AVS Forum banner

501 - 520 of 1220 Posts

·
Registered
Joined
·
173 Posts
@Soulnight

I've done a little testing and am now optimising everything with the new version, but so far I think I prefer the FALL algo too. If anything it seems a lot more predictable and you get a consistent result between similar looking scenes whereas before sometimes the targets could be all over the place and it wasn't obvious (from simply looking at the image) why. Plus it should give more similar targets between scenes within chapters so a shorter rolling average can work. I'm tossing up between a compression limit of 250 and 300, brighter movies look better with 300 but dimmer ones look better at 250 so I'll probably stick with 250 as generally brighter movies don't dip too far bellow 300 anyway, it just helps to keep the average up. I like 200 for dynamic tuning, I tried increments of 50 between 100 and 300 and I'm pretty happy with 200 overall, with 150 as a close second.

These are the settings I'm currently using:

Minimal target: 200
Maximal target: 1500
Dynamic tuning: 200
No compression limit: 250
Rolling average: 120
Merge scenes: 350
Minimal chapter duration: 24
Merge chapters: 200

Edit: 120 might be a touch fast for the average, so I'll go back to 240
 

·
Registered
Joined
·
22 Posts
@Soulnight Many thanks to you and Anna for the continuous flow of improvements!

Since you ignored my bug report of the multiple paths issue I am assuming that you and Anna have no idea what the cause may be? It's OK to just say that, btw. :)

If that's the case I may try and move all my 'TV' UHD files to be in the same folder so I don't have to process a folder at a time.
Hi Ian,

Sorry, we didn't look into this yet, because multiple paths work fine for us.

If you want me to try and recreate this, you have to give me more information. Is it a certain textbox that doesn't work? Did you try to put your first path there? Is it a problem with a certain path? Does it work if you put this path in the first place and leave all other textboxes cleared?

Did you manage to run the tool for any other multiple path selection? Does a restart of the program help?

Also, screenshots can be very useful 😊

Anna
 

·
Registered
Joined
·
787 Posts
@Soulnight

I've done a little testing and am now optimising everything with the new version, but so far I think I prefer the FALL algo too. If anything it seems a lot more predictable and you get a consistent result between similar looking scenes whereas before sometimes the targets could be all over the place and it wasn't obvious (from simply looking at the image) why. Plus it should give more similar targets between scenes within chapters so a shorter rolling average can work. I'm tossing up between a compression limit of 250 and 300, brighter movies look better with 300 but dimmer ones look better at 250 so I'll probably stick with 250 as generally brighter movies don't dip too far bellow 300 anyway, it just helps to keep the average up. I like 200 for dynamic tuning, I tried increments of 50 between 100 and 300 and I'm pretty happy with 200 overall, with 150 as a close second.

These are the settings I'm currently using:

Minimal target: 200
Maximal target: 1500
Dynamic tuning: 200
No compression limit: 250
Rolling average: 120
Merge scenes: 350
Minimal chapter duration: 24
Merge chapters: 200
Great job guys, so what would you guys recommend for 50nits guy? :p

Minimal target: 50

Maximal target: 1500
Dynamic tuning: 50
No compression limit: 200
Fix: 5

Rolling average: 120
Merge scenes: 100
Minimal chapter duration: 24
Merge chapters: 50
 

·
Premium Member
Joined
·
1,754 Posts
Hi Ian,

Sorry, we didn't look into this yet, because multiple paths work fine for us.

If you want me to try and recreate this, you have to give me more information. Is it a certain textbox that doesn't work? Did you try to put your first path there? Is it a problem with a certain path? Does it work if you put this path in the first place and leave all other textboxes cleared?

Did you manage to run the tool for any other multiple path selection? Does a restart of the program help?

Also, screenshots can be very useful 😊

Anna
Hi Anna,

Thanks for responding!

What I've learned is that the actual path string doesn't matter (e.g. the length of the path or characters used is not relevant). The exception occurs when I click the OK button of the Choose Paths panel (i.e. after having selected paths).



I can click CONTINUE and try again and no matter what path I select, once I hit the OK button at the bottom of your CHOOSE PATHS panel, the same exception occurs.

However: I've now learned that if I hit QUIT and then restart your tool, the path IS THERE. I can edit it (change it) and run the tool and it works as expected. Of course if I try to ADD another path, the exception occurs again. At this point, I can simply hit QUIT and restart the app and the path is there.

To me it seems like the problem is when I add a new path, the index (into the array of path strings?) is invalid for some reason, but I can now work around this.

Thanks again,
Ian
 

·
Registered
Joined
·
173 Posts
Great job guys, so what would you guys recommend for 50nits guy? :p

Minimal target: 50

Maximal target: 1500
Dynamic tuning: 50
No compression limit: 200
Fix: 5

Rolling average: 120
Merge scenes: 100
Minimal chapter duration: 24
Merge chapters: 50
The chapter logic shouldn't be effected by your actual display nits, so you can just copy any of the settings people are using, but I wouldn't mix and match. Or come up with your own.

For me the best scenes I found to dial in the dynamic tuning value were scenes with a large range of highlight and dark areas, like a lot of the scenes in Pacific Rim, just increase/decrease the value until those scenes look good to you and then it should translate to everything else.

Just start with the defaults and see if you like it, most of the variables really come down to taste so what looks good to me might not look good to you.
 

·
Registered
Joined
·
22 Posts
To me it seems like the problem is when I add a new path, the index (into the array of path strings?) is invalid for some reason, but I can now work around this.

Thanks again,
Ian
Ok, I know that problem and thought I had solved it. Obviously I haven't... I'll check again this evening.
 

·
Registered
Joined
·
76 Posts
@Soulnight congratulations again for the development of your tool.

I wanted to ask you about the behavior of some mesurement files after the first dynamic measurement.

For example look at this:



I had successful measurement the first time (19th January). But every time I do it again with different settings I have a succesful log with the current data, but you can see the measurement file still shows the first day data...

This happens with some files only. And I'm not sure if its generating a new one or not. Restoring original measurement gives same results. I'm using the current dynamicclipping version.

Regards.
 

·
Registered
Joined
·
80 Posts
Ok.

Our tool has ZERO issues! :)
We just reproduced what you have experienced.

If you change manually the shortcut to profile with fixed target nits, you are correct that it does overwrite the dynamic target nits even if the OSD still shows the "dynamic target nits".

I think this is new from madvr version v39. @madshi can you confirm?

Anyway, if you don't use any shortcuts to change to a predefined profile with a fixed target nits, the dynamic target used are used properly.

;)

Ps: we even used your own optimized measurement file with our mkv from Blade runner 2049 and it worked as intended. (With the exception of being able to switch to a different target nits of course).

Ps2: to convince yourself that our dynamic target nits is getting used, try to put something crazy like 2000 nits as real display nits / minimum target nits. Open the movie and you will see a veryyyy dark movie with 2000target nits getting apllied. ;-)
Thx. for looking into it again!
Two things still not clear to me, though:
A) If I understand correctly, this should apply to all movies then and not just this one. All other movies though work fine.
B) Why does my file process way too fast on your machine when you re-optimize it? You wrote, it took only 2s, indication something was wrong with it.

In any case, I will try and drop the shortcuts in madvr and see if it fixed things.
 

·
Registered
Joined
·
22 Posts
@Soulnight congratulations again for the development of your tool.

I wanted to ask you about the behavior of some mesurement files after the first dynamic measurement.

For example look at this:



I had successful measurement the first time (19th January). But every time I do it again with different settings I have a succesful log with the current data, but you can see the measurement file still shows the first day data...

This happens with some files only. And I'm not sure if its generating a new one or not. Restoring original measurement gives same results. I'm using the current dynamicclipping version.

Regards.
The field you are looking at must be the file creation time stamp. If you hover with the mouse on the measurements file you should see when it was last modified. The details file is always deleted and then created again, so it makes sense that it has a new creation time stamp.
 

·
Registered
Joined
·
1,479 Posts
Discussion Starter #511
Edit: something is strange with your original measurement file. It takes only 2s to optimize (too short). And with our measurement file from blade runner 2049, it takes 7s...

Edit2: nevermind... I just put your file on my ssd... and that's why optimizing took only 2s :) It does take 7s on my local harddrive. Anna found nothing suspicious in debug mode as well.
@(((atom)))

I already explain yesterday the 2s story. ;)
 

·
Registered
Joined
·
80 Posts
OK, so I deleted all my profiles and now it is working for none of my movies anymore. I could need a little assistance here now.

When I first tried your tool I just ran it and magically switching the profiles didn't change the brightness anymore, so I knew it was working. Also I could change the brightness with your tool, visibly.

Now all is deleted and I don't find back. What are the conditions to be met in order to make it work?

The problem with profiles is that I use them for controlling my PJ via RS232, so I'd have to find a way to keep them.
 

·
Registered
Joined
·
1,479 Posts
Discussion Starter #514 (Edited)
OK, so I deleted all my profiles and now it is working for none of my movies anymore. I could need a little assistance here now.

When I first tried your tool I just ran it and magically switching the profiles didn't change the brightness anymore, so I knew it was working. Also I could change the brightness with your tool, visibly.

Now all is deleted and I don't find back. What are the conditions to be met in order to make it work?

The problem with profiles is that I use them for controlling my PJ via RS232, so I'd have to find a way to keep them.
The only one who can help you/us is @madshi .

There seem to be a strange combination of rules, target nits shortscuts which seem to be able to override the dynamic target nits.
Until now, I also always observed this behaviour which I found perfectly logical:
- manual target nits or dynamic target nits defined in the measurement file, all "target nits" profile are ignored na matter what you switch to.

But in some cases, madVR seem to be lost. Nothing to do with our tool or files.
Only have to do with how madshi handles those ""special combinations" in madVR code.


As for you and anybody else using our tool, I would recommend to remove your profile and only keep 1.

-->One static target nits profile with something crazy like 10000 so that you know when this gets activated or not. (crazy dark).

:)

Just to be safe. :D

EDIT: well, with a single target nits profile of 10000nits without any shortcut or rules to it, the BUG always happen.
The OSD shows a varying dynamic target nits according to our file, but the picture is just veeeeeeeeery dark ~10000 target nits


So we tried, the opposite, only a single profile without shortcut or rules, but with 100 target nits instead.
With this is works perfectly, always! :)
Blade Runner 2049, The Meg etc... We even tried putting "maximum target nits" with a crazy low target nits like 20nits, and the picture did react accordingly to the measurement file and was clearly brighter than when we tried with 100 "maxim target nits".

So, please make sure that you only have one target nits profile, no rule, and set to 100nits in madVR, if you want to be 100% sure to avoid the current madVR BUG which seems to happen a bit randomly.

-->The bug is: "you believe" that dynamic target nits is working because you can see it changing in the OSD BUT in fact, it get stuck at a fixed target nits internally
 

·
Registered
Joined
·
80 Posts
The only one who can help you/us is @madshi .

There seem to be a strange combination of rules, target nits shortscuts which seem to be able to override the dynamic target nits.
Until now, I also always observed this behaviour which I found perfectly logical:
- manual target nits or dynamic target nits defined in the measurement file, all "target nits" profile are ignored na matter what you switch to.

But in some cases, madVR seem to be lost. Nothing to do with our tool or files.
Only have to do with how madshi handles those ""special combinations" in madVR code.


As for you and anybody else using our tool, I would recommend to remove your profile and only keep 1.

-->One static target nits profile with something crazy like 10000 so that you know when this gets activated or not. (crazy dark).

:)

Just to be safe. :D

EDIT: well, with a single target nits profile of 10000nits without any shortcut or rules to it, the BUG always happen.
The OSD shows a varying dynamic target nits according to our file, but the picture is just veeeeeeeeery dark ~10000 target nits


So we tried, the opposite, only a single profile without shortcut or rules, but with 100 target nits instead.
With this is works perfectly, always! :)
Blade Runner 2049, The Meg etc... We even tried putting "maximum target nits" with a crazy low target nits like 20nits, and the picture did react accordingly to the measurement file and was clearly brighter than when we tried with 100 "maxim target nits".

So, please make sure that you only have one target nits profile, no rule, and set to 100nits in madVR, if you want to be 100% sure to avoid the current madVR BUG which seems to happen a bit randomly.

-->The bug is: "you believe" that dynamic target nits is working because you can see it changing in the OSD BUT in fact, it get stuck at a fixed target nits internally
Yes! Thank you so much, now everything is working as expected, eben the good ol' Bladerunner, finally!
 

·
Registered
Joined
·
1,479 Posts
Discussion Starter #518

·
Premium Member
Joined
·
1,754 Posts
Thanks very much, Anna & Flo!
 
  • Like
Reactions: Neo-XP and AKJ89

·
Registered
Joined
·
96 Posts
...

EDIT: well, with a single target nits profile of 10000nits without any shortcut or rules to it, the BUG always happen.
The OSD shows a varying dynamic target nits according to our file, but the picture is just veeeeeeeeery dark ~10000 target nits..
...

So, please make sure that you only have one target nits profile, no rule, and set to 100nits in madVR, if you want to be 100% sure to avoid the current madVR BUG which seems to happen a bit randomly.

Hope Madshi will solve it soon ... It is very frustrating not to know if the measurement file is active ....
 
501 - 520 of 1220 Posts
Top