On Thu, May 1, 2014 at 1:42 PM, James N. hart <
james.h...@schneider-electric.com> wrote:
> Thanks for the response Dan, see inline.
>
>
> On 4/30/2014 11:55 AM, Dan Dennedy wrote:
>
> On Wed, Apr 30, 2014 at 10:19 AM, Dan Dennedy <d...@dennedy.org> wrote:
>
>> Why not simple math?
>> 200 - 120 = 80
>> qmelt 1.mov out=80 -blank 120 -consumer ...
>> yields, actually, 202 frames because out points are frame numbers (or
>> time value), which are are 0-based and not durations. Therefore, out=80 is
>> a duration of 81 frames. Thus, to really get 200 frames you need to do:
>> qmelt 1.mov out=79 -blank 119 -consumer ...
>>
>> I didn't realize it was 0 based, makes sense though, thanks. Are
> you also saying that it can take a time value? I didn't know that, what is
> the format of the time value the out and in properties can take?
>
>
This is linked to from Documentation > Miscellaneous > Time Values:
http://mltframework.blogspot.com/2012/04/time-properties.html
And in API reference, under mlt_time_format:
http://www.mltframework.org/doxygen/mlt__types_8h.html#a088f972a1bd4a9cbc2d9125958bd1325
Shotcut's project file format is MLT XML with time clock values.
> In the long run we should be able to do the math. Right now we don't have
> anything which will easily give us the number of frames, or time length, of
> the video we are using in the construction. I'm basically trying to cheat
> for demo purposes but
>
MLT can give you that, of course. You probably do not want to use the API
right now, so you can parse the XML output of 'melt file -consumer xml'
which outputs to stdout. If you want to output to a file: -consumer
xml:filename.
> limiting the final video size was also going to be a feature. It is
> looking like that would be difficult only dealing with command line
> generation from everything you have indicated.
>
> The framework and modules, by design, do not provide much in the way
>> of authoring as much as they simply facilitate it. You need to make your
>> own layer that generates the composition as melt command lines, MLT XML, or
>> by custom program using the API.
>>
>> Right now I have a DSL builder in Java which makes creating the
> complex command lines easy. I know it probably isn't the correct solution
> for the long run, but so far it has been the quickest to prove out for 90%
> of what we have to do.
>
>
>
> Now, with that said, and with your claim that you just need a simple way
> to limit the number of frames in the output, if the consumer honored an out
> property or melt let you target the track or tractor properties, then you
> could accomplish that. I am just not sure how you want to determine the
> length of the video clip vs. the number of blank frames.
>
> For the demo these are hard coded and passed to the builder directly.
> Ultimately I need the clip's meta information and the immediate need for
> limiting the output will be gone. I wanted the size of the clips, in the
> demo, not to matter but it looks like I will need to do the math ahead of
> time and pass out parameters for each clip.
>
>
> melt 1.mov -blank 119 -consumer xml:test.mlt no_meta=1
> edit test.mlt so that playlist out=199
> melt test.mlt -consumer avformat:demo.mp4
>
> So, the solution may be to generate MLT XML, perhaps with a template
> engine, instead of complex melt command lines. You can even put a consumer
> element into the XML:
> <consumer mlt_service="avformat" target="demo.mp4"/>
> Add any other encoding properties as XML attributes.
>
> Yes, I think this is essentially where we will end up as soon as we can't
> exploit the command line directly for a requirement. I would start there,
> but we probably don't have time to create and test the xml generator before
> release.
>
------------------------------------------------------------------------------
"Accelerate Dev Cycles with Automated Cross-Browser Testing - For FREE
Instantly run your Selenium tests across 300+ browser/OS combos. Get
unparalleled scalability from the best Selenium testing platform available.
Simple to use. Nothing to install. Get started now for free."
http://p.sf.net/sfu/SauceLabs
_______________________________________________
Mlt-devel mailing list
Mlt-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/mlt-devel
