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 4th January 2017, 16:41   #41841  |  Link
Asmodian
Registered User
 
Join Date: Feb 2002
Location: San Jose, California
Posts: 4,406
Out of those, 23. 23 is actually about 23.976 Hz.

Not MHz, that is a million times faster than the refresh rate.
__________________
madVR options explained
Asmodian is offline   Reply With Quote
Old 4th January 2017, 16:42   #41842  |  Link
huhn
Registered User
 
Join Date: Oct 2012
Posts: 7,903
with reclock it doesn't really matter if you are using 24p or 23p.

reclock default setting should be fine.

23p is 24000/1001 same counts for 29p and 59p.
huhn is offline   Reply With Quote
Old 4th January 2017, 16:47   #41843  |  Link
Oguignant
Registered User
 
Oguignant's Avatar
 
Join Date: Nov 2016
Posts: 181
Quote:
Originally Posted by Asmodian View Post
Out of those, 23. 23 is actually about 23.976 Hz.

Not MHz, that is a million times faster than the refresh rate.
jaja yes, Hz... and reclock is necesary?
__________________
"To infinity, and beyond!"
Oguignant is offline   Reply With Quote
Old 4th January 2017, 16:57   #41844  |  Link
sneaker_ger
Registered User
 
Join Date: Dec 2002
Posts: 5,565
Necessary? No. If you have frame repeats/drops because of a slight fps mismatch it can fix those. But even without ReClock you may not have any or only very few (e.g. 1 every hour) and you probably wouldn't notice anyways.
sneaker_ger is offline   Reply With Quote
Old 4th January 2017, 17:03   #41845  |  Link
oddball
Registered User
 
Join Date: Jan 2002
Posts: 1,264
Quote:
Originally Posted by CruNcher View Post
Huh is it losing Hardware acceleration does it accelerate it at all ?

Is it only this 4K HEVC Demo that collapses with your scaling chain in MadVR on your RX480 ?

But yeah it's the most complex one currently out of Atemes Titan



I actually wonder how Samsungs Hardware Decoder handles it

Would be nice if you could push your RX480 Decoding result of it (without MadVR overhead) in the HEVC Decoder thread we have no information on that one yet.
I will try and sort something

Here ya go.

http://forum.doom9.org/showthread.ph...68#post1791968

Last edited by oddball; 4th January 2017 at 17:28.
oddball is offline   Reply With Quote
Old 4th January 2017, 17:08   #41846  |  Link
Oguignant
Registered User
 
Oguignant's Avatar
 
Join Date: Nov 2016
Posts: 181
Quote:
Originally Posted by sneaker_ger View Post
Necessary? No. If you have frame repeats/drops because of a slight fps mismatch it can fix those. But even without ReClock you may not have any or only very few (e.g. 1 every hour) and you probably wouldn't notice anyways.
Using 23 hz and de-judder option looks amazing. Much better than with svp.
__________________
"To infinity, and beyond!"
Oguignant is offline   Reply With Quote
Old 4th January 2017, 17:21   #41847  |  Link
Oguignant
Registered User
 
Oguignant's Avatar
 
Join Date: Nov 2016
Posts: 181
Quote:
Originally Posted by sneaker_ger View Post
Necessary? No. If you have frame repeats/drops because of a slight fps mismatch it can fix those. But even without ReClock you may not have any or only very few (e.g. 1 every hour) and you probably wouldn't notice anyways.
23 hz is the same as 24p of the tv?
__________________
"To infinity, and beyond!"
Oguignant is offline   Reply With Quote
Old 4th January 2017, 21:02   #41848  |  Link
70MM
X Cinema Projectionist NZ
 
Join Date: Feb 2006
Location: Auckland NZ
Posts: 310
Does using de banding, as it softens the banding, can it also soften the image?
Do most use it or not?
What are the most preferred levels?
70MM is offline   Reply With Quote
Old 4th January 2017, 21:09   #41849  |  Link
baii
Registered User
 
Join Date: Dec 2011
Posts: 180
Quote:
Originally Posted by huhn View Post
if you want to scale to UHD you will need to buy a 1050 ti i don't see a problem with the 1050 for FHD.

you need more than 2 Gb of Vram for UHD.
Have a Rx460, don't seem to have much problem with ngu 1080p to 4k.

By the way, was there a update in madvr or dispcal that made 3dlut calibration "cheaper" to run?
baii is offline   Reply With Quote
Old 4th January 2017, 21:12   #41850  |  Link
Asmodian
Registered User
 
Join Date: Feb 2002
Location: San Jose, California
Posts: 4,406
Debanding on low does very little softening of the image, I use low/low by default on SD material and med/med if there is still banding with low/low. Only high has too much smoothing to use all the time (high is only useful for fades). I usually don't run debanding at a higher strength for fades because fade detection triggers on more often then I would like and I don't have an issue with banding during fades.

Quote:
Originally Posted by Oguignant View Post
23 hz is the same as 24p of the tv?
Sometimes, the TV will probably report 24p when in both 23Hz and 24Hz modes, but with Windows 24 Hz is usually 24.000 while 23 Hz is 23.976.
__________________
madVR options explained
Asmodian is offline   Reply With Quote
Old 4th January 2017, 21:23   #41851  |  Link
JackCY
Registered User
 
Join Date: Jun 2015
Posts: 15
Quote:
Originally Posted by Asmodian View Post
While it makes no real difference I like:
Code:
if      (deintFps < 26) and (srcWidth <= 960) and (srcHeight <= 540)  "540p25"
else if (deintFps < 61) and (srcWidth <= 960) and (srcHeight <= 540)  "540p60" 
else if (deintFps < 26) and (srcWidth <= 1280) and (srcHeight <= 720)  "720p25"
else if (deintFps < 61) and (srcWidth <= 1280) and (srcHeight <= 720)  "720p60"
else if (deintFps < 26) and (srcWidth <= 1920) and (srcHeight <= 1088) "1080p25"
else if (deintFps < 61) and (srcWidth <= 1920) and (srcHeight <= 1088) "1080p60"
else if (deintFps < 26) and (srcWidth <= 4096) and (srcHeight <= 2160) "2160p25"
else if (deintFps < 61) and (srcWidth <= 4096) and (srcHeight <= 2160) "2160p60"
Sometimes things I watch are very cropped so I want to use both dimensions when picking a profile. Resolutions fit together more then frame rates for me... not sure why.

I also tend to add an else "fast" at the end. If madVR hits the end of a profile script without picking a profile it uses the first profile in the settings menu, which is sometimes not what you want with a video that deinterlaces to 120 fps.
Exactly, most of these sample scripts here miss the ELSE part for sources that do not match any of your profiles.
Personally I see no need to have many many profiles, for common displays size it can all be done with 3 max: SD, HD+FHD, highres+highFPS.

As such I use "OR" not "AND" and ">" not "<" or "<=". Also using 1920, 1080 or other specific numbers is a bad idea since many videos are not 1920 but 1916, not 1080 but 1072 or 1200 etc. So I use dimensions that are in between the common resolutions.

But if you like managing 10 profiles with the same settings and clicking endlessly for a single change that affects many profiles, have fun and make 20 profiles of which 10 have the same settings most likely.

---

NGU... an interesting blend but NNEDI3 seems more efficient even when used with SuperRes and gives less unnatural image. Sure sometimes one looks "better" sometimes the other. NGU decent on quality source but can be a little too "vectorizing" on low resolution low quality source where the sudden vector like edges can look out of place. Nice addition though.
Less flexibility with the new GUI options and the scalers to be used after doubling are still not being respected 100% of the time. Try source 4:3 that gets displayed as 16:9, madVR still happily uses Bicubic for scaling one of the luma axis and chroma when everything is set to Jinc and SSIM it simply ignores it as in older versions. The new options only work with sources that have equal AR for the source and display.

Don't remove superXBR, it is a decent scaler that runs fast and can be useful especially on the very common machines that do not run an expensive enthusiast class GPU.

Quality wise, I guess only waifu2x or similar can do anything better than NNEDI3 but then running waifu2x realtime is a no go XD

Last edited by JackCY; 4th January 2017 at 23:44.
JackCY is offline   Reply With Quote
Old 4th January 2017, 21:28   #41852  |  Link
Neo-XP
Registered User
 
Neo-XP's Avatar
 
Join Date: Mar 2016
Location: Switzerland
Posts: 140
Quote:
Originally Posted by 70MM View Post
Does using de banding, as it softens the banding, can it also soften the image?
Do most use it or not?
What are the most preferred levels?
It only really soften the image if you use it on high IMO.
Its use is recommended, because banding is noticeable and distracting, and the debanding process is very effective and doesn't have any negative effect on a image without banding.

For films : I recommend medium for the default strength and high for the strength during fade in/out.
If you want to compare with a source with a lot of banding (off/low/medium/high) :



I notice that there is still banding with the low value and the high value is too strong (you lose some details).
I don't know the best values for anime (I don't watch anime), but I guess high for both.

Last edited by Neo-XP; 4th January 2017 at 21:41.
Neo-XP is offline   Reply With Quote
Old 4th January 2017, 21:42   #41853  |  Link
Clemens
Registered User
 
Join Date: Sep 2016
Posts: 11
Problem with HEVC 10 bit color

When I use my DVBViewer (or MPC-HC with records) in combination with the latest LAV Decoder and MadVR, HEVC 10 bit UHD channels on Astra 19.2 (HDR and non HDR) get wrong colors, e.g. the skin color is a kind of blue instead of normal. It happens on all HEVC UHD channels.

Is this a known issue? With other renderers (e.g. the EVR custom renderer of the DVBViewer) the colors are fine (but the image isn't as nice as with MadVR beside the color).
My GPU is a GTX 960

Here an image (sorry for the content , a sample video I will update later (not at home right now)).

I read in the bug tracker about that magenta bug...maybe it's something similar? I guess you can say more when you have my sample, it will come next friday if I don't forget about it. (the image I made a short while ago for explaining someone how to configure MPC-HC (for streaming of a Vu+ and using the hardware decoder eihch VLC didn't do for HEVC) and I told him to use MadVR, but not for UHD yet because of this bug)

Kind regards,
Clemens
Attached Images
 

Last edited by Clemens; 4th January 2017 at 21:54.
Clemens is offline   Reply With Quote
Old 4th January 2017, 21:50   #41854  |  Link
sneaker_ger
Registered User
 
Join Date: Dec 2002
Posts: 5,565
Press ctrl+j to see what matrix/transfer/primaries madvr is using and where that info comes from. Press the appropriate hotkeys to adjust if necessary. (Or file name tagging. Not sure if profiles also allow adjusting.)
sneaker_ger is offline   Reply With Quote
Old 4th January 2017, 23:10   #41855  |  Link
Asmodian
Registered User
 
Join Date: Feb 2002
Location: San Jose, California
Posts: 4,406
Quote:
Originally Posted by JackCY View Post
As such I use "OR" not "AND" and ">" not "<" or "<=". Also using 1920, 1080 or other specific numbers is a bad idea since many videos are not 1920 but 1916, not 1080 but 1072 or 1200 etc. So I use dimensions that are in between the common resolutions.

But if you like managing 10 profiles with the same settings and clicking endlessly for a single change that affects many profiles, have fun and make 20 profiles of which 10 have the same settings most likely.
The way I wrote the profile rules there is nothing wrong with using 1920 or 1080, etc. 1916 would be less than or equal to 1920 but more than 1280, exactly as we want.

How do you start your rules then? You cannot use "OR" instead when using "<=". I see no reason to use slightly off values for 1920, in fact wouldn't that cause problems?

Code:
if      (srcWidth > 1024) OR (srcHeight > 576) "720p"
else if (srcWidth > 1280) OR (srcHeight > 720) "1080p"
else if (srcWidth > 1920) OR (srcHeight > 1088) "UHD"
else "SD"
__________________
madVR options explained
Asmodian is offline   Reply With Quote
Old 5th January 2017, 01:48   #41856  |  Link
austonrush
Registered User
 
Join Date: Oct 2015
Posts: 33
I've seen a few post on here in regards to profiling. I'll post a copy of my scaling profile auto select rules. Back when I was using NNEDI3 for everything I needed different profiles for 60fps for performance reasons. Many of you who don't have the highest end cards will be in the same boat even with ngu. My profile has 6 scaling settings (I do not have a 4k TV or I would have 8).

Anyone who wants to use this feel free. It works perfectly

if (srcHeight <= 711) and (refreshrate < 31) "SD 24fps"
else if (srcHeight <= 711) or (fileName = "*.avi") "SD 60fps"
else if ((fileName = "*720p*") or (fileName = "*1280x720*")) and (refreshrate < 31) "HD 24fps"
else if ((fileName = "*720p*") or (fileName = "*1280x720*")) "HD 60fps"
else if ((fileName = "*1080p*") or (fileName = "*1920x1080*")) and(refreshrate < 31) "FullHD 24fps"
else if ((fileName = "*1080p*") or (fileName = "*1920x1080*")) "FullHD 60fps"
// For release groups that does not tag their files with video resolutions.
else if ((fileName = "*[FBI]*") or (fileName = "*[Vivid]*")) "HD 24fps"
else "FullHD 24fps"
austonrush is offline   Reply With Quote
Old 5th January 2017, 04:12   #41857  |  Link
Georgel
Visual Novel Dev.
 
Georgel's Avatar
 
Join Date: Nov 2015
Location: Bucharest
Posts: 200
Found a free tool that might come in handy for some of us!!!

https://vvvv.org/contribution/windows-system-timer-tool
Georgel is offline   Reply With Quote
Old 5th January 2017, 04:27   #41858  |  Link
70MM
X Cinema Projectionist NZ
 
Join Date: Feb 2006
Location: Auckland NZ
Posts: 310
Quote:
Originally Posted by Neo-XP View Post
It only really soften the image if you use it on high IMO.
Its use is recommended, because banding is noticeable and distracting, and the debanding process is very effective and doesn't have any negative effect on a image without banding.

For films : I recommend medium for the default strength and high for the strength during fade in/out.
If you want to compare with a source with a lot of banding (off/low/medium/high) :



I notice that there is still banding with the low value and the high value is too strong (you lose some details).
I don't know the best values for anime (I don't watch anime), but I guess high for both.
Thank you for that Neo-XP thats very interesting. I have looked at your images up close and large but really I see no banding on any plus they all seem to look just as detailed to me. Could it be the monitor Im using?

Can you point out something to look at up close in your images please so Im on the right track?
70MM is offline   Reply With Quote
Old 5th January 2017, 05:31   #41859  |  Link
ryrynz
Registered User
 
ryrynz's Avatar
 
Join Date: Mar 2009
Posts: 3,646
Quote:
Originally Posted by 70MM View Post
Can you point out something to look at up close in your images please so Im on the right track?
Oh, you're on the right track..

Quote:
Originally Posted by 70MM View Post
I have looked at your images up close and large but really I see no banding on any plus they all seem to look just as detailed to me. Could it be the monitor Im using?
That's because there is no banding to speak of..

These images are for all intents and purposes visually identical, the image is being softened throughout the gradients by such a small amount (One of two increments in R,G,B value) That I can confidently say
that debanding this accomplishes next to nothing.



Last edited by ryrynz; 5th January 2017 at 05:43.
ryrynz is offline   Reply With Quote
Old 5th January 2017, 14:07   #41860  |  Link
Neo-XP
Registered User
 
Neo-XP's Avatar
 
Join Date: Mar 2016
Location: Switzerland
Posts: 140
Quote:
Originally Posted by ryrynz View Post
Oh, you're on the right track..



That's because there is no banding to speak of..

These images are for all intents and purposes visually identical, the image is being softened throughout the gradients by such a small amount (One of two increments in R,G,B value) That I can confidently say
that debanding this accomplishes next to nothing.


You have to open the images in their original size if you want to compare them properly.

Click on the thumbnails on postimg and then you should see the abyssal differences between all of them by just switching from one to another in two tabs, if your screen is correctly calibrated.

You don't need any tool or zoom to see this.

Of course the RGB values are almost identical, debanding only smooth color gradations so you don't notice them.

Last edited by Neo-XP; 5th January 2017 at 14:16.
Neo-XP 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 13:15.


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