Firmware 4.10.3.103 (Pan) and 4.9.4.103 (v2) released to Beta!

Hi Beta testers,

Happy 4th holiday! We just released new firmware for Pan (4.10.3.103) and Cam v2 (4.9.4.103) to Beta. Here are the updates from last beta release:

  • Updated AI algorithm for improved accuracy
  • Minimized the ‘fast forward’ video issue caused by high CPU usage
  • A few bug fixes

Areas to test:

  • AI accuracy (person vs. no person)
  • Any serious ‘fast forward’ videos
  • Overall firmware robustness

Please let us know if you have any issues on the latest firmware, especially the detection accuracy and other issues on the camera stability. Looking forward to your feedback. Thank you very much!

8 Likes

Need to update the “What’s new” page.
First try worked for a short few minute default 3s time lapse
Bombed at 261MB when trying process a previously recorded 418MB time lapse. Tried twice, same result.
/edit
When I tried to upgrade my test V2 which was NOT next to the router but rather far enough away that I suspected signal strength may have been a problem it locked up.
Moved it closer and after several tries it succeeded

Does this mean it won’t detect every bug at night? :wink:

The motion detection is useless at night on my backyard camera.

4 Likes

Thank you! Currently we only put official release note into ‘What’s New’ since we change Beta firmware more frequently. We need to think about how to inform beta users better in the future.

2 Likes

You can turn on person detection for the camera so you can choose to receive person notifications only. It may not fit all your need but it is a bit better.

1 Like

Both software bugs and biological bugs are trouble maker :smile:

3 Likes

My update shows 4.10.3.103 on my Pan… is the above 4.9.3.103 a typo… or did I get some kind of mutant update? :slightly_smiling_face:

1 Like

Another try on Time Lapse - successfully ran a 3 hour time lapse that took up 193.17MB w/ 3 second interval.
Going for more.

1 Like

Thanks for catching that. I have corrected the title for both cameras. :slight_smile:

2 Likes

You’re welcome… :slightly_smiling_face:

1 Like

Thanks for posting about the new update. I always check to see the release notes, but most of the time there is nothing, so let’s PLEASE keep this trend going.

Have to say since switching from person/vehicle tagging to just person tagging it SUCKS! Does NOT pick up people like it should and they have to be very close it seems to be picked up. What should we expect with this new update?

Also, I notice that people tagging is only cloud rather than triggering motion recording to SD Card. Is this per design.

Of course it’s nice to see continued progress made and it’s easier to pick up on the issues rather than anything. There has always been a lot of good though too which is why we own quite a few of these guys…

The IR lights attract many bugs (insects/arachnids).

One idea to reduce this is to use a external IR illuminator and turn off the camera IR leds.

I think this has to do with the cool down period.
It detect motion (someone arriving in a car) and then by the time the cool down has happened, the person has left the scene.

This defeats the savings of getting the Wyze camera.

I’d like to see an option to exclude motion under X second(s).
All my bugs are quick sub-second blasts.

2 Likes

@mfroader is probably right about the cooldown period. It is supposed to be about five minutes…
I assume from your post you are using event recording instead of continuous to sd card. I use continuous and just use the events as markers to check on things I want to see. I’ve been able to pick up things before and after events I would have not had with event only recording. (Including 3 guys breaking into my car). just a thought.

2 Likes

Tried a 4 hr, it bombed download at 282.12MB

Wow, stress testing! :slight_smile:

1 Like

@tom and @tomp
I think you are onto something with the cool down period, however has it been stated that person tagging is a cloud only feature and NOT relating to event recording in SD cards? I suppose that it’s in the “events” area suggests this as a fact.
I literally have “person” tags only when someone is really close to the camera such as leaving in the AM and coming home in the evening. It’s facing the street from the porch and never records a person tag on the sidewalk 40ft away…
If we can’t control the cloud event motion sensitivity or area, then the cloud events will ALL be eaten up by leaves moving in the wind with the cool down period. When the camera is inside it’s less effected, but things like the sun moving throughout the days, shadows triggers as well.
I’ve raised the motion sensitivity with no avail.

I definitely have had SD card recording set to event and not continuous. While that’s been dissapointing, I guess spent less time looking for nothing. At least now that I just changed it to continuous it’s going to be recorded, but finding it will be another story. Would be nice to have a motion graph to view based on higher levels then again probably won’t make a single difference…
Being able to select specific parts of view to track motion rather than just a square will be more effective. Basically blacklist any bush tree which is affected by the wind, etc.
Regardsles, seemed tagging was much more effective when it was both people/vehicles…

@WyzeTao, talk about it - once again it’s creating LARGE .xlog files in WyzePlatformKit. I’ve 4 files that sucked up 9.09 GB.
You know, a gig here, a gig there and pretty soon it adds up to no memory on the device.

There are several threads relating to this on the Forum wishlist. Please go there and vote for all that seem relevent. The wish under one-second-wish has been relegated tp"maybe later "status, so every vote helps