0

Problem with RVSequence example in docs

Hi,

I'm trying to follow an example from the docs ( 13.7.1 RVSequence ) - I am using RV 3.10.

When I save the code as a .rv file and run it with rv, I would expect it to construct an edit for clips a and b in this order: ababb. Instead it just plays a clip from a followed by a clip from b. It seems to take some of the data from the sequence node, but node all of it. I am I misunderstanding what the sequence node should do, or am I doing something wrong? Any help would be much appreciated!

 

sourceGroup000000_source : RVSource (0) { media { string movie =  "a.mov" } }
sourceGroup000001_source : RVSource (0) { media { string movie = "b.mov" } }
connections : connection (1)
{
evaluation
{
string lhs = [ "sourceGroup000000"
"sourceGroup000001"
"sourceGroup000002" ]

string rhs = [ "sequence"
"sequence"
"sequence" ]
}
}

rv : RVSession (1)
{
session
{
int[2] range = [ [ 1 40 ] ]
int marks = [ 1 11 21 31 40 ]
float fps = 24
string viewNode = "sequence"
}
}

sequence : RVSequence (0)
{
edl
{
int frame = [ 1 11 21 31 41 ]
int source = [ 0 1 0 1 1 ]
int in = [ 1 1 11 11 21 ]
int out = [ 10 10 21 21 21 ]
}

mode
{
int autoEDL = 0
}
}

2 comments

  • 0
    Avatar
    Seth Rosenthal

    Hey Henry,

    OK, the short answer is that the example you are looking at will work in more recent versions of RV (3.12.13 or newer). It also used to work in older versions of RV (3.8 generation). It was not used much and some of the upgrades in 3.10 that made RV much more flexible unfortunately overrode the behavior--we restored it in 3.12.13.

    The expected behavior is not quite what you are expecting. It should be ABAB. There is not an extra 'B' on the end. The last column of numbers in the EDL are there to define the end of the last clip. So if you look at the frame item data:

    int frame  = [  1 11 21 31 41 ]

    it basically means that the sequence will have clips:

    Frame 1 to 10

    Frame 11 to 20

    Frame 21 to 30

    Frame 31 to 40

    This documentation is quite old and we see now that it has a few other mistakes in it. We'll clean it up for the next release. But for now here are a few points:

     

    1. This is not a complete gto file. To make it into one, it needs the first line to say:

    GTOa (3)

    2. The edl.out data is not 'wrong' but it's a bit misleading. The data in the frame property defines the length of the clips in the sequence. In this example, the in/out points for the third and fourth clip are 11-21. This would be an 11 frame clip, but the actual clip lengths are determined by the 'frame' property and so you get 10 frames for each of those clips.

    3. There is a cool and useful thing you can do when testing rv sessions this way which is to use procedural media so you don't have to make temp movies for a.mov b.mov etc. RV does this something called a 'movieproc' which uses the filename to generate media on the fly. In this case, you don't actually need any media files. So the sources could look like this:

    sourceGroup000000_source : RVSource (0) { media { string movie = "solid,red=1.0,start=1,end=100.movieproc" } }
    sourceGroup000001_source : RVSource (0) { media { string movie = "solid,blue=1.0,start=1,end=100.movieproc" } }

    You can read more about .movieproc here: http://www.tweaksoftware.com/static/documentation/rv/current/html/rv_manual.html#toc-Subsection-13.1.7

    I hope that helps. Let us know if  you have more questions.

    Cheers,

    Seth

  • 0
    Avatar
    Alan Trombla

    Hi Henry,

    Sorry for the confusion in the docs.  I've update the online versions to hopefully accurately represent RV's behavior in this case.  Those are here.

    Cheers,

    Alan

Please sign in to leave a comment.