Improving Madvr HDR to SDR mapping for projector - Page 257 - AVS Forum | Home Theater Discussions And Reviews
Forum Jump: 
 8866Likes
Reply
 
Thread Tools
post #7681 of 9401 Old 11-06-2019, 07:49 AM
AVS Forum Special Member
 
Neo-XP's Avatar
 
Join Date: Jun 2018
Location: Switzerland
Posts: 1,208
Mentioned: 220 Post(s)
Tagged: 0 Thread(s)
Quoted: 928 Post(s)
Liked: 1323
The log curve is sexy too:

Hill (100-175-225-261-288-308-325-339-350-368-381-400)



Log (100-175-231-264-287-305-320-333-344-362-377-400)


It is very slightly less detrimental on highlights than Hill, with the most extreme content only. No other visible changes whatsoever.

I don't think anyone will ever see any difference anyway. I had to pause + take screenshots + zoom on specific parts of both images + switch from one to the other to see the 0.1% improvement

But any improvement is good to take I guess
SamuriHL, Manni01, *Mori* and 3 others like this.

Last edited by Neo-XP; 11-06-2019 at 01:26 PM.
Neo-XP is offline  
Sponsored Links
Advertisement
 
post #7682 of 9401 Old 11-06-2019, 08:13 AM
Member
 
RioBar4U's Avatar
 
Join Date: Dec 2011
Location: Switzerland
Posts: 90
Mentioned: 6 Post(s)
Tagged: 0 Thread(s)
Quoted: 27 Post(s)
Liked: 39
I will try it tonight...
Neo-XP likes this.
RioBar4U is offline  
post #7683 of 9401 Old 11-06-2019, 10:56 AM
Member
 
BennyTurbo's Avatar
 
Join Date: Apr 2012
Location: Germany
Posts: 102
Mentioned: 2 Post(s)
Tagged: 0 Thread(s)
Quoted: 51 Post(s)
Liked: 33
Quote:
Originally Posted by Neo-XP View Post
For @woofer , @AKJ89 , @AmigoHD , @Reezepees , @Icaro , @CoryW , @BennyTurbo , @Boris Zatserkovnyy only:

Can live with L-80 (Manni01's version): Soulnight

Sorry for late answer. Yes ok for me.
Manni01 and Neo-XP like this.
BennyTurbo is offline  
Sponsored Links
Advertisement
 
post #7684 of 9401 Old 11-06-2019, 10:59 AM
AVS Forum Special Member
 
Neo-XP's Avatar
 
Join Date: Jun 2018
Location: Switzerland
Posts: 1,208
Mentioned: 220 Post(s)
Tagged: 0 Thread(s)
Quoted: 928 Post(s)
Liked: 1323
Quote:
Originally Posted by BennyTurbo View Post
Sorry for late answer. Yes ok for me.
Well then every user who preferred L-63 can live with L-80 (Manni01's version).
stevenjw and Manni01 like this.

Last edited by Neo-XP; 11-06-2019 at 11:27 AM.
Neo-XP is offline  
post #7685 of 9401 Old 11-06-2019, 11:21 AM
Senior Member
 
dimi123's Avatar
 
Join Date: Jan 2012
Posts: 274
Mentioned: 2 Post(s)
Tagged: 0 Thread(s)
Quoted: 221 Post(s)
Liked: 80
@Neo-XP 's Hill curve with @Manni01 's settings for me.


Epson 9300W (6040), anamorphic lens, P3 filter, 2.4 Gamma, 135" scope screen, 150 real nits
Manni01 and Neo-XP like this.

PJ: Epson EH-TW9300W | Screen: 135" 2.40 Stewart StudioTek 100 | Anamorphic lens: Prismasonic Cinomorph P-100M | AVR: Denon AVR-X7200WA | Amp: 3x Naim Nait 5si Stereo Amplifier| HTPC: Intel Core i7, GTX1080Ti | BR Player: OPPO UDP-203, OPPO BDP-103D, OPPO BDP-95 | Speakers: JBL Studio L Series, 7.2.4
dimi123 is offline  
post #7686 of 9401 Old 11-06-2019, 11:43 AM
Advanced Member
 
ddgdl's Avatar
 
Join Date: Mar 2007
Posts: 686
Mentioned: 6 Post(s)
Tagged: 0 Thread(s)
Quoted: 379 Post(s)
Liked: 360
So at this point, I think we suggest to madshi a drop down box with 3 options, hill, L80, and mirror. Names can be something like:

High contrast display/flat panel (hill)
Medium contrast display/projector (medium strength)
Medium contrast display/projector (high strength)

Now, does it make sense to spend the rest of the week playing with the parameters of mirror and hill a little to tweak them for minor improvements? Or should we start figuring out sky detection and shadow boost with these set parameters for HSTM?

My vote would be everyone taking their preferred HSTM setting of those 3, and seeing what are the best sky and shadow settings to use, given the fixed HSTM settings.
LTD_666 likes this.
ddgdl is offline  
post #7687 of 9401 Old 11-06-2019, 12:50 PM
AVS Forum Special Member
 
Neo-XP's Avatar
 
Join Date: Jun 2018
Location: Switzerland
Posts: 1,208
Mentioned: 220 Post(s)
Tagged: 0 Thread(s)
Quoted: 928 Post(s)
Liked: 1323
Quote:
Originally Posted by ddgdl View Post
Now, does it make sense to spend the rest of the week playing with the parameters of mirror and hill a little to tweak them for minor improvements? Or should we start figuring out sky detection and shadow boost with these set parameters for HSTM?
I am not sure if Mirror makes much sense for a high strength, because it has less strength on some parts of the image than L-80 sometimes...

Example:

HSTM Off / Log / L-80 (Manni's version) / Mirror


Pacific Rim - Frame 39868

It goes way too far IMO.

Last edited by Neo-XP; 11-06-2019 at 01:29 PM.
Neo-XP is offline  
post #7688 of 9401 Old 11-06-2019, 02:42 PM - Thread Starter
AVS Forum Special Member
 
Soulnight's Avatar
 
Join Date: Dec 2011
Location: Germany
Posts: 1,479
Mentioned: 235 Post(s)
Tagged: 0 Thread(s)
Quoted: 1132 Post(s)
Liked: 1863
Quote:
Originally Posted by Neo-XP View Post
I am not sure if Mirror makes much sense for a high strength, because it has less strength on some parts of the image than L-80 sometimes...

Example:

HSTM Off / Log / L-80 (Manni's version) / Mirror


Pacific Rim - Frame 39868

It goes way too far IMO.
Thank you for this comparison

In your example, Mirror clearly goes further than L-80 and bring more pop (face/background) : Stronger strength. Which part of the image has for you less strength ?

In any case "going too far" is a matter of taste as your poll as proven.

Also I m not sure than your are the most adequate person for choosing a high mode since you already dislike the balanced L-80 settings that so many user like but that you find already too strong compared to the Hill (which may be related to you being an OLED owner).

Plus it seems like the oled owners would not be the primary users of the HSTM linear curve, and even less of a potential high mode from the Preference Poll. It would be more the low nits projector users on big screen probably with DCI coverage which would be the potential user of such a high HSTM mode.
stevenjw, Manni01, AKJ89 and 1 others like this.

Last edited by Soulnight; 11-06-2019 at 02:46 PM.
Soulnight is offline  
post #7689 of 9401 Old 11-06-2019, 03:12 PM
AVS Forum Special Member
 
Neo-XP's Avatar
 
Join Date: Jun 2018
Location: Switzerland
Posts: 1,208
Mentioned: 220 Post(s)
Tagged: 0 Thread(s)
Quoted: 928 Post(s)
Liked: 1323
Quote:
Originally Posted by Soulnight View Post
Which part of the image has for you less strength ?
The armors of course, the hands of the two people in the foreground, and almost everything in the right side including the big mechanical thing and the guy.

"high contrast" is probably a better description in comparison to L-80. It certainly does not have more depth than L-80, at least here, to me.

It also depends on how HSTM will be named for new users...
Neo-XP is offline  
post #7690 of 9401 Old 11-06-2019, 03:15 PM
AVS Forum Special Member
 
Manni01's Avatar
 
Join Date: Sep 2008
Posts: 9,747
Mentioned: 384 Post(s)
Tagged: 0 Thread(s)
Quoted: 5925 Post(s)
Liked: 6507
Quote:
Originally Posted by Neo-XP View Post
The main issue I have with linear (any version) is what it does on highlights (especially on faces, but not only):

[snip]

This is with [email protected] and highlight recovery "high", it is worse without highlight recovery ofc.

I could not live with linear for this reason alone, but I also don't like what it does on frames like this:

Hill / L-80 (Manni01's version)



Thanks! I will try it.
Thanks for taking the time to do this. Timecodes?

Quote:
Originally Posted by Neo-XP View Post
Thanks for your vote! I will fight my best for Hill to stay
I don't think there is anyrisk of Hill not to stay and if there was a chance for this to happen you'd get my vote and many others. By the way, I haven't voted for the hill curve yet so please add my vote now

Quote:
Originally Posted by stevenjw View Post
A rose by any other name would smell as sweet.

I lovL-Diff as you called it is simple L-63 +17 (80-105-130-155-180-205-230-255-280-330-380-480) and not far off from L-80, so a decent compromise between the two.

If I had to name it, it would be LV-426.
I love the name but don't you fear that some might find it a bit alienating?
stevenjw, woofer and Broags like this.

Batch Utility V4.02 May 16 2019 to automate measurements files for madVR with support for BD Folders
JVC Macro feature on Vertex/Vertex2/Integral2/Maestro/Diva
Manni01 is offline  
post #7691 of 9401 Old 11-06-2019, 04:10 PM
AVS Forum Special Member
 
Neo-XP's Avatar
 
Join Date: Jun 2018
Location: Switzerland
Posts: 1,208
Mentioned: 220 Post(s)
Tagged: 0 Thread(s)
Quoted: 928 Post(s)
Liked: 1323
Quote:
Originally Posted by Manni01 View Post
Thanks for taking the time to do this. Timecodes?
I didn't note the timecodes, sorry. I just did this in 15 minutes, randomly jumping from scenes to scenes. But it is visible for me and disturbing on every content with bright highlights on faces.

A couple more from Lucy (with frames number this time ^^):

Log / L-80 (Manni's version)


Lucy - Frame 6524


Lucy - Frame 80814

Quote:
Originally Posted by Manni01 View Post
I don't think there is anyrisk of Hill not to stay and if there was a chance for this to happen you'd get my vote and many others. By the way, I haven't voted for the hill curve yet so please add my vote now
Thanks. I also add my vote for L-80 (maybe one day I will have a projector, who knows ).
Manni01 likes this.

Last edited by Neo-XP; 11-07-2019 at 11:57 AM.
Neo-XP is offline  
post #7692 of 9401 Old 11-06-2019, 04:45 PM
AVS Forum Special Member
 
SamuriHL's Avatar
 
Join Date: Jan 2005
Posts: 5,340
Mentioned: 12 Post(s)
Tagged: 0 Thread(s)
Quoted: 321 Post(s)
Liked: 410
I don't think I officially registered my secondary vote for L80, as well. But here I am, registering my secondary vote for L80. Still loving the Hill!
Manni01 likes this.

LG C8 OLED | Pioneer VSX-LX303 | HTPC (9900k, RTX 2070) | nVidia SHIELD 2019 | Panasonic UB820
SamuriHL is online now  
post #7693 of 9401 Old 11-06-2019, 05:56 PM
Member
 
jasonwc07's Avatar
 
Join Date: Jan 2010
Posts: 173
Mentioned: 3 Post(s)
Tagged: 0 Thread(s)
Quoted: 96 Post(s)
Liked: 151
Quote:
Originally Posted by Killroy View Post
No compromise items checked? Even on my RTX 2070 I have to check a few to get rendering times in the high 30's. If I leave it in the default state (first 7 or 8 checked) I can get them down to high 20's to mid 30's. The last few betas completely taxed the RTX 2070 to the point that I have been waiting to see a performance boost before I can start using it again.
Quote:
Originally Posted by Manni01 View Post
As I said, no compromise. There were performance issues with builds 104-106 due to a bug, but this was fixed in 107b. No performance issue since.

If you still have performance issues with 107b or 108, you need to debug your performance issues in a support thread, this isn’t the place for it.

Look at the obvious, like leaving the GPU performance settings in optimal instead of adaptive or maximum, using d3d11 native if you can, dropping NGU chroma to medium, using ordered dithering instead of error diffusion, disabling all filters and enhancements (unnecessary in UHD), etc.

Of course this means using profiles so that you have different settings for UHD 23p and lower res or higher refresh rates. I’m only talking about settings for UHD 23p, with a custom res (CRU) for exact 23.976 frame rate.

However a 2070 might not be powerful enough. I’m talking about a 1080ti. I would expect a 2080 to be fine, a 2070 might need some compromise.
With a GTX 1070 SC, I see 25ms rendering times with "Compromise on HDR luminance channel quality" checked (default setting; presumably for the performance impact) and 34 ms without this setting - using the latest 110 beta. I have every other trade-off disabled except for "don't render frames when fade/in out is detected" (which caused framedrops when disabled). I use ordered dithering (10 bit output set in madVR, 12 bit in NVIDIA panel), Chroma upscaling with NGU AA, medium, 3DLUT calibration for BT.2020, and have all artifact removal options disabled except reduce banding artifacts (low for default, high for fade in/out). An obvious culprit for poor performance is the "Reduce compression artifacts" setting. However, if you enable "activate only if it comes for free (as part of NGU sharp)", it will automatically be disabled for 2160p content. This setting has a significant impact on performance, even on my 1080Ti desktop system, with little gain for UHD BDs. I've posted my HDR tonemapping settings below. I also saw no benefit to Error Diffusion when outputting 10 bit, so this is another area to save performance.

TLDR: For those having performance issues:
- Try enabling "Compromise on HDR luminance channel quality" for a huge performance boost
- Disable "Reduce Compression artifacts for 2160p playback by setting the "activate only if it comes for free" option
- Try Ordered Dithering rather than Error Diffusion

madshi,

How significant is the quality impact of the "Compromise on HDR luminance channel quality" setting? I notice it is enabled by default and it has a relatively huge impact on performance, relative to other settings.

EDIT: It appears Neo-XP answered the same question earlier in this thread. I'm posting his response here in case others have the same question I did. The difference is certainly apparent in these examples, but dynamic tonemapping is a night-and-day difference versus static tonemapping. So, for those that can't use dynamic tonemapping without this setting, it's worth the compromise.

Quote:
Originally Posted by Neo-XP View Post
@mkohman Hi,

Yes and no: it eats up a lot of power, but also makes a lot of difference on scenes with fires/explosions, lights, lasers, etc.

Some random examples:

No compromise / Compromise on HDR luminance channel quality







If you have enough power, go for it
Attached Thumbnails
Click image for larger version

Name:	hdr.PNG
Views:	56
Size:	107.2 KB
ID:	2637374   Click image for larger version

Name:	trade quality.PNG
Views:	49
Size:	97.4 KB
ID:	2637376   Click image for larger version

Name:	dithering.PNG
Views:	45
Size:	89.3 KB
ID:	2637378   Click image for larger version

Name:	chroma.PNG
Views:	47
Size:	90.9 KB
ID:	2637380   Click image for larger version

Name:	artifact removal.PNG
Views:	45
Size:	89.0 KB
ID:	2637382  

xton and clark17 like this.

Last edited by jasonwc07; 11-06-2019 at 08:00 PM.
jasonwc07 is offline  
post #7694 of 9401 Old 11-06-2019, 07:57 PM
Advanced Member
 
CoryW's Avatar
 
Join Date: Aug 2001
Location: Austin, TX, USA
Posts: 727
Mentioned: 12 Post(s)
Tagged: 1 Thread(s)
Quoted: 206 Post(s)
Liked: 195
Quote:
Originally Posted by Neo-XP View Post
The main issue I have with linear (any version) is what it does on highlights (especially on faces, but not only):

I could not live with linear for this reason alone, but I also don't like what it does on frames like this:

Hill / L-80 (Manni01's version)


Well, I agree; I like his face better with Hill. Looks more 3D / better depth.

However, look at this example also in The Meg, where the clouds really lose detail with Hill vs. L80:

Hill at DT75:




L80 at DT75:



Is there a way to balance the 2? More depth for faces with Hill, yet still maintain the bright detail that L80 presents?

*If I change the strength for both to 65, it's very noticeable:

Hill at HSTM 65 strength:




L80 at HSTM 65 strength:

Manni01, Broags and Neo-XP like this.

Thanks!
Cory
Audio: Ascend Acoustics, Sennheiser HD 598, Denon AVR-S710W
Video: JVC RS400, 92" DES Abyss screen, HTPC/RTX 2080/i7 8700k 4.9GHz/madVR, Panasonic DMP-UB900 for streaming

Last edited by CoryW; 11-06-2019 at 08:03 PM.
CoryW is offline  
post #7695 of 9401 Old 11-06-2019, 10:03 PM
Member
 
RioBar4U's Avatar
 
Join Date: Dec 2011
Location: Switzerland
Posts: 90
Mentioned: 6 Post(s)
Tagged: 0 Thread(s)
Quoted: 27 Post(s)
Liked: 39
Quote:
Originally Posted by Neo-XP View Post
The log curve is sexy too:

Hill (100-175-225-261-288-308-325-339-350-368-381-400)



Log (100-175-231-264-287-305-320-333-344-362-377-400)


It is very slightly less detrimental on highlights than Hill, with the most extreme content only. No other visible changes whatsoever.

I don't think anyone will ever see any difference anyway. I had to pause + take screenshots + zoom on specific parts of both images + switch from one to the other to see the 0.1% improvement

But any improvement is good to take I guess
To be honest, I used a couple of minutes to find a scenne where Icould see a difference. But in these rare moments I like Log more than Hill. So congrats!
Manni01 and Neo-XP like this.
RioBar4U is offline  
post #7696 of 9401 Old 11-06-2019, 10:22 PM
Member
 
Broags's Avatar
 
Join Date: Nov 2010
Location: Adelaide
Posts: 95
Mentioned: 9 Post(s)
Tagged: 0 Thread(s)
Quoted: 49 Post(s)
Liked: 81
Quote:
Originally Posted by stevenjw View Post
A rose by any other name would smell as sweet..... If I had to name it, it would be LV-426.




With my behavioral inhibitor, it is impossible for me to behave


edit: @Neo-XP , do we need a new comparison chart - anything added/removed?
stevenjw, Manni01 and Neo-XP like this.

`
curve_gen_and_analyse_1.2_compare : http://www.mediafire.com/file/6cxixx...pare.xlsx/file ( thanks to 3ll3d00d for the heavy lifting )

Last edited by Broags; 11-06-2019 at 10:25 PM.
Broags is offline  
post #7697 of 9401 Old 11-07-2019, 12:37 AM
AVS Forum Special Member
 
Neo-XP's Avatar
 
Join Date: Jun 2018
Location: Switzerland
Posts: 1,208
Mentioned: 220 Post(s)
Tagged: 0 Thread(s)
Quoted: 928 Post(s)
Liked: 1323
Quote:
Originally Posted by CoryW View Post
However, look at this example also in The Meg, where the clouds really lose detail with Hill vs. L80
Quote:
Originally Posted by CoryW View Post
Is there a way to balance the 2? More depth for faces with Hill, yet still maintain the bright detail that L80 presents?
Yes, I guess if you use more strength for highlight recovery, you can have both with Hill/Log:

HSTM Off


The Meg - Frame 8663


HSTM On - Log / L-80




HSTM On - Log + highlight recovery "are you nuts!?" / L-80 + highlight recovery "are you nuts!?"



You don't have to go that high with highlight recovery strength though, "very high" works well too (and looks more natural IMO).

Quote:
Originally Posted by Broags View Post
@Neo-XP , do we need a new comparison chart - anything added/removed?
Yes, you can add the Log one, my new favorite

Log: 100-175-231-264-287-305-320-333-344-362-377-400
Manni01 likes this.

Last edited by Neo-XP; 11-08-2019 at 01:59 AM.
Neo-XP is offline  
post #7698 of 9401 Old 11-07-2019, 01:03 AM
Member
 
Broags's Avatar
 
Join Date: Nov 2010
Location: Adelaide
Posts: 95
Mentioned: 9 Post(s)
Tagged: 0 Thread(s)
Quoted: 49 Post(s)
Liked: 81
No problem Neo,

Comparison Chart:


`
curve_gen_and_analyse_1.2_compare : http://www.mediafire.com/file/6cxixx...pare.xlsx/file ( thanks to 3ll3d00d for the heavy lifting )
Broags is offline  
post #7699 of 9401 Old 11-07-2019, 01:49 AM
AVS Forum Special Member
 
Manni01's Avatar
 
Join Date: Sep 2008
Posts: 9,747
Mentioned: 384 Post(s)
Tagged: 0 Thread(s)
Quoted: 5925 Post(s)
Liked: 6507
Quote:
Originally Posted by Neo-XP View Post
Yes, I guess if you use more strength for highlight recovery, you can have both with Hill/Log:

You don't have to go that high with highlight recovery strength though, "very high" works well too (and looks more natural IMO).
Agreed, this is precisely why I have "very high" for highlights recovery in my recommended settings for L80.

I experimented a bit with "are you nuts?" for a while and then backed down to very high a while ago.

I don't know how many people test curves without testing them with their recommended settings as well. I know you do test curve+recommended settings, but it would be interesting for those who hesitate (or wonder) to try the following procedure:

When you post your settings with a new curve, I try all of them (including peak brightness and DT, so all your settings exactly as posted) and compare with a "test" profile that has my favourite curve using all your settings (including peak brightness and DT). That way I'm sure I'm comparing the curves, and nothing else, and I see your curve at work as close as possible to the way you see it, though on a projector with slighter less brightness than your OLED (110 vs 150nits) and a larger screen (88" vs 55"). The picture is a bit too dark, but not that much and completely watchable, especially with your DT which is usually significantly lower than mine.

Then I do the opposite: I compare your curve with "adapted settings" to match my peak brightness and preferred settings, to my normal settings with my preferred curve of the moment.

This is just to make sure that I don't miss a discovery because my usual settings are wrong for the curve you suggest.

Doing it that way, I frequently changed my "current favourite" to your new one, because I could see the improvements. Then I would fine-tune my settings to make the most of it here.

Of course, you need to use profiles to do this, and be able to change instantly from one set of settings to the other, with a moving or still picture.

Otherwise, given the relatively small differences between L-80 and hill on most content, it's difficult most of the time to spot differences (even when present), unless you test specific frames / shots and know what to look for.
SamuriHL, xton, LTD_666 and 3 others like this.

Batch Utility V4.02 May 16 2019 to automate measurements files for madVR with support for BD Folders
JVC Macro feature on Vertex/Vertex2/Integral2/Maestro/Diva

Last edited by Manni01; 11-07-2019 at 01:54 AM.
Manni01 is offline  
post #7700 of 9401 Old 11-07-2019, 01:58 AM
AVS Forum Special Member
 
Manni01's Avatar
 
Join Date: Sep 2008
Posts: 9,747
Mentioned: 384 Post(s)
Tagged: 0 Thread(s)
Quoted: 5925 Post(s)
Liked: 6507
Quote:
Originally Posted by Broags View Post
No problem Neo,

Comparison Chart:
Given the minute differences, I'm sure all the votes for your latest Hill can automatically be transferred to "Log One", so that Neo-XP can have his preferred one, unless someone strongly objects. I would need to see comparison screenshots though to believe that such an objection is valid.
SamuriHL, Broags and Neo-XP like this.

Batch Utility V4.02 May 16 2019 to automate measurements files for madVR with support for BD Folders
JVC Macro feature on Vertex/Vertex2/Integral2/Maestro/Diva
Manni01 is offline  
post #7701 of 9401 Old 11-07-2019, 02:33 AM
Advanced Member
 
chros73's Avatar
 
Join Date: Jan 2015
Posts: 780
Mentioned: 21 Post(s)
Tagged: 0 Thread(s)
Quoted: 489 Post(s)
Liked: 243
Quote:
Originally Posted by Soulnight View Post
In any case "going too far" is a matter of taste as your poll as proven.
Not just that, Mirror can produce problem:

Quote:
Originally Posted by chros73 View Post
Mirror: it produced the same black/white images e.g. @ 00:15 Pacific Rim (link above) with the attached settings. Probably the starting value is too low? Same happened with the one that started with 0.
Did anyone try to reproduce it? (settings is attached to that post)

Ryzen 5 2600,Asus Prime b450-Plus,16GB,MSI GTX 1060 Gaming X 6GB(v385.28),Win10 LTSB 1607,MPC-BEx64+LAV+MadVR,Yamaha RX-A870,LG OLED65B8(04.10.25+PC4:4:[email protected]/24/25/29/30/50/59/60Hz)
chros73 is offline  
post #7702 of 9401 Old 11-07-2019, 02:35 AM
AVS Forum Special Member
 
Manni01's Avatar
 
Join Date: Sep 2008
Posts: 9,747
Mentioned: 384 Post(s)
Tagged: 0 Thread(s)
Quoted: 5925 Post(s)
Liked: 6507
Quote:
Originally Posted by chros73 View Post
Not just that, Mirror can produce problem:


Did anyone try to reproduce it? (settings is attached to that post)
Link to the post?

Batch Utility V4.02 May 16 2019 to automate measurements files for madVR with support for BD Folders
JVC Macro feature on Vertex/Vertex2/Integral2/Maestro/Diva
Manni01 is offline  
post #7703 of 9401 Old 11-07-2019, 03:03 AM
AVS Forum Special Member
 
Neo-XP's Avatar
 
Join Date: Jun 2018
Location: Switzerland
Posts: 1,208
Mentioned: 220 Post(s)
Tagged: 0 Thread(s)
Quoted: 928 Post(s)
Liked: 1323
Quote:
Originally Posted by Manni01 View Post
Link to the post?
This one I guess:

Quote:
Originally Posted by chros73 View Post
Here is a 1 minute sample (along with others): filename starts with "10-*pr.mkv"
Trying to watch some Pacific Rim scenes with Mirror, I have another problem: strong flickering on this sequence (when the camera is on Charlie Hunnam): https://www.mediafire.com/file/0fcgz...issue.mkv/file

Pacific Rim, frames 139837 to 140563. No problem (that I could see) with Log or L-80.

Can someone please double check and confirm?
Manni01 and chros73 like this.

Last edited by Neo-XP; 11-07-2019 at 03:11 AM.
Neo-XP is offline  
post #7704 of 9401 Old 11-07-2019, 03:06 AM
Member
 
Broags's Avatar
 
Join Date: Nov 2010
Location: Adelaide
Posts: 95
Mentioned: 9 Post(s)
Tagged: 0 Thread(s)
Quoted: 49 Post(s)
Liked: 81
Latest chart :

Manni01 and *Mori* like this.

`
curve_gen_and_analyse_1.2_compare : http://www.mediafire.com/file/6cxixx...pare.xlsx/file ( thanks to 3ll3d00d for the heavy lifting )
Broags is offline  
post #7705 of 9401 Old 11-07-2019, 03:07 AM
Advanced Member
 
chros73's Avatar
 
Join Date: Jan 2015
Posts: 780
Mentioned: 21 Post(s)
Tagged: 0 Thread(s)
Quoted: 489 Post(s)
Liked: 243
Quote:
Originally Posted by Manni01 View Post
Link to the post?
There's a link button beside of the username when you quote.

Ryzen 5 2600,Asus Prime b450-Plus,16GB,MSI GTX 1060 Gaming X 6GB(v385.28),Win10 LTSB 1607,MPC-BEx64+LAV+MadVR,Yamaha RX-A870,LG OLED65B8(04.10.25+PC4:4:[email protected]/24/25/29/30/50/59/60Hz)
chros73 is offline  
post #7706 of 9401 Old 11-07-2019, 03:56 AM
Senior Member
 
*Mori*'s Avatar
 
Join Date: Feb 2009
Posts: 366
Mentioned: 25 Post(s)
Tagged: 0 Thread(s)
Quoted: 380 Post(s)
Liked: 576
Quote:
Originally Posted by Neo-XP View Post
Yes, you can add the Log one, my new favorite

Log: 100-175-231-264-287-305-320-333-344-362-377-400
@Neo-XP
That's interesting. I had a Hill with 100-178-227-259-282-301-317-330-341-360-375-400

That looks very similar to your Log. Unfortunately I don't remember why I discarded it. You might have a look at that as well. If there is a difference in the pictures you might deduce something out of it.
Manni01 and Neo-XP like this.
*Mori* is offline  
post #7707 of 9401 Old 11-07-2019, 04:10 AM
Advanced Member
 
CoryW's Avatar
 
Join Date: Aug 2001
Location: Austin, TX, USA
Posts: 727
Mentioned: 12 Post(s)
Tagged: 1 Thread(s)
Quoted: 206 Post(s)
Liked: 195
Quote:
Originally Posted by Neo-XP View Post
Yes, I guess if you use more strength for highlight recovery, you can have both with Hill/Log:

You don't have to go that high with highlight recovery strength though, "very high" works well too (and looks more natural IMO).
Well, that's very strange. I was already using "are you nuts?", and the difference is showing much stronger for me. Will you please post your current settings? Here are mine:

Thanks!
Cory
Audio: Ascend Acoustics, Sennheiser HD 598, Denon AVR-S710W
Video: JVC RS400, 92" DES Abyss screen, HTPC/RTX 2080/i7 8700k 4.9GHz/madVR, Panasonic DMP-UB900 for streaming
CoryW is offline  
post #7708 of 9401 Old 11-07-2019, 04:50 AM
AVS Forum Special Member
 
Neo-XP's Avatar
 
Join Date: Jun 2018
Location: Switzerland
Posts: 1,208
Mentioned: 220 Post(s)
Tagged: 0 Thread(s)
Quoted: 928 Post(s)
Liked: 1323
Quote:
Originally Posted by CoryW View Post
Well, that's very strange. I was already using "are you nuts?", and the difference is showing much stronger for me. Will you please post your current settings?
I am using these atm:



+ everything unchecked on the "trade quality..." page.

If the "are you nuts!?" strength is not enough for you, you can ask madshi to add an even higher strength to get more depth on the clouds (it is totally possible in theory to go higher). For instance, "are you nuts!?" would become "extreme" and the higher strength would become "insane".

Also, maybe the "don't desaturate" option makes you lose some details, it adds a weird yellow tint on my side:

Don't desaturate / Desaturate 1



Quote:
Originally Posted by *Mori* View Post
@Neo-XP
That's interesting. I had a Hill with 100-178-227-259-282-301-317-330-341-360-375-400

That looks very similar to your Log. Unfortunately I don't remember why I discarded it. You might have a look at that as well. If there is a difference in the pictures you might deduce something out of it.
The numbers are extremely close. I will check, but I doubt I will find the slightest difference between the two.
Manni01, *Mori* and neo_2009 like this.

Last edited by Neo-XP; 11-07-2019 at 04:54 AM.
Neo-XP is offline  
post #7709 of 9401 Old 11-07-2019, 05:01 AM
Senior Member
 
*Mori*'s Avatar
 
Join Date: Feb 2009
Posts: 366
Mentioned: 25 Post(s)
Tagged: 0 Thread(s)
Quoted: 380 Post(s)
Liked: 576
@Neo-XP

Yes, likely too close. FYI: The numbers were derived from a midpoint x=2 and a Hill coefficient of n=0.5. My first good suggestion was x=2 and n=1.
*Mori* is offline  
post #7710 of 9401 Old 11-07-2019, 05:15 AM
Newbie
 
Join Date: Nov 2019
Posts: 1
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 0
Hi all,

Great topic here, i'm newbie here and i would like to ask some 'basic' question.
I'm using this setup : PC > AV Ampllifier > Xiaomi 4K UST Projector + ALR Screen Gain 0.8

Xiaomi is HDR compliant, but results are not very nice when using HDR passtrough. I'm trying to 'play' with tone mapping capabilities of Madvr.
My questions are :
- Should i activate 'Output video in HDR Format' ?
- How can i 'manually' guess 'Real Nit' values to fill in the settings (I have no calibration tool) ? I have a 100' inch screen, VP max Lumens is 2500, let's say that i'm around 1700 (Cinema mode), it is said that nits should be set to 700...sounds weird to me, because, if i see people settings, 'real' nits value are around 100/150...
What are 'common' values for a projector ?

Do we have any Wiki / tuto who just explained 'Where to start' ?

Thanks for your reply !
khordo595 is offline  
Sponsored Links
Advertisement
 
Reply Digital Hi-End Projectors - $3,000+ USD MSRP

Tags
dynamic tone mapping , hdr , madvr , sdr , ton mapping

Thread Tools
Show Printable Version Show Printable Version
Email this Page Email this Page


Forum Jump: 

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
Trackbacks are Off
Pingbacks are Off
Refbacks are Off