[
https://issues.apache.org/jira/browse/FLEX-33374?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Erik Thomas updated FLEX-33374:
-------------------------------
Description:
The spark VideoDisplay on Mobile devices delays streaming video from FMS 4.5
for about a minute before it begins playing.
I've attached a project that displays this behavior. Run it in AIR on your
desktop or in a simulator, then run it on a device that's using WiFi with the
same bandwidth as your computer to see the difference.
This is probably a problem in OSMF but I'm posting it here for visibility to
the problem.
UPDATE 2/1/2013 - I wrote my own video controller and display classes that just
use NetConnection, NetStream, and Video classes. This is a more optimized
solution for mobile anyway, to not use heavyweight OSMF classes underneath the
spark VideoPlayer and VideoDisplay classes. So I'm not waiting on this fix and
suggest anyone building a mobile video player should do what I did anyway. Of
course we have to do our own bandwidth testing which requires a server-side asc
on FMS, but that's actually working better than VideoDisplay's adaptive
bitrates anyway that resize the video way too often.
was:
The spark VideoDisplay will not stream video from FMS 4.5, instead it appears
to do progressive download instead, and it buffers the file for about one
minute before it begins to play.
However, the same code running in FlashPlayer on web, or in an AIR mobile
simulator, the video properly streams and displays within seconds of playing.
I understand VideoDisplay is not "optimized" for mobile, but that should not
cause this behavior. I am in the process of debugging the spark codebase to
find the root cause, but so far have been unsuccessful.
So I attached a project that displays this behavior. Just unzip the file to a
directory, start FB and point the workspace to the directory, open the project,
build and run it first in an AIR simulator for an iOS mobile device, you'll see
video instantly. Then deploy it in debug or release to an actual device. When
you launch the project, you have to wait approximately a minute before the
video will start playing.
I'm assuming it's failing over to progressive download from streaming in some
way, but I really have no idea what is causing the issue.
UPDATE 2/1/2013 - I've downgraded the priority on this since I wrote my own
video controller and display classes that just use NetConnection, NetStream,
and Video classes. I guess for mobile that's the more optimized solution
anyway, as VideoDisplay and VideoPlayer seem pretty heavyweight in comparison.
So I'm not waiting on this fix and suggest anyone building a mobile video
player should do what I did anyway. Of course we have to do our own bandwidth
testing which requires a server-side asc on FMS, but that's actually working
better than VideoDisplay's adaptive bitrates anyway that resize the video way
too often, so it's really a win not to depend on OSMF.
> spark VideoDisplay and VideoPlayer fails to stream video on mobile devices
> --------------------------------------------------------------------------
>
> Key: FLEX-33374
> URL: https://issues.apache.org/jira/browse/FLEX-33374
> Project: Apache Flex
> Issue Type: Bug
> Components: Spark: VideoPlayer
> Affects Versions: Adobe Flex SDK 4.6 (Release)
> Environment: FlashBuilder 4.7 on Windows 7, iPad 3, iPhone 4, remote
> FMS 4.5 server, FLV Video source in 112K bitrate.
> Reporter: Erik Thomas
> Priority: Minor
> Attachments: VideoDisplay.jpg, VideoPlayback.zip
>
>
> The spark VideoDisplay on Mobile devices delays streaming video from FMS 4.5
> for about a minute before it begins playing.
> I've attached a project that displays this behavior. Run it in AIR on your
> desktop or in a simulator, then run it on a device that's using WiFi with the
> same bandwidth as your computer to see the difference.
> This is probably a problem in OSMF but I'm posting it here for visibility to
> the problem.
> UPDATE 2/1/2013 - I wrote my own video controller and display classes that
> just use NetConnection, NetStream, and Video classes. This is a more
> optimized solution for mobile anyway, to not use heavyweight OSMF classes
> underneath the spark VideoPlayer and VideoDisplay classes. So I'm not waiting
> on this fix and suggest anyone building a mobile video player should do what
> I did anyway. Of course we have to do our own bandwidth testing which
> requires a server-side asc on FMS, but that's actually working better than
> VideoDisplay's adaptive bitrates anyway that resize the video way too often.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
