Talk:Video Chat: Difference between revisions

From OLPC
Jump to navigation Jump to search
No edit summary
Line 23: Line 23:


A quote from the page says "Version 9 of Videochat has 640x480 resolution and 7.5 frames per second. This results in about 400kBs throughput. We are looking how we can reduce this. (Pia Waugh - Feb 2009)". I'm currently working on a compression scheme which uses techniques from machine learning/pattern recognition to compress video/audio. It works on a sliding scale and would be able to compress to a greater or lesser extent based on the available cpu time. I'm happy to release the code, once it is completed, under the MIT license and would be more than happy to help integrate it. It's written in C++. If this would be useful you can contact me via email on Da.Waterworth AT googlemail.com - Daniel Waterworth 31/8/09.
A quote from the page says "Version 9 of Videochat has 640x480 resolution and 7.5 frames per second. This results in about 400kBs throughput. We are looking how we can reduce this. (Pia Waugh - Feb 2009)". I'm currently working on a compression scheme which uses techniques from machine learning/pattern recognition to compress video/audio. It works on a sliding scale and would be able to compress to a greater or lesser extent based on the available cpu time. I'm happy to release the code, once it is completed, under the MIT license and would be more than happy to help integrate it. It's written in C++. If this would be useful you can contact me via email on Da.Waterworth AT googlemail.com - Daniel Waterworth 31/8/09.

== Video for Deaf users ==

Just stating the probably obvious: Sign language requires a relatively high frame rate and resolution. Choppy / blurry video will make sign-based communication close to impossible. A possible way to mitigate some of the effects would be to offer a turn-taking option so that only one signer is broadcasting at a time, though that's not very natural.

Revision as of 20:49, 18 February 2010

The digital whiteboard functionality would be especially useful as a literal chalkboard replacement. If the activity supports a broadcast mode, then a teacher could use her XO as a whiteboard, and all the children could watch on their own screens. This even solves one of the problems of chalkboards: not being able to face your audience and draw at the same time. Additionally, if it is possible to record the drawing over time simultaneously with audio, this would provide an easy way to record a lecture for later playback. Such functionality would also be useful to students, to review the lecture material while doing an assignment. Finally, allowing students to draw on top of the whiteboard without rebroadcasting their modifications would permit a kind of note-taking/personalization of the lecture. Ben 11:46, 18 May 2007 (EDT)

api frame buffer for pre and post processing

direct pixle buffer on video stream for pre and post processing, including subtraction of background, visual target identification (object), and related capabilities with injection into stream

more chat power needed

Id love to see/use this! Is it working / the latest repository change says the activity version was upgraded, but I dont see the xo for it... 22:09, 4 April 2008 (EDT)

help me

cant download video chat i dont know how to talk to someone through this & i would like to know.

Connect for video chat with distant parents/friends

Request a link to connect to other system existing video chat services and graceful bandwidth degrade with those services.

Story version: Children with parents/family who have had to emigrate, for example a family with father moved to another country to remit money home, would appreciate ability to connect to video chat with that family member who is using a different machine and under the child's local bandwidth limit. This includes children too small to type back and forth, and parents who are sitting in some internet café somewhere. Right now this is happening somewhat with cell phones.

Compression

A quote from the page says "Version 9 of Videochat has 640x480 resolution and 7.5 frames per second. This results in about 400kBs throughput. We are looking how we can reduce this. (Pia Waugh - Feb 2009)". I'm currently working on a compression scheme which uses techniques from machine learning/pattern recognition to compress video/audio. It works on a sliding scale and would be able to compress to a greater or lesser extent based on the available cpu time. I'm happy to release the code, once it is completed, under the MIT license and would be more than happy to help integrate it. It's written in C++. If this would be useful you can contact me via email on Da.Waterworth AT googlemail.com - Daniel Waterworth 31/8/09.

Video for Deaf users

Just stating the probably obvious: Sign language requires a relatively high frame rate and resolution. Choppy / blurry video will make sign-based communication close to impossible. A possible way to mitigate some of the effects would be to offer a turn-taking option so that only one signer is broadcasting at a time, though that's not very natural.