Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[blgentry]

Anyone interested in a View to help identify duplicates and mark false positives?

Let me start by saying that the intention of this project is to locate true duplicates.  It's not to try to find every version of a particular song or other "fuzzy matches".  I *might* try to add that ability later.  For now I'm more interested in real duplicates.

The view I'm working on has a check box field so you can mark songs as being "not duplicate" and then (optionally) hide those from the view.  I'll attach a screen shot so you can see what it looks like today.

Brian.

[Arindelle]

good idea Brian, Thanks!

Right now I have three dupe views .. never thought about adding a "mode" column; now I can combine all three into one. 

One of the things I'm getting hung up on though is finding an easier method to cull my compilation or "greatest hits" type of albums as I acquire the original recordings.  Let's say There is an Artist, Song there are 4 versions of the same song. 3 are legit in that one is on the original studio album, one is on another album, but it is a longer jam version, and the last is a live version.  The fourth is on a compilation or greatest hits album that I want to remove (or at least hide).  The obvious thing to add is duration, but the problem is that on Greatest hits albums the song is exactly the same (might be a different master but forget that for the moment), but there is a couple of seconds difference in the duration.

So what would help is if I could look for dupes on Artist, Name and Duration that is the same  or not more than 3 seconds difference in length. Like "in the range of" +/- 3 seconds -- just can't get the expression right -- anything you, Brian or anyone else can pull form their bag of expression tricks for this?  Using just artist/name brings up 5k tracks worth of stuff to wade through -- 80% are false positives.

[blgentry]

So what would help is if I could look for dupes on Artist, Name and Duration that is the same  or not more than 3 seconds difference in length. Like "in the range of" +/- 3 seconds -- just can't get the expression right -- anything you, Brian or anyone else can pull form their bag of expression tricks for this?  Using just artist/name brings up 5k tracks worth of stuff to wade through -- 80% are false positives.

I just built a "duration rounded" expression that rounds to the nearest 10 second mark.  I think that will work for you.

Code: [Select]

math(formatnumber(math([duration,0]/10), 0) * 10)
You'll probably need to make it a custom field though because ~dup only operates on fields as far as I can tell.  It puked when I pasted in my expression.

Brian.

[Arindelle]

I just built a "duration rounded" expression that rounds to the nearest 10 second mark.  I think that will work for you.

Code: [Select]

math(formatnumber(math([duration,0]/10), 0) * 10)
You'll probably need to make it a custom field though because ~dup only operates on fields as far as I can tell.  It puked when I pasted in my expression.

Brian.

thanks!    I'll check it later tonight

[Arindelle]

hmm .. most of the results come out to 0 using that code ... can't figure out why using your expression using  durations of 3:26 and 3:21 yields 210 and 3:32=0     stumped 

[blgentry]

^ Hmm, not sure what to say.  I just checked and my entire library has normal looking values using Duration Rounded.  3:32 rounds here to 210.

I wonder if there's something odd about the math on some of those decimal representations of [Duration].  Here's one more you can try, but it shouldn't make a difference:

Code: [Select]

math(formatnumber(math(formatnumber([duration,0],0)/10), 0) * 10)
Brian.

[Arindelle]

hey that worked ... have just 3 files that come up with zero (out of 105K!) and they are under 4 seconds long and I don't care.

Thanks much brian 

[blgentry]

Well this got way off track...

If anyone is interested I'll post the view I've built.

Brian.

[ohbothered]

This is a great idea and I love the View! Are you going to be putting up a little tutorial on how to build it? I'd love to re-create that, but my JRiver skills aren't all that comprehensive right now.

Also, I'm trying to weed out duplicate video/tv files. Is it possible to search by media content, say, comparing frames at 3 seconds in, 10 seconds in, 13 seconds, then 1 minute? And, then compare to fields and duration, etc?

Thanks!

[Music_Man]

Yes Brian, a post of your final view would be appreciated....

[blgentry]

This is a great idea and I love the View! Are you going to be putting up a little tutorial on how to build it?

Sorry I missed this when it was posted.  I think I was out of town and didn't see it when I returned.

Also, I'm trying to weed out duplicate video/tv files. Is it possible to search by media content, say, comparing frames at 3 seconds in, 10 seconds in, 13 seconds, then 1 minute?

I don't know of any way to do that in MC.  It might be possible, but I don't know how.

I'm attaching a copy of the view to this message.  To use it, you'll need to load this as a new view.  You'll also probably want to create a library field called "Not Duplicate" with Data type Integer, and Edit type Check.

Someone should probably write up a short tutorial on how to load views.  The basic idea is to put JVI files, like the one I've attached here, in the Saved Views directory of your Library.  Your library is located at:

https://wiki.jriver.com/index.php/Library#Location_on_Disk

After you put the JVI file there, when you create a new view, you should see all of the standard templates, and at the bottom you should see new ones for any JVI files you have put into your Saved Views folder.  This one is called Duplicate Finder.  Just click on it and create the new view.   I know this isn't a lot of detail, but I'm not feeling up to writing much more.  Maybe when I'm feeling better.

Good luck.

Brian.

[RoderickGI]

Just bumping this to say I will install your view Brian, when I get a bit of time to do so. Not that it will take long to do, but I have to wind up my brain first!

[8139david]

Thanks, Brian.

For some reason, it only worked if I put your view there:
C:\Program Files (x86)\J River\Media Center 21\Data\Saved Views

It didn't work there:
C:\Users\<user name>\AppData\Roaming\J River\Media Center 21\Library

Edit: It's because I hadn't saved any view. Once I did it, I saw that they are put there by default:
C:\Users\<user name>\AppData\Roaming\J River\Media Center 21\Library\Saved Views

Personnally, I'm using this Duplicate view (see pic).
It finds exact duplicates in Audio.

PS: Using it plus variations on yours, I did find some legitimate duplicates

[8139david]

I wonder if you have more views to share, Brian?
It would be great to see them.

[marko]

Nice work Brian.
For me, this is why MC is so good. JRiver do not say "Here's a duplicates view for you to use" They say "Here are the tools for you to create your own views"

The attached is how I do mine.
It's playlist based rather than library field based. I use the four panes to the right of "Status" to see how much of an album I've actually 'hidden', should I be curious about that. I added them there as I'm only ever curious when I'm working on dupes and thought it better to have the info in the same view rather than switching away to another.
The "Possible Duplicates" returns only files that I have not worked on, basically, excludes those on the 'hidden' or 'allowed' playlists, then checks what's left for duplicates. I just go in and clear that up every few months or so.

Select file(s) and tick against the desired playlist to add those files.
After, whenever building a smartlist or view where I don't want those marked as dupes to be played, adding the custom rule "-p=hidden" takes care of that.
"Allowed" is used for live recordings, accoustic versions and the like, where either or will do, and so allows the "Possible Duplicates" entry to work as desired, namely, listing any new duplicates that have arrived since the last time I cleared them up.

I could probably do this just as well using list placed, inline check box library fields, but they didn't exist when I set this up, and as it wasn't broken, and there's no real difference in either approach, I just kept what I had.

My head couldn't cope with the "Not Duplicate" that way around I would have to reverse that so the ticks were against the files I considered duplicates. Always fascinates me how we are all wired slightly differently like that...

[blgentry]

I wonder if you have more views to share, Brian?
It would be great to see them.

Not a lot that most people would be interested in.  But you might enjoy my Cover Art Review view.

Here's a screen shot of it.  If you like it, you can download it.

Both taken from this thread here:

https://yabb.jriver.com/interact/index.php?topic=103486.msg719041#msg719041

Brian.

[8139david]

Nice, thanks Brian!

[8139david]

I thought you had some interesting Views for classical music as well?

[blgentry]

I know very little about Classical music cataloging and organization.  My Classical collection is very very small.  I think Listener is the one who has posted his classical view several times.  But my memory could be off.

Brian.