Author |
Message |
rando
|
Posted: Fri Feb 09, 2007 11:47 am |
|
Joined: Sat Feb 11, 2006 5:26 pm
Posts: 282
Location:
Winnipeg - Canada
|
Hi there, firstly I will apologize for not providing quite all of the relevant info here, I'm sure there is something I'm going to miss, and I'd be happy to provide aditional details as questions arise.
Here is a link to a small file (13mb) that demonstrates the problem I am about to describe.
Right Click me and Save as
This is an mpg2 file, I simply created cut points around the problem area and then used the lossless transcoder to chop out everything else. It's a 22second file, with my "problem" occuring at about 8 seconds.
This is something I'm seeing in several of my recordings, very often I will not get it, but it's happening frequently enough that it certainly is annoying, especially if it's happening at a key moment during a show. I don't really know how to describe it other than saying that it's "glitching" or "pixalizing" for a second. When this happens the audio ends up jumping forward a split second (as does the video).
My setup is posted below in my sig, but I'll just quickly recap here:
R5E50
IVTV - will be whatever version R5E50 puts in (I don't know how to check my current version)
2 X PVR 150 (one is MCE, the other is not)
320GB WD HD
Abit NF7 Socket A with an Athlon 3000+ and 1gb of RAM
I'm not sure if these glitches are always happening on shows recorded on one specific tuner. Is there a way for me to check which tuner a particular show recorded on? Is there any relevant info I can give from any of my log files? Is this an IVTV thing? or a hardware problem on my cards? or something altogether different? Any help is muchly appreciated.
_________________ Currently Running:
Too lazy to update this with my current hardware, I'll redo it during my next install =)
|
|
Top |
|
|
L0o0ky
|
Posted: Fri Feb 09, 2007 3:57 pm |
|
Joined: Mon Jan 16, 2006 4:08 pm
Posts: 86
|
All of this is prefaced by my "I'm a non-expert" warning. I had some similar issues. Here's the few possibilities I've explored.
You might try the following (search the forums for instructions if you don't know how).
http://mysettopbox.tv/phpBB2/viewtopic.php?t=13257 (this one seemed to have the some impact for me on some channels, but not all)
check here http://mysettopbox.tv/phpBB2/viewtopic.php?t=13108 if you haven't already.
Try aggressive buffering on your audio.
Try using the video as a timebase.
Try upping the bitrate you use when recording your video (this one seemed to have the biggest impact).
Make sure DMA is enabled.
Good Luck. If you find something in this list of suggestions that help, please let me know so we can spread the word.
L0o0ky
|
|
Top |
|
|
rando
|
Posted: Fri Feb 09, 2007 4:25 pm |
|
Joined: Sat Feb 11, 2006 5:26 pm
Posts: 282
Location:
Winnipeg - Canada
|
I have read up on the first link (the one with the settings for the temporal filtering) and both of my tuners are set and working properly as far as that is concerned. I did change that while watching live TV and could certainly see the difference, although that isn't what I'm seeing here. A good thought though.
I'll try the agressive buffering and the video as timebase, though it'll be tricky to nail down as this is a pretty infrequent glitch. Who knows though...
_________________ Currently Running:
Too lazy to update this with my current hardware, I'll redo it during my next install =)
|
|
Top |
|
|
sgunther
|
Posted: Fri Feb 09, 2007 4:38 pm |
|
Joined: Thu Apr 27, 2006 1:07 am
Posts: 54
|
Can you post the contents of; Code: tail -n30 /var/log/kern.log and Code: tail -n50 /var/log/mythtv/mythbackend.log
|
|
Top |
|
|
rando
|
Posted: Fri Feb 09, 2007 4:39 pm |
|
Joined: Sat Feb 11, 2006 5:26 pm
Posts: 282
Location:
Winnipeg - Canada
|
This particular recording happened far enough back that it won't have occured during the time period that would show up in those logs..... do you want me to try and grab a chunk from during the time of the recording, or are you looking for something else?
_________________ Currently Running:
Too lazy to update this with my current hardware, I'll redo it during my next install =)
|
|
Top |
|
|
rando
|
Posted: Fri Feb 09, 2007 4:44 pm |
|
Joined: Sat Feb 11, 2006 5:26 pm
Posts: 282
Location:
Winnipeg - Canada
|
Here you go:
Code: root@backend:/usr/bin# tail -n30 /var/log/kern.log Feb 8 12:00:34 backend kernel: ivtv0: All encoder MPEG stream buffers are full. Dropping data. Feb 8 12:00:34 backend kernel: ivtv0: Cause: the application is not reading fast enough. Feb 8 20:02:34 backend kernel: ivtv0: All encoder MPEG stream buffers are full. Dropping data. Feb 8 20:02:34 backend kernel: ivtv0: Cause: the application is not reading fast enough. Feb 8 21:01:34 backend kernel: ivtv0: All encoder MPEG stream buffers are full. Dropping data. Feb 8 21:01:34 backend kernel: ivtv0: Cause: the application is not reading fast enough. Feb 8 22:36:34 backend kernel: ivtv1: All encoder MPEG stream buffers are full. Dropping data. Feb 8 22:36:34 backend kernel: ivtv1: Cause: the application is not reading fast enough. Feb 8 23:37:34 backend kernel: ivtv1: All encoder MPEG stream buffers are full. Dropping data. Feb 8 23:37:34 backend kernel: ivtv1: Cause: the application is not reading fast enough. Feb 9 00:30:34 backend kernel: ivtv1: All encoder MPEG stream buffers are full. Dropping data. Feb 9 00:30:34 backend kernel: ivtv1: Cause: the application is not reading fast enough. Feb 9 01:07:34 backend kernel: ivtv1: All encoder MPEG stream buffers are full. Dropping data. Feb 9 01:07:34 backend kernel: ivtv1: Cause: the application is not reading fast enough. Feb 9 11:00:25 backend kernel: ivtv1: All encoder MPEG stream buffers are full. Dropping data. Feb 9 11:00:25 backend kernel: ivtv1: Cause: the application is not reading fast enough. Feb 9 11:00:26 backend kernel: ivtv0: All encoder MPEG stream buffers are full. Dropping data. Feb 9 11:00:26 backend kernel: ivtv0: Cause: the application is not reading fast enough. Feb 9 11:30:10 backend kernel: ivtv1: All encoder MPEG stream buffers are full. Dropping data. Feb 9 11:30:10 backend kernel: ivtv1: Cause: the application is not reading fast enough. Feb 9 12:00:09 backend kernel: ivtv0: All encoder MPEG stream buffers are full. Dropping data. Feb 9 12:00:09 backend kernel: ivtv0: Cause: the application is not reading fast enough. Feb 9 13:30:08 backend kernel: ivtv0: All encoder MPEG stream buffers are full. Dropping data. Feb 9 13:30:08 backend kernel: ivtv0: Cause: the application is not reading fast enough. Feb 9 14:00:08 backend kernel: ivtv0: All encoder MPEG stream buffers are full. Dropping data. Feb 9 14:00:08 backend kernel: ivtv0: Cause: the application is not reading fast enough. Feb 9 15:00:08 backend kernel: ivtv0: All encoder MPEG stream buffers are full. Dropping data. Feb 9 15:00:08 backend kernel: ivtv0: Cause: the application is not reading fast enough. Feb 9 16:00:08 backend kernel: ivtv0: All encoder MPEG stream buffers are full. Dropping data. Feb 9 16:00:08 backend kernel: ivtv0: Cause: the application is not reading fast enough. that doesn't look so good.... what does this mean? Code: root@backend:/# tail -n50 /var/log/mythtv/mythbackend.log 2007-02-09 16:33:03.415 Scheduled 347 items in 1.6 = 0.04 match + 1.54 place 2007-02-09 16:33:10.126 Reschedule requested for id 0. 2007-02-09 16:33:11.731 Scheduled 347 items in 1.6 = 0.08 match + 1.53 place 2007-02-09 16:36:07.507 MainServer::HandleAnnounce Monitor 2007-02-09 16:36:07.513 adding: backend as a client (events: 0) 2007-02-09 16:36:12.758 MainServer::HandleAnnounce Monitor 2007-02-09 16:36:12.763 adding: backend as a client (events: 0) 0: start_time: 0.036 duration: 33.156 1: start_time: 0.026 duration: 33.137 stream: start_time: 0.289 duration: 368.512 bitrate=5089 kb/s 2007-02-09 16:36:13.339 AFD: Opened codec 0x8411db0, id(MPEG2VIDEO) type(Video) 2007-02-09 16:36:13.376 AFD: Opened codec 0x84f3470, id(MP2) type(Audio) 2007-02-09 16:36:13.539 MainServer::HandleAnnounce FileTransfer 2007-02-09 16:36:13.542 adding: backend as a remote file transfer 0: start_time: 0.028 duration: 1.964 1: start_time: 0.028 duration: 1.929 stream: start_time: 0.308 duration: 21.822 bitrate=5128 kb/s 2007-02-09 16:36:14.692 AFD: Opened codec 0x8411db0, id(MPEG2VIDEO) type(Video) 2007-02-09 16:36:14.700 AFD: Opened codec 0x84f3470, id(MP2) type(Audio) 2007-02-09 16:36:14.984 MainServer::HandleAnnounce FileTransfer 2007-02-09 16:36:14.987 adding: backend as a remote file transfer [mpeg @ 0xb73c57f0]Parser not found for Codec Id: 94210 ! 0: start_time: 0.036 duration: 349.198 1: start_time: 1107.644 duration: 349.196 2: start_time: 0.026 duration: 349.181 stream: start_time: 0.289 duration: 16186.824 bitrate=1225 kb/s 2007-02-09 16:36:15.152 AFD: Opened codec 0x8411db0, id(MPEG2VIDEO) type(Video) 2007-02-09 16:36:15.153 AFD: Opened codec 0x8544a00, id(MP2) type(Audio) 2007-02-09 16:36:15.943 MainServer::HandleAnnounce FileTransfer 2007-02-09 16:36:15.950 adding: backend as a remote file transfer [mpeg @ 0xb73c57f0]Parser not found for Codec Id: 94210 ! 0: start_time: 0.036 duration: 348.937 1: start_time: 3881.200 duration: 348.937 2: start_time: 0.026 duration: 348.909 stream: start_time: 0.289 duration: 47001.228 bitrate=421 kb/s 2007-02-09 16:36:16.225 AFD: Opened codec 0x8411db0, id(MPEG2VIDEO) type(Video) 2007-02-09 16:36:16.229 AFD: Opened codec 0x84b5660, id(MP2) type(Audio) 2007-02-09 16:36:17.010 MainServer::HandleAnnounce FileTransfer 2007-02-09 16:36:17.023 adding: backend as a remote file transfer 0: start_time: 0.028 duration: 7.634 1: start_time: 0.030 duration: 7.590 stream: start_time: 0.308 duration: 84.818 bitrate=4927 kb/s 2007-02-09 16:36:17.302 AFD: Opened codec 0x8411db0, id(MPEG2VIDEO) type(Video) 2007-02-09 16:36:17.305 AFD: Opened codec 0x84f3470, id(MP2) type(Audio) 2007-02-09 16:36:17.564 MainServer::HandleAnnounce FileTransfer 2007-02-09 16:36:17.592 adding: backend as a remote file transfer 2007-02-09 16:36:24.168 MainServer::HandleAnnounce Monitor 2007-02-09 16:36:24.174 adding: backend as a client (events: 0) 2007-02-09 16:36:27.139 MainServer::HandleAnnounce Monitor 2007-02-09 16:36:27.150 adding: backend as a client (events: 0) root@backend:/#
Take note that I haven't confirmed that the recordings that took place during this period are having the same glitching, but certainly there does seem to be some info in these logs that may be relevant.
_________________ Currently Running:
Too lazy to update this with my current hardware, I'll redo it during my next install =)
|
|
Top |
|
|
tjc
|
Posted: Fri Feb 09, 2007 6:32 pm |
|
Joined: Thu Mar 25, 2004 11:00 am
Posts: 9551
Location:
Arlington, MA
|
Searching the forums for "All encoder MPEG stream buffers are full" turns up this thread: http://mysettopbox.tv/phpBB2/viewtopic.php?t=13534&highlight=encoder+mpeg+stream+buffers+full Which includes a simple work around (I knew this had been discussed recently).
I don't know why this problem has become so much more prevalent in 0.20 but I've even seen this a couple times myself. So far not badly enough to kill a recording, but at least enough to make them glitch.
|
|
Top |
|
|
sgunther
|
Posted: Fri Feb 09, 2007 9:55 pm |
|
Joined: Thu Apr 27, 2006 1:07 am
Posts: 54
|
The thread tjc mentions is one I started. I have seen this issue which is related to the mySQL database writes causing temporary delays on the computer causing the capture cards to overfill the buffers and drop data. The result is the artifacts you are seeing. I have played with a lot of the fixes for this that I found on the net. If you read the issue on the MythTV SVN site there is a patch but you will see that the development path is going away from the patch and to move the mySQL write to a different thread. Hopefully this will be fixed in 0.21. That being said with the patch I was starting to see some weird database issues where the backend was loosing connectivity from the mySQL databse... Obviously this is not good....
The best thing I found to do is (You probably are not going to like this) download the latest 0.20-fixes branch from SVN and compile. I have seen dramaticly less buffer over-runs in the latest build. That being said the fix on the SVN link in my other post is not yet in the 0.20-fixes branch so I am not quite sure why there is such an improvement. Good luck and let me know how it goes.....
On a side note....
If you follow the directions on the Knoppmyth Wiki for compiling from SVN the ./configure commands are somewhat out of date with the latest options... Your best bet is to ./configure --help and make you own choices.
My question to the community....
If you do the SVN compile and load you have no uninstall (make uninstall is so-so) as you would with a deb package. I see that checkinstall is loaded but I have had a lot of difficulty with getting a proper deb to build with all three packages (mythtv, myththemes and mythplugins) is there a trick to it?
If I could get the debs to build properly I could post them and you can skip all of these steps....
One last word of caution....
If you do a build and install from SVN or any myth source you will break a few things that Cecil has customized in the KnoppMyth distribution (a few of the MythWeb pages are the biggest one that come to mind). If you do this know that you will still have to do some minor patches/updates post SVN compiled code install.
Cecil-
1. What commands are you issuing to ./configure in your build for RE50? (Are you building two seperate versions i586/i686 that are installed based on the inputs of the user at installation?)
2. What checkinstall (if that is what you are using) commads are you using to build your debs?
Hope this helps... Sorry I don't have the silver bullet.... I have spent a lot of time researching this issue and I am not sure that there is a real easy answer out there.
-sgunther
|
|
Top |
|
|
rando
|
Posted: Sat Feb 10, 2007 1:36 pm |
|
Joined: Sat Feb 11, 2006 5:26 pm
Posts: 282
Location:
Winnipeg - Canada
|
sgunther thank you very much for your detailed reply. I certainly appreciate the shared knowledge.
I'll be honest when i say that updating as you've advised scares me a little, i've never worked with doing updates via SVN and am not sure I'd be able to track down the fixes to the mythweb pages afterwards. I would really hate to get halfway (or even all the way) through this only to find that i've somehow borked my system.
A simplified method (deb files?) would be extremely appealing to me, especially if they were tailored to keep all the other elements of KnoppMyth running smoothly with minimal effort.
If you do get to the point where you have a deb file or something to halfway automate this process I'd be more than happy to test it out for you before you 'let it loose' on the public.
I may get brave and attempt the directions posted above..... we'll see. The problem is not so severe here as to make the system unusable, but it is a bit annoying.
_________________ Currently Running:
Too lazy to update this with my current hardware, I'll redo it during my next install =)
|
|
Top |
|
|
tjc
|
Posted: Sat Feb 10, 2007 2:04 pm |
|
Joined: Thu Mar 25, 2004 11:00 am
Posts: 9551
Location:
Arlington, MA
|
You can always try the simple work around of increasing the amount of buffer space allocated, especially since you've got 1Gb. It's not a real fix since you're just lowering the probablity of a glitch, but it might be enough to improve your user experience...
|
|
Top |
|
|
sgunther
|
Posted: Sat Feb 10, 2007 5:06 pm |
|
Joined: Thu Apr 27, 2006 1:07 am
Posts: 54
|
The buffer increase should not be an issue with 1GB... I only have 448 Mb and I was running into issues... It is an easy temporary fix that is easy enough to undo if it causes you any issues....
|
|
Top |
|
|
rando
|
Posted: Sat Feb 10, 2007 5:50 pm |
|
Joined: Sat Feb 11, 2006 5:26 pm
Posts: 282
Location:
Winnipeg - Canada
|
Sounds good, I'll do that, is this all I need to do then?
Code: options ivtv yuv_buffers=4 mpg_buffers=16 vbi_buffers=4 pcm_buffers=4
Is it specifically the mpg_buffers that are relevant here or will bumping up any of the others make a difference?
What command can I type in to see what my current buffer settings are?
Any idea what the max size I can enter is?
Will entering this command once take effect for both of my tuners?
_________________ Currently Running:
Too lazy to update this with my current hardware, I'll redo it during my next install =)
|
|
Top |
|
|
sgunther
|
Posted: Sat Feb 10, 2007 8:18 pm |
|
Joined: Thu Apr 27, 2006 1:07 am
Posts: 54
|
You can Code: modinfo ivtv to see all of the options.... I never did experiment with every possibe combination. Let me know how it goes.
|
|
Top |
|
|
rando
|
Posted: Sun Feb 11, 2007 7:26 pm |
|
Joined: Sat Feb 11, 2006 5:26 pm
Posts: 282
Location:
Winnipeg - Canada
|
I've bumped my buffers and crossed my fingers. Time will tell from here.
_________________ Currently Running:
Too lazy to update this with my current hardware, I'll redo it during my next install =)
|
|
Top |
|
|
rando
|
Posted: Mon Feb 12, 2007 9:33 am |
|
Joined: Sat Feb 11, 2006 5:26 pm
Posts: 282
Location:
Winnipeg - Canada
|
Things are looking good so far, I've had maybe a half dozen recordings and haven't seen any notices of troubles in my syslog. I haven't watched the recordings yet, but I suppose the syslog probably "tells all" so I think I'm good.
I put my mpg buffers up to 16 as indicated above..... I guess I could go higher if the problem still does occur at times, how high can I crank that setting, anybody know?
_________________ Currently Running:
Too lazy to update this with my current hardware, I'll redo it during my next install =)
|
|
Top |
|
|