Re: Score follower

[Mike Solomon]

Konstantin,

I have a metadata patch floating around that can help you kludge your way 
through this. It is not too hard, but will require some work:

1) Create a sort of metadata property for whatever grob you wanna trace.
2) for those grobs, create an override for the print method that first calls 
the correct print method and then uses the generated stencil for xy coordinate 
info, returning this stencil at the end of the function after having harvested 
its coordinates. You can have pretty prints to the command line of the grobs' 
metadata and coordinates, or you can write these to a file using Scheme.

If this sounds like Chinese, lemme know and I can try to cook up a simple proof 
of concept depending on what you want to do.

Cheers,
Mike

On Jan 5, 2011, at 11:40, Carl Sorensen <address@hidden> wrote:

> 
> 
> 
> On 1/5/11 8:44 AM, "Konstantin" <address@hidden> wrote:
> 
>> James <james.lowe <at> datacore.com> writes:
>> 
>> [snip]
>>> http://code.google.com/p/lilypond/issues/detail?id=965
>>> 
>>> I have tried this - I don't really use Linux but I managed to get this
>>> working sort (as my Linux machine is in a VM I used similar tools on a
>>> Mac), however I am not sure this is what you want?
>> [snap]
>> 
>> Hello James,
>> 
>> thank you for this link. The problem there is: the "timing points" are not
>> extracted from lilypond. They are manually created (with help of a python
>> script). Second problem: the smallest graphical object is a complete system,
>> but
>> I need graphical (and acoustical) informations about every single scores.
> 
> Unfortunately, the information you wish to have is not available.
> 
> If you will search the threads about musicXML output, you will see that the
> inability to couple musical information with graphical information is the
> main obstacle to musicXML output.
> 
> The music information is available in the stream events.
> 
> The graphical information is available in the final output.
> 
> At present there is no way to connect the two.  Ideas have been given before
> on the -devel lists.
> 
> I'm sorry there's not a better answer for you.
> 
> Carl
> 
> 
> _______________________________________________
> lilypond-devel mailing list
> address@hidden
> http://lists.gnu.org/mailman/listinfo/lilypond-devel