[theora-dev] variable frame rate for Theora

[Silvia Pfeiffer]

On Fri, Nov 14, 2008 at 6:35 AM, Pablo Rodríguez <oinos at web.de> wrote:
> Thanks for the answer, Ralph.
>
> Ralph Giles wrote:
>> On Tue, Nov 11, 2008 at 11:45 AM, Pablo Rodríguez <oinos at web.de> wrote:
>>
>>> [...]. But I'd ask you to consider implementing
>>> variable frame rates.
>>>
>>> The reason for that implementation would come from the much smaller size
>>>  in screencasts and presentations.
>>
>> We don't have any plans to add variable framerates to theora.
>>
>> There is a work around. Duplicating the previous frame is relatively
>> cheap, so an encoder could set something like 5 fps, but only update
>> the image when there's a change. The current encoder isn't smart
>> enough to do this automatically, but a screencast app could force it.
>
> Thanks for the workaround. What would it happen if the fps are set to 10
> or higher (100 or even 1000)?
>
>> The other point is that theora isn't the best format for most
>> presentations. I'd like to store them as an image stream, which can be
>> variable framerate, or use a queue track like cmml to trigger slide
>> advance in an adjacent viewer that holds the whole presentation. For
>> presentations and screencasts with a lot of motion or animation video
>> makes sense though.
>
> As I wrote on my previous message, I'm not a programmer and I'm not sure
> I understand the previous paragraph.
>
> Sure SWF, PDF or Moonlight would be better for presentations. But in
> some cases presentations should be delivered as movies (Blip.tv, Youtube).
>
>> Are you working on anything with screencasts and theora?
>
> No, I'm very interested in converting PDF presentations to movies with
> audio and synced slides (I have already done it with SWF, but I'd like
> to convert it to a movie using free codecs).
>
> How can I get that? (I have the audio, PDF document and slide sync times)

I think Ralph was referring to the fact that video is for moving
images and converting the slides into video is often times not
efficient because your an turning static images into moving ones, for
which the encoder will create difference frames and the like. So,
something like OggMNG or OggKate would be more useful to you.

OggMNG unfortunately doesn't have much tool support. Kate
(http://wiki.xiph.org/index.php/OggKate) is however getting there and
you can encapsulate the pdf text together with the images in kate. I
think mplayer and vlc have kate support. Try giving that a shot maybe.

Cheers,
Silvia.