How can the buffer size be reduced? The server is on the local network so it's difficult to see a use for it and when browsing channels the wasted time adds up pretty quickly.
|
3 hours ago
Where exactly are you talking in the web app? Are you talking about browsing a specific page, or starting video, or something else?
The web app in general seems to load quickly here.
3 hours ago
Yes, the pages load normally, it's when starting the video. I was just timing the loading and a SD channel takes more than 10 seconds but noticed that it is not fixed and the counter sometimes goes to 3 mega but other times goes to 4, with the same channel.
3 hours ago
(This post was last modified: 3 hours ago by mvallevand.)
I think he means the trranscoding buffer with the data byte count incrementing at the bottom when direct play is not available. I expect SD gets transcoded to HD.
Martin
3 hours ago
pirole, how quick does video start for you via this page: http://localhost:8866/uiclient.html
Yes, that's it.
On that client I didn't notice much difference, around 10 seconds, more if the channel belongs to a different frequency than the previous. There is no counter so I can't start counting when the data starts loading (if that's what the counter counts). If there were no buffer the video would load closer to the speeds of the non transcoding apps.
3 hours ago
I know that the webapp works on ts fragments but I also know that it is possible to encode mp4 in a way that it can be played right after receiving the header, if that is the issue.
3 hours ago
You cannot generate fast start mp4 for Live TV NextPVR already allows post processing mp4 for recordings.
Martin
3 hours ago
Ah ok. Thanks anyway. Still great as it is.
3 hours ago
You could try changing <AnalyzeDuration>default</AnalyzeDuration> the ffmpeg default is 5000000 This could impact other factors I think 10 seconds is pretty good for transcoded SD which will rarely be web format a 1 or 2 second change won't change much, just use a direct play client when you need more speed.
Martin |
|
|