GeForce Experience 2.1.1.0 issue on start

[dreadloke]

Hi,

First, i'd like to say how much I appreciate the work you've done yet, this is an amazing project.

Sadly, it seems like limelight-pi v0.8 is not compatible with the latest version of GeForce Experience. I get an RTSP OPTIONS request failed: 505 error on start at the RTSP handshake step.

I just wanted to report the issue, even if i'm guessing you're already aware of it.

Kind regards

[dreadloke]

My bad, just had a look at the XDA thread and read your comment there... Shame on me

[beakersoft]

Yeah I noticed that last night as well. Thought i'd broken something until I realized the driver had updated. Same error.

[dusda]

I'm getting the same :(.

java -jar limelight-pi.jar 192.168.1.96
Starting app
Starting RTSP handshake
java.io.IOException: RTSP OPTIONS request failed: 505
    at com.limelight.nvstream.rtsp.RtspConnection.doRtspHandshake(Unknown Source)
    at com.limelight.nvstream.NvConnection.doRtspHandshake(Unknown Source)
    at com.limelight.nvstream.NvConnection.establishConnection(Unknown Source)
    at com.limelight.nvstream.NvConnection.access$300(Unknown Source)
    at com.limelight.nvstream.NvConnection$1.run(Unknown Source)
    at java.lang.Thread.run(Thread.java:744)
RTSP OPTIONS request failed: 505
Connection Error Starting RTSP handshake failed

[cgutman]

It's fixed in the latest limelight-common. @irtimmer just needs to rebuild limelight-pi with it.

[H3llK33p3r]

That's right :). I just test and the communication is working :D. Just need to wait irtimmer because the version of irtimmer (his fork) is different than yours limelight-common repo xD (in my side the app boots but the screen is freezed, but it's the first time I'm looking to this code and my merge is not very good I think ;)).

[irtimmer]

The difference between my fork and limelight-common are very small (smaller buffers to avoid bufferbloat and prevent high latency). I've committed some changes but because I'm on vacation I can't test it. So just compile from source and hopefully it's works again.

[H3llK33p3r]

Same issue than with my version. The screen on the pi is freezed. The console prints many :"Your device is processing the A/V data too slowly. Try lowering stream resolution and/or frame rate." But I'm using 720p resolution and 30fps ;) !

May be someone else can try... (in case of the issue is only in my side)

[beakersoft]

Has anyone complied the fixes yet, just saves me having to do it! ;-)

[H3llK33p3r]

@beakersoft : http://dl.free.fr/ngFMQAZzI

[H3llK33p3r]

I think I have located the issue. The video renderer uses a pool of DecodeUnit Object. The pool (PopulatedBufferList) is initialized with the VideoDepacketizer but all objects are unstacked but no one is added in the free pool list (freeList attribute of PopulatedBufferList).

By adding "depacketizer.freeDecodeUnit(decodeUnit);" in the AbstractVideoRenderer thread, the video seems to be OK. Moreover le sound lag issue monitored seems to less important (see ticket https://github.com/irtimmer/limelight-pi/issues/26). It's may be due to the update of commons lib. I haven't test more than this.

@beakersoft : you can try the patch ? : http://dl.free.fr/tUlY6jCVA

[irtimmer]

I think that would indeed fix this issue (it was atleast missing) so I've merged it. If anyone else can confirm the issue is solved I can close it and release a new version.

[cgutman]

If you want, you can pull up to the latest common again to get a few changes I made to handle out of order packets, drop the bandwidth requirements back to GFE 2.1 levels, and properly disable resolution scaling which can cause the video decoder to crash (not sure if you handle that in your OMX renderer). It's up to you whether you want to pull those in. I'll probably do an Android release later today with them.

[cybern0id]

I just have to say; you guys are brilliant! Thanks for such a great app and for fixing the issues.

I can confirm that this video freezing issue is fixed with the recent video patch and as an added bonus, as @HellKeeper62 mentioned, the severe audio stuttering issues I was seeing have also been fixed (there are now only insignificant audio compression anomalies). My ticket #26 can be closed or I will close it if you want me to.