Jump to content

Show Intro Skip Option


Liquidfire88

Recommended Posts

Cheesegeezer
12 minutes ago, chef said:

Then we should put a hard limit at 4.

@Cheesegeezer, @rbjtech

It makes sense to have quality over quantity... I think.

If a lower scan rate is proving more successful, then should we cap the threading?

We would still have the xml config, for extreme cases, could be edited to increase the detection rate manually. However, the UI should cap at 4.

I don't think a UI hard limit. Should be installed.  I can't remember what our default is (2 or 4).  We do have a disclaimer that says low power machines should set to 2.  However, if you are like @rbjtech with a fancy new 12700 processor, then those would like to be able to up the limit without editing xml.  we could set a hard limit of 4 with a toggle switch to "increase default Values"  if switched a popup dialog can then have information stating some known scenarios and the effects it could have on the scan, with a continue and cancel button.

In the future processors will get faster and cheaper people will upgrade and the limit of 4 will be redundant as the new limit will be 8 for example.

9 minutes ago, Painkiller8818 said:

But i have "only" 8 physical cores, maybe someone has a server with 24 cores. But yes, i have much better results with 4 sessions

Do we know if its the detection that causes the issues having this set to 8 or if it was the FP process that didn't write the Chroma data well enough.  

I will do a check on the Chroma values.  If they are identical with it set to 2 parallel and 8 parallel then we can see that it's the detection that falls down when it hits the cpu limitiations.  There are too many variables to content with in terms of hardware configurations.

Sorry that was my brain dumping out there.

 

  • Like 2
Link to comment
Share on other sites

Painkiller8818
2 minutes ago, Cheesegeezer said:

Do we know if its the detection that causes the issues having this set to 8 or if it was the FP process that didn't write the Chroma data well enough.  

I uploaded a log but haven't got any response on it.

i watched both NAS and neither CPU nor DISK usage was high (this is where my movies are)

If the Emby Servers local Disk is used for this, there is absolutely nothing, i have a very fast m.2 Disk with read/write speed of about 4-6K MB/s

Link to comment
Share on other sites

33 minutes ago, Cheesegeezer said:

I don't think a UI hard limit. Should be installed.  I can't remember what our default is (2 or 4).  We do have a disclaimer that says low power machines should set to 2.  However, if you are like @rbjtech with a fancy new 12700 processor, then those would like to be able to up the limit without editing xml.  we could set a hard limit of 4 with a toggle switch to "increase default Values"  if switched a popup dialog can then have information stating some known scenarios and the effects it could have on the scan, with a continue and cancel button.

In the future processors will get faster and cheaper people will upgrade and the limit of 4 will be redundant as the new limit will be 8 for example.

Do we know if its the detection that causes the issues having this set to 8 or if it was the FP process that didn't write the Chroma data well enough.  

I will do a check on the Chroma values.  If they are identical with it set to 2 parallel and 8 parallel then we can see that it's the detection that falls down when it hits the cpu limitiations.  There are too many variables to content with in terms of hardware configurations.

Sorry that was my brain dumping out there.

 

This all makes sense. We could look at environment processors, and have a popup if the user increases the value beyond the cores / 2.

Or the toggle to increase the value beyond recommend values.

"Ludacris speed" 😆 "going plaid"... Anyone?... Spaceballs?

 

Link to comment
Share on other sites

I did this:

Those two settings are locked by default:

introlock2.png.d754dffcdaf0ceb18a156562ba79e9f8.png

 

You can unlock them, then edit the values:

introlock.png.7078996b33fb30e90406779eea581ac4.png

 

But, if you lock the setting again, the settings go back to default (value of 2).

introlock2.png.d754dffcdaf0ceb18a156562ba79e9f8.png

 

I was able to use the value of 6 on my server, and get great results.

I think that in most cases, when users unlock the setting, we should recommend 6, unless they have a massive server that can handle 8 or 16.

Low power machines should just leave the setting locked, which is a default of 2.

 

I don't know if this will help, but, at least there is a toggle that will reset to defaults if things are not working for people.

Edited by chef
  • Like 1
Link to comment
Share on other sites

rbjtech

The two tasks have different hardware dependencies - FP is disk I/O, Detect is CPU with a little disk I/O for the end credit.

It's impossible for us to cater for every scenario - I think we just need to have text with the above info, and let the user make the choice but keep the defaults low (as they are now). imo.

 

  • Like 1
Link to comment
Share on other sites

fallenwitch3r

@chefi have the same problem with the chapter insertation again. I have add a new series and all intros are correct identified but the chapter insertation did not work. Even refresh metadata has not worked.

The intro starts at 00:00 and end at 01:44 on almost all episodes.

Edited by fallenwitch3r
Link to comment
Share on other sites

1 minute ago, fallenwitch3r said:

@chefi have the same problem with the chapter insertation again. I have add a new series and all intros are correct identified but the chapter insertation did not work. Even refresh metadata has not worked.

@Cheesegeezer is the chapter guy. He knows that part of the code the best.

We might need to see a log. 

Link to comment
Share on other sites

Cheesegeezer
2 hours ago, fallenwitch3r said:

@chefi have the same problem with the chapter insertation again. I have add a new series and all intros are correct identified but the chapter insertation did not work. Even refresh metadata has not worked.

The intro starts at 00:00 and end at 01:44 on almost all episodes.

Can you check the Chapters Tab and see if you have chapter errors pop up please.

Thanks

Link to comment
Share on other sites

fallenwitch3r

I added season folder by season folder and not the whole folder which includes all episodes and maybe this is the error. I cleaned up the folder from all meta files after moving the folder to the main dir. I let the system do a rescan on the series folder. Then I copied the folder back and after adding all metadata by the system the intro skip plugin did his work like before. All timestamps and chapters are now correctly generated.

Edited by fallenwitch3r
Link to comment
Share on other sites

I havent read all the chapter code.

Interesting that is was picked up after a complete refresh.

Link to comment
Share on other sites

Cheesegeezer
2 hours ago, chef said:

I havent read all the chapter code.

Interesting that is was picked up after a complete refresh.

There isn't much to it chef.  just lots of organising the chapter positions.

There was nothing in the logs either. 

  • Like 1
Link to comment
Share on other sites

fallenwitch3r

Is that bug which causes the image loading on the chapter section fixed? Or are there any news about an upcoming fix?

I went back to the previous beta after the bug gets public.

Link to comment
Share on other sites

Cheesegeezer
30 minutes ago, fallenwitch3r said:

Is that bug which causes the image loading on the chapter section fixed? Or are there any news about an upcoming fix?

I went back to the previous beta after the bug gets public.

I asked for more information to look at it, however you have not replied to the PM or on here. You just went ahead and found a work around and now I have no idea what caused the issue in the first place.

 

  • Like 1
Link to comment
Share on other sites

fallenwitch3r
17 hours ago, fallenwitch3r said:

I added season folder by season folder and not the whole folder which includes all episodes and maybe this is the error. I cleaned up the folder from all meta files after moving the folder to the main dir. I let the system do a rescan on the series folder. Then I copied the folder back and after adding all metadata by the system the intro skip plugin did his work like before. All timestamps and chapters are now correctly generated.

This is what I did. I don't know why or how it worked. 

Link to comment
Share on other sites

Hey! I have credit sequences skipping :)

While doing so, I found a bug. I'm just going to make sure I can track it back to where the problem starts. Then I can post a dll.

  • Like 4
Link to comment
Share on other sites

It's close to a release, we just have to fix a couple things, and build a language dictionary. 

  • Like 4
  • Thanks 2
Link to comment
Share on other sites

  • 2 weeks later...
blackstar88
On 12/10/2021 at 9:07 PM, chef said:

It's close to a release, we just have to fix a couple things, and build a language dictionary. 

Pm me if u need swedish translation :)

  • Thanks 1
Link to comment
Share on other sites

magicgg91

Hello,

Would you think the possibility to improve the exclusion capability ?

I have almost 150 series, and I wanted to try this plugin with 2 or 3 to get an idea on how it works and see what are the limits I can set.

But as it's actually available, it's a pain to add one by one series to exclusion list. It should be good to have a button to add a full library to exclusion list.

And when I want to remove some series from exclusion, the list is cropped, I can't see full list, but only around 50 series. I think it's only a display bug, because if I search with browser for one serie that I konw excluded, it found 1 result, but can't display it, it's too low on the page.

Thanks for this great plugin ! I wish all of you my best wishes for Christmas !

  • Like 2
Link to comment
Share on other sites

On 12/24/2021 at 1:45 PM, magicgg91 said:

Hello,

Would you think the possibility to improve the exclusion capability ?

I have almost 150 series, and I wanted to try this plugin with 2 or 3 to get an idea on how it works and see what are the limits I can set.

But as it's actually available, it's a pain to add one by one series to exclusion list. It should be good to have a button to add a full library to exclusion list.

And when I want to remove some series from exclusion, the list is cropped, I can't see full list, but only around 50 series. I think it's only a display bug, because if I search with browser for one serie that I konw excluded, it found 1 result, but can't display it, it's too low on the page.

Thanks for this great plugin ! I wish all of you my best wishes for Christmas !

We have a release that allows you to add all the items to the exclude list, then you can remove the ones you want to scan.

Link to comment
Share on other sites

We are  finishing a wiki over on GitHub.

https://github.com/chefbennyj1/Emby.IntroSkip.Release/wiki/Emby-Intro---Credit-Skip

 

Here is the latest beta build of the plugin.

It is not perfect. It has come a long way.

There are a couple known UI issues.... but I think overall it is a beta worth checking out.

Please clear your browsing data before navigating to the configuration page.

IntroSkip_v0.0.1.55.zip

 

Edited by chef
  • Like 2
  • Thanks 1
Link to comment
Share on other sites

Initial notes: Please don't decompile our project for the sake of using the code in other projects. We  have many open source emby plugin projects on github to read and borrow from.  :)

 

V.0.0.1.55 release notes:

  • Auto skip (you'll have to check it out to see what it is... there is a lot of features there).
  • Add all series to exclude list
  • Configuration UI option to cache images to the file system (beware: image cache can get very very big, very quickly)
  • Configuration UI caches images locally in the browser  for speedy loading after the UI loads a season the first time. (different from the file system cache).
  • Local browser Cache is globally scoped, so as long as the emby dashboard is open, you can browse back to the intro skip configuration and the cache will persist for fast loading.
  • Setting a "true" detection to "false", will reset the Extracted Thumb image, and set it's timestamp to "00:00:00". Don't forget to "Save".
  • Pagination added to the Sequence table, so we will only load (and attempt thumbnail extractions) for five sequences at a time. This helps loading times, as well as limits CPU usage while editing sequence data.

 

Known limitations:

  • Linux users on Emby beta release can not see extracted thumbnail images for sequences (problem with linux ffmpeg on beta release)
  • spinning circle does not disappear after season reset (fixable)
  • Localization is not being used for credit skip
  • Some localization strings need updating for intro skip message.
  • Sending a message during intro causes some websocket issues if the client reports the capability to receive messages but actually can not.
  • Emby API says we can control the message duration (API: "SendMessage") I don't believe this is true. Option in the intro skip plugin for message duration may not actually affect the duration.
Edited by chef
  • Like 4
  • Thanks 1
Link to comment
Share on other sites

Guest
This topic is now closed to further replies.
×
×
  • Create New...