Jump to content


Photo

Air Video Server HD 2.1.1-beta1


  • Please log in to reply
29 replies to this topic

#21 Sunrise

Sunrise

    Advanced Member

  • Members
  • PipPipPip
  • 116 posts

Posted 21 March 2015 - 07:08 PM

Sunrise, we don't really support H.264 in AVI. Putting H.264 in AVI is a bad idea. The presentation timestamps are missing in such case and our pipeline doesn't currently restore PTS for H.264 (since in *all* other containers the timestamps are present). I'd suggest to remux the file to MKV using mkvtoolnix, after which the file should play.

 

Yes, you're right, that didn't really cross my mind, I was just a bit confused that VLC and LAV played it fine, but I didn't really realize the fact that it's actually H.264 in AVI and not MP4 (you can see in my above post that I even confusedly wrote .MP4).

 

H.264 in AVI is a bad idea, I fully agree, not sure why such files even exist in the first place.

 

BTW, I have just seen in the ffmpeg changelogs that ffmpeg now directly supports NVENC from Nvidia. Does this mean (were you using ffmpeg directly in combination with gstreamer or was is libav?) that we also get Nvidia Hardware Encoder support sometime in the future?



#22 whatUwant

whatUwant

    Advanced Member

  • Members
  • PipPipPip
  • 44 posts

Posted 22 March 2015 - 03:52 PM

This is most likely related to change in how we handle stale bookmark data. It seems that OS X handles the bookmark data for your ramdisk as stale. How exactly did you create it? I'll see if i can reproduce it.

I use this program to automatically create a ramdisk on start up http://inkscribbles.com/apps/tmpdisk/



#23 Sunrise

Sunrise

    Advanced Member

  • Members
  • PipPipPip
  • 116 posts

Posted 24 March 2015 - 07:52 PM

I have a question that I am wondering about for quite a while now.

 

What does the "Connection Speed Test" actually measure? Is it HTTP download speed to the router or to the source PC?

 

I am asking, since I get a relatively stable 30.000 Kbps throughput from where my AppleTV 3 is located when doing the test on my iPhone 5s. However, at the very same time, when there is a lot of WiFi activity around my house, I sometimes have problems to sustain the 10MBit or even the 5MBit preset, even though the throughput test shows me a lot higher values (always between 20-30.000 Kbps), which also vary when I do the test several times in succession.

 

As such, I get the impression that it's not really a good reliable indicator that sufficient bandwidth is available to stream at the preset I currently selected. Do you understand what I mean?

 

Could this measurement be improved somehow, e.x. by using a dummy file at the source and mimic a real world playback for like 10 seconds? Maybe you have some other idea that could improve it, I just wanted you to know, since today I had the very same symptoms appearing and it's a bit confusion from an end users standpoint.



#24 admin

admin

    Administrator

  • Administrators
  • 2582 posts

Posted 24 March 2015 - 08:07 PM

Hi,

 

the throughput test measures actual throughput between server and client, which means it generates several megabytes of random data and tests how long does it take to transfer the data. So in most situations it should be quite accurate. The problem often is that WiFi throughput can get rather variable. So throughput for the small period of time while the test is running is high enough, but the network can sustain the throughput for longer periods of time.

 

If possible I'd recommend connecting both the computer and Apple TV using ethernet. When there is transfer between two WiFi devices, the bandwidth gets divided. So just by connecting your computer to ethernet you basically double the speed. Same goes for Apple TV.



#25 Sunrise

Sunrise

    Advanced Member

  • Members
  • PipPipPip
  • 116 posts

Posted 26 March 2015 - 02:56 PM

Yes, it seems in this case using DLAN or connecting directly over RJ45 ethernet is the only solution then. Luckily this is not my main place to stream, I just wanted to be sure that the "Connection Speed Test" is as accurate as possible (which according to your explanation it is, because it already creates random data, basically emulating real-world playback).

 

One more question:

 

I read today that the coder of LAV filters has included a newly written open-source DTS decoder (see the links), which not only can decode every variant of DTS, but can now also decode DTS-MA streams perfectly and losslessly. He also wrote a wrapper for ffmpeg so others using ffmpeg can use it, too. Can you include that in an upcoming release so that Air Video HD can use the higher quality DTS-MA tracks to transcode (instead of only the core)?

 

Links:

https://github.com/foo86/dcadec (decoder developer)

http://forum.doom9.o...458#post1714458 (details about decoder included now in LAV filters)



#26 micleeso

micleeso

    Member

  • Members
  • PipPip
  • 12 posts

Posted 27 March 2015 - 08:29 PM

Does this version support video created by foscam h.264 cameras? The files they produce are playable on multiple players. An example would be VLC that is on the same host AVS is executing on.



#27 admin

admin

    Administrator

  • Administrators
  • 2582 posts

Posted 27 March 2015 - 08:30 PM

Really depends on stream details. If you upload sample file here

 

https://www.mediafir...54ce133b8c0cb0b

 

I can take a look.



#28 admin

admin

    Administrator

  • Administrators
  • 2582 posts

Posted 27 March 2015 - 08:34 PM

Yes, it seems in this case using DLAN or connecting directly over RJ45 ethernet is the only solution then. Luckily this is not my main place to stream, I just wanted to be sure that the "Connection Speed Test" is as accurate as possible (which according to your explanation it is, because it already creates random data, basically emulating real-world playback).

 

One more question:

 

I read today that the coder of LAV filters has included a newly written open-source DTS decoder (see the links), which not only can decode every variant of DTS, but can now also decode DTS-MA streams perfectly and losslessly. He also wrote a wrapper for ffmpeg so others using ffmpeg can use it, too. Can you include that in an upcoming release so that Air Video HD can use the higher quality DTS-MA tracks to transcode (instead of only the core)?

 

Links:

https://github.com/foo86/dcadec (decoder developer)

http://forum.doom9.o...458#post1714458 (details about decoder included now in LAV filters)

 

Well, WiFi speed variations is the single most frustrating thing when playing high bitrate video over local network. I'll take a look at the decoder. Maybe we could (optionally) increase directshow priority for DTS so that the server prefers directshow decoder for DTS (just like we do for VC1)



#29 micleeso

micleeso

    Member

  • Members
  • PipPip
  • 12 posts

Posted 01 April 2015 - 04:30 PM

Really depends on stream details. If you upload sample file here

 

https://www.mediafir...54ce133b8c0cb0b

 

I can take a look.

Did you recieve the file?



#30 admin

admin

    Administrator

  • Administrators
  • 2582 posts

Posted 05 April 2015 - 04:36 PM

Beta 2 is out

 

http://forums.inmeth...r-hd-211-beta2/






0 user(s) are reading this topic

0 members, 0 guests, 0 anonymous users