openmeetings-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "" <>
Subject Re: GSoC project ideas wanted
Date Mon, 11 Feb 2013 22:16:12 GMT
Regarding point No2:
Compression ratio in the screensharing is somehow limited to the available
codecs we can use.
We can use SHA1 and SHA2 (Adobe Screen Sharing Codec).
I think we currently use SHA1, SHA2 might be a nice project for a student
that is interested in coding some codec.
And I think we can provide a reference implementation somewhere.
I think this project has good chances and a well defined scope.

@Daniel: Screenshot tool is also a nice idea, but it would end up to be
part of the screensharing Web-Start application. Is that what you intend to


2013/2/12 <>

> Hi Ed,
> thanks for your ideas.
> About idea no 1:
> That is an interesting idea, however it won't be possible that you provide
> a "free to choose" bandwidth for each user.
> The background is: Every stream that any client consumes has to be created
> somewhere.
> So what could be realized is that every stream that is broadcasted from
> one user via webcam to Red5/OpenMeetings will be re-transcoded into
> multiple streams (high, middle, low) bandwidth.
> So there might be some limitations to that:
>  - "high" quality will never be better then the original material. We
> can't make a picture better then the original. So all re-transcoding will
> only make the original to lower quality, never to higher.
>  - Re-transcoding has to happen on the server side (and number of streams
> are limited, we can't provide a stream on the required bandwidth
> "on-demand" for each user, or only with very big effort)
>  - it will require real-time transcoding on server side which is possible
> with FFMPEG and some integration into Red5. But we would need a very
> specialized student that is keen and very motiviated as there is hardly any
> documentation on that available in the internet.
> What a project makes a success is if all participant know the potential
> outcome and the tools and methods that are needed to realize that. I would
> be happy to put this project on our list but it will be difficult to find
> somebody with the needed skills.
> Sebastian
> 2013/2/12 BBS Technik <>
>> Hi all,
>> I think, one of the gratest liminations for satisfactory video
>> conferencing with om is the limited bandwidth of internet connections of
>> the clients .
>> Therefore I would like to suggest the following ideas for a GSoC project :
>> 1. The image size of the videos transferred from the om server to the
>> clients should be adapted to the video window size set in the recipient
>> client.
>> Thus the recipient client itself could influence the transferred amount
>> of data to it.
>> Then all the participants achieve the best possible result for them.
>> 2. A second proposal concerns that the screensharing  bandwidth
>> requirements has an great impact on the overall quality of the video
>> conference.
>> Here, in a project the existing function of sreen sharing could be
>> expanded and enhanced.
>> For example, the possibility for the transfer on only one application
>> window, regardless of its size.
>> Or the possibility of shared browsing with a locally installed browser.
>> Moreover, certainly an improvement of the used compression method would
>> be a very good project topic.
>> I would be  happy if the subject of bandwidth consumption would plays a
>> role in the selected GSoC project .
>> Best regards
>> Ed
>> -------- Original-Nachricht --------
>> > Datum: Tue, 12 Feb 2013 08:44:54 +1300
>> > Von: "" <>
>> > An: dev <>
>> > CC:
>> > Betreff: GSoC project ideas wanted
>> > Google Summer of Code is about to start soon!
>> > Google sponsors every student with 4500USD. Plus 500 for the Apache
>> > Foundation.
>> >
>> > We are searching for ideas what porential students can do.
>> > Ideas from Non-Developers are welcome too!
>> >
>> > We will add the ideas to JIRA then with a special label so students can
>> > find it.
>> >
>> > Sebastian
> --
> Sebastian Wagner

Sebastian Wagner!/dead_lock

View raw message