UltraII on both AVS and the Reclock forum....he'll come up w/ questions until the end of days, so better create a separate thread now IMHO |
|
13th October 2011, 18:02 | #10094 | Link | ||
Registered User
Join Date: Apr 2009
Posts: 1,019
|
Quote:
I would want to do as much calibration as possible—especially the lower-end of the greyscale & gamma—using the display's own controls. I hope Madshi manages to figure out how to get 10-bit over HDMI in full-screen exclusive mode, as it would make such a difference for those of us using 10-bit native panels. (LCoS/SXRD, higher-end LCDs) Quote:
Display gamma is absolutely not intended to get a linear result from this. The short answer is that studio CRTs were as close to 2.40 gamma as they could be made, and the new generation of OLED monitors are exactly 2.40 gamma. Anything else for video/film content is wrong. (but you may have to deviate from correct if you aren't watching in a dim/dark room) PC/Game content uses sRGB, which is essentially 2.2 gamma. (technically there is an sRGB transfer function but I don't know of anything that uses it) This is good advice, and probably what he should do. If the projector has been well calibrated, it's almost certainly not worth paying €250 just to get a profile made. That price is fairly ridiculous as profiling just involves setting up the meter and leaving it to run unattended for 20+ minutes, depending on the package used. |
||
13th October 2011, 18:14 | #10095 | Link | |||||
( ≖‿≖)
Join Date: Jul 2011
Location: BW, Germany
Posts: 380
|
Quote:
Quote:
Quote:
Quote:
Quote:
__________________
Forget about my old .3dlut stuff, just use mpv if you want accurate color management |
|||||
13th October 2011, 19:53 | #10096 | Link |
Registered User
Join Date: Jan 2007
Posts: 530
|
madshi, small cosmetic issue:
dual-mon, playback on second mon, FSE mode (Win7 x64), MPC-HC. With the OSD turned on, if I then hover the mouse over the "Show Desktop" area of the toolbar (on the first mon), the seek bar then shows on the playback mon. Turning OSD off does not exhibit this behavior. |
13th October 2011, 21:37 | #10097 | Link | |||
Registered User
Join Date: Apr 2009
Posts: 1,019
|
Quote:
2.40 gamma is not displaying things "darker than they should", it's displaying things as intended. Quote:
Reference gamma is 2.40, period. If you want to see the content as it was created and intended to be viewed, that is what's required on your display. Quote:
Unless you are using pro-grade actively cooled meters in a temperature-controlled room (consumer-grade displays do not regulate their temperature) you are never going to get accurate results if it's going to take you 10 hours for measurements. 5% steps is arguably too much data for a LUT, and with some displays, even 10% steps is too much data to get a good calibration result. (it may look nice when you only plot those 10-20 points, but not when you look at the whole greyscale) At most, profiling a display—that is the computer taking automated measurements, not an interactive calibration—should take an hour. Most packages with newer meters take half that time or less. |
|||
14th October 2011, 02:43 | #10098 | Link |
Registered User
Join Date: Dec 2007
Posts: 652
|
Cross post from the J.River Beta Forum as I "think" it is a madVR bug but...
1st Issue: madVR Windowed mode = Lip Sync issues on 23.776/24hz (the video is delayed so the sound come first) There is no issue with: - EVR - madVR Exclusive mode - madVR Windowed mode at 50hz or 60hz I've tried different combinations and the following make no difference to the lipsync issues: - Decoding Vs Bitstreaming - Video Clock on or off - Normalize vol on or off - HW Accell on or off (eg CUVID vs FFDSHOW) - madVR Settings of "delay playback start until render queue is full" on or off - madVR Settings of "use a separate device for presentation" - Pause/Play or a Seek makes no difference - it is consistently out - Disabling auto refresh rate changes - Manually pre changing refresh rates My only conclusion at this stage is that it is madVR Windowed mode that delays the video when the refresh rate is at 23.976/24hz ...unless there is other suggestions on what to try. Last edited by jmone; 14th October 2011 at 06:09. |
14th October 2011, 02:47 | #10099 | Link |
Registered User
Join Date: Dec 2007
Posts: 652
|
2nd Issue: "Exclusive Mode Failed" : A few of us are having this issue on the HTPC where the Player software may not be restarted for days. At some point madVR will throw up an "Exclusive Mode Failed" error msg, and from that point on will only use Windowed Mode till the player is closed and restarted (J.River Media Center) which fixes the issue every time (but has a low WAF). Any suggestions on how to eliminate "Exclusive Mode Failed" errors?
|
14th October 2011, 05:54 | #10100 | Link |
Registered User
Join Date: Aug 2004
Location: Canada
Posts: 860
|
Does the 'Use D3D11 for presentation' need a DirectX 11 card? I enabled it and assumed it wouldn't operate, but seems to play normally.
Also, I think the 'overshoot max...' option has zeroed my presentation glitches. I thought I had it sorted before, but I noticed them cropping up again My settings now would be: 5/8 for windowed and exclusive buffers with 'overshoot max frame latency' tweak Last edited by JarrettH; 14th October 2011 at 06:08. |
Tags |
direct compute, dithering, error diffusion, madvr, ngu, nnedi3, quality, renderer, scaling, uhd upscaling, upsampling |
|
|