BUG: False recordings of gabage

Hi

This might be a duplicate of other bug reports, not really sure, if so just mark this duplicate.

I got from a nice working setup on motion v.3.2.8 with near to no false recordings, upgraded to 3.2.11 and ran into a lot of trouble.

In the last 2 days i had more than hundred false recordings, a few examples can be seen from http://www.dam.adsl.dk/badrec/

At first i would blame ffmpeg but when noticed that the .jpg itself had this gray areas i dont think this is a course of ffmpeg anyway?

I have several times had recordings where the recording continues long time after the motion has ended and the change in pictures is near zero, from my eyes to see. Right now the only rec og this i have left is the last in the above url.

I am using 4 Trendnet TV-IP100, cam1-4. Cam1 & 3 is of one harware version, same firmware.

Cam2 & 4 is another hardware version, running same firmware.

Cam 5 is a Grand IP Cam, very nightvision and bad collored, only used now and then to test with.

All cam's running identical settings as listet below, only diference is the threshold.

I have not have one single error recording from cam2, only a few from cam1, a lot from cam 3 & 4, and cam 5 is holding the record.

Nothing is seen in the log, i usualy did see 2-3 'Error getting image header' or something like that, pr. cam a day. by reading this site a can understand that those errors maybe is ignored for some reason. I have never seen so many errors in the log in the old versions as i now have had false recordings. Actualy i would like to have these errors logged becourse that if there were many from one of the cam's i could do a reboot on the cam and everything were all well again. Things to notice in the examples is of course the gray areas coursing the false event but also.

From cam3: some pixels change in a strange way 1 sec after the gray area disappear.

From cam5, event 36: I have left over the 6 first, and the last 4 pictures. interresting is that even there is 'movement' in the last pictures after a while of nothing, the post frames seem to be missing.

From cam5, event 45: there sure is more than the 3 post frames after the gray is over.

I really hope you know what is making this strange behavior and a fix is out soon as motion is more or less useless for my needs in this version. I have now, a few hours ago made the version 3.2.9 from /tags/ and the trouble seem to be over for now.

Sorry for my spelling, and messy report, hope some of you understand the meaning. B)

/Klaus

norm 0
width 640
height 480
framerate 3
netcam_url http://192.168.1.x/video.cgi (Cam1-4)
netcam_url http://192.168.1.x/Webcam.jpg (Cam5)
auto_brightness off
brightness 0
contrast 0
saturation 0
hue 0

threshold - cam 1 & 2 2500, cam 3 3200, cam4 & 5 2650
threshold_tune off
noise_level 32
noise_tune on
despeckle EedDl
smart_mask_speed 0
lightswitch 0
minimum_motion_frames 2
pre_capture 2
post_capture 3
gap 60
max_mpeg_time 0
output_all off

Environment

Motion version:	3.2.11
ffmpeg version:	ffmpeg-cvs-20060306-certified-with-motion-3.2.5
Shared libraries:	ffmpeg, mysql
Server OS:	Fedora 11 / 2.6.29.5-191.fc11.i686.PAE

-- TheOtherBug - 23 Jul 2009

Follow up

Did you read all changes from 3.2.8 to 3.2.11 ?

There's some new options affecting netcams :

http://www.lavrsen.dk/foswiki/bin/view/Motion/UpgradingFromOlderVersion

You can see detailed info of those options :

http://www.lavrsen.dk/foswiki/bin/view/Motion/ConfigFileOptions

About logging you can increase / decrease level using -d X from command line.

3.2.11.1 will be realeased soon , you can apply a patch already :

http://www.lavrsen.dk/foswiki/bin/view/Motion/BugFixesToUpgradeVersion

Last but not least we have fixing many issues in our 3.3.0 version , you can get it from trunk , but notice that many options in

config file has change you can see here :

http://www.lavrsen.dk/foswiki/bin/view/Motion/IntuitiveOptionNamesDiscussion

-- AngelCarpintero - 23 Jul 2009

---

Follow up

>Did you read all changes from 3.2.8 to 3.2.11 ?

Yes I allways do before upgrading anything. Just looked thrugh the Changelog in 3.2.11 again, I can not see anything giving me clues of what is happening in my setup. There's some writing about jpeg palettes and so, i have no clue if that has something to do with my cam's serving mjpeg streams and working for years without other trouble than the old camera lost error fixed years ago.

>There's some new options affecting netcams :

I can just see one thing that allready has been checked out, netcam_tolerant_check. reading on this looks more like a discussion about naming of some option, nowhere including the changelog I can not see that this thing is implementet.

Not really sure what it does i'll guess that i should add the keyword and set the value to the opposide of default, but i am not sure at all what this does and why! Still wondering why i now is getting so many false recordings if bad mjpeg streams is caursing this, as notet initialy i only have had a few entrys in the log a day, about error reading of images, those did not result in the results seen now.

>You can see detailed info of those options

I know that too and normaly would look there but as just claimed that would require something in the changelog would give me clue about something that involve my setup has changed. When starting motion 3.2.11 the first time it told me about options it did not know about anymore, those got marked out and a new option added and everything looked fine - that was until I should use hours on looking and deleting all those buggy recordings.

>About logging you can increase / decrease level using -d X from command line.

That is said by motion to be about debugging, i would have no clue where/what to debug i would only use that in case some of you developers is asking about doing that to post the result. I'm just a user of this brilliant software and would expect to see in the log if/when there is trouble and there is nothing to see inthere except of when motion is starting, setting up some buffer stuff (that i would suspect be the troublemaker, beeing the only change i can figure out involves my setup) and when stopping.

Actualy it's not even me who would normaly start motion, my server is living a life of it's own with a lot of automatic scripting partly broken of other reasons, u know.

>3.2.11.1 will be realeased soon , you can apply a patch already

Patches and i are not good friends, that patch i also allready has looking at but at first reading i just noticed somthing about trunk, debian and suse and did not see it as something to use time on. by now closer reading i can see that maybe it's worth a try sp thati'll do and give a fedback in this report.

>Last but not least we have fixing many issues in our 3.3.0 version , you can get it from trunk

That has allready been tried but when startet the output and all the changes got me to drop that idea at the same speed as the config errors scrolled up the screen - hehe - again the word trunk in the bigpatch has not giving me balls to try yet.

I really like this software and would dislike to should find something else but i must admit i really was close to look in the zoneminder way allready when the trouble coursed by the webinterface change hit me and i'll stil claim that the help is missplaced and that if all the trouble grom this report could have been avoided just by adding a 'netcam_tolerant_check on' that should have ben written in the change.log.

Im sorry if im sounding rude, that is absolutely not my intention. I'll try apatch and a config keyword and get back.

/Klaus

---

Follow up

Sad to tell that it was not enough, even it until now seems to have reduced the number of false recs.

Patch applied, cnofig option netcam_tolerant_check on added in motion.conf, expecting it to bu functionel in the threadfiles like most of my settings, motion startet with 'motion -d 2'

The log says:

Jul 23 23:22:36 sommergw motion: [0] Processing thread 0 - config file /usr/local/etc/motion.conf
Jul 23 23:22:36 sommergw motion: [0] Processing config file /usr/local/etc/cam1.conf
Jul 23 23:22:36 sommergw motion: [0] Processing config file /usr/local/etc/cam2.conf
Jul 23 23:22:36 sommergw motion: [0] Processing config file /usr/local/etc/cam3.conf
Jul 23 23:22:36 sommergw motion: [0] Processing config file /usr/local/etc/cam4.conf
Jul 23 23:22:36 sommergw motion: [0] Processing config file /usr/local/etc/cam5.conf
Jul 23 23:22:36 sommergw motion: [0] Motion 3.2.11.1 Started
Jul 23 23:22:36 sommergw motion: [0] Motion running as daemon process
Jul 23 23:22:36 sommergw motion: [0] ffmpeg LIBAVCODEC_BUILD 3344128 LIBAVFORMAT_BUILD 3277568
Jul 23 23:22:36 sommergw motion: [0] Thread 1 is from /usr/local/etc/cam1.conf
Jul 23 23:22:36 sommergw motion: [0] Thread 2 is from /usr/local/etc/cam2.conf
Jul 23 23:22:36 sommergw motion: [0] Thread 3 is from /usr/local/etc/cam3.conf
Jul 23 23:22:36 sommergw motion: [0] Thread 4 is from /usr/local/etc/cam4.conf
Jul 23 23:22:36 sommergw motion: [0] Thread 5 is from /usr/local/etc/cam5.conf
Jul 23 23:22:36 sommergw motion: [1] Resizing pre_capture buffer to 1 items
Jul 23 23:22:36 sommergw motion: [1] Resizing pre_capture buffer to 4 items
Jul 23 23:22:36 sommergw motion: [1] netcam_next called with no data in buffer
Jul 23 23:22:37 sommergw motion: [4] Resizing pre_capture buffer to 1 items
Jul 23 23:22:37 sommergw motion: [4] Resizing pre_capture buffer to 4 items
Jul 23 23:22:37 sommergw motion: [4] netcam_next called with no data in buffer
Jul 23 23:22:37 sommergw motion: [3] Resizing pre_capture buffer to 1 items
Jul 23 23:22:37 sommergw motion: [3] Resizing pre_capture buffer to 4 items
Jul 23 23:22:37 sommergw motion: [3] netcam_next called with no data in buffer
Jul 23 23:22:37 sommergw motion: [2] Resizing pre_capture buffer to 1 items
Jul 23 23:22:37 sommergw motion: [2] Resizing pre_capture buffer to 4 items
Jul 23 23:22:37 sommergw motion: [2] netcam_next called with no data in buffer
Jul 23 23:22:38 sommergw motion: [5] Resizing pre_capture buffer to 1 items
Jul 23 23:22:38 sommergw motion: [5] Resizing pre_capture buffer to 4 items
Jul 23 23:22:38 sommergw motion: [5] netcam_next called with no data in buffer
Jul 23 23:22:38 sommergw motion: [5] netcam_next called with no data in buffer
Jul 23 23:22:38 sommergw motion: [5] netcam_next called with no data in buffer
Jul 23 23:22:39 sommergw motion: [5] netcam_next called with no data in buffer
Jul 23 23:22:39 sommergw motion: [5] netcam_next called with no data in buffer
Jul 23 23:22:39 sommergw motion: [5] netcam_next called with no data in buffer
Jul 24 00:40:00 sommergw motion: [3] ffmpeg_avcodec_log: warning, clipping 1 dct coefficients to -255..255#012 - flag 1
Jul 24 00:40:00 sommergw motion: [1] ffmpeg_avcodec_log: warning, clipping 1 dct coefficients to -255..255#012 - flag 1
Jul 24 00:40:00 sommergw motion: [4] ffmpeg_avcodec_log: warning, clipping 1 dct coefficients to -255..255#012 - flag 1
Jul 24 00:40:00 sommergw motion: [5] ffmpeg_avcodec_log: warning, clipping 1 dct coefficients to -255..255#012 - flag 1
Jul 24 00:53:08 sommergw motion: [1] ffmpeg_avcodec_log: warning, clipping 1 dct coefficients to -127..127#012 - flag 1
Jul 24 00:53:08 sommergw motion: [1] ffmpeg_avcodec_log: warning, clipping 1 dct coefficients to -127..127#012 - flag 1
Jul 24 00:53:08 sommergw motion: [1] ffmpeg_avcodec_log: warning, clipping 1 dct coefficients to -127..127#012 - flag 1
Jul 24 00:53:08 sommergw motion: [1] ffmpeg_avcodec_log: warning, clipping 1 dct coefficients to -127..127#012 - flag 1
Jul 24 00:53:08 sommergw motion: [1] ffmpeg_avcodec_log: warning, clipping 1 dct coefficients to -127..127#012 - flag 1
Jul 24 00:53:08 sommergw motion: [1] ffmpeg_avcodec_log: warning, clipping 1 dct coefficients to -127..127#012 - flag 1
Jul 24 00:53:08 sommergw motion: [1] ffmpeg_avcodec_log: warning, clipping 1 dct coefficients to -127..127#012 - flag 1
Jul 24 00:53:08 sommergw motion: [1] ffmpeg_avcodec_log: warning, clipping 1 dct coefficients to -127..127#012 - flag 1
Jul 24 00:53:08 sommergw motion: [1] ffmpeg_avcodec_log: warning, clipping 1 dct coefficients to -127..127#012 - flag 1
Jul 24 00:53:08 sommergw motion: [1] ffmpeg_avcodec_log: warning, clipping 1 dct coefficients to -127..127#012 - flag 1
Jul 24 00:53:08 sommergw motion: [1] ffmpeg_avcodec_log: warning, clipping 1 dct coefficients to -127..127#012 - flag 1
Jul 24 00:53:08 sommergw motion: [1] ffmpeg_avcodec_log: warning, clipping 1 dct coefficients to -127..127#012 - flag 1
Jul 24 00:53:08 sommergw motion: [1] ffmpeg_avcodec_log: warning, clipping 1 dct coefficients to -127..127#012 - flag 1
Jul 24 00:53:08 sommergw motion: [1] ffmpeg_avcodec_log: warning, clipping 1 dct coefficients to -127..127#012 - flag 1
Jul 24 00:53:08 sommergw motion: [1] ffmpeg_avcodec_log: warning, clipping 1 dct coefficients to -127..127#012 - flag 1
Jul 24 00:53:08 sommergw motion: [1] ffmpeg_avcodec_log: warning, clipping 1 dct coefficients to -127..127#012 - flag 1
Jul 24 00:53:08 sommergw motion: [1] ffmpeg_avcodec_log: warning, clipping 1 dct coefficients to -127..127#012 - flag 1
Jul 24 00:53:08 sommergw motion: [1] ffmpeg_avcodec_log: warning, clipping 1 dct coefficients to -127..127#012 - flag 1
Jul 24 00:53:08 sommergw motion: [1] ffmpeg_avcodec_log: warning, clipping 1 dct coefficients to -127..127#012 - flag 1
Jul 24 00:53:08 sommergw motion: [1] ffmpeg_avcodec_log: warning, clipping 1 dct coefficients to -127..127#012 - flag 1
Jul 24 00:53:08 sommergw motion: [1] ffmpeg_avcodec_log: warning, clipping 1 dct coefficients to -127..127#012 - flag 1
Jul 24 00:53:08 sommergw motion: [1] ffmpeg_avcodec_log: warning, clipping 1 dct coefficients to -127..127#012 - flag 1
Jul 24 00:53:08 sommergw motion: [1] ffmpeg_avcodec_log: warning, clipping 1 dct coefficients to -127..127#012 - flag 1
Jul 24 00:53:08 sommergw motion: [1] ffmpeg_avcodec_log: warning, clipping 1 dct coefficients to -127..127#012 - flag 1 
Added: a lot more like this.
Jul 24 01:28:44 sommergw motion: [4] ffmpeg_avcodec_log: warning, clipping 1 dct coefficients to -127..127#012 - flag 1

The result: http://www.dam.adsl.dk/badrec/20090724005307-01-cam1.avi

added: http://www.dam.adsl.dk/badrec/20090724012843-02-cam4.avi

-- TheOtherBug - 23 Jul 2009

---

Follow up

25 more bad recordings over the night, they are all in the above mentined place too watch including 2 regular recordings for the joy.

I'll again go back to 3.2.9, please tell if there is something to try, i'll keep an eye on this thread.

-- TheOtherBug - 24 Jul 2009

---

Fix record