Posts: 112
Joined: Jan 2012
Reputation:
3
Thanks janbar. I been at work all day but I just managed to get you a log of a show transition. It freezes the frontend while the backend continues to record. I sent the log via email.
Posts: 112
Joined: Jan 2012
Reputation:
3
2012-10-09, 07:53
(This post was last modified: 2012-10-09, 07:56 by Aubrien.)
Janbar,
I compiled a fresh XBMC as well as your current master with all of your changes from this evening and it looks like we are almost there. As you know, the channel change delay is not in your master but the buffer delay is. As I was hoping, the buffer delay code also fixes the channel change delay since it is the same thing as the delay between shows. I can now tune and it fills the buffer until it can start playback.
I am having a couple issues still though. When I start playback from nothing it can take 45s for it to tune while it very very slowly fills the buffer. That is way higher than it should take. Once tuned, the buffer is really low and the slightest disturbance will make it buffer. When it is having to re-buffer, it is really slow (1% every 2s or so). As soon as it starts playback again (it will start playback when the buffer is at ~10%) it has to buffer again in about for 1s for about 15s. The cycle just continues..buffer 15s...playback for <1s...loop. It is buffering over 95% of the time. I think there is something making the buffering code be very slow which is causing all of the issues I am seeing.
It looks to be buffering when it needs to now, just at a very slow pace. Once we get the buffering working at a regular rate I may have additional feedback, but it is hard to do much since it has to buffer almost the entire time. An additional idea may be to have a min buffer size, so if it does have to buffer it will grab at least a certain amount of data before starting playback again. It may increase tuning time to have no buffer, but it is just too unstable imo. I'm not even talking about a big min buffer size either. Just something though. I feel like as soon as it has a byte of data it plays it back and re-empties the buffer again causing the slow re-buffer.
As always, thanks for the hard work. I can tell we are almost there.
Posts: 53
Joined: Apr 2012
Reputation:
0
Hey jskube,
What kind of channel change time are you seing with this new build? is it comparable to mythfrontend?
Thanks,
Kevin
Posts: 397
Joined: Jan 2011
Reputation:
3
Could someone compile a version for Windows 7 (64 bit)? I'd love to test this out on my box.
HTPC(s): All running LibreELEC
- AMD 2200G APU on Gigabyte AB350N-Gaming WIFI-CF
- RPI3 x2 | RPI2 x2
NAS: FreeNAS (Latest Stable) | NFS/CIFS
Posts: 112
Joined: Jan 2012
Reputation:
3
I see alot of you are using the Ceton cards and HDHomeRun. I can't use those because DISH Network doesn't have CableCards. Actually, only terrestrial cable is required to provide those so nobody with sattellite can get an mcard. But before I switched to DISH for all of the extra channels and lower prices I was looking into one of these CableCard tuners when I had Time Warner and Comcast. I was a little reluctant to invest in them though since you can only use channels that are flagged as copy freely. It also looked like any channel labeled as copy freely can change at any time if the cable company wants to as well. There is basically no promises that it will work tomorrow. I didn't like the sound of that. Also all of my premium channels wouldn't be able to be tunned to with those capture devices using xbmc and mythbackend since they are flagged copy once/never. Are any of you experiencing this issue? For example can you watch HBO, Cinemax, Shotime, etc. with this setup? I'm currently using the analog loophole with a set top box and HD PVR to strip the DRM, but I have a friend with TWC asking what card to get. If I tell him that he cant get some channels to work without being forced to do playback through an "approved media extender" (not xbmc/mythbackend) then he wont be happy. Thanks