Commercial Advance Beta Release

[Jere_Jones]

Originally Posted by julesjohn

As I said I do use Sagewake, but maybe I need to note to the creator of Sagewake about its interaction or it needing to be aware of Comskip. I did so here. And his response is here but here it is quoted too in case you find it useful...


Thanks.

VERY interesting. I knew that there must be a mechanism fo stopping a machine from going to sleep, but I didn't know what it was.

Unfortunately, this doesn't help comskip. Comskip, as a console application, has no awareness of windows messages.

I will investigate this and may incorporate this ability.

Thanks,
Jere

[stanger89]

Jere,

Way of the road here, but how hard would it be for comskip to determine the AR of the recording, ie figure out if it were letterbox or not?

[fcollman]

its already in his algorithm... it uses it as one of the factors in determining what are the commercials... its pretty straight forward, just look at the frame and see if the top and bottom sections have an avg brightness below some threshold, and the center section has some brightness above some threshold (to filter out those times that you have just a entirely black screen, or just a logo in the center.

One of the things that I think might make the comskip application better, is if its output were a bit more informative than just the commercial frames... if you seperate out the gathering data about frames, and maybe creating blocks, and then left the deciding which blocks were commercials to a seperate application (maybe integrated, or linked into sage) then it would be possible to analyze a file, but then fiddle with parameters to have it detect commercials better, or do things like you seem to be alluding too... add extra functionality. would make tuning the parameters to make the system work the best on your system a bit easier too, as its extra the data that takes the most time.

[stanger89]

Originally Posted by fcollman

its already in his algorithm...

Alrighty then, how about adding a spot to the output file specifying if it's letterbox or not?

[Crashless]

Originally Posted by stanger89

Alrighty then, how about adding a spot to the output file specifying if it's letterbox or not?

You want automatic AR selection don't you. You dog....



Great idea.

[stanger89]

Well, it started with the DVD metadata, DVD Profiler provides AR for each disk and whether it's anamorphic or not, based on that, it wouldn't be hard to make the media player switch to a given AR based on a flag in the metadata. The tricky part is Sage recordings, since they're all 4x3 or non-anamorphic. But Jere already does AR detection it sounds like, so if he could spit out a flag (letterbox = yes/no) it wouldn't be too hard to extend it to everything.

[Jere_Jones]

Originally Posted by stanger89

Well, it started with the DVD metadata, DVD Profiler provides AR for each disk and whether it's anamorphic or not, based on that, it wouldn't be hard to make the media player switch to a given AR based on a flag in the metadata. The tricky part is Sage recordings, since they're all 4x3 or non-anamorphic. But Jere already does AR detection it sounds like, so if he could spit out a flag (letterbox = yes/no) it wouldn't be too hard to extend it to everything.

This would be easy enough. The problem lies in how. Should comskip output another file? It already puts out a .txt, .log, .logo.txt, .ccyes, .ccno, .Vproj, etc. I think I am single handedly responsible for the pollution of computer directories around the world!!!

I could put the information inside the .txt file, but then using the information becomes MUCH harder.

Of course, all my musing is for naught. The real questions are: Will Cayars implement such a beast? If so, how does HE want the information?

Jere

[Stuntman]

I just suggested this over on Cayars board.. would it be possible to implement a MySql database to store all this info in? I too have so many text files of various types that my directories are a mess.. seems like it should be possible to make a database entry for each show and then have all the necessary data for each of those shows contained within that object. I'd imagine one could then create .proj files or whatever dynamically by querying the database as needed for burning or whatever, then clean them up again at the end of the project.

I believe MythTV already uses MySQL for some of this.. seems like a logical way to store data.. too many files just gets too messy..

[stanger89]

Originally Posted by Jere_Jones

This would be easy enough. The problem lies in how. Should comskip output another file? It already puts out a .txt, .log, .logo.txt, .ccyes, .ccno, .Vproj, etc. I think I am single handedly responsible for the pollution of computer directories around the world!!!

I could put the information inside the .txt file, but then using the information becomes MUCH harder.

Of course, all my musing is for naught. The real questions are: Will Cayars implement such a beast? If so, how does HE want the information?

Jere

I had figured just put it in the TXT file, somewhere where it wouldn't screw stuff up.

[owilsky]

Hi Jere,

I need you help.

For me the old comskip.exe 0.21 works much better than your newer versions.
What has changed since then? Did the old 0.21 version also did all the tests like black frame, logo, scene change, fuzzy logic - check?
I want to use the new versions, mainly because if Cayars plans to outsource comskip off the stv, but the old works so much better.
What settings do you suggest? I live in Germany, so I could not use CC.
My settings: (the important ones)

v.21:

 
fps=25
border=10
max_brightness=120
test_brightness=80  
max_avg_brightness=15	
max_commercialbreak=700	
min_commercialbreak=30	
max_commercial_size=180
min_commercial_size=30
require_div5=0		 
div5_tolerance=.75

v.64:

detect_method=15
fps=25
border=10 
max_brightness=180      
test_brightness=60      
max_avg_brightness=60
max_commercialbreak=700	
min_commercialbreak=30	
max_commercial_size=180	
min_commercial_size=30
require_div5=0		 
give_up_logo_search=600	
two_pass_logo=1		;start over after finding logo?

Thanks for your help,

Oliver

[Jere_Jones]

If you set detectmethod = 1, then .21 and .65 should find exactly the same commercials. Version .21 only did checks for black frames.

I'd like to get my hands on some recordings from Europe to see why comskip doesn't work over there since the fundamentals should be the same. Then I could probably help you better.

Jere

Originally Posted by owilsky

Hi Jere,

I need you help.

For me the old comskip.exe 0.21 works much better than your newer versions.
What has changed since then? Did the old 0.21 version also did all the tests like black frame, logo, scene change, fuzzy logic - check?
I want to use the new versions, mainly because if Cayars plans to outsource comskip off the stv, but the old works so much better.
What settings do you suggest? I live in Germany, so I could not use CC.

If you set detectmethod = 1, then 0.21 and 0.65 should find the same commercials. Version .21 only did checks for black frames.

I'd sure would like to get my hands on some recordings from Europe to see why comskip doesn't work over there since the fundamentals should be the same. Then I could probably help you better.

Jere

[owilsky]

Originally Posted by Jere_Jones

If you set detectmethod = 1, then .21 and .65 should find exactly the same commercials. Version .21 only did checks for black frames.

Thanks for the info.

Originally Posted by Jere_Jones

I'd like to get my hands on some recordings from Europe to see why comskip doesn't work over there since the fundamentals should be the same. Then I could probably help you better.

I could upload some mpegs to my server, then you can download them... Do you have broadband? The files will be very big... but I can record them in less quality ... or will that influence commercial detection?
Let me know if I should upload some shows for you, I will PM you the link when done so.

Oliver

[Juncti]

I've been trying to use this, I can get it to run by manually doing it at the command prompt (but can't get sage to play it like that). I have yet to find it inside Sage. I imported the stv in Sage but don't see it anywhere. I've seen mention to look for Custom Menus then settings, but I don't see custom menus anywhere. Is there a keyboard shortcut I need to do to pull up that menu?

I'm running Sage 2.1 and the only stv I've imported was this one.

Thanks

[malore]

Follow the instructions on Malore's STV Homepage. I've update the first post to indicate this and that the latest version of my STV updated for use with SageTV 2.1 can be found there.

Note: Several other custom STV files have added support for comskip. So feel free to try them as well.

[nigfink]

What SVT's support the live tv realtime processing of commskip? Are there any that work with 2.0.1? I don't want to have to upgrade if I don't have too, my system works great. Thanks.

[mikesm]

Is there a version of COMSKIP available for Linux? I am thinking about deploying a customized sage, but all my media is stored on a linux system and remote mounted, and it would be much more efficient to run comskip on the fileserver than on the Sage machine.

Also, now that Myth is releasing an improved commercial skip code base, will these changes be put into comskip?

Thanks,
Mike

[Jere_Jones]

Originally Posted by mikesm

Is there a version of COMSKIP available for Linux? I am thinking about deploying a customized sage, but all my media is stored on a linux system and remote mounted, and it would be much more efficient to run comskip on the fileserver than on the Sage machine.

Also, now that Myth is releasing an improved commercial skip code base, will these changes be put into comskip?

Thanks,
Mike

I don't use linux so I, personally, will not be releasing a linux version. But the code should be easily portable. I would, but I don't run any linux machines.

And, BTW, the improved commercial skip code base that MythTV is releasing came from the comskip that Sage users have had for a while now. Unless there is something that I didn't notice, there isn't anything new to add.

Jere

[Captain_Murdoch]

Originally Posted by Jere_Jones

I don't use linux so I, personally, will not be releasing a linux version. But the code should be easily portable. I would, but I don't run any linux machines.

And, BTW, the improved commercial skip code base that MythTV is releasing came from the comskip that Sage users have had for a while now. Unless there is something that I didn't notice, there isn't anything new to add.

Jere

To clarify,

The improved code base in MythTV did NOT come from comskip, I put a comment in my source saying that the edge detection IDEAS for logo detection came from comskip, but none of the code did and the other ideas that have gone into the improved detection base in Myth have been on my TODO list for a couple years now. I've been working on the commercial detection code in Myth for over 2 years now off and on, and I think I'm safe in saying that most of the detection methods you have implemented have been mentioned on the Myth mailing lists at least 3-4 times and probably on the net in other locations more than that. Not trying to get mean or anything, but I don't like the inference that my code is a copy of comskip source. I've got a list a page long of detection ideas that I've come up with or that MythTV users have mentioned on the lists, so just because a feature is added to Myth doesn't mean that the code was copied from your source (which was a direct copy of my original Myth code I might add).

Chris Pinkham

[Captain_Murdoch]

Originally Posted by heffe2001

Anybody build this program on a linux box yet? I remember somewhere towards the front of this thread where it was mentioned that it would be possibly done, but can't remember seeing anything about it after that.

Here's a link to a patch you can apply to get compskip compiled under Linux. I just updated this to the current version in subversion but haven't even run it to test so I can't say that I caught everything, but it does compile so you can play with it if you want.

The patch includes a Makefile and README.Linux file that should be enough to get you going if you do want to do anything with it.

http://www4.infi.net/~cpinkham/tmp/comskip_linux.patch

[Jere_Jones]

Originally Posted by Captain_Murdoch

The improved code base in MythTV did NOT come from comskip

Hmmm... it would seem that I am guilty of a common crime - assumption. Having not looked at the source code for MythTV's commercial detection, I assumed that comskip was the source of the "ideas" if not the code.

Originally Posted by Captain_Murdoch

Not trying to get mean or anything, but I don't like the inference that my code is a copy of comskip source.

My sincerest apologies. I was responding (before I did any research) because I didn't like the inference that all I have done is copy code from MythTV.

Originally Posted by Captain_Murdoch

(which was a direct copy of my original Myth code I might add).

This is very true of the original comskip code, but I have rewritten or discarded (the PNG scanning sequence comes to mind) almost all of it. So I feel a little protective of my code, as do you.

These are my comments on each new feature and why I made the comment I did:

Major Enhancement (Commercial Flagging)

• Code should now detect actual Blank frames rather than Black frames as was previously done. The method of detecting blank frames changed to detect the difference between min/max brightness rather than just checking for frames with a max brightness below a certain level. This seems to eliminate a lot of false positives and do better detection on dark shows and shows with black frames interspersed throughout the show. Tested with some problem episodes of shows such as CSI, Law & Order, Alien, and LAX. Works well on all of these except when CBS and/or the local affiliate decide to put their station logo on the blank frames before/after commercials. For that, when logo information is detected, it is accounted for, and the frames are detected as being blank even with the logo present. This is especially useful in the new "All" detection method mentioned below.

I took this as a form of the "intelligent black threshold" that I added a long time ago.

• Tightened up some of the commercial length checks for the blank frame detection method.

Many moon ago, I changed the tolerance from a blanket set time to a tolerance based on the number of black frames surrounding a block. I figured this was the same thing.

• Bumped up border size for area that is ignored during commercial detection from a 10 pixel border to 20.

Border size has been customizable almost from the beginning.

• Added new "All" detection method. This detection method uses information from blank-frame, scene change, and logo detection and can easily be modified to take advantage of other detection information. It works in a fundamentally different way than the normal blank-frame and scene change methods. The code groups the frames into blocks separated by blank frames and rates each block based upon several factors such as the rate of scene change, what percentage of frames in the block have a logo present, how long/short the block is, etc.. Indications that a block is part of a commercial (such as very high rates of scene changes) lower the block's score, while indications that a block is part of the show (such as having a logo present on most frames) raise the block's score. In the end the blocks are looked at as a group and various bits of logic applied to make the final determination where commercials start and end. On most of the shows tested, this method seems to be doing a much better job than just blank-frame or blank-frame plus scene-change.
• Simply put, the new "All" method detects almost every commercial. It has been successfully tested on (previously) problem shows like CSI, West Wing, Enterprise, etc.

This was the big one. This is the addition/change/upgrade that made me think "Hey, cool, he's using some of my ideas." As far as I could tell everyone previously did commercial detection in a linear fashion. I especially keyed off on the words "block" and "score" which are all over the my source code for this. If you didn't copy the idea for this from me, my apologies. If you did, thank you for finding one of my ideas useful.

• Add code to check for 5 & 10 second commercials to try to catch some of those spots for things like local news and SciFi blurbs.

This can also be found in my code since I tried to get away from the "divisible by 5" idea.

Having calmed down and given this some thought, here is my comment:

Comskip was born from MythTV's commercial detection routines and owes it's existence to that code. Comskip has since grown and morphed into it's own identity that (IMHO) is innovative in some areas.

As far as I can tell, some of the improvements to MythTV's commercial detection code have previously been incorporated into comskip. The other's are not relevant to Sage's use of comskip. If I'm missing something, please let me know.

Regarding the linux patch, I have been unable to get it to successfully merge with any windows patching tools, but I will patch it manually and then commit it to the sage-community website.

Jere