Latest App update has another mistake!

I have turned off all of the options in AI and rely on the ordinary motion events as the AI never managed to differentiate between another cat, a fox, a Badger, s rabbit etc. I just use a more accurate and reliable Human Intelligence.

In the past I have received a notification of movement after the thing that triggered it had left the field of view. I could quickly rewind and see which animal triggered it. If it was a fox then I could go outside with a flashlight which was enough to scare away the fox before it got to the hens.

This means that I am not going outside for every event and trying to scare rabbits and other non-threat animals away.

Live view is no use for this as the trigger has left the field of view, and waiting for the event screen to show is too slow. If I managed to put a camera near the hens then I would see a live view of the fox doing the damage which is no use at all.

It all worked fine until the last update and I am now relying on listening through an open window - at least I can hear a disturbance albeit later than when I could rewind in the event screen before the last update.

Basically, I tested one cam to see if it would work as I wanted before buying seven more - now they do not work as they did when I tested them.

The whole idea was to place cameras to monitor the runs that the foxes always use to get to the hens - unfortunately every other animal also uses the same runs. Monitoring the runs gave me sufficient time to take action if needed.

It looks like I am going to have to change the fencing around the hens once again - but it doesn’t take a fox long find a way in. The technology was a substitute to physical barriers akin to Fort Knox - but not any more.

For my purposes the cameras may as well be broken.

The hens by the way are pets - so it matters to me that they are safe.

I am sure there are other people who have requirements for a similar system and also bought Wyze - just to find out that the cams don’t perform as they did before the last update.

1 Like

I have a front door camera that keeps watch for people. It also keeps an eye on my cats. I have a doorbell but many people don’t know it’s there or choose not to ring it. Especially delivery drivers. They drop packages off on my door step but I also have medication delivered that requires a signature. So imagine a delivery driver standing at my door waiting for a signature but I don’t get notified until after the event is finished recording. The driver could have stepped out of frame while waiting because I have a large front porch. The recording time could very well take several minutes depending on the motion. By the time I walk to the front door, the delivery person or mailman could have already hopped back into their vehicle and drove away. It’s nobody’s business why I need to back up that live event capture, I just need to continue doing it because Wyze roped us in with that feature and then unceremoniously took it away.

2 Likes

Yes - that’s exactly my point as well.

It was a selling point for me - much more so than the AI which I didn’t find very accurate when I did actually use it.

Another observation I have only just noticed - I have been plagued with two notifications for each event ever since the AWS outage - I have reported it to Wyze but nothing happened about it.

Now I see the point of having two notifications - the first one arrives when the event starts and the second one arrives when the event had been recorded and I can see the event view rather than the live view when I tap on the thumbnail … I can’t imagine Wyze planned it to work like that!

Below is a screenshot:

Time on phone clock = 14.52

Time notification arrived = 14.48

It is still showing “live” - therefore I cannot rewind to the start of the event for over 4 minutes… Ok it was a long recording (and not of a fox this time) … But over 4 minutes before I can look at what triggered the event. Seriously - is this acceptable??

1 Like

You don’t.

Previously, it WAS a near real time ‘live’ view of video currently recording to AWS with the slider that allowed you to ‘rewind’ within that same UI to review the previously recorded footage of the currently recording video on AWS KVS (Amazon Kenesis Video Stream). That functionality was lost with the move to TUTK (ThroughTek Co., LTD Taiwan) which now only streams P2P from the cam in love view rather than from the AWS recorded footage in near real time.

Others have different preferences as they have different application scenarios\landscapes and have been successfully using the old functionality for some time without the so called stability issues. The loss of usefully functionality is an issue to some. Perhaps for those who did rely on this functionality, it is a game changer issue.

You are assuming users are not getting full length video in CamPlus which can be up to 5 minute videos.


So, my original hypothesis didn’t go deep enough. The move was not to reduce the near real time AWS download streams of still recording events launched from notifications, it was to completely terminate them all together in favor of real time P2P streaming thru TUTK.

1 Like

It is for me - but it will take me ages to replace all eight as they are built into wall lights so unless I can find another make with the same dimensions then I will have to modify some more lights.

I don’t think the change is due to stability - more likely the new provider is cheaper than AWS.

1 Like

The one in the screenshots above is over 4 minutes and still It is messing around in “live” … Useless!

All this mess in the name of stability - It is easier to dump the old functionality rather than find s fix for instability ???

2 Likes

Exactly!

1 Like

Those are good points, I didn’t think about it for those use cases, and clearly Wyze didn’t either.

I think Wyze still uses KVS for events, and has used TUTK for live view. The only thing that has changed is that they now use it for viewing notifications as well.

Events are processed like this:
Camera sees motion, starts recording.
After 6 seconds, the clips is sent to AWS for ai processing and notification. This is done after 6 seconds so you don’t have to wait the full length of the video before getting the notification.
Then it keeps recording, until max length is up, or it doesn’t see motion. The full event video is now uploaded to AWS, where it’s accessible in the app.

This makes it so if you were to view the events page right as a notification is sent, the video won’t be done recording, and it won’t be uploaded so you get an error. It can only be a few seconds if it’s a short video, but if it’s long your gunna have to wait a long time.

The new system was supposed to improve this by taking you to live view, but I don’t think it was considered that the subject may only be visible for the first few seconds.

As you suggested, an option to switch between both would be the best bet since there’s obv a real use case for both of them.

Everything I said is just how I think the system works, but I’m probably not 100% right on all of it. Im not sure how you were able to watch the event video as soon as you had clicked the notif, unless the video was only a few seconds long?

1 Like

It took approximately 5s to load the video but once I saw what triggered it then I deleted the video. I never got to the end of the event video after I realised that it wouldn’t contain any more information. All I needed to know was fox, cat, rabbit etc.

The chances are that the event videos were not very long as a fox can run through the field of view in 2- 3 seconds and then there is no more movement for the cameras to record.

I do hope Wyze takes on board the option to choose which system to use - then everyone is happy.

1 Like

Yea, that would be ideal, a per-camera setting for events or live view.

1 Like

just like their services… now they have the customers, lets turn the screws… this is how drug pushers work… lots of freebies… then comes the paying for it… don’t like being pushed into cam plus lite… even if i select $0 (which they really try to make you feel guilty about)… they rarely implement changes on their Wishlist page… ive spent many $100’s of dollars on their stuff, and just wish id gone to a different company.

1 Like

Same here - I even upgraded my broadband connection and added a mesh wi-fi system to make sure the cameras had a good wi-fi connection. Plus fitting 6x 940nm I-R illuminators so the cameras can see in the dark without having any light showing to human it animal eyes. In my case it was £ but still hundreds of them!

Then the time taken to modify some wall lights to fit the cameras into them so that they are effectively invisible to human intruders so they may not bother hiding their faces.

1 Like

Incorrect.

With Event Recording set to Detects Motion, once the Cam’s Light Pixelation Algorithm detects motion the video is immediately sent to the AWS server for credentialing, servicing and storage, it does not wait 6 sec to send the video to AWS. With AI Smart settings on in Event Recording for CP\CPL, the AI Bot immediately begins interrogation as soon as the CP\CPL credentials are verified.

If Notifications is on for All Motion Events, the user will be sent the general motion notification as soon as the server is hit with the video, not 6 seconds later. (Only if the notorious Wyze Notification Latency problem isn’t active). If a user has AI notifications on, a notification will be sent to the user when the AI Bot confirms the AI object tagging…BUT… Only if this happens within the first 6 seconds of the interrogation. If it takes the bot more than 6 sec to tag AI, no AI notification is ever sent. This was the Fix-it Friday issue brought up because with All Motion notifications off and AI notifications on, any event taking longer than 6 sec to tag never produces a notification and is therefore defeating the purpose of having CP\CPL in the first place.

The reason it worked before is because the AWS KVS service allows for bidirectional video stream in near real time. Using AWS KVS, clicking a notification or event that was still in the record process did not take you to a P2P live stream to the cam, it took you to a near real time ‘very close to live’ microsecond delay download stream of the recorded video that was still being uploaded, rendered and saved. An expensive service for sure. You weren’t watching live video, you were watching recorded video from the cloud on an incredibly short delay.

Wyze claims the change was because of stability issues for those clicking in so fast that it couldn’t render and return. But, rather than coding in a 1 or 2 second delay for the notification or event posting to insure enough video had been uploaded and rendered, they instead opted for the total termination of the AWS KVS and instead went to straight P2P live stream. Now it takes what amounts to an eternity compared to the previous method to actually see what triggered the event in the first place. Moreover, the move significantly reduces the bandwidth throughput on the AWS servers thereby reducing Wyze costs while at the same time reducing services and functionality to customers who paid for that service.

2 Likes

I bet you will never get Wyze to admit it is all a cost saving exercise - I thought Wyze was different to the run of the mill tech companies… I was wrong.

Over here in the UK we have British Telecom (BT) - My broadband went wrong, giving me 5mb download. It took them 7 weeks to mend it - asking me to run tests and reboot the router regularly. Admitting something was broken would cost them money so eventually I stopped paying them. Within a week it was mended - the engineer said he had to replace a circuit board in the exchange - but BT never admitted there was a fault. Everything is to try and save money which never results in a good service

2 Likes

Remember we all loose if Wyze goes out of business… But I do agree, it would be nice if this feature could come back as an option for some usecases.

1 Like

Hello ShardlowM,

Thanks for the feedback. It is a deep discussion about the live event feature on the specific scenario. Let me show more tech details about the feature.
1, For most cases, the average spent time between the motion detected by the camera and the notification arriving at the phone is around 6 seconds.
2, For an alerted user, when he/she hears the notification alarm,take out the phone, unlock it, tap the notification bar and the phone builds a connection to play the live video, it would take another 9 seconds.
3, The camera will finish the CamPlus event recording process after 8 seconds if there is no motion can be detected.
The attached timeline will be more straightforward. And I think this model can be used to discuss the urgent scenario the users need to figure out what was and is happening.

For this scenario, under the current TUTK live solution, the user will miss the first 15 seconds of the event, which means he/she can start watching the event from the 15th second.
If the motion is still ongoing, such as the fox wandering in front of the camera, it should be fine for this case. The owner will get an alert from the live view and yell the fox away. – This case is not your top concern, right?
If the motion is finished before the TUTK live establishing, which means the fox is running out of the scope of the camera lens and the user didn’t see anything from the live view, the user ought to leave the live page and go to the Event tab to check the latest event that is uploaded to the AWS storage. The event will play it from the very first beginning. I think this is the case you suggested we consider. We thought about it for sure: for this case, how many seconds the user has to wait to know what was happening from the app. It should be 30 seconds at most.
The longest waiting time will take place when the motion stop right before 15th second. And the camera will include all captures from 0 to 23 seconds to the cloud. It might take another 7 seconds to finalize the upgrading process and download it to the phone app. The user will be able to access to the playback event from the 30th second.

On top of my head is we should set up a shortcut on the live page to direct users to enter the Event tab to watch saved events.

Hopefully, my explanation won’t confuse you or complicated the application of the feature.

To echo your suggestion, I guess it would slow down the response of the app if we set up 2 live streams for the feature.

To fill the 30ish second gap for the urgent event, personally, I advise the users to set up the camera in a better place at a better angle, leaving more cushion to track the motions. For example, if the camera directs the gate of the lobby, it only can track 2-3 seconds of motions when people get through it. If the camera is set at a corner of the lobby, it will track 15ish seconds when people walk through the lobby.

Hopefully, my explanation can solve your questions.

Best wishes,

Xiaolong

2 Likes

That makes sense, and I thought the same thing, on the live view page have a shortcut to the latest event, so then the user can choose what they want every time.

1 Like

@Xchen

So no change from the present situation then.

I am not interested in seeing the live view of the event - the fox has passed by and it is out of the field of view in 2-5 seconds.

I need to see what triggered the event - just like I could before you made this change.

How does changing the angle of the cameras apply in my situation?? In 30s a fox can be 100 yards away - that is why I need to see what triggered the event - that’s why I need to rewind the event video.

It is plain to see that Wyze have made a decision on this - unfortunately just another example of tech-centred people providing what they think users want.

What makes it worse is that the functionality that myself and other users need was provided and then taken away - all in the name of stability etc …

@Xchen - did you read my earlier post with the screen shot? One event yesterday was still going to live view 4 minutes after the initial trigger … How is that acceptable? Ok, it was unusual for it to be that long but it shows that it is possible.

1 Like

This is so frustrating - The latest event will still have to do it’s processing and I still will not be able to see what triggered it using the slider.

What happens after the trigger is useless - live view is useless- animals move quickly and are gone from the field of view … So live view sees nothing. Moving the camera angle doesn’t work - the fox has passed by and causing damage and mayhem.

By the way - I am referring to the fox in the singular when in fact there are several. When I tested Wyze it had the functionality to allow me to manage the situation but this is no longer the case.

When this change happened at the last update I thought it was a mistake as it is illogical to me to remove a function purposely. But now I see that it is not an accidental error - it is a money-saving measure.

I will start my search for something more functional and hopefully rebuild my system based on another company.

The cameras were the core product of Wyze but now the product range is so large that I feel the cameras are now an add-on to the scales, the door locks, the sprinklers, the gun safes etc etc etc. I know you have to progress your business model but it seems you are doing this at the expense of the cameras.

You obviously have removed the functionality from the cameras to save money - and nothing is going to change that.

I am writing this in the early morning - it is quite cold here yet I am sat with a window open because at least I will know if something is happening to my hens, or cat, or my neighbours cat. The technological solution is now useless.

1 Like

@ShardlowM It’s frustrating to get anywhere with these guys. It’s like talking to a concrete wall as my dad would say. Just let me Rewind my current live event to see what triggered the motion. The robber’s at the door. Or the coyote has attacked my cat. Or the wild pig has entered my property and wants to eat my plants. So yea don’t make me wait minutes until the video is finished recording to let me see what set it off. How much plainer can we get? And don’t tell us to re-angle our cameras or purchase more to account for wider views. You really don’t want to tell people that. Really.

2 Likes