Jump to content

Show Intro Skip Option (Plugin)


Liquidfire88

Recommended Posts

rbjtech
5 minutes ago, crusher11 said:

I hit save on every episode I edited, yes.

I don't know how well it worked previously as I was avoiding digging too deep into editing while we waited for core implementation. It's possible I've simply stumbled on a bug that has always been there. 

Possibly.

I've used the editor a reasonable amount but I've never had issues.

Let me do some further testing and come back to you ...

Link to comment
Share on other sites

crusher11

I also have a handful of shows that simply don't show up in the plugin at all.

Link to comment
Share on other sites

rbjtech
5 minutes ago, crusher11 said:

I also have a handful of shows that simply don't show up in the plugin at all.

Assuming they are of type TV Shows, then there is zero reason why they shouldn't ..

Link to comment
Share on other sites

crusher11
4 minutes ago, rbjtech said:

Assuming they are of type TV Shows, then there is zero reason why they shouldn't ..

There's an excellent reason why they shouldn't: they don't show up anywhere in Emby at all.

I have a folder from my PC mounted in my NAS, because the latter ran out of space and I'm using that as a stopgap. Or rather, had a folder from my PC mounted in my NAS. It no longer is. I have no idea how that happened and can't remember how to set it up.

Shit.

Link to comment
Share on other sites

Banquero

Buenas tardes, voy a hacer una pregunta tonta...El plugin Skip Intro y Intro/Credit Skip, que aparece en los datos avanzados del Servidor, es lo mismo? Gracias...

Link to comment
Share on other sites

Painkiller8818

Maybe something people don't know. It is also working on Libs with mixed content.

I have a Documentation library and it is skipping intros for the multi part documentations.

really nice

  • Like 1
Link to comment
Share on other sites

Cheesegeezer
Just now, Painkiller8818 said:

Maybe something people don't know. It is also working on Libs with mixed content.

I have a Documentation library and it is skipping intros for the multi part documentations.

really nice

Only if its seen as an episode 

Link to comment
Share on other sites

Skyfay

I have now downloaded the plugin to test it, first of all thank you very much for the work!
Well, in my setup, the media and Emby are separate, so one server for Emby and one server for the media.
Now I started the audio fingerprinting and it immediately started loading data onto the Emby server at over 100 MB/s.
Well, does he only do that with the beginning of the episodes or does he really just load 20-30TB over to test? 😅

Link to comment
Share on other sites

DarKni8
Posted (edited)

Initially i suppose

Edited by DarKni8
Link to comment
Share on other sites

Cheesegeezer
46 minutes ago, Skyfay said:

I have now downloaded the plugin to test it, first of all thank you very much for the work!
Well, in my setup, the media and Emby are separate, so one server for Emby and one server for the media.
Now I started the audio fingerprinting and it immediately started loading data onto the Emby server at over 100 MB/s.
Well, does he only do that with the beginning of the episodes or does he really just load 20-30TB over to test? 😅

so all that is happening is that a small audio fingerprint bin file is being created for each episode (found in programdata/plugins/audio - I think) this is then used to allow for detection. Once detection completes the fingerprint file is removed. 
 

hope that makes sense.

Link to comment
Share on other sites

Skyfay
Posted (edited)
4 minutes ago, Cheesegeezer said:

so all that is happening is that a small audio fingerprint bin file is being created for each episode (found in programdata/plugins/audio - I think) this is then used to allow for detection. Once detection completes the fingerprint file is removed. 
 

hope that makes sense.

Yeah! That makes sense. My question is, however, what exactly does emby call up from the media? 
The whole files in full size or just the beginning of the media like 2 minutes or so?

Edited by Skyfay
Link to comment
Share on other sites

Cheesegeezer
2 minutes ago, Skyfay said:

Yeah! That makes sense. My question is, however, what exactly does emby call up from the media? 
The whole files in full size or just the beginning of the media like 2 minutes or so?

If the episode is 30mins or less, 15mins of audio only is taken, if its more then 20mins of audio only.  
i can’t remember file sizes exactly but they are extremely small as the are converted to mono streams as we are only interested in waveform comparison.

Link to comment
Share on other sites

rbjtech
47 minutes ago, Cheesegeezer said:

If the episode is 30mins or less, 15mins of audio only is taken, if its more then 20mins of audio only.  
i can’t remember file sizes exactly but they are extremely small as the are converted to mono streams as we are only interested in waveform comparison.

The FP files are kept actually cheese as we need them for future comparisons - but as you say, they are very small - 36K is the biggest file (per episode).

The other factor is the parallelism settings - if these are set high, then ffmpeg (the Chromaprint engine) will be doing MULTIPLE reads on the remote share/disk - and if this is limited by ethernet (at 1Gb/sec) then this becomes your bottleneck - at approx 112Mbytes/sec.  In this scenario, it is probably better to set the parallelism low to maybe 1 for FP and 1 for Detect - it will take a lot longer, but it won't impact the day-day running of emby.   

The task will carry on where it left off btw - so you can stop it at anytime and resume - so you could time to do it out of hours for example. 

Link to comment
Share on other sites

Skyfay
38 minutes ago, rbjtech said:

The FP files are kept actually cheese as we need them for future comparisons - but as you say, they are very small - 36K is the biggest file (per episode).

The other factor is the parallelism settings - if these are set high, then ffmpeg (the Chromaprint engine) will be doing MULTIPLE reads on the remote share/disk - and if this is limited by ethernet (at 1Gb/sec) then this becomes your bottleneck - at approx 112Mbytes/sec.  In this scenario, it is probably better to set the parallelism low to maybe 1 for FP and 1 for Detect - it will take a lot longer, but it won't impact the day-day running of emby.   

The task will carry on where it left off btw - so you can stop it at anytime and resume - so you could time to do it out of hours for example. 

Thanks for the Information!
How can i pause and resume the detection?
When i click there, i think the task stops completly right?
image.png.a2c233d14cfb3f085b7d8b69f91c5ae1.png

Link to comment
Share on other sites

rbjtech
16 minutes ago, Skyfay said:

Thanks for the Information!
How can i pause and resume the detection?
When i click there, i think the task stops completly right?
image.png.a2c233d14cfb3f085b7d8b69f91c5ae1.png

Sure, it stops, but when you re-start the task, the task will start at 1.8% already complete if that makes sense ?

Link to comment
Share on other sites

Skyfay
10 minutes ago, rbjtech said:

Sure, it stops, but when you re-start the task, the task will start at 1.8% already complete if that makes sense ?

Yes, that makes perfect sense.
I've let the whole thing run in default settings mode for an hour now.

Time Percentage data consumption
10:07 1.0% 5.1 TB Down 
11:07 2.3% 5.4 TB Down

So in one hour he did 1.3% and had to retrieve about 300GB of data.

Link to comment
Share on other sites

Cheesegeezer
5 minutes ago, Skyfay said:

Yes, that makes perfect sense.
I've let the whole thing run in default settings mode for an hour now.

Time Percentage data consumption
10:07 1.0% 5.1 TB Down 
11:07 2.3% 5.4 TB Down

So in one hour he did 1.3% and had to retrieve about 300GB of data.

Wow!!! as @rbjtech says we actually do keep the files now.  But my largest is 38Kb as you can see.  How many episodes do you have in total? I think in debug logging mode, there is a line that says how many episodes it has to process.  Maybe Rbj can elaborate - he is the logging king lol 🤣

image.png.c0424ed3fbeb1e3533224a5dcc8aac00.png

Link to comment
Share on other sites

Skyfay
Posted (edited)
10 minutes ago, Cheesegeezer said:

Wow!!! as @rbjtech says we actually do keep the files now.  But my largest is 38Kb as you can see.  How many episodes do you have in total? I think in debug logging mode, there is a line that says how many episodes it has to process.  Maybe Rbj can elaborate - he is the logging king lol 🤣

image.png.c0424ed3fbeb1e3533224a5dcc8aac00.png

I think you misunderstood me 😅
This 300GB of data are the traffic from my media to emby. That's why I asked if all the files had to be read or just the beginning. 
If the entire length of a sequence is called up, then that will end up being around 20-30TB of data traffic.
And that doesn't work for me via the internal network but via the Internet. That's a huge amount of data, that's why I asked. 😆
The files generated by Emby are probably only kb in size.

Edited by Skyfay
Link to comment
Share on other sites

rbjtech

It reads two sets of info - for the Intro, depending on episode duration, it may read UP TO 20 mins worth.  For the credit's it reads the last 2 minutes.

So keeping it simple for the maths and assuming it's a constant bitrate - if you have a 9Gb/60min long episode file - then it would read 3 Gbytes for the intro and 0.3 Gbytes for the credits.

I need to check on the credit black frame detection - as that also reads, but I'm not 100% sure on exactly what it reads - it may be 2 mins worth again but I think it is a much shorter duration than this.  @chef @Cheesegeezer Can you kindly confirm ?

ffmpeg only reads what it needs - it does not need to read the entire file.

To note - this is why the Intro FP length is important - if the Intro is in the first 30 seconds of the episode, we have wasted the vast majority of the read.  However, if the Intro is in the last 1 minute of the 20 minute duration (a real possibility these days) then any shorter and we would have missed it.

Some sort of 'manual' or metadata assisted value would be ideal - or some sort of 'repeat' value increasing the duration might be an idea.  ie - try once with a 5 minute FP, on fail, try again with a 10 minutes, then 20 minute - after this, assume no Intro.     As the majority of Intro's do actually happen in the first 5 mins, then this 'may' actually be quicker overall even with the multiple tries on those episodes needing longer duration's. 

Link to comment
Share on other sites

rbjtech
1 hour ago, Cheesegeezer said:

  Maybe Rbj can elaborate - he is the logging king lol 🤣

Yes - The summary before scanning is in Info mode - so it's always written - the logging is very informative and in Debug mode, it will be listing all the items found, timings, breakdown of each item etc.   

Class is always 'Intro Skip' - do you can filter on it and see everything that is going on within the Plugin.

@Cheesegeezer - we should probably change the 'Skip:Chapters' class to 'Intro Skip' as that is done by the Plugin ... but no big deal - the info is there .. ;)

image.png.20d8f4c41621a65b9e257d2c303c8762.png

Link to comment
Share on other sites

Hi!

Black frame detection ran the last 3 1/2 minutes of the file. If I remember correctly 😃

 

  • Thanks 1
Link to comment
Share on other sites

rbjtech
48 minutes ago, chef said:

Hi!

Black frame detection ran the last 3 1/2 minutes of the file. If I remember correctly 😃

 

thanks Chef !

So yes, this is also a very intensive task from a CPU perspective - but on the example above - it will also be dowloading an extra 0.525Gb from the source file as well.

In summary, remote storage of the media is great for streaming, but if you need to 'process' it in any way - then you are going to have to partially download it - there is no getting away from that if you process it locally.

Link to comment
Share on other sites

2 hours ago, rbjtech said:

For the credit's it reads the last 2 minutes.

 

2 hours ago, chef said:

Black frame detection ran the last 3 1/2 minutes of the file

Yeah, I can think of a number of shows that have at least 3 minutes of credits (maybe more)....

Link to comment
Share on other sites

Cheesegeezer
2 minutes ago, ebr said:

 

Yeah, I can think of a number of shows that have at least 3 minutes of credits (maybe more)....

Vikings Valhala is actually over 3 mins, it's the longest one i know so far

@chef what do you think about doing a 2min check and if nothing found moved to 4mins?

image.png.6ca1ef490dce7babe249e12e07ab9e01.png

Link to comment
Share on other sites

4 minutes ago, ebr said:

 

Yeah, I can think of a number of shows that have at least 3 minutes of credits (maybe more)....

Especially when shows have alternate langauge audio. Most of those shows will do an extended credit sequence that lists all the voice actors for each language.

I've seen it on D+, as well as Nflix. 

The newest season of Stranger Things is a good example of that. Each episode has extended credit sequences. 

  • Agree 1
Link to comment
Share on other sites

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