motion-user Mailing List for motion (Page 2)

On Thu, 24 Apr 2025 15:21:11 -0500
Barry Martin <bar...@gm...> wrote:

> Now here comes the part I'm uncertain of.  Can I use a PoE
> (extension) but (temporarily) still power the Pi's with their wall
> warts?  

I believe you can use either or.  That is, you can have the PoE hat
connected to only the pi and not ethernet while powering the pi over
its usb power port.  One power source at a time only.


> The key word in my question (though I might be running in to a power 
> limit too) is can I plug the two Motion Pi's in to the PoE extension 
> while still using the wall warts? 
no, from the PoE+ Hat documentation (
https://datasheets.raspberrypi.com/poe/poe-plus-hat-product-brief.pdf )

"When the Raspberry Pi PoE+ HAT is connected to your Raspberry Pi, your
Raspberry Pi should only be powered through the ethernet cable. Do not
use any additional method to power the Raspberry Pi"


There are redundant ways of powering the pi's over USB that people have
come up with, I suggest exploring that route.

Gentlemen (and any ladie):

> No wifi here, all copper, and top shows motion running. 

Here's a question with a probably obvious answer but I'd rather ask and 
avoid 'letting the smoke out'.  Two of the three (current) Motion 
systems are Raspberry Pi 4's,USB cameras, using the 5G WiFi.  The two 
are side-by-side.  What I'm considering is switching them from WiFi to 
Ethernet.

Now here comes the part I'm uncertain of.  Can I use a PoE (extension) 
but (temporarily) still power the Pi's with their wall warts?  I do have 
a PoE Injector rated 802.3af/at and 54v, 30 and 60 W -- if 802.3at = 30W 
how double the output wattage??  Anyway, the injector powers a PoE 
extension with a couple of Pi's on it. I'm thinking of using the 
currently unused port on this extension to power a new PoE extension to 
eventually power the 2 Motion Pi's.

The key word in my question (though I might be running in to a power 
limit too) is can I plug the two Motion Pi's in to the PoE extension 
while still using the wall warts?  I know there are HATs to extract the 
power and data but I'd like to test is wired fixes the glitch problem 
before going to the extra expense.  The price of the PoE extension is 
fine for the test, just the PoE + HATs + new cases is getting up there.

Thanks!

Barry

Hi Stewart!

> I have seen that exact behavior several times. The pattern of when it 
> occurred, is that a vehicle moved across the field of view, leaving a 
> 10 second trail of disruption until the pixels are unchanging long 
> enough to End the event. Usually motion records the entire 10 seconds 
> as changed pixels, but sometimes does not. Then this behavior is, in 
> that 10 second period if a second vehicle appears, it is not counted 
> as a new change, and gets ignored.
>
Yes, I have seen that here also, I'm talking when a person is walking up 
the front walkway or across the field of view and they are suddenly 
10'/3m from where they were.  The walking motion is continuous and fast 
enough not to have Motion time-out.  The embedded timestamps also 
indicate the skip yet we haven't figured out the cause.  And of course 
it doesn't occur constantly but often enough.


> No wifi here, all copper, and top shows motion running. But it is 
> running 3 cameras, not 1. Watching the cpu load at the exact time 
> would be a challenge.
>
Just capturing the load on the CPU puts a load on the CPU! <g>  Here 
using WiFi admittedly is a convenience over running another Ethernet 
cable.  Most of the computers are hard-wired; the Motion cameras are not 
just because of physical location.  The signal strength seems to be more 
than sufficient -- not anywhere near borderline.  No noted RF 
Interference (QRM/QRN is Ham-talk, though I'm not one)....

Thanks for the suggestions!

Barry

I have seen that exact behavior several times. The pattern of when it 
occurred, is that a vehicle moved across the field of view, leaving a 10 
second trail of disruption until the pixels are unchanging long enough 
to End the event. Usually motion records the entire 10 seconds as 
changed pixels, but sometimes does not. Then this behavior is, in that 
10 second period if a second vehicle appears, it is not counted as a new 
change, and gets ignored.

That was during the motion v4.5.x time frame, and since I upgraded to 
4.6 in 2023-Nov, I can't say that I have seen it in a while. Since it is 
not repeatable predictably, I ignored it.

No wifi here, all copper, and top shows motion running. But it is 
running 3 cameras, not 1. Watching the cpu load at the exact time would 
be a challenge.

Stewart


On 4/12/25 08:05, Barry Martin wrote:
>
>
> “Slightly” off-topic but involving my Motion cameras. Randomly the 
> recording will skip several seconds. There is motion going on, so 
> should record, but while the motion is occurring (and being recorded) 
> the video will skip several seconds: no grey, no pixelation, just as 
> if someone took scissors to the film (showing my age! <g>). Does not 
> occur constantly/consistently.
>
>

I got the answer - my resolution is 4608 x 2592.

I need to use the text_scale-option.

My fault - sorry



Am 24.04.2025 um 08:10 schrieb tosiara:
> Are you using pass-through option?
> 
> On Wed, Apr 23, 2025 at 9:44 PM seb...@de...
> <seb...@de...> wrote:
>>
>> Hi,
>>
>> i know i could put some information in the picture by using for example
>> "text_left %Y-%m"
>>
>> but how do i get the same text into the video files?
>>
>> The Video ist very clear and had a very good quality, but there is no
>> text inside.
>>
>> Tnx
>>
>> Sebastian
>>
>>
>> _______________________________________________
>> Motion-user mailing list
>> Mot...@li...
>> https://lists.sourceforge.net/lists/listinfo/motion-user
>> https://motion-project.github.io/
>>
>> Unsubscribe: https://lists.sourceforge.net/lists/options/motion-user

Are you using pass-through option?

On Wed, Apr 23, 2025 at 9:44 PM seb...@de...
<seb...@de...> wrote:
>
> Hi,
>
> i know i could put some information in the picture by using for example
> "text_left %Y-%m"
>
> but how do i get the same text into the video files?
>
> The Video ist very clear and had a very good quality, but there is no
> text inside.
>
> Tnx
>
> Sebastian
>
>
> _______________________________________________
> Motion-user mailing list
> Mot...@li...
> https://lists.sourceforge.net/lists/listinfo/motion-user
> https://motion-project.github.io/
>
> Unsubscribe: https://lists.sourceforge.net/lists/options/motion-user

Hi,

i know i could put some information in the picture by using for example 
"text_left %Y-%m"

but how do i get the same text into the video files?

The Video ist very clear and had a very good quality, but there is no 
text inside.

Tnx

Sebastian

Hi Jack!
>
> Sounds similar to an issue I had. Motion was waiting for the 
> on_movie_end script to end. Fixed by commit 035b1af.
>
> See https://github.com/Motion-Project/motion/issues/1856
>
Initially looks to be part of Roger's suggestion  to save to SSD.    
...Generally /tmp is on a drive, not RAM. but easy enough to tell where 
to (temporarily) store the files.

In my case it seems the final storage location (the NAS's HDD) is the 
issue based on Roger's message.  If the glitches were independent (at 
different times) I'd think your suggestion was more correct but with the 
glitch at almost exactly the same time the HDD issue seems more correct.

I'm definitely not excluding your suggestion -- will be keeping it for 
reference -- so thank you!

Barry

Sounds similar to an issue I had. Motion was waiting for the 
on_movie_end script to end. Fixed by commit 035b1af.

See https://github.com/Motion-Project/motion/issues/1856

On 4/12/25 11:05, Barry Martin wrote:
>
>
> Hi!
>
> “Slightly” off-topic but involving my Motion cameras. Randomly the 
> recording will skip several seconds. There is motion going on, so 
> should record, but while the motion is occurring (and being recorded) 
> the video will skip several seconds: no grey, no pixelation, just as 
> if someone took scissors to the film (showing my age! <g>). Does not 
> occur constantly/consistently.
>
>
> Appears has something to do with either the WiFi (5G) between the 
> Motion units (so RF transmission/reception) and/or the actual 
> recording to the hard drive on the NAS. Looking for a clue how to 
> figure out and hopefully correct the problem. Yesterday checked and 
> the glitch occurred on two cameras with overlapping fields of view and 
> recording simultaneously:
>
> Unit #201 14:15:41 <skip> 14:16:25
>
> Unit #202 14:15:43 <skip> 14:16:25
>
>
> Nothing significant about the time; nobody near the Motion units nor 
> WiFi router, so not like someone walked in the RF path. LIS, the 
> glitch occurs randomly; just listed the times should the ~45 seconds 
> of missing recording have a potential troubleshooting meaning. (I 
> don’t know the times of previous glitches.)
>
>
>
> Both units:
>
>  *
>
>     Motion Version 4.5.1
>
>  *
>
>     Raspberry Pi 4B, with one USB camera each
>
>  *
>
>     Individual power supply on UPS
>
>  *
>
>     Uptime: (several days ago)
>
>
> The WiFi and NAS are also on their own UPSs, and no power issues.
>
>
> Thanks for the help!
>
> Barry
>
>
>
>
> _______________________________________________
> Motion-user mailing list
> Mot...@li...
> https://lists.sourceforge.net/lists/listinfo/motion-user
> https://motion-project.github.io/
>
> Unsubscribe:https://lists.sourceforge.net/lists/options/motion-user

I use an SSD to initially save my images and then daily I move those
images to spinning disks.

Without doing that any slight issue with the disk will cause dropouts.
  An error on the disk will cause a 3-30second blip depending on the
what the disks default timeout is.

this is a sample of smartctl output:
smartctl --xall /dev/sda | grep -iE 'sector|UNC'
Sector Sizes:     512 bytes logical, 4096 bytes physical
  5 Reallocated_Sector_Ct   PO--CK   200   200   140    -    0
197 Current_Pending_Sector  -O--CK   200   200   000    -    0
198 Offline_Uncorrectable   ----CK   200   200   000    -    0
SMART           Log Directory Version 1 [multi-sector log support]
SMART Extended Comprehensive Error Log Version: 1 (6 sectors)
    COUNT  = Count (was: Sector Count) Register
    LL     = LBA Low (was: Sector Number) Register     ]
  40 -- 51 00 00 00 00 bc 9c 57 f8 40 00  Error: UNC at LBA =
0xbc9c57f8 = 3164362744
  40 -- 51 00 00 00 00 55 57 f0 88 40 00  Error: UNC at LBA =
0x5557f088 = 1431826568
  40 -- 51 00 00 00 01 36 58 20 70 40 00  Error: UNC at LBA =
0x136582070 = 5206712432
  40 -- 51 00 00 00 00 47 86 ce 78 40 00  Error: UNC at LBA =
0x4786ce78 = 1200017016
  40 -- 51 00 00 00 01 2b e1 e0 00 40 00  Error: UNC at LBA =
0x12be1e000 = 5031190528
  40 -- 51 00 00 00 00 87 62 10 48 40 00  Error: UNC at LBA =
0x87621048 = 2271350856
SMART Extended Self-test Log Version: 1 (1 sectors)

The realllocated/pending counts rising will cause a several second
blip, and the UNC errors will cause blips.

I have a cron job that saves a log for each disk daily so I can rule
out a disk issue causing a blip.

On Sat, Apr 12, 2025 at 11:05 AM Barry Martin <bar...@gm...> wrote:
>
>
> Hi JMB!
> >
> > Try using 'top' to monitor the system load.  Maybe 'disk' writes (to
> > the NAS) are piling up.  Maybe some background process is consuming
> > excessive CPU?  Did it start after a recent 'sudo aptitude safe-update'?
>
> Will try your suggestions -- right now not sure how to correlate (right
> word?) the time of the skip with top's activity reporting but will
> experiment and learn!  Most Motion recordings are fine (no skip/glitch)
> -- I did forget to mention that originally. Skip/glitch problem has been
> occurred for some time, so can't blame a specific upgrade.
>
> Will report back!
>
> Barry
>
>
>
>
>
> _______________________________________________
> Motion-user mailing list
> Mot...@li...
> https://lists.sourceforge.net/lists/listinfo/motion-user
> https://motion-project.github.io/
>
> Unsubscribe: https://lists.sourceforge.net/lists/options/motion-user

Hi JMB!
>
> Try using 'top' to monitor the system load.  Maybe 'disk' writes (to 
> the NAS) are piling up.  Maybe some background process is consuming 
> excessive CPU?  Did it start after a recent 'sudo aptitude safe-update'?

Will try your suggestions -- right now not sure how to correlate (right 
word?) the time of the skip with top's activity reporting but will 
experiment and learn!  Most Motion recordings are fine (no skip/glitch) 
-- I did forget to mention that originally. Skip/glitch problem has been 
occurred for some time, so can't blame a specific upgrade.

Will report back!

Barry

Hello,

Try using 'top' to monitor the system load.  Maybe 'disk' writes (to the 
NAS) are piling up.  Maybe some background process is consuming 
excessive CPU?  Did it start after a recent 'sudo aptitude safe-update'?

Hope this helps

On 4/12/25 11:05, Barry Martin wrote:
>
> Hi!
>
> “Slightly” off-topic but involving my Motion cameras. Randomly the 
> recording will skip several seconds. There is motion going on, so 
> should record, but while the motion is occurring (and being recorded) 
> the video will skip several seconds: no grey, no pixelation, just as 
> if someone took scissors to the film (showing my age! <g>). Does not 
> occur constantly/consistently.
>
> Appears has something to do with either the WiFi (5G) between the 
> Motion units (so RF transmission/reception) and/or the actual 
> recording to the hard drive on the NAS. Looking for a clue how to 
> figure out and hopefully correct the problem. Yesterday checked and 
> the glitch occurred on two cameras with overlapping fields of view and 
> recording simultaneously:
>
> Unit #201 14:15:41 <skip> 14:16:25
>
> Unit #202 14:15:43 <skip> 14:16:25
>
> Nothing significant about the time; nobody near the Motion units nor 
> WiFi router, so not like someone walked in the RF path. LIS, the 
> glitch occurs randomly; just listed the times should the ~45 seconds 
> of missing recording have a potential troubleshooting meaning. (I 
> don’t know the times of previous glitches.)
>
> Both units:
>
>  *
>
>     Motion Version 4.5.1
>
>  *
>
>     Raspberry Pi 4B, with one USB camera each
>
>  *
>
>     Individual power supply on UPS
>
>  *
>
>     Uptime: (several days ago)
>
> The WiFi and NAS are also on their own UPSs, and no power issues.
>
> Thanks for the help!
>
> Barry
>
>
>
>
> _______________________________________________
> Motion-user mailing list
> Mot...@li...
> https://lists.sourceforge.net/lists/listinfo/motion-user
> https://motion-project.github.io/
>
> Unsubscribe:https://lists.sourceforge.net/lists/options/motion-user

Hi!

“Slightly” off-topic but involving my Motion cameras. Randomly the 
recording will skip several seconds. There is motion going on, so should 
record, but while the motion is occurring (and being recorded) the video 
will skip several seconds: no grey, no pixelation, just as if someone 
took scissors to the film (showing my age! <g>). Does not occur 
constantly/consistently.


Appears has something to do with either the WiFi (5G) between the Motion 
units (so RF transmission/reception) and/or the actual recording to the 
hard drive on the NAS. Looking for a clue how to figure out and 
hopefully correct the problem. Yesterday checked and the glitch occurred 
on two cameras with overlapping fields of view and recording simultaneously:

Unit #201 14:15:41 <skip> 14:16:25

Unit #202 14:15:43 <skip> 14:16:25


Nothing significant about the time; nobody near the Motion units nor 
WiFi router, so not like someone walked in the RF path. LIS, the glitch 
occurs randomly; just listed the times should the ~45 seconds of missing 
recording have a potential troubleshooting meaning. (I don’t know the 
times of previous glitches.)



Both units:

  *

    Motion Version 4.5.1

  *

    Raspberry Pi 4B, with one USB camera each

  *

    Individual power supply on UPS

  *

    Uptime: (several days ago)


The WiFi and NAS are also on their own UPSs, and no power issues.


Thanks for the help!

Barry

Motionplus does work but you have to compile it from source since the 
library names changed since the last debs were created.

MrDave

On 2/3/2025 7:15 AM, Jack Christensen wrote:
> Thank you very much! I have been thinking that I should try 
> MotionPlus; this may be an opportunity to do that. I wonder if 
> MotionPlus will also have an issue as it uses the same libraries as 
> Motion.
>
> On 2/3/25 01:35, MrDave wrote:
>> Confirmed. I submitted this as a bug to the libcamera project.
>>
>> https://bugs.libcamera.org/show_bug.cgi?id=253
>>
>> On 2/2/2025 12:52 PM, Jack Christensen wrote:
>>> Attached is another log file. Not sure if I missed something with 
>>> the previous one, but note the "Permission denied" messages.
>>>
>>> Also, I am able to capture images with rpicam-still and video with 
>>> rpicam-vid successfully, but when running these commands, I see the 
>>> following:
>>>
>>> [0:14:44.064597744] [782] ERROR IPAModule ipa_module.cpp:171 Symbol 
>>> ipaModuleInfo not found
>>> [0:14:44.064989743] [782] ERROR IPAModule ipa_module.cpp:291 
>>> v4l2-compat.so: IPA module has no valid info
>>>
>>> On 2/2/25 14:24, Jack Christensen wrote:
>>>> Hello all,
>>>>
>>>> It seems a recent software update may have broken motion on this 
>>>> headless Raspberry Pi Zero W.
>>>>
>>>> I noticed the following package updates:
>>>>   libcamera0.4:armhf (0.4.0+rpt20250127-3, automatic)
>>>>   rpicam-apps-lite:armhf (1.5.3-1, 1.6.0-1)
>>>>   libcamera-v4l2:armhf (0.3.2+rpt20241119-1, 0.4.0+rpt20250127-3)
>>>>   libcamera-dev:armhf (0.3.2+rpt20241119-1, 0.4.0+rpt20250127-3)
>>>>   libcamera-ipa:armhf (0.3.2+rpt20241119-1, 0.4.0+rpt20250127-3)
>>>>   libcamera-tools:armhf (0.3.2+rpt20241119-1, 0.4.0+rpt20250127-3)
>>>>
>>>> OS: Raspbian GNU/Linux 12 (bookworm) armv6l
>>>> Host: Raspberry Pi Zero W Rev 1.1
>>>> Kernel: 6.6.74+rpt-rpi-v6
>>>> Motion: 4.7.0+git20240909-8d69cbf
>>>>
>>>> I had originally built motion 4.7.0 from source. I went through 
>>>> that process again, starting with a fresh clone of the repo, but 
>>>> that made no difference.
>>>>
>>>> Looks like motion is crashing. Motion's pid file is still present 
>>>> but that pid is not running. Attached is a level 9 log from motion. 
>>>> The only messages I see in syslog are these:
>>>>
>>>> Feb 02 14:03:05 cam1 motion[3022]: [0:motion] [NTC] [ALL] 
>>>> conf_load: Processing thread 0 - config file 
>>>> /home/jack/.motion/motion.conf
>>>> Feb 02 14:03:05 cam1 motion[3022]: [0:motion] [NTC] [ALL] 
>>>> motion_startup: Logging to file (/home/jack/motion.log)
>>>>
>>>> Wondering if anyone else has noticed this, or if someone may have a 
>>>> suggestion.
>>>>
>>>> Thanks in advance,
>>>>
>>>
>>>
>>> _______________________________________________
>>> Motion-user mailing list
>>> Mot...@li...
>>> https://lists.sourceforge.net/lists/listinfo/motion-user
>>> https://motion-project.github.io/
>>>
>>> Unsubscribe: https://lists.sourceforge.net/lists/options/motion-user
>>
>>
>>
>> _______________________________________________
>> Motion-user mailing list
>> Mot...@li...
>> https://lists.sourceforge.net/lists/listinfo/motion-user
>> https://motion-project.github.io/
>>
>> Unsubscribe: https://lists.sourceforge.net/lists/options/motion-user

Confirmed.  I submitted this as a bug to the libcamera project.

https://bugs.libcamera.org/show_bug.cgi?id=253

On 2/2/2025 12:52 PM, Jack Christensen wrote:
> Attached is another log file. Not sure if I missed something with the 
> previous one, but note the "Permission denied" messages.
>
> Also, I am able to capture images with rpicam-still and video with 
> rpicam-vid successfully, but when running these commands, I see the 
> following:
>
> [0:14:44.064597744] [782] ERROR IPAModule ipa_module.cpp:171 Symbol 
> ipaModuleInfo not found
> [0:14:44.064989743] [782] ERROR IPAModule ipa_module.cpp:291 
> v4l2-compat.so: IPA module has no valid info
>
> On 2/2/25 14:24, Jack Christensen wrote:
>> Hello all,
>>
>> It seems a recent software update may have broken motion on this 
>> headless Raspberry Pi Zero W.
>>
>> I noticed the following package updates:
>>   libcamera0.4:armhf (0.4.0+rpt20250127-3, automatic)
>>   rpicam-apps-lite:armhf (1.5.3-1, 1.6.0-1)
>>   libcamera-v4l2:armhf (0.3.2+rpt20241119-1, 0.4.0+rpt20250127-3)
>>   libcamera-dev:armhf (0.3.2+rpt20241119-1, 0.4.0+rpt20250127-3)
>>   libcamera-ipa:armhf (0.3.2+rpt20241119-1, 0.4.0+rpt20250127-3)
>>   libcamera-tools:armhf (0.3.2+rpt20241119-1, 0.4.0+rpt20250127-3)
>>
>> OS: Raspbian GNU/Linux 12 (bookworm) armv6l
>> Host: Raspberry Pi Zero W Rev 1.1
>> Kernel: 6.6.74+rpt-rpi-v6
>> Motion: 4.7.0+git20240909-8d69cbf
>>
>> I had originally built motion 4.7.0 from source. I went through that 
>> process again, starting with a fresh clone of the repo, but that made 
>> no difference.
>>
>> Looks like motion is crashing. Motion's pid file is still present but 
>> that pid is not running. Attached is a level 9 log from motion. The 
>> only messages I see in syslog are these:
>>
>> Feb 02 14:03:05 cam1 motion[3022]: [0:motion] [NTC] [ALL] conf_load: 
>> Processing thread 0 - config file /home/jack/.motion/motion.conf
>> Feb 02 14:03:05 cam1 motion[3022]: [0:motion] [NTC] [ALL] 
>> motion_startup: Logging to file (/home/jack/motion.log)
>>
>> Wondering if anyone else has noticed this, or if someone may have a 
>> suggestion.
>>
>> Thanks in advance,
>>
>
>
> _______________________________________________
> Motion-user mailing list
> Mot...@li...
> https://lists.sourceforge.net/lists/listinfo/motion-user
> https://motion-project.github.io/
>
> Unsubscribe: https://lists.sourceforge.net/lists/options/motion-user

Attached is another log file. Not sure if I missed something with the 
previous one, but note the "Permission denied" messages.

Also, I am able to capture images with rpicam-still and video with 
rpicam-vid successfully, but when running these commands, I see the 
following:

[0:14:44.064597744] [782] ERROR IPAModule ipa_module.cpp:171 Symbol 
ipaModuleInfo not found
[0:14:44.064989743] [782] ERROR IPAModule ipa_module.cpp:291 
v4l2-compat.so: IPA module has no valid info

On 2/2/25 14:24, Jack Christensen wrote:
> Hello all,
>
> It seems a recent software update may have broken motion on this 
> headless Raspberry Pi Zero W.
>
> I noticed the following package updates:
>   libcamera0.4:armhf (0.4.0+rpt20250127-3, automatic)
>   rpicam-apps-lite:armhf (1.5.3-1, 1.6.0-1)
>   libcamera-v4l2:armhf (0.3.2+rpt20241119-1, 0.4.0+rpt20250127-3)
>   libcamera-dev:armhf (0.3.2+rpt20241119-1, 0.4.0+rpt20250127-3)
>   libcamera-ipa:armhf (0.3.2+rpt20241119-1, 0.4.0+rpt20250127-3)
>   libcamera-tools:armhf (0.3.2+rpt20241119-1, 0.4.0+rpt20250127-3)
>
> OS: Raspbian GNU/Linux 12 (bookworm) armv6l
> Host: Raspberry Pi Zero W Rev 1.1
> Kernel: 6.6.74+rpt-rpi-v6
> Motion: 4.7.0+git20240909-8d69cbf
>
> I had originally built motion 4.7.0 from source. I went through that 
> process again, starting with a fresh clone of the repo, but that made 
> no difference.
>
> Looks like motion is crashing. Motion's pid file is still present but 
> that pid is not running. Attached is a level 9 log from motion. The 
> only messages I see in syslog are these:
>
> Feb 02 14:03:05 cam1 motion[3022]: [0:motion] [NTC] [ALL] conf_load: 
> Processing thread 0 - config file /home/jack/.motion/motion.conf
> Feb 02 14:03:05 cam1 motion[3022]: [0:motion] [NTC] [ALL] 
> motion_startup: Logging to file (/home/jack/motion.log)
>
> Wondering if anyone else has noticed this, or if someone may have a 
> suggestion.
>
> Thanks in advance,
>

Hello all,

It seems a recent software update may have broken motion on this 
headless Raspberry Pi Zero W.

I noticed the following package updates:
   libcamera0.4:armhf (0.4.0+rpt20250127-3, automatic)
   rpicam-apps-lite:armhf (1.5.3-1, 1.6.0-1)
   libcamera-v4l2:armhf (0.3.2+rpt20241119-1, 0.4.0+rpt20250127-3)
   libcamera-dev:armhf (0.3.2+rpt20241119-1, 0.4.0+rpt20250127-3)
   libcamera-ipa:armhf (0.3.2+rpt20241119-1, 0.4.0+rpt20250127-3)
   libcamera-tools:armhf (0.3.2+rpt20241119-1, 0.4.0+rpt20250127-3)

OS: Raspbian GNU/Linux 12 (bookworm) armv6l
Host: Raspberry Pi Zero W Rev 1.1
Kernel: 6.6.74+rpt-rpi-v6
Motion: 4.7.0+git20240909-8d69cbf

I had originally built motion 4.7.0 from source. I went through that 
process again, starting with a fresh clone of the repo, but that made no 
difference.

Looks like motion is crashing. Motion's pid file is still present but 
that pid is not running. Attached is a level 9 log from motion. The only 
messages I see in syslog are these:

Feb 02 14:03:05 cam1 motion[3022]: [0:motion] [NTC] [ALL] conf_load: 
Processing thread 0 - config file /home/jack/.motion/motion.conf
Feb 02 14:03:05 cam1 motion[3022]: [0:motion] [NTC] [ALL] 
motion_startup: Logging to file (/home/jack/motion.log)

Wondering if anyone else has noticed this, or if someone may have a 
suggestion.

Thanks in advance,

-- 
Jack Christensen
Sent from Linux Mint 22.1 using Mozilla Thunderbird

I'm not using KDE, that repo contains ARM NEON CPU optimizations

On Sat, 28 Dec 2024, 18:24 Zamar Ac, <za...@gm...> wrote:

> Hi tosiara, thank you for pointing to your repo. I've Bodhi Linux 7.0,
> derived from Ubuntu 22.04 LTS (Jammy Jellyfish) 64x base currently
> installed and running with latest Motion. Will try your suggested config,
> and possibly compile ffmpeg if Motion can't open my cam files for
> processing. Would be nice if your sample config include some zone mask. I
> need to ignore in processing the frame area with luminescent sealing lamp
> (just left of the opposite door: see avi sample linked), since it gives
> plenty of false positives in motion detection, especially at sunny day
> time.
>
> Is there any reason you chosen KDE Neon to run Motion? Can you test the
> linked video sample to see if Motion is able to open the file and detect
> motion in it, and how long it takes? Not sure if email file attachments are
> possible in mailing lists. if Motion can't open the file, what would you
> suggest to do, i.e. compile latest ffmpeg, use a different motion detection
> package, etc. Re-encoding video files is NOT an option for me, since I've
> hundreds of video clips, and it'll take forever to re-encode.
>
> MotionTest.avi (
> https://drive.google.com/file/d/1M8K5gr3L9JiFE3z_5cwW2cchpg7A6oTM/view )
>
> Thank you!
>
>
> On 12/28/2024 5:25 AM, tosiara wrote:
>
> Motion can process offline files using netcam config option and file://
> protocol.
>
> You can check an example:
> https://github.com/tosiara/motion/wiki/Processing-of-offline-video-files
>
> On Sat, 28 Dec 2024, 08:27 Zamar Ac, <za...@gm...> wrote:
>
>> Hello, my security camera has basic motion detection in hardware, but
>> saves many false positive 2min motion clips continuously when the scene
>> lighting or other conditions keep changing. The Motion package allows to
>> re-analyze such static video sets for true motion presence. One can use
>> scripting options like *on_event_end*, *on_motion_detected* etc, which
>> allow to invoke external scripts.
>>
>> I'm not a coder, and want to ask if current Motion developer or some
>> experienced users can offer bash script examples for typical scenarios of
>> Motion use for static video analysis. I need to run Motion fast on a loop
>> through a large folder with monthly and daily subfolders storing hundreds
>> of 1080p 25fps HEVC 2min video clips from a security camera, to identify
>> video clips with true motion, and delete remaining false positive videos.
>> It might be possible to speedup video processing by setting higher video
>> FPS value than original, and adding a frame zone mask where to look for
>> motion.
>>
>> Once true motion event is detected in a video file, its name should be
>> added to MotionDetectedList.txt, and the file moved to a MotionDetected
>> folder with its date&time added to the file name. Motion should then start
>> processing the next video until all are analyzed.
>>
>> I tried a similar small Windows package VideoMotionDetect
>> <https://github.com/shuuryou/videomotiondetect> (
>> https://github.com/shuuryou/videomotiondetect ) based on OpenCVSharp
>> with embedded ffmpg to analyze near 800 continuous daily 720P 25fps 1min
>> AVI video clips from security camera for motion presence, and it took about
>> 30min to process such batch with high detection accuracy, with near 10-15%
>> of clip names with motion detected added to a text file. Sadly, that app
>> can't open HEVC 1080P videos from my another security camera.
>>
>> I'd really appreciate any help with this scripting task, or please
>> suggest an existing similar Linux or Windows app able to run such batch
>> motion detection task fast enough on a typical home PC. What clip
>> processing speed one would expect from Motion package in such scenario?
>>
>> Thank you!
>> _______________________________________________
>> Motion-user mailing list
>> Mot...@li...
>> https://lists.sourceforge.net/lists/listinfo/motion-user
>> https://motion-project.github.io/
>>
>> Unsubscribe: https://lists.sourceforge.net/lists/options/motion-user
>
>
>
> _______________________________________________
> Motion-user mailing lis...@li...://lists.sourceforge.net/lists/listinfo/motion-userhttps://motion-project.github.io/
>
> Unsubscribe: https://lists.sourceforge.net/lists/options/motion-user
>
> _______________________________________________
> Motion-user mailing list
> Mot...@li...
> https://lists.sourceforge.net/lists/listinfo/motion-user
> https://motion-project.github.io/
>
> Unsubscribe: https://lists.sourceforge.net/lists/options/motion-user

Hi tosiara, thank you for pointing to your repo. I've Bodhi Linux 7.0, 
derived from Ubuntu 22.04 LTS (Jammy Jellyfish) 64x base currently 
installed and running with latest Motion. Will try your suggested 
config, and possibly compile ffmpeg if Motion can't open my cam files 
for processing. Would be nice if your sample config include some zone 
mask. I need to ignore in processing the frame area with luminescent 
sealing lamp (just left of the opposite door: see avi sample linked), 
since it gives plenty of false positives in motion detection, especially 
at sunny day time.

Is there any reason you chosen KDE Neon to run Motion? Can you test the 
linked video sample to see if Motion is able to open the file and detect 
motion in it, and how long it takes? Not sure if email file attachments 
are possible in mailing lists. if Motion can't open the file, what would 
you suggest to do, i.e. compile latest ffmpeg, use a different motion 
detection package, etc. Re-encoding video files is NOT an option for me, 
since I've hundreds of video clips, and it'll take forever to re-encode.

MotionTest.avi ( 
https://drive.google.com/file/d/1M8K5gr3L9JiFE3z_5cwW2cchpg7A6oTM/view )

Thank you!


On 12/28/2024 5:25 AM, tosiara wrote:
> Motion can process offline files using netcam config option and 
> file:// protocol.
>
> You can check an example: 
> https://github.com/tosiara/motion/wiki/Processing-of-offline-video-files
>
> On Sat, 28 Dec 2024, 08:27 Zamar Ac, <za...@gm...> wrote:
>
>     Hello, my security camera has basic motion detection in hardware,
>     but saves many false positive 2min motion clips continuously when
>     the scene lighting or other conditions keep changing. The Motion
>     package allows to re-analyze such static video sets for true
>     motion presence. One can use scripting options like
>     /on_event_end/, /on_motion_detected/ etc, which allow to invoke
>     external scripts.
>
>     I'm not a coder, and want to ask if current Motion developer or
>     some experienced users can offer bash script examples for typical
>     scenarios of Motion use for static video analysis. I need to run
>     Motion fast on a loop through a large folder with monthly and
>     daily subfolders storing hundreds of 1080p 25fps HEVC 2min video
>     clips from a security camera, to identify video clips with true
>     motion, and delete remaining false positive videos. It might be
>     possible to speedup video processing by setting higher video FPS
>     value than original, and adding a frame zone mask where to look
>     for motion.
>
>     Once true motion event is detected in a video file, its name
>     should be added to MotionDetectedList.txt, and the file moved to a
>     MotionDetected folder with its date&time added to the file name.
>     Motion should then start processing the next video until all are
>     analyzed.
>
>     I tried a similar small Windows package VideoMotionDetect
>     <https://github.com/shuuryou/videomotiondetect> (
>     https://github.com/shuuryou/videomotiondetect ) based on
>     OpenCVSharp with embedded ffmpg to analyze near 800 continuous
>     daily 720P 25fps 1min AVI video clips from security camera for
>     motion presence, and it took about 30min to process such batch
>     with high detection accuracy, with near 10-15% of clip names with
>     motion detected added to a text file. Sadly, that app can't open
>     HEVC 1080P videos from my another security camera.
>
>     I'd really appreciate any help with this scripting task, or please
>     suggest an existing similar Linux or Windows app able to run such
>     batch motion detection task fast enough on a typical home PC. What
>     clip processing speed one would expect from Motion package in such
>     scenario?
>
>     Thank you!
>
>     _______________________________________________
>     Motion-user mailing list
>     Mot...@li...
>     https://lists.sourceforge.net/lists/listinfo/motion-user
>     https://motion-project.github.io/
>
>     Unsubscribe: https://lists.sourceforge.net/lists/options/motion-user
>
>
>
> _______________________________________________
> Motion-user mailing list
> Mot...@li...
> https://lists.sourceforge.net/lists/listinfo/motion-user
> https://motion-project.github.io/
>
> Unsubscribe:https://lists.sourceforge.net/lists/options/motion-user

Motion can process offline files using netcam config option and file://
protocol.

You can check an example:
https://github.com/tosiara/motion/wiki/Processing-of-offline-video-files

On Sat, 28 Dec 2024, 08:27 Zamar Ac, <za...@gm...> wrote:

> Hello, my security camera has basic motion detection in hardware, but
> saves many false positive 2min motion clips continuously when the scene
> lighting or other conditions keep changing. The Motion package allows to
> re-analyze such static video sets for true motion presence. One can use
> scripting options like *on_event_end*, *on_motion_detected* etc, which
> allow to invoke external scripts.
>
> I'm not a coder, and want to ask if current Motion developer or some
> experienced users can offer bash script examples for typical scenarios of
> Motion use for static video analysis. I need to run Motion fast on a loop
> through a large folder with monthly and daily subfolders storing hundreds
> of 1080p 25fps HEVC 2min video clips from a security camera, to identify
> video clips with true motion, and delete remaining false positive videos.
> It might be possible to speedup video processing by setting higher video
> FPS value than original, and adding a frame zone mask where to look for
> motion.
>
> Once true motion event is detected in a video file, its name should be
> added to MotionDetectedList.txt, and the file moved to a MotionDetected
> folder with its date&time added to the file name. Motion should then start
> processing the next video until all are analyzed.
>
> I tried a similar small Windows package VideoMotionDetect
> <https://github.com/shuuryou/videomotiondetect> (
> https://github.com/shuuryou/videomotiondetect ) based on OpenCVSharp with
> embedded ffmpg to analyze near 800 continuous daily 720P 25fps 1min AVI
> video clips from security camera for motion presence, and it took about
> 30min to process such batch with high detection accuracy, with near 10-15%
> of clip names with motion detected added to a text file. Sadly, that app
> can't open HEVC 1080P videos from my another security camera.
>
> I'd really appreciate any help with this scripting task, or please suggest
> an existing similar Linux or Windows app able to run such batch motion
> detection task fast enough on a typical home PC. What clip processing speed
> one would expect from Motion package in such scenario?
>
> Thank you!
> _______________________________________________
> Motion-user mailing list
> Mot...@li...
> https://lists.sourceforge.net/lists/listinfo/motion-user
> https://motion-project.github.io/
>
> Unsubscribe: https://lists.sourceforge.net/lists/options/motion-user

Hello, my security camera has basic motion detection in hardware, but 
saves many false positive 2min motion clips continuously when the scene 
lighting or other conditions keep changing. The Motion package allows to 
re-analyze such static video sets for true motion presence. One can use 
scripting options like /on_event_end/, /on_motion_detected/ etc, which 
allow to invoke external scripts.

I'm not a coder, and want to ask if current Motion developer or some 
experienced users can offer bash script examples for typical scenarios 
of Motion use for static video analysis. I need to run Motion fast on a 
loop through a large folder with monthly and daily subfolders storing 
hundreds of 1080p 25fps HEVC 2min video clips from a security camera, to 
identify video clips with true motion, and delete remaining false 
positive videos. It might be possible to speedup video processing by 
setting higher video FPS value than original, and adding a frame zone 
mask where to look for motion.

Once true motion event is detected in a video file, its name should be 
added to MotionDetectedList.txt, and the file moved to a MotionDetected 
folder with its date&time added to the file name. Motion should then 
start processing the next video until all are analyzed.

I tried a similar small Windows package VideoMotionDetect 
<https://github.com/shuuryou/videomotiondetect> ( 
https://github.com/shuuryou/videomotiondetect ) based on OpenCVSharp 
with embedded ffmpg to analyze near 800 continuous daily 720P 25fps 1min 
AVI video clips from security camera for motion presence, and it took 
about 30min to process such batch with high detection accuracy, with 
near 10-15% of clip names with motion detected added to a text file. 
Sadly, that app can't open HEVC 1080P videos from my another security 
camera.

I'd really appreciate any help with this scripting task, or please 
suggest an existing similar Linux or Windows app able to run such batch 
motion detection task fast enough on a typical home PC. What clip 
processing speed one would expect from Motion package in such scenario?

Thank you!

I would do a grep 8088 on all of your conf file and see if another
conf file has the wrong port defined.

Best guess is you have 2 conf files and one overrides the other or you
are using the wrong conf file.

On Tue, Dec 10, 2024 at 12:03 PM Richard Bown <ri...@g8...> wrote:
>
> Hi
>
> this is weird
>
> I have 2 odroid N2 computers both running motion both on 192.168.1.0/24
> subnet, Ubuntu 24
>
> each has 5 ip cams , one also has a USB cam, both are running ubuntu24 ,
> both headless
>
> odroidN2 , x.x. 1.130,  odroidN24 x.x.1.145, port range 8080 - 8089
>
> the cams on odroidN2 are OK, but on odroidN24 camera9 , port 8089
> something very strange
>
> if I use VLC with the ip address of the camera as listed in camera9.conf
> on odroidN24, I see the stream from the correct camera
>
> Yet If I use the mini http server on x.x.1.145:8089 I see the camera on
> odroidN2   x.x.1.130:8088
>
> I uses port 7770 on both computers as the control port,
> http://x.x.1.130:7770 shows all the correct cams connected to the
> version of motion running on that computer
>
> on my laptop with firefox browser http://x.x.1.145:7770  shows all the
> correct except one which is connected to x.x.1.130:8088
>
> from camera9.conf
> # Target base directory for pictures and films
> # Recommended to use absolute path. (Default: current working directory)
> ;target_dir /home/odroid/Pictures/motion9/
> target_dir /home/SDC/motion9/
>
> # The mini-http server listens to this port for requests (default: 0 = disabled)
> stream_port 8089
>
> two separate computers with diffenernt ip addresses
> ther ip cam connected to odroidN24, x.x.1.145  is on x.x.1.88 stream port 8089
>
> the ip cam on odroidN2  x.x.1.130  is on x.x.1.103   steam port 8088
>
> version on both 4.6.0
>
> odroidN24 has very recently been upgraded to ubuntu 24
>
> If I stop motion on odroidN2 x.x.1.130  and run http://x.x.1.145:7770 it
> still shows the wrong camera, so it must be something on the config
>
> of motion on odroidN24, but the url of the camera is correct and
> verified using VLC and the stream port is set to 8089
>
> The only thing the ip cams have in common is they are both ie-geek cams,
> but the both show correctly in their app on my android phone.
>
> the same config files were fine on ubuntu22 running motion 4.2
>
> I've even tried changing the port numbers on odroidN24 to 9080 ~ 9089
>
> any ideas ????
>
> Thanks Richard
>
>
>
>
> _______________________________________________
> Motion-user mailing list
> Mot...@li...
> https://lists.sourceforge.net/lists/listinfo/motion-user
> https://motion-project.github.io/
>
> Unsubscribe: https://lists.sourceforge.net/lists/options/motion-user

Hi

this is weird

I have 2 odroid N2 computers both running motion both on 192.168.1.0/24 
subnet, Ubuntu 24

each has 5 ip cams , one also has a USB cam, both are running ubuntu24 , 
both headless

odroidN2 , x.x. 1.130,  odroidN24 x.x.1.145, port range 8080 - 8089

the cams on odroidN2 are OK, but on odroidN24 camera9 , port 8089  
something very strange

if I use VLC with the ip address of the camera as listed in camera9.conf 
on odroidN24, I see the stream from the correct camera

Yet If I use the mini http server on x.x.1.145:8089 I see the camera on 
odroidN2   x.x.1.130:8088

I uses port 7770 on both computers as the control port, 
http://x.x.1.130:7770 shows all the correct cams connected to the 
version of motion running on that computer

on my laptop with firefox browser http://x.x.1.145:7770  shows all the 
correct except one which is connected to x.x.1.130:8088

from camera9.conf
# Target base directory for pictures and films
# Recommended to use absolute path. (Default: current working directory)
;target_dir /home/odroid/Pictures/motion9/
target_dir /home/SDC/motion9/

# The mini-http server listens to this port for requests (default: 0 = disabled)
stream_port 8089

two separate computers with diffenernt ip addresses
ther ip cam connected to odroidN24, x.x.1.145  is on x.x.1.88 stream port 8089

the ip cam on odroidN2  x.x.1.130  is on x.x.1.103   steam port 8088

version on both 4.6.0

odroidN24 has very recently been upgraded to ubuntu 24

If I stop motion on odroidN2 x.x.1.130  and run http://x.x.1.145:7770 it 
still shows the wrong camera, so it must be something on the config

of motion on odroidN24, but the url of the camera is correct and 
verified using VLC and the stream port is set to 8089

The only thing the ip cams have in common is they are both ie-geek cams, 
but the both show correctly in their app on my android phone.

the same config files were fine on ubuntu22 running motion 4.2

I've even tried changing the port numbers on odroidN24 to 9080 ~ 9089

any ideas ????

Thanks Richard

I have camera installed.  On my Raspberry Pi,  I am able to connect and view the camera with url rtsp://admin:password@cameraip:554.  I installed motion and used the same url.  Message in browser says unable to connect to camera.   ?
Richard Robbin

WA8RR

Sent from my iPhone

Try to restore the module back, reboot and check the output:

ls -l /dev/v4l/by-id/

How many devices does it show?

On Sun, Dec 1, 2024 at 10:48 PM Richard Bown <ri...@g8...> wrote:
>
> after a bit of googling I found the cam needs the uvc module, so I have blacklisted the meson_vdec module
>
> now lsusb does not show the cam aaaagh!
>
> On 01/12/2024 20:18, Richard Bown wrote:
>
> Hi
>
> using Ubuntu22 with the C270  USB camera works, v 4.3
>
> upgrading to ubuntu 24 with motion 4.6 it fails with log entry of camera does not capture.
>
> I looked at the loaded libs between the two earlier, and it seemed to be more v4l was load on UB 22
>
> I had to upgrade today to Ub24 on the second Odroid N2, it needs two to handle 11 ip-cams, ubuntu22 died, so got upgraded
>
> and still stuck with this camera down.
>
> [1:ml1:Front porch] [WRN] [ALL] [Dec 01 20:13:30] mlp_retry: Retrying until successful connection with camera
> [1:ml1:Front porch] [NTC] [VID] [Dec 01 20:13:30] vid_start: Opening V4L2 device
> [1:ml1:Front porch] [DBG] [ALL] [Dec 01 20:13:30] util_parms_add: Parsed: >palette< >17<
> [1:ml1:Front porch] [DBG] [ALL] [Dec 01 20:13:30] util_parms_add: Parsed: >input< >-1<
> [1:ml1:Front porch] [DBG] [ALL] [Dec 01 20:13:30] util_parms_add: Parsed: >norm< >0<
> [1:ml1:Front porch] [DBG] [ALL] [Dec 01 20:13:30] util_parms_add: Parsed: >frequency< >0<
> [1:ml1:Front porch] [NTC] [VID] [Dec 01 20:13:30] v4l2_device_open: Using videodevice /dev/video0 and input -1
> [1:ml1:Front porch] [DBG] [VID] [Dec 01 20:13:30] v4l2_device_capability: ------------------------
> [1:ml1:Front porch] [DBG] [VID] [Dec 01 20:13:30] v4l2_device_capability: cap.driver: "meson-vdec"
> [1:ml1:Front porch] [DBG] [VID] [Dec 01 20:13:30] v4l2_device_capability: cap.card: "Amlogic Video Decoder"
> [1:ml1:Front porch] [DBG] [VID] [Dec 01 20:13:30] v4l2_device_capability: cap.bus_info: "platform:meson-vdec"
> [1:ml1:Front porch] [DBG] [VID] [Dec 01 20:13:30] v4l2_device_capability: cap.capabilities=0x84204000
> [1:ml1:Front porch] [DBG] [VID] [Dec 01 20:13:30] v4l2_device_capability: ------------------------
> [1:ml1:Front porch] [NTC] [VID] [Dec 01 20:13:30] v4l2_device_capability: - STREAMING
> [1:ml1:Front porch] [ERR] [VID] [Dec 01 20:13:30] v4l2_device_capability: Device does not support capturing.
> [1:ml1:Front porch] [DBG] [VID] [Dec 01 20:13:30] v4l2_device_close: close
>
> Logitech C270
>
> Technical Specifications
>
> Max Resolution: 720p/30fps
> Camera mega pixel: 0.9
> Focus type: fixed focus
> Lens type: plastic
> Built-in mic: Mono
> Mic range: Up to 1 m
> Diagonal field of view (dFoV): 55°
>
> Any help appreciated please
>
> --
> --
>  Best wishes
>  Richard Bown
>
>  Email : ri...@g8...
>
>  HTTP  :  https://www.g8jvm.com
>
>  ###################################################################
>
>  Ham Call: G8JVM . QRV: HF + 50-432 MHz + Microwave
>  Maidenhead QRA: IO82SP38, LAT. 52.6620 N LONG. 2.4696 W
>  VHF: 6mtrs 200W, 4mtrs 150W, 2mtrs 300W, 70cms 200W,
>  Microwave: 23cms 110W, 13cms 100W, 9cms 15W, 6cms 10W & 3cm 5W
>  OS: Linux Mint 22 Wilma, on a Dell Latitude 5590 laptop
>  ###################################################################
>
>
>
> _______________________________________________
> Motion-user mailing list
> Mot...@li...
> https://lists.sourceforge.net/lists/listinfo/motion-user
> https://motion-project.github.io/
>
> Unsubscribe: https://lists.sourceforge.net/lists/options/motion-user
>
> --
> --
>  Best wishes
>  Richard Bown
>
>  Email : ri...@g8...
>
>  HTTP  :  https://www.g8jvm.com
>
>  ###################################################################
>
>  Ham Call: G8JVM . QRV: HF + 50-432 MHz + Microwave
>  Maidenhead QRA: IO82SP38, LAT. 52.6620 N LONG. 2.4696 W
>  VHF: 6mtrs 200W, 4mtrs 150W, 2mtrs 300W, 70cms 200W,
>  Microwave: 23cms 110W, 13cms 100W, 9cms 15W, 6cms 10W & 3cm 5W
>  OS: Linux Mint 22 Wilma, on a Dell Latitude 5590 laptop
>  ###################################################################
>
> _______________________________________________
> Motion-user mailing list
> Mot...@li...
> https://lists.sourceforge.net/lists/listinfo/motion-user
> https://motion-project.github.io/
>
> Unsubscribe: https://lists.sourceforge.net/lists/options/motion-user