(2015-10-29, 00:44)couto27 Wrote: [ -> ]I few years ago i made 2 simple calibration tutorials (in portuguese language ) but you can use some browser translation to english it would be comprehensive the most part.
I have seen those tutorials before. It is comprehensive and very well written. Nice work!
As for the picture quality comparisons, it is really not about color errors. It is the sharpness of the decoded image that is in question. SATV seems to output a softer image. At least, that is what I have noticed visually.
frenshprince may have something else to add.
(2015-10-29, 01:40)wesk05 Wrote: [ -> ] (2015-10-29, 00:44)couto27 Wrote: [ -> ]I few years ago i made 2 simple calibration tutorials (in portuguese language ) but you can use some browser translation to english it would be comprehensive the most part.
I have seen those tutorials before. It is comprehensive and very well written. Nice work!
As for the picture quality comparisons, it is really not about color errors. It is the sharpness of the decoded image that is in question. SATV seems to output a softer image. At least, that is what I have noticed visually. frenshprince may have something else to add.
I switch from an HTPC to the SATV and my first impression after playing a movie on the shield with Kodi was that the image was not as sharp as on the HTPC...
It is not a BIG problem as the SATV has a good image and eventually I get use to it... but not as sharp as maybe it could be.
Since there is good harware in the SATV is there a possibilty that we have more image options and parameters to try to improve or tweak it ? Like MadVR do for Windows...
(2015-10-29, 01:40)wesk05 Wrote: [ -> ] (2015-10-29, 00:44)couto27 Wrote: [ -> ]I few years ago i made 2 simple calibration tutorials (in portuguese language ) but you can use some browser translation to english it would be comprehensive the most part.
I have seen those tutorials before. It is comprehensive and very well written. Nice work!
As for the picture quality comparisons, it is really not about color errors. It is the sharpness of the decoded image that is in question. SATV seems to output a softer image. At least, that is what I have noticed visually. frenshprince may have something else to add.
No I don't, you got it all.
I understand Couto27's point, but it's not relevant in this case.
A good calibration can change perceptual sharpness, but can not display that much difference.
My JVC projector is calibrated with a DisplayPro, and when I switch device, the impact on my Delta is very slight.
And again, Absolute Quality doesn't matter here, since we are comparing two devices in
source direct, with no external variable.
And we are not comparing colors error, but sharpness.
Of course, even the most strict procedure to take photo can not reproduce the reality.
But it's enough if you do it well.
I'm thinking to a new method for comparison :
Capturing via Profesional HDMI card in uncompressed format, with a HDCP stripper.
(2015-10-28, 15:00)tdbegley Wrote: [ -> ]So your mysql database will have a table called MyVideos94 which Jarvis Alpha 1 and 2 will use however Alpha 3 and the newest nightly builds are using MyVideos97 so it will not see the database you were using in the previous builds.
P.Sto use a mysql setup ALL devices must run the same version of kodi.
If you logging to your mysql setup and run the command "show databases;"
You will see numerous MyVideos% and my Music% databases.
When you move onto a full jarvis build and end up using MyVideos97 you can then drop the rest of the databases from mysql unless you ever intend on going backwards to isengard.
You could always use mysql to copy the old database to the new database so that you don't need to scrape it again.
Alpha 1 +2 uses 94 and Alpha 3 & 4 uses 97 so use this command
This is why they always say when using a mysql setup all devices MUST run the same version of Kodi.
Thanks for all the detailed info, I had a suspicion it was MySQL related. However the first an current Jarvis build I have installed was/is Alpha 4, I never had Alpha 1, 2 or 3 prior.
Although I do have Kodi installed on a total of 4 devices, it is really only used on the Shield. I fired up the laptop to check if it was doing the same on another device. The laptop was updated to Jarvis before it had a chance to interact with MySQL. The other two devices have not been turned on since I started using Jarvis on the Shield.
This is what it currently looks like, so I guess I just keep Video97 and Music54.
On the Database table I don't see Video99 and Music55 or Music56 which I have also
I can see my the File section of Video99 the last played shows as Oct 18 when I first noted this issue on here, so I gather when I first installed a later build than the 7th.
I have exported the entire MySQL list, so I guess I can delete all and just keep Video97 and Music54 and should be right going forward for all current and future Jarvis nightly builds?
(2015-10-29, 12:59)Hiphopopotamus Wrote: [ -> ] (2015-10-28, 15:00)tdbegley Wrote: [ -> ]So your mysql database will have a table called MyVideos94 which Jarvis Alpha 1 and 2 will use however Alpha 3 and the newest nightly builds are using MyVideos97 so it will not see the database you were using in the previous builds.
P.Sto use a mysql setup ALL devices must run the same version of kodi.
If you logging to your mysql setup and run the command "show databases;"
You will see numerous MyVideos% and my Music% databases.
When you move onto a full jarvis build and end up using MyVideos97 you can then drop the rest of the databases from mysql unless you ever intend on going backwards to isengard.
You could always use mysql to copy the old database to the new database so that you don't need to scrape it again.
Alpha 1 +2 uses 94 and Alpha 3 & 4 uses 97 so use this command
This is why they always say when using a mysql setup all devices MUST run the same version of Kodi.
Thanks for all the detailed info, I had a suspicion it was MySQL related. However the first an current Jarvis build I have installed was/is Alpha 4, I never had Alpha 1, 2 or 3 prior.
Although I do have Kodi installed on a total of 4 devices, it is really only used on the Shield. I fired up the laptop to check if it was doing the same on another device. The laptop was updated to Jarvis before it had a chance to interact with MySQL. The other two devices have not been turned on since I started using Jarvis on the Shield.
This is what it currently looks like, so I guess I just keep Video97 and Music54.
On the Database table I don't see Video99 and Music55 or Music56 which I have also
I can see my the File section of Video99 the last played shows as Oct 18 when I first noted this issue on here, so I gather when I first installed a later build than the 7th.
I have exported the entire MySQL list, so I guess I can delete all and just keep Video97 and Music54 and should be right going forward for all current and future Jarvis nightly builds?
That is correct. I would bookmark the version page as when it changes next time you will be able to see what its changed to and import the libery to that database and delete the old one again to keep ontop of things.
(2015-10-29, 12:55)frenshprince Wrote: [ -> ]I'm thinking to a new method for comparison :
Capturing via Profesional HDMI card in uncompressed format, with a HDCP stripper.
I was thinking of doing the exact same thing. I have a Blackmagic Design DecLink 4K Extreme 12G card. I was going to capture a movie sequence and then open the file in a video editor to grab the exact same frame. May be, I will do it this weekend.
Just to chime into the picture quality discussion of the Shield......
Our old Patriot PBO box produced noticeably better output for our Samsung JS8500 than the Nvidia Shield.
It turned out that the difference was due to PBO->HDMI/1080p versus Nvidia Shield->HDMI/4k (i.e. driving TV in 1080p versus 4k)
Meaning that PBO was effectively having the JS8500 do the up scaling versus Nvidia Shield/Kodi.
I had to end up buying an HDMI splitter + 1080p HDMI dongle to force the Nvidia Shield to output at 1080p and this resolved the problem.
It was a quite noticeable difference (and hat off to Samsung for their upscaling/shapening algorithms).
Sure that this doesn't explain the difference?
(2015-10-29, 19:07)samboy Wrote: [ -> ]Just to chime into the picture quality discussion of the Shield......
Our old Patriot PBO box produced noticeably better output for our Samsung JS8500 than the Nvidia Shield.
It turned out that the difference was due to PBO->HDMI/1080p versus Nvidia Shield->HDMI/4k (i.e. driving TV in 1080p versus 4k)
Meaning that PBO was effectively having the JS8500 do the up scaling versus Nvidia Shield/Kodi.
I had to end up buying an HDMI splitter + 1080p HDMI dongle to force the Nvidia Shield to output at 1080p and this resolved the problem.
It was a quite noticeable difference (and hat off to Samsung for their upscaling/shapening algorithms).
Sure that this doesn't explain the difference?
That example might mean the Shield is just doing a proper upscale with no processing. That is, making one pixel in 1080p four pixels in "4k". Where as the tv does that plus, as you said, go through a sharpening algorithm. Which would get into the debate of "true" vs "better"?
Ultimately the only way to settle it would be what wesk05 said he may do this weekend. Capture a specific frame directly from the boxes hdmi output and compare. 1 from the Shield, 1 from whatever other device, and 1 from a software decoder we know doesn't do anything but decode the frame.
You can complain all you want about picture quality but there's nothing we can do I think.
Go complain on the nvidia forums.
(2015-10-29, 20:26)Martijn Wrote: [ -> ]You can complain all you want about picture quality but there's nothing we can do I think.
Go complain on the nvidia forums.
That was my question
Thank you.
Allow me one more :
On Shield's version of Kodi, we have only two upscalling (bilinear and nearest), and not so much deinterlacing method.
Is a limitation of Shield's version of Kodi, Android 5.1, or the Shield ?
Thank you
(2015-10-29, 19:19)drhill Wrote: [ -> ]Ultimately the only way to settle it would be what wesk05 said he may do this weekend. Capture a specific frame directly from the boxes hdmi output and compare. 1 from the Shield, 1 from whatever other device, and 1 from a software decoder we know doesn't do anything but decode the frame.
It will be the ultimate way to convince those who have doubts.
But facts are already settled.
Now we have only hopes that Nvida do something about it (if they can).
(2015-10-29, 21:11)frenshprince Wrote: [ -> ] (2015-10-29, 19:19)drhill Wrote: [ -> ]Ultimately the only way to settle it would be what wesk05 said he may do this weekend. Capture a specific frame directly from the boxes hdmi output and compare. 1 from the Shield, 1 from whatever other device, and 1 from a software decoder we know doesn't do anything but decode the frame.
It will be the ultimate way to convince those who have doubts.
But facts are already settled.
Now we have only hopes that Nvida do something about it (if they can).
No facts are evident. You took photos with a camera of two different devices with no confirmation of the same frame and you included no reference frame which you can guarantee doesn't mess with the decoded frame. For all anyone knows the device you are using is sharpening the image before output and the Shield is outputting a true image. If that is the case then don't mess up the Shield for those who don't want to distort the image.
But Martijn is right. Wrong place for this.
From NVIDIA:
Quote:So the usual questions I ask are
- are they on a 4k TV
- are they in same mode on all (res, yuv vs rgb, etc.). 4k we might be losing something in scaling, yuv420 vs yuv/rgb 444 clearly drops bits..
- have they tried latest nightly with surface changes
- have they connected to same Hdmi port, or triple checked picture settings same on all ports
And please guys go complain at NVIDIA forums about this as it's just filling the thread without anything we can do.
We will not be mediators for this. The only ones that could potentially fix this are NVIDIA and they don't read this thread.
BTW, don't forget to mention if you're using 15.2 or 16 nightly, as the upscalers are different (egl for 15.2, android for 16 with mediacodec surface).
(2015-10-28, 07:11)knives of ice Wrote: [ -> ] (2015-10-28, 06:29)Tinwarble Wrote: [ -> ] (2015-10-28, 06:19)knives of ice Wrote: [ -> ]the problem is the only way i can think of to get it to confluence is manually removing silvo from my addons folder because it won't boot on these newer nighly's. it won't move past the boot screen, but i can hear the remote moving sounds behind the jarvis overlay. i'm just guessing on what the issue is at this point really the skin may not be the problem, not sure why it would be since it is jarvis compatible
I'm not really sure what you mean. All you should have to do is go to Appearance>Skin and change it to Confluence while your on the old version of Kodi. Then update and see if you're having the same issue, if you don't then it's the Silvo skin and you'll have to talk to the author.
If you are already on the latest nightly, and need a way to get back to Confluence, then just use ES File Explorer and backup your skins settings, then delete it. If you need to you can just copy it back and you'll be up and running as before.
However, you shouldn't need to remove the Silvo skin to find out.
ahhh...i see what you mean now. use a version that boots up and switch to confluence on it then load the latest nightly. good idea. thanks i should have thought of that as well haha
Or edit guisettings.xml without kodi running.